Compare commits

..
407 Commits
Author SHA1 Message Date
Laurence IslaandGitHub 39e3246623 Upgrade protolude to 0.3.1 (#2306) 2022-06-08 16:15:56 -05:00
steve-chavez 120ed06134 bump version to 9.0.1
* Format CHANGELOG
2022-06-03 22:19:16 -05:00
steve-chavez a8452e7a2e fix: using CPU while idle 2022-06-03 22:19:16 -05:00
steve-chavez f7a2220159 fix: disable parallel GC for perf on high-core CPU 2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez d8085d41fb ci: run on branches starting with "rel-" 2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez 13d343bdbf ci: Run ARM builds on releases 2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez 6fc47aa525 ci: Fix Windows build using Stack
Added PostgreSQL binaries to the PATH and removed the use of msys2-keyring
2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 02eb3f0612 build(deps): bump codecov/codecov-action from 3.0.0 to 3.1.0
Bumps [codecov/codecov-action](https://github.com/codecov/codecov-action) from 3.0.0 to 3.1.0.
- [Release notes](https://github.com/codecov/codecov-action/releases)
- [Changelog](https://github.com/codecov/codecov-action/blob/master/CHANGELOG.md)
- [Commits](https://github.com/codecov/codecov-action/compare/v3.0.0...v3.1.0)

---
updated-dependencies:
- dependency-name: codecov/codecov-action
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez 3119c52cb7 ci: remove armv7 architectures for ARM builds
Now only builds aarch64 on the remote server without Docker
2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez 74678391c6 Add CI for ARM architectures (#2127) 2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez d16972e8da Fix misleading disambiguation error where relationship looks like valid syntax
* Add columns for the m2m relationship
2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 0f5caa50c6 build(deps): bump codecov/codecov-action from 2.1.0 to 3.0.0
Bumps [codecov/codecov-action](https://github.com/codecov/codecov-action) from 2.1.0 to 3.0.0.
- [Release notes](https://github.com/codecov/codecov-action/releases)
- [Changelog](https://github.com/codecov/codecov-action/blob/master/CHANGELOG.md)
- [Commits](https://github.com/codecov/codecov-action/compare/v2.1.0...v3.0.0)

---
updated-dependencies:
- dependency-name: codecov/codecov-action
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 176db06168 build(deps): bump actions/cache from 2.1.7 to 3
Bumps [actions/cache](https://github.com/actions/cache) from 2.1.7 to 3.
- [Release notes](https://github.com/actions/cache/releases)
- [Commits](https://github.com/actions/cache/compare/v2.1.7...v3)

---
updated-dependencies:
- dependency-name: actions/cache
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez 3ee3bcd7ae Bypass FreeBSD builds due to timeouts 2022-06-03 22:19:16 -05:00
Ezequiel AlvarezandSteve Chavez 113c487db2 fix: json/jsonb columns should not have type "string" in OpenAPI spec (#2203)
* Switch to no type for json/jsonb
2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 6876fbb0aa build(deps): bump actions/checkout from 2.4.0 to 3 (#2186) 2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez b8b486740c build(deps): bump actions/upload-artifact from 2.3.1 to 3 (#2187) 2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 83496fbcb8 build(deps): bump actions/download-artifact from 2.1.0 to 3 (#2185) 2022-06-03 22:19:16 -05:00
Laurence IslaandSteve Chavez ebc26d0c09 fix: Using GET with certain Content-Type headers now correctly calls the no parameter function if it exists
Using GET with text/plain or application/octet-stream as Content-Type headers no longer returns 404 Not Found when a function with no parameters exists
2022-06-03 22:19:16 -05:00
Steve Chavez dd8a2c8979 fix: Keep working when EMFILE is reached (#2158)
Done by upgrading to warp 3.3.19
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 7cc8502a64 fix: Remove aggregates, procedures and window functions from schema cache and OpenAPI output
Aggregates and Window functions can't be called as RPCs in a useful way.

Procedures are not supported right now, but might be added later.

Resolves #2101

Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 5a88161ff7 fix: Take PG version into account in --dump-schema
The PG version is only read by the Connection Worker, which is not used in the case dump-schema. Now, the pg version is read in the schema cache queries directly, avoiding this problem in all cases.

Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 973c22a102 ci: Fix windows build failing because of outdated keyring
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 2c001010df fix: Remove trigger functions from schema cache and OpenAPI output
Trigger functions can't be called directly from SQL and can't be called
via the /rpc prefix either - it makes no sense to expose them in the
OpenAPI output. And we don't need to cache them in the schema cache
either.

Best practice would be to keep the trigger functions in a non-exposed
schema anyway.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez a36c5ce52d ci: Add -with-rtsopts=-KxK option to prevent introducing space leaks
Resolves #387
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 8bfbf9eb89 Align columns in postgrest.cabal 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 2e334db31b fix: Read database configuration properly when = is present in value
Resolves #2120
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez d1851a6966 refactor: Reformat db-config query 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 76fdf29f5c refactor: Move each Wai Middleware to a separate file 2022-06-03 22:19:16 -05:00
Steve Chavez 490f631011 cov: empty coverage overlay (#2119)
The overlay depended on the version number, which caused errors when upgrading.

Also update cabal description.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 685acccf6d test: Reorganize test/ folder into one subdirectory for each test type 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 9378de558f Remove unused dependencies from postgrest.cabal 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 747b0234ae cov: Add coverage overlay for impossible case in corsPolicy 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez ce578ec1b8 refactor: simplify logger middleware 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 7272d8894c cov: Remove unreachable static icon 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez de8253facc cov: Add tests for request logging 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez a15118576d test: Remove unused unix-socket.config file
Unix sockets are the default in io tests and we are testing ip connections explicitly.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 0a44125d92 test: Improve io test perfomance
Local test improves performance by 15%.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 9c9e25b2af test: Move io-tests SQL fixtures to a separate file 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez ce6ea74ee1 test: Improve log output for failing io-tests 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez e8e3349a71 test: Refactor invalid_role_claim_key_notify_reload test to properly read all available input 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 06a9794448 test: Use .encode() end .decode default values in io-tests
Those default to "utf-8" anyway.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 30a6ae6241 nix: Fix nested postgrest-with-postgresql-x calls when -f is used 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez c228c37619 Remove outdated heroku manifest 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 9f09a6547d cov: Improve jwt claims test-cases 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 130dcc4f66 fix: Fix wrong CORS header Authentication -> Authorization
Also refactors defaultCorsPolicy and corsPolicy and cleans up CORS tests
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 7db4a75146 fix: Make recursive view parsing work with XMLTABLE + DEFAULT 2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez eedfe957b5 build(deps): bump actions/upload-artifact from 2.3.0 to 2.3.1 (#2094) 2022-06-03 22:19:16 -05:00
Steve Chavez ef240ef05a fix: is not working with upper/mixed case values (#2081) 2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 89e93ef6f5 build(deps): bump actions/upload-artifact from 2.2.4 to 2.3.0
Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 2.2.4 to 2.3.0.
- [Release notes](https://github.com/actions/upload-artifact/releases)
- [Commits](https://github.com/actions/upload-artifact/compare/v2.2.4...v2.3.0)

---
updated-dependencies:
- dependency-name: actions/upload-artifact
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
2022-06-03 22:19:16 -05:00
dependabot[bot]andSteve Chavez 1307d00df2 build(deps): bump actions/download-artifact from 2.0.10 to 2.1.0
Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 2.0.10 to 2.1.0.
- [Release notes](https://github.com/actions/download-artifact/releases)
- [Commits](https://github.com/actions/download-artifact/compare/v2.0.10...v2.1.0)

---
updated-dependencies:
- dependency-name: actions/download-artifact
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 9c7e6070f0 fix: Dump db-config option without quotes as it's a boolean 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 3c8e309a32 nix(feat): Add postgrest-git-hooks to enable pre-commit and pre-push hooks 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 7940e73a12 nix: Add weeder to postgrest-coverage
Removes a bit of dead code from SpecHelper.hs
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 9b397878dc refactor: Make asJsonSingleF "safe", by accessing only the first element of a json_agg result 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez eda59c241e fix: Execute deferred constraint triggers when using Prefer: tx=rollback
Resolves #2020
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 4d64aae870 fix: Link to latest docs in pre-release openapi output
Resolves #2018
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez e337891c89 ci: Use matrix strategy to test against postgres versions 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez d8f03e37ae ci: Run IO tests against all postgres versions
Resolves #1820.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 0d99789938 ci: Disable update of dockerhub description on release
This is not allowed via personal access token right now and could be re-enabled in the future, once solved by docker hub.
2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 88674c5af6 nix(feat): Add bash completion for git references 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez 1aebe62a03 nix(feat): Add postgrest-loadtest 2022-06-03 22:19:16 -05:00
Wolfgang WaltherandSteve Chavez f7b3608d4e nix(feat): Add postgrest-with-git and postgrest-with-pgrst tools 2022-06-03 22:19:16 -05:00
Steve ChavezandGitHub c9f017c632 Bump to v9 (#2045)
* Reorganize the CHANGELOG
* Add missing change for dropping support for embedding hints used with '.'
* Move the pg 14 compatibility entry to Added
2021-11-26 14:10:57 -05:00
steve-chavez 132ecd8d74 Correct count query when doing inner joins 2021-11-26 10:39:19 -05:00
Wolfgang Walther 19c3ee2143 ci: Fix target commit for report workflow 2021-11-26 08:01:00 +01:00
4fb0d1279c test: Add test to confirm ASSERT will return 500 Internal Server Error
Co-authored-by: Dmitry Wagin <dmitry.wagin@ya.ru>
2021-11-25 22:56:49 +01:00
Wolfgang Walther 81f9bf290c ci: Fix permissions for report workflow 2021-11-24 23:01:33 +01:00
Wolfgang WaltherandWolfgang Walther dbef11869c ci: Add report workflow ahead of #1812 2021-11-24 21:23:57 +01:00
Wolfgang WaltherandWolfgang Walther b38564e623 nix(refactor): Make checking for required environment variables a no-op 2021-11-24 21:23:57 +01:00
Wolfgang WaltherandWolfgang Walther 2a2a4bdeac nix: remove unused ghr import 2021-11-24 21:23:57 +01:00
Wolfgang WaltherandWolfgang Walther a65ac9b0f6 feat: drop support for pg 9.5 2021-11-24 19:42:53 +01:00
Wolfgang WaltherandWolfgang Walther 7ed8b10776 test: Remove conditional for pgVersion95 in JsonOperatorSpec
pgVersion95 is the minimum already.
2021-11-24 19:42:53 +01:00
dependabot[bot]andWolfgang Walther 2f239b9048 build(deps): bump actions/cache from 2.1.6 to 2.1.7
Bumps [actions/cache](https://github.com/actions/cache) from 2.1.6 to 2.1.7.
- [Release notes](https://github.com/actions/cache/releases)
- [Commits](https://github.com/actions/cache/compare/v2.1.6...v2.1.7)

---
updated-dependencies:
- dependency-name: actions/cache
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
2021-11-24 18:55:38 +01:00
Wolfgang WaltherandWolfgang Walther c3854e7c74 ci: Add dependabot config for auto updates of github actions 2021-11-24 18:41:32 +01:00
Wolfgang WaltherandRemo b7dcc63e37 nix: Simplify with_tmp_db 2021-11-24 18:06:56 +01:00
Laurence IslaandGitHub 62243852b6 feat: Improve error message for ambiguous embedding
- Adds a relevant hint that includes unambiguous embedding suggestions.
- Joins origin and target into one single embedding key
2021-11-23 22:23:29 -05:00
steve-chavez 1b48531369 remove the db-embed-default-join config 2021-11-23 12:32:51 -05:00
Wolfgang WaltherandWolfgang Walther 1981160536 ci: Use include in stack build matrix 2021-11-21 21:53:24 +01:00
Wolfgang WaltherandWolfgang Walther f3eb8a317d ci: Disable fail-fast in stack matrix to keep other builds alive on failure 2021-11-21 21:53:24 +01:00
Wolfgang WaltherandWolfgang Walther 7aedebecee ci: Visually separate jobs in workflow file 2021-11-21 18:40:42 +01:00
Wolfgang WaltherandWolfgang Walther d5afdc7c53 ci: Use matrix strategy to build with stack 2021-11-21 18:40:42 +01:00
Wolfgang WaltherandWolfgang Walther b7c261f520 ci: Update actions 2021-11-21 18:40:42 +01:00
Wolfgang WaltherandWolfgang Walther 942391c508 ci: Add .github scripts to shellcheck 2021-11-21 18:40:42 +01:00
Wolfgang WaltherandWolfgang Walther f333305491 ci: Add setup-nix action to reuse nix-related code 2021-11-21 18:40:42 +01:00
Wolfgang WaltherandWolfgang Walther 88283801cb ci: Limit release job permissions to minimum required 2021-11-21 18:40:42 +01:00
Laurence IslaandGitHub 4e07846c5d fix: Add tests for embedding on overloaded functions with a default param
Resolves #1672

Was actually fixed in #1841 / 67c2ed7, this just adds the tests to confirm it.
2021-11-19 09:00:16 +01:00
steve-chavez bf58a3600a refactor: remove pgFmtLit 2021-11-18 17:01:58 -05:00
steve-chavez bae5afbad6 feat: allow unknown on the is operator
Also make the IS values strict at the Parser level since IS cannot be
parametrized.
2021-11-18 17:01:58 -05:00
Laurence IslaandGitHub 91689e28f0 Remove partitions from the schema cache
breaking: Embedding, OpenApi and other features related to the schema cache are no longer allowed on partitions.
2021-11-18 14:12:14 -05:00
Wolfgang WaltherandRemo 041f73cae8 nix(feat): Add actionlint to postgrest-lint 2021-11-17 18:58:34 +01:00
monacoremo 43c8bde8ce refactor: Add doctests 2021-11-13 04:07:46 +01:00
monacoremo 82dd1b732d ci: disable cirrusci stack work cache which seems to create issues 2021-11-12 20:25:22 +01:00
monacoremo 41ac0d5d93 refactor: Remove Protolude.Conv from Middleware 2021-11-12 20:25:22 +01:00
monacoremo 60db026dcc refactor: Remove Protolude.Conv from RangeQuery 2021-11-12 20:25:22 +01:00
monacoremo 47b6f04a66 refactor: Remove Protolude.Conv from AppState 2021-11-12 20:25:22 +01:00
monacoremo f228cccbae refactor: Remove Protolude.Conv from Proxy 2021-11-12 20:25:22 +01:00
monacoremo 3820590649 refactor: Remove Protolude.Conv from SqlFragment 2021-11-12 20:25:22 +01:00
monacoremo 8cc8531e49 refactor: Remove Protolude.Conv from Workers 2021-11-12 20:25:22 +01:00
monacoremo 65062364c4 refactor: Remove Protolude.Conv from Error 2021-11-12 20:25:22 +01:00
monacoremo ebe66cb985 refactor: Remove Protolude.Conv from Statements 2021-11-12 20:25:22 +01:00
monacoremo 9c172db534 refactor: Remove Protolude.Conv from CLI 2021-11-12 20:25:22 +01:00
monacoremo de531b6b41 refactor: Remove Protolude.Conv from OpenAPIw 2021-11-12 20:25:22 +01:00
monacoremo 71cb8afc0d refactor: Remove Protolude.Conv from JSPath 2021-11-12 20:25:22 +01:00
monacoremo 36dbd09459 refactor: Remove Protolude.Conv from ApiRequest 2021-11-12 20:25:22 +01:00
monacoremo fa4df95f55 refactor: Remove Protolude.Conv from Parsers 2021-11-12 20:25:22 +01:00
monacoremo 4104f5fb75 refactor: Remove Protolude.Conv from App 2021-11-12 20:25:22 +01:00
monacoremo abb6167caa refactor: Remove Protolude.Conv from DbStructure 2021-11-12 20:25:22 +01:00
monacoremo 15ea7dfe09 refactor: Remove Protolude.Conv from GucHeader 2021-11-12 20:25:22 +01:00
monacoremo a2c2544271 refactor: Remove Protolude.Conv from Config 2021-11-12 20:25:22 +01:00
monacoremo 2b22f889bd refactor: Remove GHC.Show instances from JSPath module 2021-11-09 19:13:52 +01:00
monacoremo cd3013569e refactor: Remove GHC.Show instances from QualifiedIdentifier and Config 2021-11-09 19:13:52 +01:00
monacoremo 5dc37fc8e8 refactor: Remove GHC.Show instances from Request.Types module 2021-11-09 19:13:52 +01:00
monacoremo 038d84b62d refactor: Remove GHC.Show instances in Preferences module 2021-11-09 19:13:52 +01:00
Laurence IslaandGitHub 40f9a6068a Allow overloaded functions if one has a single unnamed JSON param
Avoids the breaking change in #1927: If there's a function "my_func" having a single unnamed json param and other overloaded pairs(with any number of params), PostgREST won't be able to resolve a POST request to "my_func".
2021-11-08 18:26:33 -05:00
monacoremo 5cd7d35966 refactor: Make import aliases consistent across the codebase 2021-11-05 08:08:00 +01:00
monacoremo 71262fbc5c nix: Hook hsie check-aliases into postgrest-lint 2021-11-05 08:08:00 +01:00
monacoremo d16a4a9a2b nix: fix statix hints 2021-11-02 23:10:23 +01:00
monacoremo 2357961f48 nix: add statix to postgrest-style 2021-11-02 23:10:23 +01:00
monacoremo 162c4e38af Update stackage to LTS 18.14 (GHC 8.10.7) 2021-11-01 23:23:19 +01:00
monacoremo 0f95900822 nix: Upgrade hsie tool to GHC 8.10.7 2021-11-01 23:23:19 +01:00
monacoremo 8ac14447ef nix: Update nixpkgs to latest unstable version 2021-11-01 23:23:19 +01:00
Wolfgang WaltherandWolfgang Walther a1b3bca7fc fix: Make pfkSourceColumns query work with standard_conforming_strings = off 2021-10-28 08:21:11 +02:00
Wolfgang WaltherandWolfgang Walther 8cde90b7fe refactor: Use include_triggers argument to pg_relation_is_updatable instead of manual check 2021-10-26 19:15:02 +02:00
3b83f536ca Correct inner join to use proper aliases (#1978)
"_" was used for all the internal json_agg aliases. Now an alias
with "_" plus a local table name is used.

Co-authored-by: Cole Arendt <cole.arendt@outlook.com>
2021-10-18 17:34:31 -05:00
laurenceislaandGitHub 42d7cf8ae5 Add db-embed-default-join config parameter to the configuration example 2021-10-15 19:09:40 -05:00
laurenceislaandGitHub 627c3c34b7 feat: Make GUC names for headers, cookies and jwt claims compatible with PostgreSQL v14
Getting the value for a header GUC on PostgreSQL v14 is done using `current_setting('request.headers')::json->>'name-of-header'` and in a similar way for `request.cookies` and `request.jwt.claims`

PostgreSQL versions below 14 can opt in to the new JSON GUCs by setting the `db-use-legacy-gucs` config option to false (true by default)
2021-10-12 12:58:52 -05:00
steve-chavez ee56dd5db1 feat: filter top-level resource with embed filter
This is enabled by adding `!inner` to the embedded resource

/projects?select=*,clients!inner(*)&clients.id=eq.12

This behaviour can be enabled by default with the config option

db-embed-default-join='inner'

Which saves the need for specifying `!inner` on every request.
If this is enabled, the previous behavior can be restored
per request by specifying `!left`  on the embedded resource.

/projects?select=*,clients!left(*)&clients.id=eq.12`

Tested on M20/02M/M2M relationships, views, RPC.
2021-10-04 13:46:32 -05:00
steve-chavez bf91187e63 Drop support for embed hints used as '.'
Hints used like `select=projects.client_id(*)` were already
deprecated.

'!' should be used from now on `select=projects!client_id(*)`
2021-10-04 13:46:32 -05:00
monacoremo 2982c5de2e nix: Fix postgrest-check 2021-09-23 22:36:58 +02:00
monacoremo add95c32ca ci: Fix triggering events and release actions
* Run CI only on pull requests, push to main and push
  to tags matching v*. This avoids running the CI more than
  once for a single user action
* Change pre-releases, which are now identified by a version
  number with 4 components (e.g., 1.1.1.1)
* Change how release artifacts are packaged, ensuring they
  are executable and compressed with xz
2021-09-23 22:36:58 +02:00
laurenceislaandGitHub 243e4f0408 test: correct limit proc results tests (#1951)
Query params where being appended using `?` instead of `&`
2021-09-16 22:03:35 -05:00
steve-chavez 885ff60138 pre-release: Bump to v9.0.0-a1 2021-09-14 15:51:32 -05:00
steve-chavez 72a28a145d feat: Allow backslash escaping inside double quotes
Examples:

- ?col=in.("Double\"Quote")
- ?col=in.("Back\\slash")

Backslashed chars get passed as is

- ?col=in.("\a\b\c") = ?col=in.(abc)
2021-09-13 19:06:30 -05:00
steve-chavez 4a594d5e95 fix: using " and \ as values on the in filter
Properly escape the array literal built at the QueryBuilder
stage.
2021-09-13 19:06:30 -05:00
Remo 0ba133a0bd ci: Fix FreeBSD build 2021-09-08 19:11:00 +02:00
Wolfgang WaltherandSteve Chavez b05898d17f fix: Return array of objects for RPCs with single-column TABLE return type
This is a regression after changing responses for RPCs with INOUT or OUT arguments in #1615.

Resolves #1930.
2021-09-02 10:13:03 -05:00
steve-chavez d4c6abbaec feat: RPC POST for function w/single unnamed param
For POST on RPC, allows:

* passing a json object without using `Prefer: params=single-object`
  The function must be defined with a single unnamed json param and
  `Content-Type: application/json` must be specified.

* uploading binary to a function
  The function must be defined with a single unnamed bytea param and
  `Content-Type: application/octet-stream` must be specified.

* uploading raw text to a function
  The function must be defined with a single unnamed text param and
  `Content-Type: text/plain` must be specified.

BREAKING CHANGE If there's a function "my_func" having a single
unnamed json param and other overloaded pairs(with any number of
params), PostgREST won't be able to resolve a POST request to
"my_func". For solving this, you can name the unnamed json param.

my_func(json) -> my_func(prm json)
2021-08-30 18:17:59 -05:00
steve-chavez caaa34b5de refactor: add CallQuery type
Shortens requestToCallProcQuery, which should generate SQL in a more
direct way.
2021-08-30 18:17:59 -05:00
steve-chavez c9a60373f6 refactor: PgArg to ProcParam
Clarify the difference between arguments and parameters.
Parameters are part of the function definition, arguments are the values
passed to the function.

Also clarify the findProc function comments and error message.
2021-08-30 18:17:59 -05:00
laurenceislaandGitHub ed5072f4b1 fix: Detect default values for String types in OpenAPI output (#1928) 2021-08-27 21:59:54 -05:00
Felix YanandRemo ed98ac9d7e Allow newer base, hspec, and retry
Builds fine and all tests are passing here, with the new versions.
2021-08-19 18:44:30 +02:00
Gaurav GautamandGitHub d9f1ae6a46 feat: Add Retry-After hint (#1916)
Add Retry-After header when response status is 503.
Its value is the connection worker delay(seconds) when
it's recovering.

This closes issue #1817.
2021-08-18 10:14:29 -05:00
Wolfgang WaltherandRemo 5874482f03 nix(fix): Database cluster not cleaned up before artifact upload 2021-08-15 12:35:52 +02:00
Wolfgang WaltherandRemo 0511e99dd4 nix(refactor): Add withPath to checkedShellScript 2021-08-15 12:35:52 +02:00
laurenceislaandGitHub 4e4548d702 feat: Include partitioned tables into the schema cache
Allows embedding, UPSERT, INSERT with Location response, OPTIONS request and OpenAPI support for partitioned tables
2021-08-13 15:06:56 -05:00
Remo b3899e7aa4 ci: Set Docker Hub username 2021-08-08 06:42:59 +02:00
monacoremo 6d9fcfa09f ci: Remove unused create_test_db script 2021-08-08 06:42:59 +02:00
monacoremo 38e5b76996 ci: Remove appveyor.yml 2021-08-08 06:42:59 +02:00
monacoremo ca4539e78d ci: Remove .travis.yml 2021-08-08 06:42:59 +02:00
monacoremo a54a61d00e ci: Remove .circleci/config.yml 2021-08-08 06:42:59 +02:00
monacoremo 142fc06b15 ci: Add .github/release script that tags and pushes a new release 2021-08-08 06:42:59 +02:00
monacoremo c7b1cda3ed ci: Move codecov.yml and add thresholds as intermittent fix to random coverage changes 2021-08-08 06:42:59 +02:00
monacoremo fb2adea5e9 ci: Adjust build status badge in READMEs 2021-08-08 06:42:59 +02:00
monacoremo d9b15adb35 ci: Use GitHub Actions for CI/CD 2021-08-08 06:42:59 +02:00
steve-chavez 0bc7c034cf Bump to version v8.0.0 2021-07-25 23:12:51 -05:00
steve-chavez cbe0e5254e Reorganize the CHANGELOG
* Removed some entries that were fixes for the nigthly versions, not
stable versions.
* Reference the actual issue number(not the PR number) on some entries
* Remove duplicate entries
* Clarify some entries
* Grouped CLI entries
2021-07-25 23:12:51 -05:00
Remo RechkemmerandGitHub 68bf17118d Upgrade Nix dependencies including nixpkgs and static-haskell-nix (#1899) 2021-07-23 19:54:29 +02:00
Remo RechkemmerandGitHub fdf6c510ec Update changelog (#1897) 2021-07-17 00:14:37 +02:00
laurenceislaandGitHub db95bd1c37 feat: Add compatibility with connection poolers on transaction mode
Update the hasql-transaction library to version 1.0.1

Add hints and kill thread at configuration read when using incompatible pooling modes: statement pooling and transaction pooling with prepared statements enabled.
2021-07-15 13:39:55 -05:00
monacoremo 83cc358e1b nix: Update static-haskell-nix and fix stripped executable 2021-07-15 09:41:52 +02:00
monacoremo 11374c2fae ci: upgrade stackage to LTS 18.2 / GHC 8.10.4 2021-07-15 09:41:52 +02:00
monacoremo 27c8dabd8d nix: Use GHC 8.10.4 in Nix build environment 2021-07-15 09:41:52 +02:00
monacoremo 9568c4605a nix: Update nixpkgs and static-haskell-nix 2021-07-15 09:41:52 +02:00
Laurence Isla 6fe88ef53c Add Oblivious as a sponsor and a Sponsors section to the docker hub description 2021-07-12 19:30:35 -05:00
steve-chavez 0961a587c0 Correct openapi ignore-privs table comments 2021-07-06 11:19:54 -05:00
laurenceislaandGitHub 67c2ed7c62 feat: Show comprehensive error when an RPC is not found in a stale schema cache (#1841)
* Add 300 response for overloaded functions with same argument names but different types
* Add Path type to handle validations and errors before defining the Target
2021-07-06 09:19:26 -05:00
steve-chavez f3a184af01 Correct openapi ignore mode to filter schemas
Also rename openapi-mode options

* follow-acl -> follow-privileges
* ignore-acl -> ignore-privileges
2021-06-27 22:48:49 -05:00
Steve ChavezandGitHub 41d119b19f feat: add openapi-mode config (#1881)
* openapi-mode="follow-acl"(default): follows access control for the
JWT role.
* openapi-mode="ignore-acl": ignores access control for the JWT role.
* openapi-mode="disabled": disables OpenAPI output, the root endpoint
replies with 404 Not Found.
2021-06-25 19:20:04 -05:00
laurenceislaandGitHub 214a92f207 Change the default db-channel-enabled config value from False to True (#1880) 2021-06-25 11:12:28 -05:00
steve-chavez 1759c1c75b Update CirrusCI encrypted var 2021-06-22 20:10:43 -05:00
Steve ChavezandGitHub 2944cf94c7 feat: Add time to startup/worker logs (#1872)
BREAKING CHANGE

Sends startup/worker logs to stderr to differentiate them from
access logs, which go to stdout
2021-06-19 15:55:18 -05:00
laurenceislaandGitHub 59e0b006a5 fix: Location header with a null PK on a view (#1875)
Regression after #1475
2021-06-18 18:09:19 -05:00
laurenceislaandGitHub 1b12b112a1 fix: Add a hint and clarification to the no relationship found error (#1858) 2021-05-31 12:15:31 -05:00
Alexander LjungbergandGitHub 88a481da8a fix: drop gzip support, improve performance. (#1854)
BREAKING CHANGE

Not gzipping the output more than doubles the throughput for a simple "do nothing" RPC function.

Postgrest's gzip support was surprisingly slow, lacked in configurability, was undocumented and untested.

It was possible to achieve this before by not sending `Accept-Encoding: gzip`, but that turned into a bit of an undocumented gotcha. Most browsers do send it and by default a proxy pass would forward this header. Better to not support gzip at all, further focusing on ensuring Postgrest is a minimal layer over Postgres.

Nginx, or whatever proxy, provides critical options such as min length, vary headers and so forth, and will likely have much better performance as well.
2021-05-30 17:33:44 -05:00
steve-chavez d99909c403 feat: request.spec GUC for db-root-spec
The request.spec GUC contains the schema cache structure in json.

It's only available when the root endpoint(/) is requested and when
db-root-spec is not empty.

Also correct db-root-spec to accept a schema.
2021-05-30 16:48:08 -05:00
steve-chavez f169661ce6 refactor: move DbStructure.PgVersion to Config
Now that PgVersion is not part of DbStructure, Config is a more apt
module for it.

Also rename getDbStructure to queryDbStructure. AppState also had a
getDbStructure function for a record field.
2021-05-30 16:48:08 -05:00
steve-chavez 823348a72a refactor: configDbPreRequest to QualifiedIdentifier 2021-05-30 16:48:08 -05:00
steve-chavez 5c75f0dcc2 Update BACKERS.md
Correct wrong link
2021-05-16 14:32:51 -05:00
steve-chavez 601c802f23 Update BACKERS.md 2021-05-16 14:23:10 -05:00
Laurence IslaandWolfgang Walther b4ca70708a fix: Fix requests for overloaded functions from html forms to no longer hang
Fixes #1846
2021-05-13 10:07:49 +02:00
steve-chavez 082c91c855 refactor: remove pgVersion from DbStructure
Also use a dedicated MVar for listener
2021-04-30 10:06:07 -05:00
steve-chavez 0f6a13191c refactor: remove colFK from Column 2021-04-30 10:06:07 -05:00
Remo RechkemmerandGitHub 6670a3214b refactor: Move accept content type code from App into ApiRequest (#1830) 2021-04-25 21:58:03 +02:00
monacoremo 63e0292e23 fix: Panic when attempting to run with unix socket on non-unix host 2021-04-24 19:42:58 +02:00
monacoremo 59f0fe5b32 fix: Properly close unix domain socket on exit 2021-04-24 19:42:58 +02:00
monacoremo 387c387073 fix: Improve log message on config load 2021-04-24 19:42:58 +02:00
monacoremo acd787a5af refactor: Split main/Main.hs into library modules 2021-04-24 19:42:58 +02:00
Wolfgang WaltherandWolfgang Walther 4ded01b104 nix(fix): Fix missed devtools->devTools renaming 2021-04-24 15:25:10 +02:00
Wolfgang WaltherandWolfgang Walther c691b37f76 nix(feat): Create all temporary directories with script name template
Temporary directories are now created in $TMPDIR/postgrest/script-name-XXX. This allows recognizing the origin of a single temporary directory and uploading the whole $TMPDIR/postgrest folder as an artifact in CI.
2021-04-24 15:07:13 +02:00
Wolfgang WaltherandWolfgang Walther 14ddecc81a nix(fix): Avoid rebuilding during postgrest-coverage 2021-04-24 15:07:13 +02:00
Wolfgang WaltherandWolfgang Walther 3611eab39c nix(fix): Make postgrest-dump-schema work in nix-shell --pure 2021-04-24 15:07:13 +02:00
Wolfgang WaltherandWolfgang Walther e5350d2d50 nix(refactor): Split cabalTools from devtools
This allows to use `postgrest-build` in CI without pulling in a lot of dependencies.
2021-04-24 15:07:13 +02:00
Wolfgang WaltherandWolfgang Walther 249d4117e0 nix(style): reorder tools in default.nix 2021-04-24 15:07:13 +02:00
Wolfgang WaltherandWolfgang Walther ba47a54293 nix(refactor): Make with_tmp_db true checkedShellScript
This extends the interface of withTmpDb to allow loading different sets
of database fixtures via `postgrest-with-postgresql-xx --fixtures
<path>`.
2021-04-24 15:07:13 +02:00
laurenceislaandGitHub f6b3a5cc22 fix: Make OPTIONS consider view instead of triggers (#1824) 2021-04-23 18:12:43 -05:00
steve-chavez 6e0ef95320 Add GNUHost as a sponsor 2021-04-23 14:29:33 -05:00
steve-chavez 082da78f4c fix: disregard internal junction when embedding
Added test cases for a lone internal junction and an internal junction
exposed homonym
2021-04-21 10:28:20 -05:00
steve-chavez 21d280497b refactor: Relation to Relationship
Also remove the unused UnknownRelationship error.
2021-04-21 10:28:20 -05:00
Wolfgang WaltherandWolfgang Walther 7ca0d46936 nix(fix): Fix postgrest-push-cachix in nix-shell --pure 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther b566481477 nix(refactor): Move shell scripts to subdirectory nix/tools 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther fe497fc438 nix(refactor): Add buildToolbox to streamline module interface towards shell.nix 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther de73ced34a nix(feat): Change memoryTests shell argument to memory
Avoids confusion with `memoryTest`.
2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther bc879507ae nix(fix): Make naming of tools in postgrest-release- namespace consistent 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 280b89d87c lint: Apply shellcheck suggestions for bash scripts in test/* 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther eee5d5d482 nix(feat): Add shellcheck to postgrest-lint for bash scripts in test/* 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 2e73886adc nix(refactor): Split memoryTest from tests.nix into memory.nix 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther e199337a53 nix(feat): Check required environment variables in postgrest-release tools 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 2417cfb02f nix(feat): Force optional arguments to go before positional arguments in argbash scripts
This improves passing subcommands to postgrest-with- tools.
2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 2559d32912 nix(feat): Add argbash bash completion support 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther c2df7d8198 nix(fix): Avoid losing exit trap through subcommand in postgrest-with-all 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther c8edfac39e nix(feat): Add argbash support to checked-shell-script
Each postgrest- script now has a -h/--help option showing the
checked-shell-script.docs argument.

Additional CLI arguments can be defined through the
checked-shell-script.args argument using the argbash template syntax:
https://argbash.readthedocs.io/en/stable/guide.html
2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 39d4646a4c nix(refactor): Add withEnv to checked-shell-script 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 40b1dccc54 nix(fix): Load withTmpDb into nix to have shellcheck run on it 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther 32bac81d91 nix(refactor): Move withPostgresqlVersions and withAllVersions to new withTools 2021-04-18 13:51:39 +02:00
Wolfgang WaltherandWolfgang Walther dccce946e3 nix(refactor): Move name arguments into scripts and use global postgresql in tests.nix 2021-04-18 13:51:39 +02:00
Remo RechkemmerandGitHub 915d568257 ci: Use cachix authtoken in CI (#1821) 2021-04-17 16:25:28 +02:00
Steve ChavezandGitHub 698bfe6e7b refactor: make Junction non-recursive (#1818)
Also Cardinality now includes a Junction or ConstraintName
2021-04-16 13:12:58 -05:00
Wolfgang WaltherandWolfgang Walther 1f206a560b ci: check sha512 of downloaded codecov script 2021-04-16 14:10:49 +02:00
Wolfgang WaltherandWolfgang Walther 42f8f4fdcb ci: remove stack-test-memory 2021-04-16 14:10:49 +02:00
Wolfgang WaltherandWolfgang Walther c67cd5e6fc ci: reorder tests in nix-test to improve performance and give early coverage feedback 2021-04-16 14:10:49 +02:00
Wolfgang WaltherandWolfgang Walther 67e3886547 ci: split nix-build-test into nix-build nix-test and introduce circleci contexts 2021-04-16 14:10:49 +02:00
laurenceislaandGitHub 4b46c4eff3 feat: Add explicit headers-only POST request using Prefer header
Allows using header "Prefer=headers-only" explicitly to get a response with only a Location header.

BREAKING CHANGE: Change default for POST request from headers-only to minimal

Resolves #1656
2021-04-14 00:14:12 +02:00
Wolfgang WaltherandWolfgang Walther cd09cc6c52 ci: fix cirrus build timeout 2021-04-14 00:13:08 +02:00
Remo RechkemmerandGitHub c2c7bbe9dd refactor: remove obsoleted Private module (#1811) 2021-04-12 20:47:22 +02:00
Wolfgang WaltherandWolfgang Walther d40d704043 remove unused static assets 2021-04-12 20:31:32 +02:00
Wolfgang WaltherandWolfgang Walther 91d0c732f0 add comment to Setup.hs 2021-04-12 20:31:32 +02:00
Remo RechkemmerandGitHub 801e229c59 refactor: Carve CLI and Version modules out of Config (#1806) 2021-04-12 19:17:47 +02:00
Wolfgang WaltherandWolfgang Walther 67ca814d0b nix: properly throw away tix files for postgrest-watch, resolves #1807 2021-04-12 19:03:16 +02:00
Wolfgang WaltherandWolfgang Walther fc791a6320 nix: add withTmpDir to checked-shell-script 2021-04-12 19:03:16 +02:00
579a626de2 nix: move withtmpdb to separate file
Co-authored-by: Remo Rechkemmer <59358383+monacoremo@users.noreply.github.com>
2021-04-12 19:03:16 +02:00
Remo RechkemmerandGitHub f99fd6cbad refactor: Split up Types.hs and logically organize modules (#1793) 2021-04-11 18:28:01 +02:00
Remo RechkemmerandGitHub 8c44410ce0 remove misleadingly named emptyOnFalse and similar (#1803) 2021-04-11 10:38:35 +02:00
Wolfgang WaltherandWolfgang Walther 9118a4a780 fix random codecov/project drops in CI, resolves #1800
Those are caused by running the spec tests with `-threaded`, which results in a race to write to spec.tix.
2021-04-10 20:38:51 +02:00
Wolfgang WaltherandGitHub a5372e4713 nix: show better error messages for missing arguments in hsie scripts (#1798) 2021-04-09 18:48:00 +02:00
Jérémy PagéandGitHub 67f555d24e nix: Fix incorrect use of bash variable (#1797) 2021-04-08 14:33:18 +02:00
7f11c1a991 fix: Void functions return null instead of empty body (#1795)
* Change empty body reponse to null for scalar functions returning void or null values

* Add test for functions returning an integer with null value

Co-authored-by: Laurence Isla <lau.isla.c@gmail.com>
2021-04-08 01:45:13 +02:00
steve-chavez 9c79a4174c Add Contributor Covenant Code of Conduct v2 2021-04-06 22:22:13 -05:00
Wolfgang WaltherandWolfgang Walther 376beac22f update contributing guidelines 2021-04-06 17:07:14 +02:00
Remo RechkemmerandGitHub 65e7f9e846 nix: Add a tool for checking Haskell imports and exports (#1768) 2021-04-06 11:46:12 +02:00
Boris KorzunandSteve Chavez 9f074cecce cirrusci: update FreeBSD pkg 2021-03-26 10:13:38 -05:00
steve-chavez 71f6061d30 Add Supabase as a sponsor 2021-03-25 20:42:43 -05:00
Robert VollmertandSteve Chavez 3a466aea9a tests: don't match JSON bodies literally
The change of hash function with hashable-1.3.1.0 changes object
ordering in JSON output, causing some test failures:
https://app.circleci.com/pipelines/github/PostgREST/postgrest/805/workflows/067844c9-9ce4-49e8-8790-315625ab309b/jobs/8309
e.g.:

> expected: "[{\"b\":\"baz\",\"a\":\"bar\"}]"
>  but got: "[{\"a\":\"bar\",\"b\":\"baz\"}]"

This changes the tests to not compare against the body literally,
and instead use the ResponseMatcher instance from Test.Hspec.Wai.JSON.
2021-03-14 12:45:02 -05:00
Robert VollmertandSteve Chavez 5baec4819f circleci: update base image
Switches to the non-legacy cimg/base: https://circleci.com/developer/images/image/cimg/base

In particular, this updates the gcc version to fix the build of
cryptonite-0.27: https://github.com/haskell-crypto/cryptonite/issues/326
2021-03-14 12:45:02 -05:00
steve-chavez d3a8b5f6e1 Change db-load-guc-config to db-config
Make clear that in-db config is being read
2021-03-05 19:03:15 -05:00
steve-chavez 498e77215a Reread in-db config when recoverying connection
* Separate reading files from whole config re-read
* Only reload external file on SIGUSR2/NOTIFY
2021-03-05 19:03:15 -05:00
steve-chavez 6750a5c44d Restrict db settings to current db and global
Global settings are overriden by database specific settings if they
share common ones.
2021-03-05 19:03:15 -05:00
steve-chavez e4516ab606 Correct db settings to use "_" instead of "-"
GUC settings with dashes cannot be shown with show or current_setting.

https://www.postgresql.org/message-id/flat/20210209144059.GA21360%40depesz.com
2021-03-05 19:03:15 -05:00
Robert VollmertandSteve Chavez c3ccaf1a08 Allow lens-5.0 2021-02-28 13:02:37 -05:00
Robert VollmertandSteve Chavez a7dab6d95a cirrusci: update to FreeBSD 12.2
This fixes CI failure because FreeBSD 12.1 became EOL in Jan 2021
(3 months after release of 12.2): https://www.freebsd.org/security/#sup
2021-02-27 13:26:02 -05:00
steve-chavez b50b67491d Update CYBERTEC logo and BACKERS 2021-02-25 20:03:38 -05:00
Remo RechkemmerandGitHub e6973f966b refactor: App.hs and related changes (#1725)
* Use ExceptT to avoid 'staircasing' case analysis in App.hs
* Split large function in App.hs into individual handler functions
* Adapt API of Auth.hs, OpenApi.hs etc. to simplify the use of those modules in App.hs
* Split optional rollback functionality into Middleware
* Unify SimpleError and ApiRequestError into one Error type, so it can be used across modules
2021-02-23 22:41:48 +01:00
Remo RechkemmerandGitHub 0ddd676ef0 nix: Add a script to generate a graph of Haskell imports (#1751) 2021-02-07 22:15:44 +01:00
Steve ChavezandGitHub c93e8f9e0c Correct hardcoded postgrest_test_authenticator (#1743)
Also remove panic from Config and correct io test not running
2021-01-25 18:38:46 -05:00
steve-chavez 6557f1f9c0 correct NOTIFY config reload dying on error 2021-01-22 15:56:08 -05:00
steve-chavez 17af56adb1 refactor: config validation inside readAppConfig
Remove Either from configJwtRoleClaimKey/configServerUnixSocketMode
and remove whenLefts.
2021-01-22 15:56:08 -05:00
steve-chavez 4344cc9202 refactor: use optString and move overrideFrom 2021-01-22 15:56:08 -05:00
steve-chavez 125ea8f6d9 feat: allow reloading config with NOTIFY
Enables reloading the config by doing:
NOTIFY pgrst, 'reload config'

Adds an alias for reloading the schema cache:
NOTIFY pgrst, 'reload schema'
2021-01-19 13:49:40 -05:00
steve-chavez 9c005fc683 feat: get configuration parameters from the db
Allows configuring postgrest from the db by setting config parameters
on the connection role. For example:

ALTER ROLE postgrest_test_authenticator
SET pgrst.jwt-secret = "REALLYREALLYREALLYREALLYVERYSAFE"

The above wWill set the `jwt-secret` config option accordingly.

SUPERUSER privileges are required for ALTERing role settings,
so this might not work on some cloud-managed databases.

This feature is enabled by default, for disabling it you can add the
following to the config file:

db-load-guc-config = false
2021-01-19 13:49:40 -05:00
steve-chavez 674615041a Correct --dump-schema swallowing error
When there's no connection to pg, the following error happens
when running --dump-schema:

postgrest: user error (Pattern match failure in do expression at
main/Main.hs:361:5-14)

Now it shows a regular "could not connect to server.." error.
2021-01-19 13:49:40 -05:00
steve-chavez 15039553db Correct openapi preparing statements by default 2021-01-19 13:49:40 -05:00
Wolfgang WaltherandWolfgang Walther ba86b479f8 cov: refactor ApiRequest profile header detection 2021-01-14 16:40:29 +01:00
Wolfgang WaltherandWolfgang Walther 6dd126461e cov: Remove unused code 2021-01-14 16:40:29 +01:00
Wolfgang WaltherandWolfgang Walther 97d8456382 cov: Add io tests for basic cli commands and invalid config options
Full code coverage for Config.hs except expected overlays.
2021-01-14 16:40:29 +01:00
Wolfgang WaltherandRemo Rechkemmer ab6f90aa78 nix: Add postgrest-with-* tools to run with temporary databases of different versions 2021-01-04 23:14:44 +01:00
Wolfgang WaltherandWolfgang Walther 522308217a refactor: Combine relConstraint and relJunction in relLink 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther d173b7d8d6 cov: remove unused fields from type Column 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther b8e6450af1 refactor: simplify addXRels 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther 45ebaf0ad8 cov: Add comment on SCHEMA in test fixtures 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther 1edbef0538 Add STABLE RPCs in test fixtures to increase coverage 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther 6efa304142 refactor: Move parseArg to SQL 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther eed6015e70 refactor: Move addHasVariadic to SQL 2021-01-03 17:54:28 +01:00
Wolfgang WaltherandWolfgang Walther 94b516d074 Add flag to postgrest.cabal to disable collecting coverage data 2021-01-02 13:51:00 +01:00
Wolfgang WaltherandWolfgang Walther 56a7bc5030 nix: Fix build errors with stale .tix files present 2021-01-02 13:51:00 +01:00
Wolfgang WaltherandWolfgang Walther af2788468e Add basic error handling to --dump-schema 2021-01-02 13:50:19 +01:00
Wolfgang Walther cc54135bf0 Rename master branch to main; replace whiteList with isAllowed 2020-12-31 14:37:52 +01:00
Wolfgang WaltherandGitHub 519550e9af Add codecov to README.md 2020-12-30 17:52:15 +01:00
Wolfgang WaltherandRemo Rechkemmer 93541035f6 Fix postgrest-coverage throwing error when installed with nix-env 2020-12-30 16:38:13 +01:00
Wolfgang WaltherandRemo Rechkemmer 568477b6af Fix io tests not collecting coverage data when using run()
Before, postgrest was killed after each test and no coverage data was saved.
Now the process is terminated gracefully via SIGTERM first to allow writing the .tix file.
2020-12-30 14:51:58 +01:00
Wolfgang WaltherandGitHub c7f0d42323 Add postgrest-coverage to show and upload hpc reports to codecov 2020-12-30 13:22:25 +01:00
Wolfgang WaltherandGitHub 2015688312 Fix spec test concurrency issue related to tx=commit on simple_pk 2020-12-29 23:58:37 +01:00
Wolfgang WaltherandWolfgang Walther 8e02551956 Fix "unused-imports" warning for windows build 2020-12-29 16:07:07 +01:00
Remo RechkemmerandGitHub 96478ed016 Add docs to all checked shell scripts in nix 2020-12-28 13:23:12 +01:00
Remo RechkemmerandGitHub 2425cddaed improve io-tests
Tests are now run through unix socket by default allowing
parallel execution. Test setup configuration is done via
environment variables instead of config files.
2020-12-24 16:09:25 +01:00
Wolfgang WaltherandWolfgang Walther b5185de706 nix: add "cd $rootdir" to all shell scripts via checked-shell-script 2020-12-24 14:44:01 +01:00
Wolfgang WaltherandWolfgang Walther 56557c9c40 ci: remove broken travis coverage job 2020-12-23 19:39:40 +01:00
Wolfgang WaltherandWolfgang Walther 2cbe1ba903 feat: Add --example cli option to show example config file 2020-12-23 19:39:40 +01:00
Wolfgang WaltherandWolfgang Walther 69b459e7b6 Use environment variables for configuration in developer tooling
Avoid the need for configuration files for temporary database connections.
2020-12-23 19:39:40 +01:00
Wolfgang WaltherandWolfgang Walther add326a79d ci: skip postgrest-release-dockerhubdescription for nightly builds 2020-12-23 19:39:40 +01:00
Wolfgang WaltherandWolfgang Walther b7fc393e49 feat: Read config directly from environment variables
resolves #1624
2020-12-23 19:39:40 +01:00
monacoremoandSteve Chavez f2f639e484 update stack to lts-16.26 2020-12-23 09:39:36 -05:00
monacoremoandSteve Chavez 9cfc66a6a4 upgrade nixpkgs and apply related fixes 2020-12-23 09:39:36 -05:00
Remo RechkemmerandGitHub bf141ca13f feat: Added --dump-schema CLI option to dump JSON of dbStructure schema cache.
Added postgrest-dump-schema to nix-shell returning a YAML dump of the test fixtures.

Authored-by: monacoremo <monacoremo>
2020-12-20 10:22:52 +01:00
monacoremoandSteve Chavez fe09637711 catch ReadTimeout in IO tests 2020-12-17 17:05:03 -05:00
Wolfgang WaltherandWolfgang Walther dbe3c163bd run spec tests in parallel
reduces time to run postgrest-test-spec-all by 20-25%
2020-12-14 00:42:03 +01:00
monacoremoandWolfgang Walther a52c114ff4 add test for spec idempotence 2020-12-13 22:13:05 +01:00
Remo RechkemmerandGitHub 71fa87937b Migrate io-tests from bash to pytest
Authored-by: monacoremo <monacoremo>
2020-12-13 16:19:26 +01:00
Wolfgang WaltherandWolfgang Walther 0c25f12825 fix: fix output of RPCs returning scalar values with multiple rows
resolves #1584

BREAKING CHANGE:
Changed output of RPCs to match return type better:
* single scalar return: value (unchanged)
* setof scalar return: array of values (new; was array of objects)
* single composite return: object (new; was array of objects)
* setof composite return: array of objects (unchanged)
* void return: nothing (new; was "null")

A single OUT column is now treated as "composite" instead of "scalar",
i.e. consistent with multiple OUT columns.
2020-12-10 19:55:42 +01:00
Wolfgang Walther 093fd3c8f6 fix: fix embedding through views with subqueries inside function calls
resolves #1608

Refactors the pfkSourceColumns query to use an intermediate JSON format
for parsing. This allows much more robust extraction of source columns.
2020-12-10 19:48:05 +01:00
Steve ChavezandGitHub ebd474a7e6 fix: retry connection on failed schema cache load (#1685)
Retry the connection when the
"Failed to load the schema cache" error happens.

Also die if the schema cache query has a syntax error.
2020-12-09 14:20:00 -05:00
steve-chavez 11d62a8010 Prepared statement for set_config 2020-12-07 20:09:18 -05:00
steve-chavez 7069bb3c01 refactor: Change SET LOCAL gucs to set_config 2020-12-07 20:09:18 -05:00
Wolfgang WaltherandWolfgang Walther 9254f119f6 fix: implement robust parsing of boolean config values
resolves #1572
2020-12-06 21:59:32 +01:00
Wolfgang WaltherandWolfgang Walther ed58511de3 feat: renamed config options with prefixes; added aliases for old names
* secret-is-base64 -> jwt-secret-is-base64
* role-claim-key -> jwt-role-claim-key
* max-rows -> db-max-rows
* pre-request -> db-pre-request
* root-spec -> db-root-spec
* db-schema -> db-schemas

This is not a breaking change, because aliases are added as well.

refactor: sorted all config keys alphabetically where applicable
2020-12-06 21:59:03 +01:00
Wolfgang Walther ab3375998d ci: fix cirrus low memory error 2020-12-06 20:49:48 +01:00
Wolfgang WaltherandWolfgang Walther 8979442e27 dev: add colors to io tests 2020-12-06 13:33:25 +01:00
Wolfgang WaltherandWolfgang Walther eb46cf2662 added: cli option --dump-config prints loaded config and exits
This is most useful for automated tests for upcoming configuration
features. Can also be used for debugging.
2020-12-06 13:33:25 +01:00
Wolfgang WaltherandWolfgang Walther d218a9eaff dev: improve postgrest-watch experience
* speed up io tests
* avoid infinite loop while watching io tests
* improve auto-complete with postgrest-watch postgrest-run
* detect when files are added or deleted
2020-12-05 16:47:50 +01:00
Wolfgang WaltherandWolfgang Walther 7d6d015822 improve error output for io-tests and with_tmp_db 2020-12-03 22:32:07 +01:00
Wolfgang WaltherandGitHub 609c9aead8 Make transaction-rollback=true the default for test-suite (#1663)
Change test-suite to use db-tx-rollback-all = true by default
2020-12-03 18:54:33 +01:00
monacoremoandWolfgang Walther 10a70d4e52 add autocompletion for postgrest-watch 2020-12-01 22:13:41 +01:00
Wolfgang WaltherandWolfgang Walther de5742fc7d Make nix the one and only dev environment; remove obsolete tooling 2020-12-01 00:42:21 +01:00
Wolfgang WaltherandWolfgang Walther 1a5f6cce46 Add docker files to run nix dev environment on Windows 2020-12-01 00:42:21 +01:00
Wolfgang WaltherandGitHub f2cb91740e Use cabal v2-exec to run io tests, include io tests in nix-shell by default; remove obsolete ncat dependency (#1673)
Use cabal v2-exec to run io tests, include io tests in nix-shell by default; remove obsolete ncat dependency
2020-12-01 00:37:45 +01:00
Remo RechkemmerGitHubmonacoremo <monacoremo>
07cb47e6ac Improve dev tools in the Nix environment (#1666)
* Improve dev tools in the nix environment

Added postgrest-build, postgrest-run, postgrest-clean, postgrest-check
and postgrest-watch tools. Changed order of postgrest-test-spec-all to
run backwards from pg13 to pg9.5. Added shellcheck to lint shell scripts
in nix environment.

Co-authored-by: monacoremo <monacoremo>
2020-11-30 22:21:19 +01:00
steve-chavez 9adec12a67 cirrusci: add build timeout 2020-11-23 19:00:50 -05:00
steve-chavez 787973f323 Add db-prepared-statements config 2020-11-23 19:00:50 -05:00
steve-chavez 4bd5e6bd82 perf: enable prepared statements for GET
- Parametrize filters
- Parametrize LIMIT/OFFSET
- Parametrize JSON path(select=col->$1)
- Single parameter for IN(use ANY)
- Also enable prepared statement for the EXPLAIN
  used on the estimated count.
2020-11-23 19:00:50 -05:00
steve-chavez 8e58b56d5c refactor: Use hasql-dynamic on RPC/POST/PUT/PATCH
Change callProc/createWriteStatement to H.Snippet.

Parametrize the inputs on the same SQLFragments by taking advantage
of hasql-dynamic-statements. It's not necessary to parametrize
every input, inlining with pgFmtLit can still be used for queries
that can't be parametrized(like SET LOCALs).

Also adds hasql-dynamic-statements to Nix and stack.
2020-11-23 19:00:50 -05:00
Wolfgang WaltherandWolfgang Walther b13e95aefb Refactor tx-... config options to single tx-end option 2020-11-23 22:43:49 +01:00
Wolfgang WaltherandWolfgang Walther 3425352035 Clean up postgrest --help output 2020-11-23 17:40:04 +01:00
Wolfgang WaltherandSteve Chavez dbf99c6ac1 Add support for Prefer tx=rollback 2020-11-22 18:21:06 -05:00
Wolfgang WaltherandSteve Chavez 698fac8ff2 Add test to ignore functions unnamed arguments for rpc disambiguation, resolves #1638 2020-11-22 14:48:48 -05:00
YvanandSteve Chavez 0d5520d91d Update README.md
Added "Since" for the security row level PostgreSQL has it since 9.5
2020-11-21 15:50:27 -05:00
Wolfgang WaltherandSteve Chavez 944efb3a6a Split StructureSpec into OpenApiSpec and OptionsSpec 2020-11-20 11:51:49 -05:00
Wolfgang WaltherandSteve Chavez d91f47e582 Remove executable bit on .nix and .sql file 2020-11-20 10:48:47 -05:00
Wolfgang WaltherandSteve Chavez 2c52b96e04 Refactor tests: replace str with json QuasiQuoter where appropriate 2020-11-20 09:24:32 -05:00
Wolfgang WaltherandSteve Chavez e08bb3a197 Fix overloading of functions with unnamed arguments (specifically for prefer params=single-object) 2020-11-20 09:24:32 -05:00
Wolfgang WaltherandSteve Chavez b091586394 Refactor tests: renaming json table to prevent hiding of native json type 2020-11-20 09:24:32 -05:00
Wolfgang WaltherandSteve Chavez ca7ffd0a70 Fix RPC return type handling for domains with composite base type 2020-11-20 09:24:32 -05:00
Gergely MészárosandGitHub 3f690ec78f Removing single column limit from join table M2M mapping detection. (#1593) 2020-11-17 18:49:02 -05:00
Wolfgang WaltherandGitHub 6e04fe7454 Enable embedding through multiple layers of views recursively (#1625)
* Include hidden views from the search path. Hidden views are views in unexposed schemas that are part of a view dependency chain.

* Change allSourceColumns to only return pk and fk columns
2020-11-15 14:58:57 -05:00
steve-chavez 65968b5320 circleci: fix nightly release
Also delete unused variable in release/default.nix
2020-11-13 14:12:28 -05:00
steve-chavez ed8bf8272f Update BACKERS.md 2020-11-13 14:01:58 -05:00
steve-chavez 3830887577 ci: add nightly version to cabal file 2020-11-10 22:47:58 -05:00
steve-chavez b6d8d89b32 ci: add git sha with hours/minutes to nightly 2020-11-10 22:47:58 -05:00
steve-chavez 814da058ec cirrusci: add nightly release 2020-11-10 22:47:58 -05:00
steve-chavez 5d140f6fa0 circleci: add nightly release
For github and dockerhub
2020-11-10 22:47:58 -05:00
steve-chavez 51b43f5605 nix: change release scripts to use a version arg 2020-11-10 22:47:58 -05:00
steve-chavez 8fd9a9ab5d travisci: add nightly release 2020-11-10 22:47:58 -05:00
steve-chavez 5f3b581562 appveyor: add nightly release 2020-11-10 22:47:58 -05:00
steve-chavez 139f9b987a nix: fix postgrest-release-dockerhubdescription 2020-11-10 22:47:58 -05:00
steve-chavez 63849f1275 circleci: make release job a machine
setup_remote_docker is somehow not working and gives:

docker: command not found
2020-11-10 22:47:58 -05:00
steve-chavez 9f8d1af2cd nix: add postgrest-docker-login wrapper func 2020-11-10 22:47:58 -05:00
Wolfgang WaltherandSteve Chavez 122fea1507 Add test to insert into a view and expect location header return 2020-11-03 10:43:06 -05:00
Wolfgang WaltherandSteve Chavez a7403fecc2 refactor: split InsertSpec into InsertSpec and UpdateSpec 2020-11-03 10:43:06 -05:00
Wolfgang WaltherandSteve Chavez efc725fb24 run stylish-haskell v0.12.1.0 2020-11-02 13:22:01 -05:00
Wolfgang WaltherandSteve Chavez 4b4a622a17 fix non-variadic repeated param in variadic function 2020-10-30 10:06:53 -05:00
steve-chavez 8618ffa5fc perf: shortcut for proc with no variadic arg
Also refactor how rpc params are converted to json
2020-10-29 18:53:24 -05:00
steve-chavez 2798ced9b9 Fix charset=utf-8 appending to binary output 2020-10-29 18:53:24 -05:00
steve-chavez a6696f3ba1 refactor: add Hasql arrayColumn decoder 2020-10-29 18:53:24 -05:00
steve-chavez 04eaeec7fc refactor: reorder PgVersion(95/96)Spec tests
* move json ops tests to JsonOperatorSpec
* move phfts operator tests to QuerySpec and RpcSpec
* add custom spec for pre-request header guc tests
2020-10-29 18:53:24 -05:00
Wolfgang WaltherandSteve Chavez 302d4e15ad Allow calling variadic functions with repeated query params or JSON array in body 2020-10-26 17:49:51 -05:00
Wolfgang WaltherandSteve Chavez 18cc214c04 refactor moving findProc into ApiRequest to allow parsing parameters differently by proc 2020-10-26 17:49:51 -05:00
Wolfgang WaltherandSteve Chavez 06e85357c7 add tests to pass regular arrays properly to RPCs 2020-10-26 17:49:51 -05:00
Wolfgang WaltherandSteve Chavez eed0d3a66c ci: add coveralls.io integration 2020-10-22 12:05:25 -05:00
steve-chavez d1d0c6772a perf: change Text queries to ByteString
Improves performance by not utf8 encoding the whole query with
encodeUtf8. Only certain parts.

It's also a gradual step needed to use the Snippet type
from hasql-dynamic-statements.
2020-10-17 15:01:06 -05:00
steve-chavez 7e3e19acbb refactor: untangle addFiltersOrdersRanges
Remove convoluted function composition
2020-10-17 15:01:06 -05:00
Wolfgang WaltherandGitHub b7b66b600d ci: remove io tests from stack-test because of random segmentation faults (#1627)
io tests already run on nix-build-test
2020-10-16 13:53:43 -05:00
Wolfgang WaltherandGitHub 780970885e ci: run spec tests against all pg versions, even if one of them fails (#1628)
- when the build fails no tests are run(circleci-agent step halt)
- when the build succeeds all tests are run, even if some of them fail(when: always)
2020-10-16 11:15:20 -05:00
Wolfgang WaltherandSteve Chavez 5f33f01094 correct issue numbers in changelog 2020-10-16 10:24:41 -05:00
Wolfgang WaltherandSteve Chavez 719c4abba2 Drop support for postgres 9.4 2020-10-15 08:43:28 -05:00
Wolfgang WaltherandSteve Chavez 29858b8d0d improve cirrus build time by properly using cabal cache 2020-10-14 22:25:42 -05:00
Wolfgang WaltherandSteve Chavez b21a0f6192 fix travis build timeout swallowing error code 2020-10-14 10:49:47 -05:00
Wolfgang WaltherandSteve Chavez f5331d1a77 fix build error on windows 2020-10-14 09:26:35 -05:00
Wolfgang WaltherandSteve Chavez 5bb670b736 fix appveyour build script 2020-10-14 09:26:35 -05:00
Remo RechkemmerandSteve Chavez 2eb8083869 Simplify Nix CI jobs 2020-10-14 09:10:47 -05:00
Wolfgang WaltherandGitHub 99ecc7af70 Update stack.yaml.lock (#1616) 2020-10-13 10:50:19 -05:00
steve-chavez ccad9eb9bc Change default log-level, from info to error 2020-10-06 14:21:46 -05:00
steve-chavez d9a608d9b0 Add warn log level 2020-10-06 14:21:46 -05:00
steve-chavez f6b6abe734 Add error log level 2020-10-06 14:21:46 -05:00
steve-chavez e9efcc70a5 Add log-level config 2020-10-06 14:21:46 -05:00
Robert VollmertandSteve Chavez 60398ad538 Allow wai-middleware-static 0.9
- Only serve static files on HEAD or GET requests.
2020-10-06 12:45:40 -05:00
Remo RechkemmerandSteve Chavez 741f017a17 Use Postgres 13 in tests/docker-compose.yml 2020-10-05 13:47:53 -05:00
monacoremoandSteve Chavez e4e84e5714 Upgrade nixpkgs and add Postgres 13 support 2020-10-05 13:47:53 -05:00
Wolfgang WaltherandSteve Chavez d395bb6052 Fix error messages on connection failure for postgres on localized Windows 2020-10-03 10:11:41 -05:00
Felix YanandGitHub 17f04a886e Allow hspec-wai/hspec-wai-json 0.11 (#1595) 2020-09-27 16:04:47 -05:00
Robert VollmertandSteve Chavez 807dae1768 Allow wai-extra 3.1 2020-09-22 11:45:58 -05:00
Remo RechkemmerGitHubmonacoremo <monacoremo>
7af54c5813 Upgrade nixpkgs to current master and add an upgrade checklist (#1579)
* simplify haskell packages overrides based on new nixpkgs
* update Nix gitignore
* simplify Haskell packages overlay

Co-authored-by: monacoremo <monacoremo>
2020-09-09 19:08:16 -05:00
Robert VollmertandSteve Chavez bd2160db26 Make cabals bound more consistent
Specifically, allow point-updates to hasql-notifications, and
give a reasonable lower bound for network (the version that comess
with GHC 8.0 on stackage).
2020-08-14 21:12:58 -05:00
Robert VollmertandSteve Chavez a3f4548a81 Allow base64-bytestring 1.2 and optparse-applicative 0.16 2020-08-14 21:12:58 -05:00
Wolfgang WaltherandGitHub 8e4687fb53 Return 405 Method not Allowed for GET of volatile RPC instead of 500 (#1560) 2020-07-16 16:05:57 -05:00
Steve ChavezandGitHub 189847927e Schema cache reload with zero downtime (#1559)
* Improve error messages and comments

* Reorder Main.hs functions
2020-07-16 13:39:25 -05:00
Wolfgang WaltherandGitHub 6b2767d35c Additionally allow "bearer" instead of only "Bearer" in Authorization… (#1558) 2020-07-16 12:09:40 -05:00
RemoandSteve Chavez 1f6a824dfb Bump nixpkgs pin to unstable as of 2020-07-13 2020-07-13 18:47:45 -05:00
steve-chavez e8b4e3771c Allow config file reloading with SIGUSR2
* move config validation to Config.hs
* Add tests for jwt-secret/app.settings.*/db-schema reload
2020-07-13 11:30:16 -05:00
steve-chavez e272ea47be refactor: move logStdout/corsPolicy to Middleware
* remove configQuiet from Config

configQuiet was not an end user setting.
The logging setup is now an internal parameter.

* move proxy uri validation to Private dir
2020-07-13 11:30:16 -05:00
steve-chavez 0ff05edd16 refactor: move parseSecret out of App.postgrest
parseSecret only needs to be computed once, after the config is read.
2020-07-13 11:30:16 -05:00
steve-chavez 96a16a377f refactor: move loadDbUriFile/SecretFile to Config
* make schema cacher filler use Appconfig
* change configRoleClaimKey to Either Text JSPath
2020-07-13 11:30:16 -05:00
steve-chavez 896b79f05b refactor: separate reading file from parsing it 2020-07-13 11:30:16 -05:00
Wolfgang WaltherandGitHub 343e41c51d Location header improvements(#1475)
* Create location header only on primary key columns, closes https://github.com/PostgREST/postgrest/issues/1461

* Return location header when pk columns not selected, fixes https://github.com/PostgREST/postgrest/issues/1162
2020-07-08 14:36:40 -05:00
steve-chavez 55b4f4fbe7 Fix expired JWTs starting an empty transaction
Fixes https://github.com/PostgREST/postgrest/issues/1094.

Expired JWTs were doing an empty BEGIN/COMMIT in the db.
2020-07-03 17:23:10 -05:00
Wolfgang WaltherandSteve Chavez a5bc293372 add test cases for singular minimal 2020-07-03 13:54:39 -05:00
Steve ChavezandGitHub 43d71e95ac Allow schema cache reloading with NOTIFY (#1542)
Fixes https://github.com/PostgREST/postgrest/issues/1512

Helps on environments where you can't send unix signals(Windows, managed
containers). Also provides better UX for schema reloads - NOTIFY
can be sent from pg clients(psql, pgadmin).

`NOTIFY pgrst` - with no payload - should be done to reload the schema cache.
Notifications with a payload will be ignored.

The channel can be enabled with `db-channel-enabled`(false by default)
and its name can be configured with `db-channel`.

The LISTEN thread uses a dedicated pg connection.
This connection is recovered if it fails.
A debounce of 1ms is done in case too many NOTIFYs arrive.
2020-06-24 19:00:02 -05:00
steve-chavez 24064f8626 github: CHANGELOG reminder in PR template 2020-06-05 13:57:47 -05:00
steve-chavez 8588a42aa9 Fix broken links in .github dir
Also add a PR template
2020-06-05 13:40:03 -05:00
Steve ChavezandGitHub 0f0d617951 Allow http status override through response.status guc (#1541)
Fixes https://github.com/PostgREST/postgrest/issues/1525
2020-06-05 13:01:36 -05:00
Remo RechkemmerandGitHub 69b09e312a Nixify io and memory tests (#1538)
* Include tests for io and memory in the Nix environment.
* include spec tests in nix-shell by default
* install the io and memory tests in CI
2020-05-28 12:45:14 -05:00
Felix YanandGitHub 784ebe57d7 Allow aeson 1.5 (#1537) 2020-05-26 10:43:14 -05:00
Remo RechkemmerandGitHub 08186ea51c Nixify CircleCI setup (#1535)
* package ghr
* add nix-based release scripts
* refactor CI based on nix tests and release scripts
* update stack version in circleci
* update makefile
* remove docker directory
* add Docker Hub description
* add README on docker
2020-05-25 11:27:35 -05:00
steve-chavez d4aba5cb08 Add notice about single Linux static executable 2020-05-22 14:30:23 -05:00
Remo RechkemmerandGitHub 3a1844ec8e Fix the static build (#1534)
* openssl linking fix is required after all

* remove Cabal override (obsoleted on new nixpkgs)
2020-05-21 14:47:07 -05:00
Remo RechkemmerandGitHub 48c9ac36b1 Bump nixpkgs and simplify/clean up nix setup (#1529)
* Refactor tests and fix README

* fix linting hints from new hlint version

* postgrest-style with new nixpkgs-fmt
2020-05-20 21:30:20 -05:00
Remo RechkemmerandGitHub e6874c866d Fix UNKNOWN being shown as git commit hash (#1533)
Fixes https://github.com/PostgREST/postgrest/issues/1530
2020-05-20 14:10:53 -05:00
Remo RechkemmerandGitHub 289bb66f56 Reduce size of the static docker image by patching the Nix openssl package (#1528)
* hook docker images into the default builds
2020-05-20 11:49:11 -05:00
Steve ChavezandGitHub 67344c8e0a cabal: fix upload to hackage (#1532)
When doing:

```
cabal upload dist-newstyle/sdist/postgrest-7.0.1.tar.gz
```

The following error is shown:

```
Error: Invalid package

'ghc-options: -Wall -Werror' makes the package very easy to break with
future GHC versions because new GHC versions often add new warnings.
Use just 'ghc-options: -Wall' instead. Alternatively, if you want to
use this, make it conditional based on a Cabal configuration flag
(with 'manual: True' and 'default: False') and enable that flag
during development.
```

Put -Werror in a cabal flag to work around this restriction.
2020-05-19 13:59:33 -05:00
steve-chavez 222a53015e Fix Dockerfile
Remove libpq5 dependency since static bin already contains it.
2020-05-18 16:50:53 -05:00
267 changed files with 17863 additions and 9580 deletions
-446
View File
@@ -1,446 +0,0 @@
version: 2
jobs:
build-cache:
docker:
- image: circleci/buildpack-deps:trusty
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y --only-upgrade binutils
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests dependencies
command: |
stack build --fast -j1 --only-dependencies
stack build --fast --test --no-run-tests --only-dependencies
- save_cache:
paths:
- "~/.stack"
- ".stack-work"
key: v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
style:
docker:
- image: nixos/nix:2.3
steps:
- checkout
- run:
name: Install linting and styling scripts
command: nix-env -f default.nix -iA lint style
- run:
name: Run linter
command: |
echo 'Note: For checking this locally, use `make lint`. Or, if using `nix`, run `nix-shell --run postgrest-lint`'
postgrest-lint
- run:
name: Run style check
command: |
echo 'Note: For checking this locally, use `make style`. Or, if using `nix`, run `nix-shell --run postgrest-style`'
postgrest-style-check
build-test-9.4:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- image: circleci/postgres:9.4.26
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install ncat
command: |
# utility needed to test socket connection with curl < 7.40
sudo apt-get install nmap
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y --only-upgrade binutils
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests dependencies
command: |
stack build --fast -j1 --only-dependencies
stack build --fast --test --no-run-tests --only-dependencies
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
- run:
name: run tests
command: |
POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://circleci@localhost" postgrest_test) stack test
test/io-tests.sh
build-test-9.6:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- image: circleci/postgres:9.6.17
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
- run:
name: run tests
command: POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://circleci@localhost" postgrest_test) stack test postgrest:spec
build-test-10:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- image: circleci/postgres:10.12
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
- run:
name: run tests
command: POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://circleci@localhost" postgrest_test) stack test postgrest:spec
build-test-11:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- image: circleci/postgres:11.7
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
- run:
name: run tests
command: POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://circleci@localhost" postgrest_test) stack test postgrest:spec
build-test-12:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- image: circleci/postgres:12.2
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build src and tests
command: |
stack build --fast -j1
stack build --fast --test --no-run-tests
- run:
name: run tests
command: POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://circleci@localhost" postgrest_test) stack test postgrest:spec
build-prof-test:
docker:
- image: circleci/buildpack-deps:trusty
environment:
- PGHOST=localhost
- TERM=xterm
- image: circleci/postgres:9.6.17
environment:
- POSTGRES_USER=circleci
- POSTGRES_DB=circleci
- POSTGRES_HOST_AUTH_METHOD=trust
steps:
- checkout
- restore_cache:
keys:
- v1-stack-prof-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: install stack & dependencies
command: |
curl -L https://github.com/commercialhaskell/stack/releases/download/v2.1.3/stack-2.1.3-linux-x86_64.tar.gz | tar zx -C /tmp
sudo mv /tmp/stack-2.1.3-linux-x86_64/stack /usr/bin
sudo apt-get update
sudo apt-get install -y libgmp-dev
sudo apt-get install -y postgresql-client
stack setup
- run:
name: build dependencies with profiling enabled
command: |
stack build --profile -j1 --only-dependencies
- save_cache:
paths:
- "~/.stack"
- ".stack-work"
key: v1-stack-prof-dependencies-{{ checksum "postgrest.cabal" }}-{{ checksum "stack.yaml" }}
- run:
name: build with profiling enabled
command: |
stack build --profile -j1
- run:
name: run memory usage tests
command: |
test/create_test_db "postgres://circleci@localhost" postgrest_test
test/memory-tests.sh
release:
docker:
- image: circleci/golang:1.9
steps:
- attach_workspace:
at: /tmp/workspace
- checkout
- run:
name: add body and tars to github release
command: |
go get -u github.com/tcnksm/ghr
START=$(echo $CIRCLE_TAG | cut -c2-)
END='## \['
BODY=$(sed -n "1,/$START/d;/$END/q;p" CHANGELOG.md)
ghr -t $GITHUB_TOKEN -u $CIRCLE_PROJECT_USERNAME -r $CIRCLE_PROJECT_REPONAME -b "$BODY" --replace $CIRCLE_TAG /tmp/workspace/bin
- setup_remote_docker
- run:
name: publish docker image
command: |
docker build --build-arg POSTGREST_VERSION=$CIRCLE_TAG -t postgrest ./docker/
docker login -u $DOCKER_USER -p $DOCKER_PASS
docker tag postgrest postgrest/postgrest:$CIRCLE_TAG
docker push postgrest/postgrest:$CIRCLE_TAG
docker tag postgrest postgrest/postgrest:latest
docker push postgrest/postgrest:latest
build-nix:
machine: true
steps:
- checkout
- run:
name: Install Nix
command: |
curl -L https://nixos.org/nix/install | sh
echo "source $HOME/.nix-profile/etc/profile.d/nix.sh" >> $BASH_ENV
- run:
name: Install and use the Cachix binary cache
command: |
nix-env -iA cachix -f https://cachix.org/api/v1/install
cachix use postgrest
- run:
name: Build the dynamic PostgREST executable
command: |
# We build the dynamic postgrest executable first in order to fail fast if anything
# should be wrong with the source code.
nix-build -A postgrest
- restore_cache:
keys:
- v1-cabal-nix-dependencies-{{ checksum "postgrest.cabal" }}
- run:
name: Run the spec tests against all supported PostgreSQL versions
command: |
# Our utility scripts use cabal to run the tests, which will pick up all the libraries that
# are built with Nix. So this should only ever build the PostgREST itself and be reasonably
# quick. We accelerate that part further by caching `~/.cabal` and `dist-newstyle`.
nix-shell --run postgrest-test-spec-all
- save_cache:
paths:
- "~/.cabal"
- "dist-newstyle"
key: v1-cabal-nix-dependencies-{{ checksum "postgrest.cabal" }}
- run:
name: Build all derivations from default.nix and push results to Cachix
command: |
# Only push to the cache when CircleCI makes the CACHIX_SIGNING_KEY available (e.g. not for
# pull requests).
if [ -n "${CACHIX_SIGNING_KEY:-""}" ]; then
echo "Building and caching all derivations..."
nix-build | cachix push postgrest
else
echo "Building all derivations (caching skipped for outside pull requests)..."
nix-build
fi
build-static-bin:
machine: true
steps:
- checkout
- run:
name: Install Nix
command: |
curl -L https://nixos.org/nix/install | sh
echo "source $HOME/.nix-profile/etc/profile.d/nix.sh" >> $BASH_ENV
- run:
name: Install and use the Cachix binary cache
command: |
nix-env -iA cachix -f https://cachix.org/api/v1/install
cachix use postgrest
- run:
name: Build the PostgREST statically linked binary
command: |
nix-build -A postgrestStatic
- run:
name: compress binary
command: |
mkdir -p /tmp/workspace/bin
cd /tmp/workspace/bin
tar cvJf postgrest-$CIRCLE_TAG-linux-x64-static.tar.xz -C ~/project/result/bin postgrest
- persist_to_workspace:
root: /tmp/workspace
paths:
- bin/*
workflows:
version: 2
build-test-release:
jobs:
- build-cache:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
- style:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
- build-test-9.4:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
requires:
- build-cache
- build-test-9.6:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
requires:
- build-cache
- build-test-10:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
requires:
- build-cache
- build-test-11:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
requires:
- build-cache
- build-test-12:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
requires:
- build-cache
- build-prof-test:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
- build-static-bin:
requires:
- build-nix
- build-prof-test
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
branches:
ignore: /.*/
- release:
requires:
- build-static-bin
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
- build-nix:
filters:
tags:
only: /v[0-9]+(\.[0-9]+)*/
+11 -51
View File
@@ -1,55 +1,15 @@
freebsd_instance:
image: freebsd-12-1-release-amd64
image_family: freebsd-13-0
build_task:
env:
ASSUME_ALWAYS_YES: YES
install_script:
## TODO: Some of these workaround steps might be unneeded now
## gmake required, otherwise bsd make is used and some cryptic "No operator" errors are shown
- pkg install -y gmake
## gcc required, otherwise warning: macro expansion producing 'defined' has undefined behavior [-Wexpansion-to-defined]
- pkg install -y gcc
- pkg install -y libiconv
- pkg install -y gmp
- pkg install -y postgresql95-client
- pkg install -y pkgconf
- pkg install -y ghc
- pkg install -y hs-cabal-install
cabal_cache:
folder: ~/.cabal
fingerprint_script: cat postgrest.cabal
build_script:
- cabal v2-update
- cabal v2-build
binaries_artifacts:
path: "dist-newstyle/build/x86_64-freebsd/ghc-8.6.5/postgrest-7.0.1/x/postgrest/build/postgrest/*"
#binary_script:
#- mkdir bin
## cabal v2-install --installdir=.
## the installdir option is not working on this cabal version(form hs-cabal-install), so the binary must be extrated from the path
## this is too brittle so for now we don't use it
#- ldd /tmp/cirrus-ci-build/dist-newstyle/build/x86_64-freebsd/ghc-8.6.5/postgrest-7.0.1/x/postgrest/build/postgrest/postgrest
#- mv /tmp/cirrus-ci-build/dist-newstyle/build/x86_64-freebsd/ghc-8.6.5/postgrest-7.0.1/x/postgrest/build/postgrest/postgrest bin/
## https://cirrus-ci.org/guide/writing-tasks/#artifacts-instruction
#binaries_artifacts:
#path: bin/
cleanup_script:
- rm -rf ~/.cabal/logs
- rm ~/.cabal/config
name: Build FreeBSD (Stack)
install_script: pkg install -y postgresql13-client hs-stack
#publish_task:
#skip: $CIRRUS_TAG == ""
#env:
#TOKEN: ENCRYPTED[c99babbf04e8a33962ffbbc75bf20d3abe408f9bede5ed5d8956cd24da6fdb34266c920984cc43f8106ef3423fba0e98]
#depends_on:
#- build
#script:
## https://cirrus-ci.org/guide/writing-tasks/#artifacts-instruction
#- wget https://api.cirrus-ci.com/v1/artifact/github/$CIRRUS_REPO_FULL_NAME/build/binaries/bin/postgrest
#- tar cjf postgrest-$CIRRUS_TAG-freebsd.tar.xz postgrest
### The CIRRUS_RELEASE env var is not the same as the tag. It's an id for the release.
#- curl -X POST --data-binary @$NAME \
#-H "Authorization:token ${TOKEN}" \
#-H "Content-Type:application/octet-stream" \
#"https://uploads.github.com/repos/$CIRRUS_REPO_FULL_NAME/releases/${CIRRUS_RELEASE}/assets?name=${NAME}"
stack_cache:
folders: /.stack
fingerprint_script: cat postgrest.cabal stack.yaml.lock
reupload_on_changes: false
build_script: stack build -j 1 --local-bin-path . --copy-bins
bin_artifacts:
path: postgrest
+13 -15
View File
@@ -15,14 +15,14 @@ your contributions.
## Issues
For questions on how to use PostgREST, please use
[GitHub discussions](https://github.com/PostgREST/postgrest/discussions).
### Reporting an Issue
* Make sure you test against the latest released version. It is possible
we already fixed the bug you're experiencing.
* Also check the `CHANGELOG.md` to see if any unreleased changes affect
the issue. The very newest changes can take a little while to be released
as a new official version.
* Make sure you test against the latest [stable release](https://github.com/PostgREST/postgrest/releases/latest)
and also against the latest [nightly release](https://github.com/PostgREST/postgrest/releases/tag/nightly).
It is possible we already fixed the bug you're experiencing.
* Provide steps to reproduce the issue, including your OS version and
the specific database schema that you are using.
@@ -32,11 +32,14 @@ your contributions.
then [find your logs](http://blog.endpoint.com/2014/11/dear-postgresql-where-are-my-logs.html).
* If your database schema has changed while the PostgREST server is running,
[send the server a `SIGUSR1` signal](http://postgrest.org/en/v5.2/admin.html#schema-reloading) or restart it to ensure the schema cache
[send the server a `SIGUSR1` signal](http://postgrest.org/en/latest/admin.html#schema-reloading) or restart it to ensure the schema cache
is not stale. This sometimes fixes apparent bugs.
## Code
We have a fully nix-based development environment with many tools for a smooth development workflow available.
Check the [development docs](https://github.com/PostgREST/postgrest/blob/main/nix/README.md) on how to set it up and use it.
### Haskell Conventions
* All contributions must pass the tests before being merged. When
@@ -44,14 +47,9 @@ your contributions.
* All code must also pass [hlint](http://community.haskell.org/~ndm/hlint/) and [stylish-haskell](https://github.com/jaspervdj/stylish-haskell)
with no warnings. This helps enforce a uniform style for all committers. Continuous integration will check this as well on every
pull request. There's a useful Makefile that helps with checking this locally. You can run `make commit-check` to do this manually but
we recommend adding it to `.git/hooks/pre-commit` to automatically check this before doing a commit.
* For help building the Haskell code on your computer check out the [building from
source](https://postgrest.com/en/stable/install.html#build-from-source) docs section.
pull request. There are useful tools in the nix-shell that help with checking this locally. You can run `postgrest-check` to do this manually but
we recommend adding it to `.git/hooks/pre-commit` as `nix-shell --run postgrest-check` to automatically check this before doing a commit.
### Running Tests
For instructions on running tests, see the official docs hosted here:
https://postgrest.com/en/stable/install.html#postgrest-test-suite
For instructions on running tests, see the [development docs](https://github.com/PostgREST/postgrest/blob/main/nix/README.md#testing).
+3 -1
View File
@@ -1,6 +1,8 @@
<!--
Before reporting a bug:
If your database schema has changed while the PostgREST server is running, send the server a SIGUSR1 signal or restart it(http://postgrest.org/en/v5.2/admin.html#schema-reloading) to ensure the schema cache is not stale. This sometimes fixes apparent bugs.
If your database schema has changed while the PostgREST server is running,
send the server a SIGUSR1 signal or restart it(http://postgrest.org/en/stable/admin.html#schema-reloading)
to ensure the schema cache is not stale. This sometimes fixes apparent bugs.
-->
### Environment
+6
View File
@@ -0,0 +1,6 @@
<!--
When submitting a new feature or fix:
- Add a new entry to the CHANGELOG - https://github.com/PostgREST/postgrest/blob/main/CHANGELOG.md#unreleased
- If relevant, update the docs - https://github.com/PostgREST/postgrest-docs
-->
+21
View File
@@ -0,0 +1,21 @@
name: Setup Nix
description: Installs nix, sets up cachix and installs a subset of tooling.
inputs:
authToken:
description: Token to pass to cachix
tools:
description: Tools to install with nix-env -iA <tools>
runs:
using: composite
steps:
- uses: cachix/install-nix-action@v16
- uses: cachix/cachix-action@v10
with:
name: postgrest
authToken: ${{ inputs.authToken }}
- if: ${{ inputs.tools }}
run: nix-env -f default.nix -iA ${{ inputs.tools }}
shell: bash
+18
View File
@@ -0,0 +1,18 @@
codecov:
branch: main
require_ci_to_pass: false
comment: false
coverage:
status:
project:
default:
target: auto
threshold: 1%
only_pulls: false
patch:
default:
target: auto
threshold: 1%
only_pulls: true
+6
View File
@@ -0,0 +1,6 @@
version: 2
updates:
- package-ecosystem: github-actions
directory: /
schedule:
interval: weekly
+56
View File
@@ -0,0 +1,56 @@
#!/usr/bin/env bash
set -euo pipefail
# Get the FreeBSD PostgREST executable built by CirrusCI for the given GITHUB_COMMIT in GITHUB_REPOSITORY
# We use the GitHub API for 'check suites' to find the corresponding CirrusCI job, see:
# https://docs.github.com/en/rest/reference/checks#list-check-suites-for-a-git-reference
cirrus_artifact_name=bin
gh_accept_header="Accept: application/vnd.github.v3+json"
get_gh_check_runs_url() {
gh_checks_list_url="https://api.github.com/repos/$GITHUB_REPOSITORY/commits/$GITHUB_COMMIT/check-suites"
>&2 echo "Getting list of check-suites from $gh_checks_list_url ..."
curl --fail -H "$gh_accept_header" "$gh_checks_list_url" \
| jq -r '.check_suites[] | select(.app.slug == "cirrus-ci") | .check_runs_url'
}
wait_for_cirrusci() {
gh_check_runs_url="$(get_gh_check_runs_url)"
>&2 echo "Waiting to CirrusCI run to complete (two hours maximum)..."
for _ in $(seq 1 120); do
echo "Checking for CirrusCI task status at $gh_check_runs_url ..."
status=$(curl --fail "$gh_check_runs_url" | jq -r '.check_runs[] | .status')
if [ "$status" == "completed" ]; then
break
else
echo "CirrusCI task is still $status, waiting..."
sleep 60
fi
done
}
# The CirrusCI taskid can change if a new check run is started for the same commit,
# e.g. when pushing both a branch and tag. We make sure that we have the very
# latest taskid by re-loading the 'gh_check_runs_url' and the 'check run' itself.
get_cirrus_taskid() {
gh_check_runs_url="$(get_gh_check_runs_url)"
>&2 echo "Getting the CirrusCI task id from $gh_check_runs_url ..."
curl --fail -H "$gh_accept_header" "$gh_check_runs_url" \
| jq -r '.check_runs[] | .external_id'
}
download_artifact() {
cirrus_task_id="$(get_cirrus_taskid)"
cirrus_artifact_url="https://api.cirrus-ci.com/v1/artifact/task/$cirrus_task_id/$cirrus_artifact_name.zip"
>&2 echo "Attemping to download the CirrusCI artifact from $cirrus_artifact_url ..."
curl --fail "$cirrus_artifact_url" -o freebsd.zip
}
wait_for_cirrusci
download_artifact
echo "Unpacking executable..."
unzip freebsd.zip -d .
rm -rf freebsd.zip
+12
View File
@@ -0,0 +1,12 @@
#!/usr/bin/env bash
# Tag a release that will be built with Github Actions. The version is
# read from 'postgrest.cabal'.
version="$(grep -oP '^version:\s*\K.*' postgrest.cabal)"
echo "Tagging version v$version"
git tag -f "v$version"
echo "Pushing tag..."
git push -f origin "refs/tags/v$version"
+63
View File
@@ -0,0 +1,63 @@
#!/bin/bash
# This script builds PostgREST in a remote ARM server
[ -z "$1" ] && { echo "Missing 1st argument: PostgREST github commit SHA"; exit 1; }
[ -z "$2" ] && { echo "Missing 2nd argument: Build environment directory name"; exit 1; }
PGRST_GITHUB_COMMIT="$1"
SCRIPT_DIR="$2"
DOCKER_BUILD_DIR="$SCRIPT_DIR/docker-env"
install_packages() {
sudo apt-get update -y
sudo apt-get upgrade -y
sudo apt-get install -y git build-essential curl libffi-dev libffi7 libgmp-dev libgmp10 libncurses-dev libncurses5 libtinfo5 llvm libnuma-dev zlib1g-dev libpq-dev jq gcc
sudo apt-get clean
}
install_ghcup() {
export BOOTSTRAP_HASKELL_NONINTERACTIVE=1
export BOOTSTRAP_HASKELL_MINIMAL=1
curl --proto '=https' --tlsv1.2 -sSf https://get-ghcup.haskell.org | sh
source ~/.ghcup/env
}
install_cabal() {
ghcup upgrade
ghcup install cabal 3.6.0.0
ghcup set cabal 3.6.0.0
}
install_ghc() {
ghcup install ghc 8.10.7
ghcup set ghc 8.10.7
}
install_packages
# Add ghcup to the PATH for this session
[ -f ~/.ghcup/env ] && source ~/.ghcup/env
ghcup --version || install_ghcup
cabal --version || install_cabal
ghc --version || install_ghc
cd ~/$SCRIPT_DIR
# Clone the repository and build the project
git clone https://github.com/PostgREST/postgrest.git
cd postgrest
git checkout $PGRST_GITHUB_COMMIT
cabal v2-update && cabal v2-build
# Copy the built binary to the Dockerfile directory
PGRST_BIN=$(cabal exec which postgrest | tail -1)
cp $PGRST_BIN ~/$DOCKER_BUILD_DIR
# Move and compress the built binary
mkdir -p ~/$SCRIPT_DIR/result
mv $PGRST_BIN ~/$SCRIPT_DIR/result
cd ~/$SCRIPT_DIR
tar -cJf result.tar.xz result
+16
View File
@@ -0,0 +1,16 @@
# PostgREST docker hub image
FROM ubuntu:focal AS postgrest
RUN apt-get update -y \
&& apt install -y --no-install-recommends libpq-dev zlib1g-dev jq gcc libnuma-dev \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/*
COPY postgrest /usr/bin/postgrest
EXPOSE 3000
USER 1000
CMD postgrest
+50
View File
@@ -0,0 +1,50 @@
#!/bin/bash
# This script publishes the Docker ARM images to Docker Hub.
[ -z "$1" ] && { echo "Missing 1st argument: PostgREST github commit SHA"; exit 1; }
[ -z "$2" ] && { echo "Missing 2nd argument: Docker repo"; exit 1; }
[ -z "$3" ] && { echo "Missing 3rd argument: Docker username"; exit 1; }
[ -z "$4" ] && { echo "Missing 4th argument: Docker password"; exit 1; }
[ -z "$5" ] && { echo "Missing 5th argument: Build environment directory name"; exit 1; }
[ -z "$6" ] && { echo "Missing 6th argument: PostgREST version"; exit 1; }
PGRST_GITHUB_COMMIT="$1"
DOCKER_REPO="$2"
DOCKER_USER="$3"
DOCKER_PASS="$4"
SCRIPT_DIR="$5"
PGRST_VERSION="v$6"
IS_PRERELEASE="$7"
DOCKER_BUILD_DIR="$SCRIPT_DIR/docker-env"
clean_env()
{
sudo docker logout
}
# Login to Docker
sudo docker logout
{ echo $DOCKER_PASS | sudo docker login -u $DOCKER_USER --password-stdin; } || { echo "Couldn't login to docker"; exit 1; }
trap clean_env sigint sigterm exit
# Move to the docker build environment
cd ~/$DOCKER_BUILD_DIR
# Push final images to Docker hub
# NOTE: This command publishes a separate ARM image because the builds cannot
# be added to the manifest if they are not in the registry beforehand.
# This image must be manually deleted from Docker Hub at the end of the process.
sudo docker buildx build --build-arg PGRST_GITHUB_COMMIT=$PGRST_GITHUB_COMMIT \
-t $DOCKER_REPO/postgrest:$PGRST_VERSION-arm \
--push .
# Add the arm images to the manifest
# NOTE: This assumes that there already is a `postgrest:<version>` image
# for the amd64 architecture pushed to Docker Hub
sudo docker buildx imagetools create --append -t $DOCKER_REPO/postgrest:$PGRST_VERSION $DOCKER_REPO/postgrest:$PGRST_VERSION-arm
[ -z $IS_PRERELEASE ] && sudo docker buildx imagetools create --append -t $DOCKER_REPO/postgrest:latest $DOCKER_REPO/postgrest:$PGRST_VERSION-arm
sudo docker logout
+485
View File
@@ -0,0 +1,485 @@
name: CI
on:
push:
branches:
- main
- rel-*
tags:
- v*
pull_request:
branches:
- main
- rel-*
jobs:
Lint-Style:
name: Lint & check code style
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: style
- name: Run linter (check locally with `nix-shell --run postgrest-lint`)
run: postgrest-lint
- name: Run style check (auto-format with `nix-shell --run postgrest-style`)
run: postgrest-style-check
Test-Nix:
name: Test (Nix)
runs-on: ubuntu-latest
defaults:
run:
# Hack for enabling color output, see:
# https://github.com/actions/runner/issues/241#issuecomment-842566950
shell: script -qec "bash --noprofile --norc -eo pipefail {0}"
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: tests
- name: Run coverage (IO tests and Spec tests against PostgreSQL 14)
run: postgrest-coverage
- name: Upload coverage to codecov
uses: codecov/codecov-action@v3.1.0
with:
files: ./coverage/codecov.json
- name: Run doctests
if: always()
run: nix-shell --run postgrest-test-doctests
- name: Check the spec tests for idempotence
if: always()
run: postgrest-test-spec-idempotence
Test-Pg-Nix:
strategy:
fail-fast: false
matrix:
pgVersion: [9.6, 10, 11, 12, 13, 14]
name: Test PG ${{ matrix.pgVersion }} (Nix)
runs-on: ubuntu-latest
defaults:
run:
# Hack for enabling color output, see:
# https://github.com/actions/runner/issues/241#issuecomment-842566950
shell: script -qec "bash --noprofile --norc -eo pipefail {0}"
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: tests withTools
- name: Run spec tests
if: always()
run: postgrest-with-postgresql-${{ matrix.pgVersion }} postgrest-test-spec
- name: Run IO tests
if: always()
run: postgrest-with-postgresql-${{ matrix.pgVersion }} -f test/io/fixtures.sql postgrest-test-io
- name: Run query cost tests
if: always()
run: postgrest-with-postgresql-${{ matrix.pgVersion }} postgrest-test-querycost
Test-Memory-Nix:
name: Test memory (Nix)
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: memory
- name: Run memory tests
run: postgrest-test-memory
Build-Static-Nix:
name: Build Linux static (Nix)
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
authToken: '${{ secrets.CACHIX_AUTH_TOKEN }}'
- name: Build static executable
run: nix-build -A postgrestStatic
- name: Save built executable as artifact
uses: actions/upload-artifact@v3
with:
name: postgrest-linux-static-x64
path: result/bin/postgrest
if-no-files-found: error
- name: Build Docker image
run: nix-build -A docker.image --out-link postgrest-docker.tar.gz
- name: Save built Docker image as artifact
uses: actions/upload-artifact@v3
with:
name: postgrest-docker-x64
path: postgrest-docker.tar.gz
if-no-files-found: error
- name: Build and push everything to Cachix (main branch only)
if: ${{ github.ref == 'refs/heads/main' }}
run: |
nix-build
nix-env -f default.nix -iA devTools
postgrest-push-cachix
Build-Stack:
strategy:
fail-fast: false
matrix:
include:
- name: Linux & test
runs-on: ubuntu-latest
cache: |
~/.stack
.stack-work
test: true
pgdir: /usr/lib/postgresql
artifact: postgrest-ubuntu-x64
- name: MacOS & test
runs-on: macos-latest
cache: |
~/.stack
.stack-work
test: true
pgdir: /usr/local/Cellar/postgresql
artifact: postgrest-macos-x64
- name: Windows
runs-on: windows-latest
cache: |
~\AppData\Roaming\stack
~\AppData\Local\Programs\stack
.stack-work
deps: Add-Content $env:GITHUB_PATH $env:PGBIN
# We'd need to make test/with_tmp_db run on Windows first
# test: true
artifact: postgrest-windows-x64
name: Build ${{ matrix.name }} (Stack)
runs-on: ${{ matrix.runs-on }}
steps:
- uses: actions/checkout@v3
- name: Stack working files cache
uses: actions/cache@v3
with:
path: ${{ matrix.cache }}
key: ${{ runner.os }}-${{ hashFiles('stack.yaml.lock') }}
- name: Install dependencies
if: ${{ matrix.deps }}
run: ${{ matrix.deps }}
- name: Build with Stack
run: stack build --local-bin-path result --copy-bins
- name: Run Spec tests with Stack
if: ${{ matrix.test }}
run: |
postgresql_bin="$(find ${{ matrix.pgdir }} -maxdepth 2 -type d -name bin | head -n 1)"
echo "Using PostgreSQL binaries at $postgresql_bin ..."
PATH="$postgresql_bin:$PATH" test/with_tmp_db stack test
- name: Save built executable as artifact
uses: actions/upload-artifact@v3
with:
name: ${{ matrix.artifact }}
path: |
result/postgrest
result/postgrest.exe
if-no-files-found: error
Get-FreeBSD-CirrusCI:
name: Get FreeBSD build from CirrusCI
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Get FreeBSD executable from CirrusCI
env:
# GITHUB_SHA does weird things for pull request, so we roll our own:
GITHUB_COMMIT: ${{github.event.pull_request.head.sha || github.sha}}
run: .github/get_cirrusci_freebsd
- name: Save executable as artifact
uses: actions/upload-artifact@v3
with:
name: postgrest-freebsd-x64
path: postgrest
if-no-files-found: error
Build-Cabal-Arm:
name: Build aarch64 (Cabal)
if: ${{ github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/v') || startsWith(github.ref, 'refs/heads/rel-') }}
runs-on: ubuntu-latest
outputs:
remotepath: ${{ steps.Remote-Dir.outputs.remotepath }}
env:
GITHUB_COMMIT: ${{ github.sha }}
steps:
- uses: actions/checkout@v2.4.0
- id: Remote-Dir
name: Unique directory name for the remote build
run: echo "::set-output name=remotepath::postgrest-build-$(uuidgen)"
- name: Copy script files to the remote server
uses: appleboy/scp-action@master
with:
host: ${{ secrets.SSH_ARM_HOST }}
username: ubuntu
key: ${{ secrets.SSH_ARM_PRIVATE_KEY }}
fingerprint: ${{ secrets.SSH_ARM_FINGERPRINT }}
source: ".github/scripts/arm/*"
target: ${{ steps.Remote-Dir.outputs.remotepath }}
strip_components: 3
- name: Build ARM
uses: appleboy/ssh-action@master
env:
REMOTE_DIR: ${{ steps.Remote-Dir.outputs.remotepath }}
with:
host: ${{ secrets.SSH_ARM_HOST }}
username: ubuntu
key: ${{ secrets.SSH_ARM_PRIVATE_KEY }}
fingerprint: ${{ secrets.SSH_ARM_FINGERPRINT }}
command_timeout: 120m
script_stop: true
envs: GITHUB_COMMIT,REMOTE_DIR
script: bash ~/$REMOTE_DIR/build.sh "$GITHUB_COMMIT" "$REMOTE_DIR"
- name: Download binaries from remote server
uses: nicklasfrahm/scp-action@main
with:
direction: download
host: ${{ secrets.SSH_ARM_HOST }}
username: ubuntu
key: ${{ secrets.SSH_ARM_PRIVATE_KEY }}
fingerprint: ${{ secrets.SSH_ARM_FINGERPRINT }}
source: "${{ steps.Remote-Dir.outputs.remotepath }}/result.tar.xz"
target: "result.tar.xz"
- name: Extract downloaded binaries
run: tar -xvf result.tar.xz && rm result.tar.xz
- name: Save aarch64 executable as artifact
uses: actions/upload-artifact@v2.3.1
with:
name: postgrest-ubuntu-aarch64
path: result/postgrest
if-no-files-found: error
Prepare-Release:
name: Prepare release
if: startsWith(github.ref, 'refs/tags/v')
runs-on: ubuntu-latest
needs:
- Lint-Style
- Test-Nix
- Test-Pg-Nix
- Test-Memory-Nix
- Build-Static-Nix
- Build-Stack
#- Get-FreeBSD-CirrusCI
- Build-Cabal-Arm
outputs:
version: ${{ steps.Identify-Version.outputs.version }}
isprerelease: ${{ steps.Identify-Version.outputs.isprerelease }}
steps:
- uses: actions/checkout@v3
- id: Identify-Version
name: Identify the version to be released
run: |
tag_version="${GITHUB_REF##*/}"
cabal_version="$(grep -oP '^version:\s*\K.*' postgrest.cabal)"
if [ "$tag_version" != "v$cabal_version" ]; then
echo "Tagged version ($tag_version) does not match the one in postgrest.cabal (v$cabal_version). Aborting release..."
exit 1
else
echo "Version to be released is $cabal_version"
echo "::set-output name=version::$cabal_version"
fi
if [[ "$cabal_version" != *.*.*.* ]]; then
echo "Version is for a full release (version does not have four components)"
else
echo "Version is for a pre-release (version has four components, e.g., 1.1.1.1)"
echo "::set-output name=isprerelease::1"
fi
- name: Identify changes from CHANGELOG.md
run: |
version="${{ steps.Identify-Version.outputs.version }}"
isprerelease="${{ steps.Identify-Version.outputs.isprerelease }}"
if [ -n "$isprerelease" ]; then
echo "Getting unreleased changes..."
sed -n "1,/## Unreleased/d;/## \[/q;p" CHANGELOG.md > CHANGES.md
else
echo "Full release (no '-' in version), getting changes for version $version ..."
sed -n "1,/## \[$version\]/d;/## \[/q;p" CHANGELOG.md > CHANGES.md
fi
echo "Relevant extract from CHANGELOG.md:"
cat CHANGES.md
- name: Save CHANGES.md as artifact
uses: actions/upload-artifact@v3
with:
name: release-changes
path: CHANGES.md
if-no-files-found: error
Release-GitHub:
name: Release on GitHub
permissions:
contents: write
runs-on: ubuntu-latest
needs: Prepare-Release
env:
VERSION: ${{ needs.Prepare-Release.outputs.version }}
steps:
- uses: actions/checkout@v3
- name: Download all artifacts
uses: actions/download-artifact@v3
with:
path: artifacts
- name: Create release bundle with archives for all builds
run: |
find artifacts -type f -iname postgrest -exec chmod +x {} \;
mkdir -p release-bundle
tar cJvf "release-bundle/postgrest-v$VERSION-linux-static-x64.tar.xz" \
-C artifacts/postgrest-linux-static-x64 postgrest
# No need to release Ubuntu, as the static Linux binary built with Nix
# covers all Linux use-cases
#tar cfJv "release-bundle/postgrest-v$VERSION-ubuntu-x64.tar.xz" \
# -C artifacts/postgrest-ubuntu-x64 postgrest
tar cJvf "release-bundle/postgrest-v$VERSION-macos-x64.tar.xz" \
-C artifacts/postgrest-macos-x64 postgrest
# TODO: Fix timeouts for FreeBSD builds in Cirrus
#tar cJvf "release-bundle/postgrest-v$VERSION-freebsd-x64.tar.xz" \
# -C artifacts/postgrest-freebsd-x64 postgrest
tar cJvf "release-bundle/postgrest-v$VERSION-ubuntu-aarch64.tar.xz" \
-C artifacts/postgrest-ubuntu-aarch64 postgrest
zip "release-bundle/postgrest-v$VERSION-windows-x64.zip" \
artifacts/postgrest-windows-x64/postgrest.exe
- name: Save release bundle
uses: actions/upload-artifact@v3
with:
name: release-bundle
path: release-bundle
if-no-files-found: error
- name: Publish release on GitHub
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
isprerelease="${{ needs.Prepare-Release.outputs.isprerelease }}"
echo "Releasing version v$VERSION on GitHub (isprerelease=$isprerelease)..."
gh release delete "v$VERSION" || true
gh release create "v$VERSION" \
-F artifacts/release-changes/CHANGES.md \
${isprerelease:+"--prerelease"} \
release-bundle/*
Release-Docker:
name: Release on Docker Hub
runs-on: ubuntu-latest
needs:
- Build-Cabal-Arm
- Prepare-Release
env:
GITHUB_COMMIT: ${{ github.sha }}
DOCKER_REPO: postgrest
DOCKER_USER: stevechavez
DOCKER_PASS: ${{ secrets.DOCKER_PASS }}
VERSION: ${{ needs.Prepare-Release.outputs.version }}
ISPRERELEASE: ${{ needs.Prepare-Release.outputs.isprerelease }}
steps:
- uses: actions/checkout@v3
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: release
- name: Download Docker image
uses: actions/download-artifact@v3
with:
name: postgrest-docker-x64
- name: Publish images on Docker Hub
run: |
docker login -u "$DOCKER_USER" -p "$DOCKER_PASS"
docker load -i postgrest-docker.tar.gz
docker tag postgrest:latest "$DOCKER_REPO/postgrest:v$VERSION"
docker push "$DOCKER_REPO/postgrest:v$VERSION"
# Only tag 'latest' for full releases
if [[ -z "$ISPRERELEASE" ]]; then
echo "Pushing to 'latest' tag for full release of v$VERSION ..."
docker tag postgrest:latest "$DOCKER_REPO"/postgrest:latest
docker push "$DOCKER_REPO"/postgrest:latest
else
echo "Skipping pushing to 'latest' tag for v$VERSION pre-release..."
fi
- name: Publish images for ARM builds on Docker Hub
uses: appleboy/ssh-action@master
env:
REMOTE_DIR: ${{ needs.Build-Cabal-Arm.outputs.remotepath }}
with:
host: ${{ secrets.SSH_ARM_HOST }}
username: ubuntu
key: ${{ secrets.SSH_ARM_PRIVATE_KEY }}
fingerprint: ${{ secrets.SSH_ARM_FINGERPRINT }}
script_stop: true
envs: GITHUB_COMMIT,DOCKER_REPO,DOCKER_USER,DOCKER_PASS,REMOTE_DIR,VERSION,ISPRERELEASE
script: bash ~/$REMOTE_DIR/docker-publish.sh "$GITHUB_COMMIT" "$DOCKER_REPO" "$DOCKER_USER" "$DOCKER_PASS" "$REMOTE_DIR" "$VERSION" "$ISPRERELEASE"
# TODO: Enable dockerhub description update again, once a solution for the permission problem is found:
# https://github.com/docker/hub-feedback/issues/1927
# - name: Update descriptions on Docker Hub
# env:
# DOCKER_PASS: ${{ secrets.DOCKER_PASS }}
# run: |
# if [[ -z "$ISPRERELEASE" ]]; then
# echo "Updating description on Docker Hub..."
# postgrest-release-dockerhub-description
# else
# echo "Skipping updating description for pre-release..."
# fi
Clean-Arm-Server:
name: Remove copied files from server
needs:
- Build-Cabal-Arm
- Release-Docker
if: ${{ always() && (github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/v') || startsWith(github.ref, 'refs/heads/rel-')) }}
runs-on: ubuntu-latest
env:
REMOTE_DIR: ${{ needs.Build-Cabal-Arm.outputs.remotepath }}
steps:
- uses: actions/checkout@v2.4.0
- name: Remove uploaded files from server
uses: appleboy/ssh-action@master
with:
host: ${{ secrets.SSH_ARM_HOST }}
username: ubuntu
key: ${{ secrets.SSH_ARM_PRIVATE_KEY }}
fingerprint: ${{ secrets.SSH_ARM_FINGERPRINT }}
envs: REMOTE_DIR
script: rm -rf $REMOTE_DIR
+34
View File
@@ -0,0 +1,34 @@
name: Loadtest
on:
push:
branches:
- main
tags:
- v*
pull_request:
branches:
- main
jobs:
Loadtest-Nix:
name: Loadtest (Nix)
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
with:
fetch-depth: 0
- name: Setup Nix Environment
uses: ./.github/actions/setup-nix
with:
tools: loadtest
- name: Run loadtest
run: |
postgrest-loadtest-against main
postgrest-loadtest-report > loadtest/loadtest.md
- name: Upload report
uses: actions/upload-artifact@v3
with:
name: loadtest.md
path: loadtest/loadtest.md
if-no-files-found: error
+33
View File
@@ -0,0 +1,33 @@
name: Upload Reports
on:
workflow_run:
workflows: ["Loadtest"]
types:
- completed
jobs:
upload:
name: Loadtest
permissions:
checks: write
runs-on: ubuntu-latest
if: ${{ github.event.workflow_run.conclusion == 'success' }}
steps:
- name: Download from Artifacts
uses: dawidd6/action-download-artifact@v2
with:
workflow: ${{ github.event.workflow.name }}
run_id: ${{github.event.workflow_run.id }}
name: loadtest.md
path: artifacts
- name: Upload to GitHub Checks
uses: LouisBrunner/checks-action@v1.2.0
with:
token: ${{ secrets.GITHUB_TOKEN }}
sha: ${{ github.event.workflow_run.head_sha }}
name: Loadtest Results
conclusion: neutral
output: |
{"summary":""}
output_text_description_file: artifacts/loadtest.md
+7
View File
@@ -15,3 +15,10 @@ site
*.swp
result*
dist-newstyle
postgrest.hp
postgrest.prof
__pycache__
*.tix
coverage
.hpc
loadtest
-70
View File
@@ -1,70 +0,0 @@
## Travis is only used for building an OSX binary ,
## no tests are run here.
language: generic
sudo: false
os:
- osx
cache:
timeout: 1000
directories:
- $HOME/.stack
- $HOME/.local/bin
before_install:
- mkdir -p "$HOME/.local/bin"
- export PATH="$PATH:$HOME/.local/bin"
install:
- |
if test -f "$HOME/.local/bin/stack"
then
echo 'Stack is already installed.'
else
echo "Installing Stack..."
travis_retry curl -L https://www.stackage.org/stack/osx-x86_64 > stack.tar.gz
gunzip stack.tar.gz
tar -x -f stack.tar --strip-components 1
mv stack "$HOME/.local/bin/"
rm stack.tar
fi
- |
if test -f "$HOME/.local/bin/ghr"
then
echo 'ghr is already installed.'
else
echo "Installing ghr..."
travis_retry curl -L https://github.com/tcnksm/ghr/releases/download/v0.5.4/ghr_v0.5.4_darwin_386.zip > ghr.zip
unzip ghr.zip -d "$HOME/.local/bin"
rm ghr.zip
fi
script:
## Building the whole project can take longer than 50 minutes. Since Travis has a global timeout of 50 minutes
## we compile for 30 minutes tops(`gtimeout 1800`) and quit compiling with no error.
## Since we CACHE the compile results we can continue compiling from where we left off
## on the next commit.
- gtimeout 1800 stack build --no-terminal --only-snapshot --install-ghc || true
- |
if test ! "$TRAVIS_TAG"
then
echo 'No tag pushed. Skip building binary.'
else
stack build --no-terminal --copy-bins --local-bin-path .
fi
- |
if test ! "$TRAVIS_TAG"
then
echo 'No tag pushed. Skipping release.'
else
OWNER="$(echo "$TRAVIS_REPO_SLUG" | cut -f1 -d/)"
REPO="$(echo "$TRAVIS_REPO_SLUG" | cut -f2 -d/)"
START=$(echo $TRAVIS_TAG | cut -c2-)
END='## \['
BODY=$(sed -n "1,/$START/d;/$END/q;p" CHANGELOG.md)
strip postgrest
tar cJf postgrest-$TRAVIS_TAG-osx.tar.xz postgrest
ghr -t $GITHUB_TOKEN -u $OWNER -r $REPO -b "$BODY"--replace $TRAVIS_TAG postgrest-$TRAVIS_TAG-osx.tar.xz
fi
+28 -6
View File
@@ -8,18 +8,36 @@ PostgREST ongoing development is only possible thanks to our Sponsors and Backer
<tbody>
<tr>
<td align="center" valign="middle">
<a href="https://www.cybertec-postgresql.com/en/" target="_blank">
<img width="222px" src="static/cybertec.png">
<a href="https://www.cybertec-postgresql.com/en/?utm_source=postgrest.org&utm_medium=referral&utm_campaign=postgrest" target="_blank">
<img width="222px" src="static/cybertec-new.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://www.2ndquadrant.com/en/?utm_campaign=External%20Websites&utm_source=PostgREST&utm_medium=Logo" target="_blank">
<img width="222px" src="static/2ndquadrant.png">
<img width="296px" src="static/2ndquadrant.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://tryretool.com/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="222px" src="static/retool.png">
<img width="296px" src="static/retool.png">
</a>
</td>
</tr>
<tr></tr>
<tr>
<td align="center" valign="middle">
<a href="https://gnuhost.eu/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="static/gnuhost.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://supabase.io?utm_source=postgrest%20backers&utm_medium=open%20source%20partner&utm_campaign=postgrest%20backers%20github&utm_term=homepage" target="_blank">
<img width="296px" src="static/supabase.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://oblivious.ai/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="static/oblivious.jpg">
</a>
</td>
</tr>
@@ -28,9 +46,9 @@ PostgREST ongoing development is only possible thanks to our Sponsors and Backer
## Lead Backers
- [Daniel Babiak](https://github.com/d-babiak)
- Evans Fernandes
- [Jan Sommer](https://github.com/nerfpops)
- [Franz Gusenbauer](https://www.igutech.at/)
## Backers
@@ -38,13 +56,15 @@ PostgREST ongoing development is only possible thanks to our Sponsors and Backer
- Michel Pelletier
- Jay Hannah
- Robert Stolarz
- Kofi Gumbs
- Nicholas DiBiase
- Christopher Reid
- Nathan Bouscal
- Daniel Rafaj
- David Fenko
- Remo Rechkemmer
- Severin Ibarluzea
- Tom Saleeba
- Pawel Tyll
## Former Backers
@@ -61,3 +81,5 @@ PostgREST ongoing development is only possible thanks to our Sponsors and Backer
</table>
- [Christiaan Westerbeek](https://devotis.nl)
- [Daniel Babiak](https://github.com/dbabiak)
- Kofi Gumbs
+110 -1
View File
@@ -3,12 +3,117 @@
All notable changes to this project will be documented in this file.
This project adheres to [Semantic Versioning](http://semver.org/).
## Unreleased
## [9.0.1] - 2022-06-03
### Fixed
- #2165, Fix json/jsonb columns should not have type in OpenAPI spec - @clrnd
- #2020, Execute deferred constraint triggers when using `Prefer: tx=rollback` - @wolfgangwalther
- #2077, Fix `is` not working with upper or mixed case values like `NULL, TrUe, FaLsE` - @steve-chavez
- #2024, Fix schema cache loading when views with XMLTABLE and DEFAULT are present - @wolfgangwalther
- #1724, Fix wrong CORS header Authentication -> Authorization - @wolfgangwalther
- #2120, Fix reading database configuration properly when `=` is present in value - @wolfgangwalther
- #2135, Remove trigger functions from schema cache and OpenAPI output, because they can't be called directly anyway. - @wolfgangwalther
- #2101, Remove aggregates, procedures and window functions from the schema cache and OpenAPI output. - @wolfgangwalther
- #2153, Fix --dump-schema running with a wrong PG version. - @wolfgangwalther
- #2042, Keep working when EMFILE(Too many open files) is reached. - @steve-chavez
- #2147, Ignore `Content-Type` headers for `GET` requests when calling RPCs. - @laurenceisla
+ Previously, `GET` without parameters, but with `Content-Type: text/plain` or `Content-Type: application/octet-stream` would fail with `404 Not Found`, even if a function without arguments was available.
- #2239, Fix misleading disambiguation error where the content of the `relationship` key looks like valid syntax - @laurenceisla
- #2294, Disable parallel GC for better performance on higher core CPUs - @steve-chavez
- #1076, Fix using CPU while idle - @steve-chavez
## [9.0.0] - 2021-11-25
### Added
- #1783, Include partitioned tables into the schema cache. Allows embedding, UPSERT, INSERT with Location response, OPTIONS request and OpenAPI support for partitioned tables - @laurenceisla
- #1878, Add Retry-After hint header when in recovery mode - @gautam1168
- #1735, Allow calling function with single unnamed param through RPC POST. - @steve-chavez
+ Enables calling a function with a single json parameter without using `Prefer: params=single-object`
+ Enables uploading bytea to a function with `Content-Type: application/octet-stream`
+ Enables uploading raw text to a function with `Content-Type: text/plain`
- #1938, Allow escaping inside double quotes with a backslash, e.g. `?col=in.("Double\"Quote")`, `?col=in.("Back\\slash")` - @steve-chavez
- #1075, Allow filtering top-level resource based on embedded resources filters. This is enabled by adding `!inner` to the embedded resource, e.g. `/projects?select=*,clients!inner(*)&clients.id=eq.12`- @steve-chavez, @Iced-Sun
- #1857, Make GUC names for headers, cookies and jwt claims compatible with PostgreSQL v14 - @laurenceisla, @robertsosinski
+ Getting the value for a header GUC on PostgreSQL 14 is done using `current_setting('request.headers')::json->>'name-of-header'` and in a similar way for `request.cookies` and `request.jwt.claims`
+ PostgreSQL versions below 14 can opt in to the new JSON GUCs by setting the `db-use-legacy-gucs` config option to false (true by default)
- #1988, Allow specifying `unknown` for the `is` operator - @steve-chavez
- #2031, Improve error message for ambiguous embedding and add a relevant hint that includes unambiguous embedding suggestions - @laurenceisla
### Fixed
- #1871, Fix OpenAPI missing default values for String types and identify Array types as "array" instead of "string" - @laurenceisla
- #1930, Fix RPC return type handling for `RETURNS TABLE` with a single column. Regression of #1615. - @wolfgangwalther
- #1938, Fix using single double quotes(`"`) and backslashes(`/`) as values on the "in" operator - @steve-chavez
- #1992, Fix schema cache query failing with standard_conforming_strings = off - @wolfgangwalther
### Changed
- #1949, Drop support for embedding hints used with '.'(`select=projects.client_id(*)`), '!' should be used instead(`select=projects!client_id(*)`) - @steve-chavez
- #1783, Partitions (created using `PARTITION OF`) are no longer included in the schema cache. - @laurenceisla
- #2038, Dropped support for PostgreSQL 9.5 - @wolfgangwalther
## [8.0.0] - 2021-07-25
### Added
- #1525, Allow http status override through response.status guc - @steve-chavez
- #1512, Allow schema cache reloading with NOTIFY - @steve-chavez
- #1119, Allow config file reloading with SIGUSR2 - @steve-chavez
- #1558, Allow 'Bearer' with and without capitalization as authentication schema - @wolfgangwalther
- #1470, Allow calling RPC with variadic argument by passing repeated params - @wolfgangwalther
- #1559, No downtime when reloading the schema cache with SIGUSR1 - @steve-chavez
- #504, Add `log-level` config option. The admitted levels are: crit, error, warn and info - @steve-chavez
- #1607, Enable embedding through multiple views recursively - @wolfgangwalther
- #1598, Allow rollback of the transaction with Prefer tx=rollback - @wolfgangwalther
- #1633, Enable prepared statements for GET filters. When behind a connection pooler, you can disable preparing with `db-prepared-statements=false`
+ This increases throughput by around 30% for simple GET queries(no embedding, with filters applied).
- #1729, #1760, Get configuration parameters from the db and allow reloading config with NOTIFY - @steve-chavez
- #1824, Allow OPTIONS to generate certain HTTP methods for a DB view - @laurenceisla
- #1872, Show timestamps in startup/worker logs - @steve-chavez
- #1881, Add `openapi-mode` config option that allows ignoring roles privileges when showing the OpenAPI output - @steve-chavez
- CLI options(for debugging):
+ #1678, Add --dump-config CLI option that prints loaded config and exits - @wolfgangwalther
+ #1691, Add --example CLI option to show example config file - @wolfgangwalther
+ #1697, #1723, Add --dump-schema CLI option for debugging purposes - @monacoremo, @wolfgangwalther
- #1794, (Experimental) Add `request.spec` GUC for db-root-spec - @steve-chavez
### Fixed
- #1592, Removed single column restriction to allow composite foreign keys in join tables - @goteguru
- #1530, Fix how the PostgREST version is shown in the help text when the `.git` directory is not available - @monacoremo
- #1094, Fix expired JWTs starting an empty transaction on the db - @steve-chavez
- #1162, Fix location header for POST request with select= without PK - @wolfgangwalther
- #1585, Fix error messages on connection failure for localized postgres on Windows - @wolfgangwalther
- #1636, Fix `application/octet-stream` appending `charset=utf-8` - @steve-chavez
- #1469, #1638 Fix overloading of functions with unnamed arguments - @wolfgangwalther
- #1560, Return 405 Method not Allowed for GET of volatile RPC instead of 500 - @wolfgangwalther
- #1584, Fix RPC return type handling and embedding for domains with composite base type (#1615) - @wolfgangwalther
- #1608, #1635, Fix embedding through views that have COALESCE with subselect - @wolfgangwalther
- #1572, Fix parsing of boolean config values for Docker environment variables, now it accepts double quoted truth values ("true", "false") and numbers("1", "0") - @wolfgangwalther
- #1624, Fix using `app.settings.xxx` config options in Docker, now they can be used as `PGRST_APP_SETTINGS_xxx` - @wolfgangwalther
- #1814, Fix panic when attempting to run with unix socket on non-unix host and properly close unix domain socket on exit - @monacoremo
- #1825, Disregard internal junction(in non-exposed schema) when embedding - @steve-chavez
- #1846, Fix requests for overloaded functions from html forms to no longer hang (#1848) - @laurenceisla
- #1858, Add a hint and clarification to the no relationship found error - @laurenceisla
- #1841, Show comprehensive error when an RPC is not found in a stale schema cache - @laurenceisla
- #1875, Fix Location headers in headers only representation for null PK inserts on views - @laurenceisla
### Changed
- #1522, #1528, #1535, Docker images are now built from scratch based on a the static PostgREST executable (#1494) and with Nix instead of a `Dockerfile`. This reduces the compressed image size from over 30mb to about 4mb - @monacoremo
- #1461, Location header for POST request is only included when PK is available on the table - @wolfgangwalther
- #1560, Volatile RPC called with GET now returns 405 Method not Allowed instead of 500 - @wolfgangwalther
- #1584, #1849 Functions that declare `returns composite_type` no longer return a single object array by default, only functions with `returns setof composite_type` return an array of objects - @wolfgangwalther
- #1604, Change the default logging level to `log-level=error`. Only requests with a status greater or equal than 500 will be logged. If you wish to go back to the previous behaviour and log all the requests, use `log-level=info` - @steve-chavez
+ Because currently there's no buffering for logging, defaulting to the `error` level(minimum logging) increases throughput by around 15% for simple GET queries(no embedding, with filters applied).
- #1617, Dropped support for PostgreSQL 9.4 - @wolfgangwalther
- #1679, Renamed config settings with fallback aliases. e.g. `max-rows` is now `db-max-rows`, but `max-rows` is still accepted - @wolfgangwalther
- #1656, Allow `Prefer=headers-only` on POST requests and change default to `minimal` (#1813) - @laurenceisla
- #1854, Dropped undocumented support for gzip compression (which was surprisingly slow and easily enabled by mistake). In some use-cases this makes Postgres up to 3x faster - @aljungberg
- #1872, Send startup/worker logs to stderr to differentiate from access logs on stdout - @steve-chavez
## [7.0.1] - 2020-05-18
### Fixed
@@ -19,6 +124,10 @@ This project adheres to [Semantic Versioning](http://semver.org/).
- #1508, Fix `Content-Profile` not working for POST RPC - @steve-chavez
- #1452, Fix PUT restriction for all columns - @steve-chavez
### Changed
- From this version onwards, the release page will only include a single Linux static executable that can be run on any Linux distribution.
## [7.0.0] - 2020-04-03
### Added
+132
View File
@@ -0,0 +1,132 @@
# Contributor Covenant Code of Conduct
## Our Pledge
We as members, contributors, and leaders pledge to make participation in our
community a harassment-free experience for everyone, regardless of age, body
size, visible or invisible disability, ethnicity, sex characteristics, gender
identity and expression, level of experience, education, socio-economic status,
nationality, personal appearance, race, caste, color, religion, or sexual identity
and orientation.
We pledge to act and interact in ways that contribute to an open, welcoming,
diverse, inclusive, and healthy community.
## Our Standards
Examples of behavior that contributes to a positive environment for our
community include:
* Demonstrating empathy and kindness toward other people
* Being respectful of differing opinions, viewpoints, and experiences
* Giving and gracefully accepting constructive feedback
* Accepting responsibility and apologizing to those affected by our mistakes,
and learning from the experience
* Focusing on what is best not just for us as individuals, but for the
overall community
Examples of unacceptable behavior include:
* The use of sexualized language or imagery, and sexual attention or
advances of any kind
* Trolling, insulting or derogatory comments, and personal or political attacks
* Public or private harassment
* Publishing others' private information, such as a physical or email
address, without their explicit permission
* Other conduct which could reasonably be considered inappropriate in a
professional setting
## Enforcement Responsibilities
Community leaders are responsible for clarifying and enforcing our standards of
acceptable behavior and will take appropriate and fair corrective action in
response to any behavior that they deem inappropriate, threatening, offensive,
or harmful.
Community leaders have the right and responsibility to remove, edit, or reject
comments, commits, code, wiki edits, issues, and other contributions that are
not aligned to this Code of Conduct, and will communicate reasons for moderation
decisions when appropriate.
## Scope
This Code of Conduct applies within all community spaces, and also applies when
an individual is officially representing the community in public spaces.
Examples of representing our community include using an official e-mail address,
posting via an official social media account, or acting as an appointed
representative at an online or offline event.
## Enforcement
Instances of abusive, harassing, or otherwise unacceptable behavior may be
reported to the community leaders responsible for enforcement at support@postgrest.org.
All complaints will be reviewed and investigated promptly and fairly.
All community leaders are obligated to respect the privacy and security of the
reporter of any incident.
## Enforcement Guidelines
Community leaders will follow these Community Impact Guidelines in determining
the consequences for any action they deem in violation of this Code of Conduct:
### 1. Correction
**Community Impact**: Use of inappropriate language or other behavior deemed
unprofessional or unwelcome in the community.
**Consequence**: A private, written warning from community leaders, providing
clarity around the nature of the violation and an explanation of why the
behavior was inappropriate. A public apology may be requested.
### 2. Warning
**Community Impact**: A violation through a single incident or series
of actions.
**Consequence**: A warning with consequences for continued behavior. No
interaction with the people involved, including unsolicited interaction with
those enforcing the Code of Conduct, for a specified period of time. This
includes avoiding interactions in community spaces as well as external channels
like social media. Violating these terms may lead to a temporary or
permanent ban.
### 3. Temporary Ban
**Community Impact**: A serious violation of community standards, including
sustained inappropriate behavior.
**Consequence**: A temporary ban from any sort of interaction or public
communication with the community for a specified period of time. No public or
private interaction with the people involved, including unsolicited interaction
with those enforcing the Code of Conduct, is allowed during this period.
Violating these terms may lead to a permanent ban.
### 4. Permanent Ban
**Community Impact**: Demonstrating a pattern of violation of community
standards, including sustained inappropriate behavior, harassment of an
individual, or aggression toward or disparagement of classes of individuals.
**Consequence**: A permanent ban from any sort of public interaction within
the community.
## Attribution
This Code of Conduct is adapted from the [Contributor Covenant][homepage],
version 2.0, available at
[https://www.contributor-covenant.org/version/2/0/code_of_conduct.html][v2.0].
Community Impact Guidelines were inspired by
[Mozilla's code of conduct enforcement ladder][Mozilla CoC].
For answers to common questions about this code of conduct, see the FAQ at
[https://www.contributor-covenant.org/faq][FAQ]. Translations are available
at [https://www.contributor-covenant.org/translations][translations].
[homepage]: https://www.contributor-covenant.org
[v2.0]: https://www.contributor-covenant.org/version/2/0/code_of_conduct.html
[Mozilla CoC]: https://github.com/mozilla/diversity
[FAQ]: https://www.contributor-covenant.org/faq
[translations]: https://www.contributor-covenant.org/translations
-41
View File
@@ -1,41 +0,0 @@
.PHONY: commit-check check clean lint style test test-watch coverage circleci circleci-prof-test check-dburi prompt-clean prompt-long-process
commit-check: lint style
check: lint style test
clean: prompt-clean
stack clean --full
# For running these you'll need to install hlint and stylish-haskell first. Run:
# stack install hlint stylish-haskell
lint:
git ls-files | grep '\.l\?hs$$' | xargs stack exec -- hlint -X QuasiQuotes -X NoPatternSynonyms "$$@"
style:
git ls-files | grep '\.l\?hs$$' | xargs stack exec -- stylish-haskell -i && git diff-index --exit-code HEAD -- '*.hs' '*.lhs'
test: check-dburi
stack test
test-watch: check-dburi
stack build --file-watch --test --test-arguments '--rerun --failure-report=.TESTREPORT --rerun-all-on-success'
coverage: check-dburi clean
stack build --coverage
stack test --coverage
circleci: prompt-long-process
circleci local execute --job build-test-9.4
circleci-prof-test: prompt-long-process
circleci local execute --job build-prof-test
check-dburi:
test -n "$(POSTGREST_TEST_CONNECTION)" # Requires POSTGREST_TEST_CONNECTION environmental variable
prompt-clean:
@echo -n 'Are you sure? You will have to rebuild. [y/N] ' && read ans && [ $${ans:-N} = y ]
prompt-long-process:
@echo -n 'Are you sure? This might take a while. [y/N] ' && read ans && [ $${ans:-N} = y ]
+27 -20
View File
@@ -2,13 +2,11 @@
[![Donate](https://img.shields.io/badge/Donate-Patreon-orange.svg?colorB=F96854)](https://www.patreon.com/postgrest)
[![Donate](https://img.shields.io/badge/Donate-PayPal-green.svg)](https://www.paypal.me/postgrest)
<a href="https://heroku.com/deploy?template=https://github.com/PostgREST/postgrest">
<img src="https://img.shields.io/badge/%E2%86%91_Deploy_to-Heroku-7056bf.svg" alt="Deploy">
</a>
[![Join the chat at https://gitter.im/begriffs/postgrest](https://img.shields.io/badge/gitter-join%20chat%20%E2%86%92-brightgreen.svg)](https://gitter.im/begriffs/postgrest)
[![Docs](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://postgrest.org)
[![Docker Stars](https://img.shields.io/docker/pulls/postgrest/postgrest.svg)](https://hub.docker.com/r/postgrest/postgrest/)
[![Build Status](https://circleci.com/gh/PostgREST/postgrest/tree/master.svg?style=shield)](https://circleci.com/gh/PostgREST/postgrest/tree/master)
[![Build Status](https://github.com/postgrest/postgrest/actions/workflows/ci.yaml/badge.svg?branch=main)](https://github.com/PostgREST/postgrest/actions?query=branch%3Amain)
[![Coverage Status](https://img.shields.io/codecov/c/github/postgrest/postgrest/main)](https://app.codecov.io/gh/PostgREST/postgrest)
[![Hackage docs](https://img.shields.io/hackage/v/postgrest.svg?label=hackage)](http://hackage.haskell.org/package/postgrest)
PostgREST serves a fully RESTful API from any existing PostgreSQL
@@ -21,18 +19,36 @@ API than you are likely to write from scratch.
<tbody>
<tr>
<td align="center" valign="middle">
<a href="https://www.cybertec-postgresql.com/en/" target="_blank">
<img width="222px" src="static/cybertec.png">
<a href="https://www.cybertec-postgresql.com/en/?utm_source=postgrest.org&utm_medium=referral&utm_campaign=postgrest" target="_blank">
<img width="222px" src="static/cybertec-new.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://www.2ndquadrant.com/en/?utm_campaign=External%20Websites&utm_source=PostgREST&utm_medium=Logo" target="_blank">
<img width="222px" src="static/2ndquadrant.png">
<img width="296px" src="static/2ndquadrant.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://tryretool.com/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="222px" src="static/retool.png">
<img width="296px" src="static/retool.png">
</a>
</td>
</tr>
<tr></tr>
<tr>
<td align="center" valign="middle">
<a href="https://gnuhost.eu/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="static/gnuhost.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://supabase.io?utm_source=postgrest%20backers&utm_medium=open%20source%20partner&utm_campaign=postgrest%20backers%20github&utm_term=homepage" target="_blank">
<img width="296px" src="static/supabase.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://oblivious.ai/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="static/oblivious.jpg">
</a>
</td>
</tr>
@@ -57,7 +73,7 @@ Latest documentation is at [postgrest.org](http://postgrest.org). You can contri
## Performance
TLDR; subsecond response times for up to 2000 requests/sec on Heroku
free tier. If you're used to servers written in interpreted languages,
free tier. If you're used to servers written in interpreted languages,
prepare to be pleasantly surprised by PostgREST performance.
Three factors contribute to the speed. First the server is written
@@ -93,15 +109,6 @@ the connection cannot do anything the user themselves couldn't. Other
forms of authentication can be built on top of the JWT primitive. See
the docs for more information.
PostgreSQL 9.5 supports true [row-level
security](http://www.postgresql.org/docs/9.5/static/ddl-rowsecurity.html).
In previous versions it can be simulated with triggers and
security-barrier views. Because the possible queries to the database
are limited to certain templates using
[leakproof](http://blog.2ndquadrant.com/how-do-postgresql-security_barrier-views-work/)
functions, the trigger workaround does not compromise row-level
security.
## Versioning
A robust long-lived API needs the freedom to exist in multiple
@@ -150,8 +157,8 @@ Every donation will be spent on making PostgREST better for the whole community.
The PostgREST organization is grateful to:
- The project [sponsors and backers](https://github.com/PostgREST/postgrest/blob/master/BACKERS.md) who support PostgREST's development.
- The project [sponsors and backers](https://github.com/PostgREST/postgrest/blob/main/BACKERS.md) who support PostgREST's development.
- The project [contributors](https://github.com/PostgREST/postgrest/graphs/contributors) who have improved PostgREST immensely with their code
and good judgement. See more details in the [changelog](https://github.com/PostgREST/postgrest/blob/master/CHANGELOG.md).
and good judgement. See more details in the [changelog](https://github.com/PostgREST/postgrest/blob/main/CHANGELOG.md).
The cool logo came from [Mikey Casalaina](https://github.com/casalaina).
+1
View File
@@ -1,2 +1,3 @@
-- This file is required by Hackage.
import Distribution.Simple
main = defaultMain
-59
View File
@@ -1,59 +0,0 @@
{
"name": "PostgREST",
"description": "RESTful API for any PostgreSQL database.",
"logo": "https://avatars2.githubusercontent.com/u/15115011",
"repository": "https://github.com/PostgREST/postgrest",
"env": {
"BUILDPACK_URL": {
"description": "Heroku buildpack for deploying Haskell applications",
"value": "https://github.com/PostgREST/postgrest-heroku"
},
"POSTGREST_VER": {
"description": "Version of PostgREST to deploy",
"value": "7.0.1"
},
"DB_URI": {
"description": "Database connection string, e.g. postgres://user:pass@xxxxxxx.rds.amazonaws.com/mydb",
"required": true
},
"DB_SCHEMA": {
"description": "The database schema to expose to REST clients. Tables, views and stored procedures in this schema will get API endpoints",
"required": true,
"value": "public"
},
"DB_ANON_ROLE": {
"description": "The database role to use when executing commands on behalf of unauthenticated clients",
"required": true
},
"DB_POOL": {
"description": "Number of connections to keep open in PostgRESTs database pool",
"required": false,
"value": "10"
},
"SERVER_PROXY_URI": {
"description": "Overrides the base URL used within the OpenAPI self-documentation hosted at the API root path",
"required": false
},
"JWT_SECRET": {
"description": "The secret used to decode JWT tokens clients provide for authentication",
"required": false
},
"SECRET_IS_BASE64": {
"description": "When this is set to true, the value derived from jwt-secret will be treated as a base64 encoded secret",
"required": false,
"value": "false"
},
"JWT_AUD": {
"description": "The audience that should be validated if the JWT token contains an aud claim",
"required": false
},
"MAX_ROWS": {
"description": "A hard limit to the number of rows PostgREST will fetch from a view, table, or stored procedure",
"required": false
},
"PRE_REQUEST": {
"description": "A schema-qualified stored procedure name to call right after switching roles for a client request",
"required": false
}
}
}
-36
View File
@@ -1,36 +0,0 @@
## AppVeyor is only used for building a Windows binary, no tests are run here.
platform: x64
image: Visual Studio 2015
cache:
- "c:\\sr"
- .stack-work
- "c:\\Users\\appveyor\\AppData\\Local\\Programs\\stack"
environment:
global:
STACK_ROOT: "c:\\sr"
GOPATH: c:\gopath
TMP: "c:\\tmp"
test: off
install:
- set PATH=C:\Program Files\PostgreSQL\9.6\bin\;%PATH%
- curl -sS -ostack.zip -L --insecure http://www.stackage.org/stack/windows-x86_64
- 7z x stack.zip stack.exe
- set PATH=%GOPATH%\bin;c:\go\bin;%PATH%
- go get -u github.com/tcnksm/ghr
build_script:
- stack setup --no-terminal > nul
# Appveyor has a timeout of 60 mins, building can take longer, limit the time and make sure this succeeds,
# previous work will get cached and finish on next commit
- bash -lc "timeout 2700 'C:\projects\postgrest\stack.exe' build -j1 --copy-bins --local-bin-path . || true"
artifacts:
- path: postgrest.exe
deploy_script:
- IF DEFINED APPVEYOR_REPO_TAG_NAME 7z a -tzip postgrest-%APPVEYOR_REPO_TAG_NAME%-windows-x64.zip postgrest.exe
- IF DEFINED APPVEYOR_REPO_TAG_NAME bash -lc " exec 0</dev/null && cd $APPVEYOR_BUILD_FOLDER && ghr -t $GITHUB_TOKEN -u $APPVEYOR_ACCOUNT_NAME -r $APPVEYOR_PROJECT_NAME -b \"$(sed -n \"1,/$(echo $APPVEYOR_REPO_TAG_NAME | cut -c2-)/d;/## \[/q;p\" CHANGELOG.md)\" --replace $APPVEYOR_REPO_TAG_NAME postgrest-$APPVEYOR_REPO_TAG_NAME-windows-x64.zip"
+93 -46
View File
@@ -3,7 +3,7 @@ let
"postgrest";
compiler =
"ghc883";
"ghc8107";
# PostgREST source files, filtered based on the rules in the .gitignore files
# and file extensions. We want to include as litte as possible, as the files
@@ -26,11 +26,18 @@ let
sha256 = nixpkgsVersion.tarballHash;
};
allOverlays =
import nix/overlays;
overlays =
[
(import nix/overlays/postgresql-legacy.nix)
(import nix/overlays/gitignore.nix)
(import nix/overlays/haskell-packages)
allOverlays.build-toolbox
allOverlays.checked-shell-script
allOverlays.gitignore
allOverlays.postgresql-default
allOverlays.postgresql-legacy
allOverlays.postgresql-future
(allOverlays.haskell-packages { inherit compiler; })
];
# Evaluated expression of the Nixpkgs repository.
@@ -39,27 +46,40 @@ let
postgresqlVersions =
[
pkgs.postgresql_12
pkgs.postgresql_11
pkgs.postgresql_10
pkgs.postgresql_9_6
pkgs.postgresql_9_5
pkgs.postgresql_9_4
{ name = "postgresql-14"; postgresql = pkgs.postgresql_14; }
{ name = "postgresql-13"; postgresql = pkgs.postgresql_13; }
{ name = "postgresql-12"; postgresql = pkgs.postgresql_12; }
{ name = "postgresql-11"; postgresql = pkgs.postgresql_11; }
{ name = "postgresql-10"; postgresql = pkgs.postgresql_10; }
{ name = "postgresql-9.6"; postgresql = pkgs.postgresql_9_6; }
];
patches =
pkgs.callPackage nix/patches {};
pkgs.callPackage nix/patches { };
# Base dynamic derivation for the PostgREST package.
drv =
pkgs.haskellPackages.callCabal2nix name src {};
# Dynamic derivation for PostgREST
postgrest =
pkgs.haskell.packages."${compiler}".callCabal2nix name src { };
# Static derivation for the PostgREST executable.
drvStatic =
import nix/static { inherit nixpkgs name src compiler patches; };
# Function that derives a fully static Haskell package based on
# nh2/static-haskell-nix
staticHaskellPackage =
import nix/static-haskell-package.nix { inherit nixpkgs compiler patches allOverlays; };
lib =
pkgs.haskell.lib;
# Options passed to cabal in dev tools and tests
devCabalOptions =
"-f dev --test-show-detail=direct";
profiledHaskellPackages =
pkgs.haskell.packages."${compiler}".extend (self: super:
{
mkDerivation =
args:
super.mkDerivation (args // { enableLibraryProfiling = true; });
}
);
inherit (pkgs.haskell) lib;
in
rec {
inherit nixpkgs pkgs;
@@ -67,43 +87,70 @@ rec {
# Derivation for the PostgREST Haskell package, including the executable,
# libraries and documentation. We disable running the test suite on Nix
# builds, as they require a database to be set up.
postgrestWithLib =
lib.dontCheck drv;
# Derivation for just the PostgREST binary, where we strip all dynamic
# libraries and documentation, leaving only the executable. Note that the
# executable is static with regards to Haskell libraries, but not system
# libraries like glibc and libpq.
postgrest =
lib.justStaticExecutables postgrestWithLib;
postgrestPackage =
lib.dontCheck postgrest;
# Static executable.
postgrestStatic =
lib.justStaticExecutables (lib.dontCheck drvStatic);
lib.justStaticExecutables (lib.dontCheck (staticHaskellPackage name src));
# Docker image and loading script.
# Profiled dynamic executable.
postgrestProfiled =
lib.enableExecutableProfiling (
lib.dontHaddock (
lib.dontCheck (profiledHaskellPackages.callCabal2nix name src { })
)
);
inherit (postgrest) env;
# Tooling for analyzing Haskell imports and exports.
hsie =
pkgs.callPackage nix/hsie {
inherit (pkgs.haskell.packages."${compiler}") ghcWithPackages;
};
### Tools
cabalTools =
pkgs.callPackage nix/tools/cabalTools.nix { inherit devCabalOptions postgrest; };
# Development tools.
devTools =
pkgs.callPackage nix/tools/devTools.nix { inherit tests style devCabalOptions hsie withTools; };
# Docker images and loading script.
docker =
pkgs.callPackage nix/docker { postgrest = postgrestStatic; };
pkgs.callPackage nix/tools/docker { postgrest = postgrestStatic; };
# Environment in which PostgREST can be built with cabal, useful e.g. for
# defining a shell for nix-shell.
env =
drv.env;
# Load testing tools.
loadtest =
pkgs.callPackage nix/tools/loadtest.nix { inherit withTools; };
# Script for running memory tests.
memory =
pkgs.callPackage nix/tools/memory.nix { inherit postgrestProfiled withTools; };
# Utility for updating the pinned version of Nixpkgs.
nixpkgsUpgrade =
pkgs.callPackage nix/nixpkgs-upgrade.nix {};
nixpkgsTools =
pkgs.callPackage nix/tools/nixpkgsTools.nix { };
tests =
pkgs.callPackage nix/tests.nix
{
inherit postgresqlVersions;
postgrestBuildEnv = env;
};
# Scripts for publishing new releases.
release =
pkgs.callPackage nix/tools/release { };
# Linting and styling tools.
style =
pkgs.callPackage nix/style.nix {};
pkgs.callPackage nix/tools/style.nix { inherit hsie; };
lint =
pkgs.callPackage nix/lint.nix {};
# Scripts for running tests.
tests =
pkgs.callPackage nix/tools/tests.nix {
inherit postgrest devCabalOptions withTools;
ghc = pkgs.haskell.compiler."${compiler}";
inherit (pkgs.haskell.packages."${compiler}") hpc-codecov;
};
withTools =
pkgs.callPackage nix/tools/withTools.nix { inherit devCabalOptions postgresqlVersions postgrest; };
}
-57
View File
@@ -1,57 +0,0 @@
# To build use:
# docker build --build-arg POSTGREST_VERSION=<v5.2.0 or another version> -t postgrest ./docker/
FROM debian:buster-slim
ARG POSTGREST_VERSION
# Install libpq5
RUN apt-get -qq update && \
apt-get -qq install -y --no-install-recommends libpq5 && \
apt-get -qq clean && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# Install postgrest
RUN BUILD_DEPS="curl ca-certificates xz-utils" && \
apt-get -qq update && \
apt-get -qq install -y --no-install-recommends $BUILD_DEPS && \
cd /tmp && \
curl -SLO https://github.com/PostgREST/postgrest/releases/download/${POSTGREST_VERSION}/postgrest-${POSTGREST_VERSION}-ubuntu.tar.xz && \
tar -xJvf postgrest-${POSTGREST_VERSION}-ubuntu.tar.xz && \
mv postgrest /usr/local/bin/postgrest && \
cd / && \
apt-get -qq purge --auto-remove -y $BUILD_DEPS && \
apt-get -qq clean && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
COPY postgrest.conf /etc/postgrest.conf
ENV PGRST_DB_URI= \
PGRST_DB_SCHEMA=public \
PGRST_DB_ANON_ROLE= \
PGRST_DB_POOL=100 \
PGRST_DB_EXTRA_SEARCH_PATH=public \
PGRST_SERVER_HOST=*4 \
PGRST_SERVER_PORT=3000 \
PGRST_OPENAPI_SERVER_PROXY_URI= \
PGRST_JWT_SECRET= \
PGRST_SECRET_IS_BASE64=false \
PGRST_JWT_AUD= \
PGRST_MAX_ROWS= \
PGRST_PRE_REQUEST= \
PGRST_ROLE_CLAIM_KEY=".role" \
PGRST_ROOT_SPEC= \
PGRST_RAW_MEDIA_TYPES=
RUN groupadd -g 1000 postgrest && \
useradd -r -u 1000 -g postgrest postgrest && \
chown postgrest:postgrest /etc/postgrest.conf
USER 1000
# PostgREST reads /etc/postgrest.conf so map the configuration
# file in when you run this container
CMD exec postgrest /etc/postgrest.conf
EXPOSE 3000
-18
View File
@@ -1,18 +0,0 @@
FROM centos:centos7
RUN yum -y update
RUN yum -y install perl make automake gcc gmp-devel libffi zlib zlib-devel xz tar
RUN yum -y install yum install https://download.postgresql.org/pub/repos/yum/10/redhat/rhel-7-x86_64/pgdg-centos10-10-2.noarch.rpm
RUN yum -y install postgresql10-devel
RUN yum clean all
RUN curl -sSL https://get.haskellstack.org/ | sh
ENV PATH $PATH:/usr/pgsql-10/bin
# To disable warning when building
ENV PATH $PATH:/root/.local/bin
RUN mkdir /source
WORKDIR /source
ENTRYPOINT ["stack"]
-20
View File
@@ -1,20 +0,0 @@
FROM ubuntu:16.04
## TODO pin the stack version
#
RUN BUILD_DEPS="curl ca-certificates build-essential" && \
apt-get -qq update && \
apt-get -qqy --no-install-recommends install \
$BUILD_DEPS \
libpq-dev && \
curl -sSL https://get.haskellstack.org/ | sh && \
apt-get -qq clean && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# To disable warning when building
ENV PATH $PATH:/root/.local/bin
RUN mkdir /source
WORKDIR /source
ENTRYPOINT ["stack"]
@@ -1,20 +0,0 @@
FROM i386/ubuntu:16.04
## TODO pin the stack version
RUN BUILD_DEPS="curl ca-certificates build-essential" && \
apt-get -qq update && \
apt-get -qqy --no-install-recommends install \
$BUILD_DEPS \
libpq-dev && \
curl -sSL https://get.haskellstack.org/ | sh && \
apt-get -qq clean && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# To disable warning when building
ENV PATH $PATH:/root/.local/bin
RUN mkdir /source
WORKDIR /source
ENTRYPOINT ["stack"]
-19
View File
@@ -1,19 +0,0 @@
stgrest:
image: pg_local
ports:
- "3000:3000"
links:
- postgres:postgres
environment:
PGRST_DB_URI: postgres://app_user:password@postgres:5432/app_db
PGRST_DB_SCHEMA: public
PGRST_DB_ANON_ROLE: app_user
postgres:
image: postgres
ports:
- "5432:5432"
environment:
POSTGRES_DB: app_db
POSTGRES_USER: app_user
POSTGRES_PASSWORD: password
-19
View File
@@ -1,19 +0,0 @@
db-uri = "$(PGRST_DB_URI)"
db-schema = "$(PGRST_DB_SCHEMA)"
db-anon-role = "$(PGRST_DB_ANON_ROLE)"
db-pool = "$(PGRST_DB_POOL)"
db-extra-search-path = "$(PGRST_DB_EXTRA_SEARCH_PATH)"
server-host = "$(PGRST_SERVER_HOST)"
server-port = "$(PGRST_SERVER_PORT)"
openapi-server-proxy-uri = "$(PGRST_OPENAPI_SERVER_PROXY_URI)"
jwt-secret = "$(PGRST_JWT_SECRET)"
secret-is-base64 = "$(PGRST_SECRET_IS_BASE64)"
jwt-aud = "$(PGRST_JWT_AUD)"
role-claim-key = "$(PGRST_ROLE_CLAIM_KEY)"
max-rows = "$(PGRST_MAX_ROWS)"
pre-request = "$(PGRST_PRE_REQUEST)"
root-spec = "$(PGRST_ROOT_SPEC)"
raw-media-types = "$(PGRST_RAW_MEDIA_TYPES)"
+31 -314
View File
@@ -1,330 +1,47 @@
{-# LANGUAGE CPP #-}
module Main where
module Main (main) where
import qualified Data.ByteString as BS
import qualified Data.ByteString.Base64 as B64
import qualified Hasql.Pool as P
import qualified Hasql.Transaction.Sessions as HT
import qualified Data.Map.Strict as M
import Control.AutoUpdate (defaultUpdateSettings, mkAutoUpdate,
updateAction)
import Control.Retry (RetryStatus, capDelay,
exponentialBackoff, retrying,
rsPreviousDelay)
import Data.Either.Combinators (whenLeft)
import Data.IORef (IORef, atomicWriteIORef, newIORef,
readIORef)
import Data.String (IsString (..))
import Data.Text (pack, replace, strip, stripPrefix)
import Data.Text.IO (hPutStrLn)
import Data.Time.Clock (getCurrentTime)
import Network.Wai.Handler.Warp (defaultSettings, runSettings,
setHost, setPort, setServerName)
import System.IO (BufferMode (..), hSetBuffering)
import System.IO (BufferMode (..), hSetBuffering)
import PostgREST.App (postgrest)
import PostgREST.Config (AppConfig (..), configPoolTimeout',
prettyVersion, readOptions)
import PostgREST.DbStructure (getDbStructure, getPgVersion)
import PostgREST.Error (PgError (PgError), checkIsFatal,
errorPayload)
import PostgREST.OpenAPI (isMalformedProxyUri)
import PostgREST.Types (ConnectionStatus (..), DbStructure,
PgVersion (..), Schema,
minimumPgVersion)
import Protolude hiding (hPutStrLn, head, replace, toS)
import Protolude.Conv (toS)
import qualified PostgREST.App as App
import qualified PostgREST.CLI as CLI
import PostgREST.Config (readPGRSTEnvironment)
import Protolude
#ifndef mingw32_HOST_OS
import System.Posix.Signals
import UnixSocket
import qualified PostgREST.Unix as Unix
#endif
{-|
The purpose of this worker is to fill the refDbStructure created in 'main'
with the 'DbStructure' returned from calling 'getDbStructure'. This method
is meant to be called by multiple times by the same thread, but does nothing if
the previous invocation has not terminated. In all cases this method does not
halt the calling thread, the work is preformed in a separate thread.
Note: 'atomicWriteIORef' is essentially a lazy semaphore that prevents two
threads from running 'connectionWorker' at the same time.
Background thread that does the following :
1. Tries to connect to pg server and will keep trying until success.
2. Checks if the pg version is supported and if it's not it kills the main
program.
3. Obtains the dbStructure.
4. If 2 or 3 fail to give their result it means the connection is down so it
goes back to 1, otherwise it finishes his work successfully.
-}
connectionWorker
:: ThreadId -- ^ This thread is killed if pg version is unsupported
-> P.Pool -- ^ The PostgreSQL connection pool
-> [Schema] -- ^ Schemas PostgREST is serving up
-> IORef (Maybe DbStructure) -- ^ mutable reference to 'DbStructure'
-> IORef Bool -- ^ Used as a binary Semaphore
-> IO ()
connectionWorker mainTid pool schemas refDbStructure refIsWorkerOn = do
isWorkerOn <- readIORef refIsWorkerOn
unless isWorkerOn $ do
atomicWriteIORef refIsWorkerOn True
void $ forkIO work
where
work = do
atomicWriteIORef refDbStructure Nothing
putStrLn ("Attempting to connect to the database..." :: Text)
connected <- connectionStatus pool
case connected of
FatalConnectionError reason -> hPutStrLn stderr reason
>> killThread mainTid -- Fatal error when connecting
NotConnected -> return () -- Unreachable
Connected actualPgVersion -> do -- Procede with initialization
result <- P.use pool $ do
dbStructure <- HT.transaction HT.ReadCommitted HT.Read $ getDbStructure schemas actualPgVersion
liftIO $ atomicWriteIORef refDbStructure $ Just dbStructure
case result of
Left e -> do
putStrLn ("Failed to query the database. Retrying." :: Text)
hPutStrLn stderr . toS . errorPayload $ PgError False e
work
Right _ -> do
atomicWriteIORef refIsWorkerOn False
putStrLn ("Connection successful" :: Text)
{-|
Used by 'connectionWorker' to check if the provided db-uri lets
the application access the PostgreSQL database. This method is used
the first time the connection is tested, but only to test before
calling 'getDbStructure' inside the 'connectionWorker' method.
The connection tries are capped, but if the connection times out no error is
thrown, just 'False' is returned.
-}
connectionStatus :: P.Pool -> IO ConnectionStatus
connectionStatus pool =
retrying (capDelay 32000000 $ exponentialBackoff 1000000)
shouldRetry
(const $ P.release pool >> getConnectionStatus)
where
getConnectionStatus :: IO ConnectionStatus
getConnectionStatus = do
pgVersion <- P.use pool getPgVersion
case pgVersion of
Left e -> do
let err = PgError False e
hPutStrLn stderr . toS $ errorPayload err
case checkIsFatal err of
Just reason -> return $ FatalConnectionError reason
Nothing -> return NotConnected
Right version ->
if version < minimumPgVersion
then return . FatalConnectionError $ "Cannot run in this PostgreSQL version, PostgREST needs at least " <> pgvName minimumPgVersion
else return . Connected $ version
shouldRetry :: RetryStatus -> ConnectionStatus -> IO Bool
shouldRetry rs isConnSucc = do
let delay = fromMaybe 0 (rsPreviousDelay rs) `div` 1000000
itShould = NotConnected == isConnSucc
when itShould $
putStrLn $ "Attempting to reconnect to the database in " <> (show delay::Text) <> " seconds..."
return itShould
{-|
This is where everything starts.
-}
main :: IO ()
main = do
--
setBuffering
hasPGRSTEnv <- not . M.null <$> readPGRSTEnvironment
opts <- CLI.readCLIShowHelp hasPGRSTEnv
CLI.main installSignalHandlers runAppInSocket opts
installSignalHandlers :: App.SignalHandlerInstaller
#ifndef mingw32_HOST_OS
installSignalHandlers = Unix.installSignalHandlers
#else
installSignalHandlers _ = pass
#endif
runAppInSocket :: Maybe App.SocketRunner
#ifndef mingw32_HOST_OS
runAppInSocket = Just Unix.runAppWithSocket
#else
runAppInSocket = Nothing
#endif
setBuffering :: IO ()
setBuffering = do
-- LineBuffering: the entire output buffer is flushed whenever a newline is
-- output, the buffer overflows, a hFlush is issued or the handle is closed
--
-- NoBuffering: output is written immediately and never stored in the buffer
hSetBuffering stdout LineBuffering
hSetBuffering stdin LineBuffering
hSetBuffering stderr NoBuffering
--
-- readOptions builds the 'AppConfig' from the config file specified on the
-- command line
conf <- loadDbUriFile =<< loadSecretFile =<< readOptions
let schemas = toList $ configSchemas conf
host = configHost conf
port = configPort conf
proxy = configOpenAPIProxyUri conf
maybeSocketAddr = configSocket conf
socketFileMode = configSocketMode conf
pgSettings = toS (configDatabase conf) -- is the db-uri
roleClaimKey = configRoleClaimKey conf
appSettings =
setHost ((fromString . toS) host) -- Warp settings
. setPort port
. setServerName (toS $ "postgrest/" <> prettyVersion) $
defaultSettings
whenLeft socketFileMode panic
-- Checks that the provided proxy uri is formated correctly
when (isMalformedProxyUri $ toS <$> proxy) $
panic
"Malformed proxy uri, a correct example: https://example.com:8443/basePath"
-- Checks that the provided jspath is valid
whenLeft roleClaimKey $
panic $ show roleClaimKey
-- create connection pool with the provided settings, returns either
-- a 'Connection' or a 'ConnectionError'. Does not throw.
pool <- P.acquire (configPool conf, configPoolTimeout' conf, pgSettings)
--
-- To be filled in by connectionWorker
refDbStructure <- newIORef Nothing
--
-- Helper ref to make sure just one connectionWorker can run at a time
refIsWorkerOn <- newIORef False
--
-- This is passed to the connectionWorker method so it can kill the main
-- thread if the PostgreSQL's version is not supported.
mainTid <- myThreadId
--
-- Sets the refDbStructure
connectionWorker
mainTid
pool
schemas
refDbStructure
refIsWorkerOn
--
-- Only for systems with signals:
--
-- releases the connection pool whenever the program is terminated,
-- see issue #268
--
-- Plus the SIGHUP signal updates the internal 'DbStructure' by running
-- 'connectionWorker' exactly as before.
#ifndef mingw32_HOST_OS
forM_ [sigINT, sigTERM] $ \sig ->
void $ installHandler sig (Catch $ do
P.release pool
throwTo mainTid UserInterrupt
) Nothing
void $ installHandler sigUSR1 (
Catch $ connectionWorker
mainTid
pool
schemas
refDbStructure
refIsWorkerOn
) Nothing
#endif
-- ask for the OS time at most once per second
getTime <- mkAutoUpdate defaultUpdateSettings {updateAction = getCurrentTime}
let postgrestApplication =
postgrest
conf
refDbStructure
pool
getTime
(connectionWorker
mainTid
pool
schemas
refDbStructure
refIsWorkerOn)
-- run the postgrest application with user defined socket. Only for UNIX systems.
#ifndef mingw32_HOST_OS
whenJust maybeSocketAddr $
runAppInSocket appSettings postgrestApplication socketFileMode
#endif
-- run the postgrest application
whenNothing maybeSocketAddr $ do
putStrLn $ ("Listening on port " :: Text) <> show (configPort conf)
runSettings appSettings postgrestApplication
{-|
The purpose of this function is to load the JWT secret from a file if
configJwtSecret is actually a filepath and replaces some characters if the JWT
is base64 encoded.
The reason some characters need to be replaced is because JWT is actually
base64url encoded which must be turned into just base64 before decoding.
To check if the JWT secret is provided is in fact a file path, it must be
decoded as 'Text' to be processed.
decodeUtf8: Decode a ByteString containing UTF-8 encoded text that is known to
be valid.
-}
loadSecretFile :: AppConfig -> IO AppConfig
loadSecretFile conf = extractAndTransform mSecret
where
mSecret = decodeUtf8 <$> configJwtSecret conf
isB64 = configJwtSecretIsBase64 conf
--
-- The Text (variable name secret) here is mSecret from above which is the JWT
-- decoded as Utf8
--
-- stripPrefix: Return the suffix of the second string if its prefix matches
-- the entire first string.
--
-- The configJwtSecret is a filepath instead of the JWT secret itself if the
-- secret has @ as its prefix.
extractAndTransform :: Maybe Text -> IO AppConfig
extractAndTransform Nothing = return conf
extractAndTransform (Just secret) =
fmap setSecret $
transformString isB64 =<<
case stripPrefix "@" secret of
Nothing -> return . encodeUtf8 $ secret
Just filename -> chomp <$> BS.readFile (toS filename)
where
chomp bs = fromMaybe bs (BS.stripSuffix "\n" bs)
--
-- Turns the Base64url encoded JWT into Base64
transformString :: Bool -> ByteString -> IO ByteString
transformString False t = return t
transformString True t =
case B64.decode $ encodeUtf8 $ strip $ replaceUrlChars $ decodeUtf8 t of
Left errMsg -> panic $ pack errMsg
Right bs -> return bs
setSecret bs = conf {configJwtSecret = Just bs}
--
-- replace: Replace every occurrence of one substring with another
replaceUrlChars =
replace "_" "/" . replace "-" "+" . replace "." "="
{-
Load database uri from a separate file if `db-uri` is a filepath.
-}
loadDbUriFile :: AppConfig -> IO AppConfig
loadDbUriFile conf = extractDbUri mDbUri
where
mDbUri = configDatabase conf
extractDbUri :: Text -> IO AppConfig
extractDbUri dbUri =
fmap setDbUri $
case stripPrefix "@" dbUri of
Nothing -> return dbUri
Just filename -> strip <$> readFile (toS filename)
setDbUri dbUri = conf {configDatabase = dbUri}
-- Utilitarian functions.
whenJust :: Applicative f => Maybe a -> (a -> f ()) -> f ()
whenJust (Just x) f = f x
whenJust Nothing _ = pass
whenNothing :: Applicative f => Maybe a -> f () -> f ()
whenNothing Nothing f = f
whenNothing _ _ = pass
hSetBuffering stderr LineBuffering
-40
View File
@@ -1,40 +0,0 @@
module UnixSocket (
runAppInSocket
)where
import Network.Socket (Family (AF_UNIX),
SockAddr (SockAddrUnix), Socket,
SocketType (Stream), bind, close,
defaultProtocol, listen,
maxListenQueue, socket)
import Network.Wai (Application)
import Network.Wai.Handler.Warp
import System.Directory (removeFile)
import System.IO.Error (isDoesNotExistError)
import System.Posix.Files (setFileMode)
import System.Posix.Types (FileMode)
import Protolude
createAndBindSocket :: FilePath -> Maybe FileMode -> IO Socket
createAndBindSocket socketFilePath maybeSocketFileMode = do
deleteSocketFileIfExist socketFilePath
sock <- socket AF_UNIX Stream defaultProtocol
bind sock $ SockAddrUnix socketFilePath
mapM_ (setFileMode socketFilePath) maybeSocketFileMode
return sock
where
deleteSocketFileIfExist path = removeFile path `catch` handleDoesNotExist
handleDoesNotExist e
| isDoesNotExistError e = return ()
| otherwise = throwIO e
-- run the postgrest application with user defined socket.
runAppInSocket :: Settings -> Application -> Either Text FileMode -> FilePath -> IO ()
runAppInSocket settings app socketFileMode sockPath = do
sock <- createAndBindSocket sockPath (rightToMaybe socketFileMode)
putStrLn $ ("Listening on unix socket " :: Text) <> show sockPath
listen sock maxListenQueue
runSettingsSocket settings sock app
-- clean socket up when done
close sock
+22
View File
@@ -0,0 +1,22 @@
# This Dockerfile is only used as a development environment for
# non-nix systems, i.e. Windows.
FROM nixos/nix:latest
RUN apk --no-cache add \
wget
RUN nix-env -iA cachix -f https://cachix.org/api/v1/install \
&& cachix use postgrest
# We need an unprivileged user here, to make PG run at all.
RUN adduser --disabled-password --ingroup root nix \
&& chown -R nix:root /nix
USER nix:root
ENV USER=nix
VOLUME /nix
VOLUME /postgrest
WORKDIR /postgrest
CMD nix-shell
+123 -26
View File
@@ -12,21 +12,43 @@ their dependencies will be stored. It will also link the Nix executables like
other PostgREST dependencies from here on out. To clean up older build
artifacts from the `/nix/store`, you can run `nix-collect-garbage`.
If you are on a system that does not support nix, for example Windows, you can
run the nix development environment in a docker container. Inside the `nix/`
directory run `docker-compose run --rm nix` to start the docker container. This
will set up the binary cache and launch `nix-shell` automatically.
## Building PostgREST
To build PostgREST from your local checkout of the repository, run:
```bash
nix-build --attr postgrest
nix-build --attr postgrestPackage
```
This will create a `result` directory that contains the PostgREST binary at
`result/bin/postgrest`. The `--attr` parameter (or short: `-A`) tells Nix to
build the `postgrest` attribute from the Nix expression it finds in our
build the `postgrestPackage` attribute from the Nix expression it finds in our
`default.nix` (see below for details). Nix will take care of getting the right
GHC version and all the build dependencies.
## Binary cache
We recommend that you use the PostgREST binary cache on
[cachix](https://cachix.org/):
```bash
# Install cachix:
nix-env -iA cachix -f https://cachix.org/api/v1/install
# Set cachix up to use the PostgREST binary cache:
cachix use postgrest
```
Without cachix, your machine will have to rebuild all the dependencies that are
derived on top of `Musl` for the static builds, which can take a very long time.
## Developing
A development environment for PostgREST is available with `nix-shell`. The
@@ -42,32 +64,61 @@ Within `nix-shell`, you can run Cabal commands as usual. You can also run
stack with the `--nix` option, which causes stack to pick up the non-Haskell
dependencies from the same pinned Nixpkgs version that the Nix builds use.
## Aside: Working with `nix-shell` and the PostgREST utility scripts
## Working with `nix-shell` and the PostgREST utility scripts
The PostgREST utilities available in `nix-shell` all have names that begin
with `postgrest-`, so you can use tab completion (`postgrest-<tab>`) in
`nix-shell` to see all that are available:
The PostgREST utilities available in `nix-shell` all have names that begin with
`postgrest-`, so you can use tab completion (typing `postgrest-` and pressing
`<tab>`) in `nix-shell` to see all that are available:
```
```bash
# Note: The utilities listed here might not be up to date.
[nix-shell:~/Projects/postgrest]$ postgrest-
postgrest-lint postgrest-test-spec-postgresql-10.12
postgrest-style postgrest-test-spec-postgresql-11.7
postgrest-style-check postgrest-test-spec-postgresql-12.2
postgrest-test-all postgrest-test-spec-postgresql-9.4.24
postgrest-test-spec postgrest-test-spec-postgresql-9.5.21
postgrest-test-spec-all postgrest-test-spec-postgresql-9.6.17
[nix-shell]$ postgrest-<tab>
postgrest-build postgrest-test-spec
postgrest-check postgrest-watch
postgrest-clean postgrest-with-all
postgrest-coverage postgrest-with-postgresql-10
postgrest-lint postgrest-with-postgresql-11
postgrest-run postgrest-with-postgresql-12
postgrest-style postgrest-with-postgresql-13
postgrest-style-check postgrest-with-postgresql-9.6
postgrest-test-io
...
[nix-shell:~/Projects/postgrest]$
[nix-shell]$
```
Some additional modules like `memory`, `docker` and `release`
have large dependencies that would need to be built before the shell becomes
available, which could take an especially long time if the cachix binary cache
is not used. You can activate those by passing a flag to `nix-shell` with
`nix-shell --arg <module> true`. This will make the respective utilites available:
```bash
$ nix-shell --arg memory true
[nix-shell]$ postgrest-<tab>
postgrest-build postgrest-test-spec
postgrest-check postgrest-watch
postgrest-clean postgrest-with-all
postgrest-coverage postgrest-with-postgresql-10
postgrest-lint postgrest-with-postgresql-11
postgrest-run postgrest-with-postgresql-12
postgrest-style postgrest-with-postgresql-13
postgrest-style-check postgrest-with-postgresql-9.6
postgrest-test-io
postgrest-test-memory
...
```
Note that `postgrest-test-memory` is now also available.
To run one-off commands, you can also use `nix-shell --run <command>`, which
will lauch the Nix shell, run that one command and exit. Note that the tab
completion will not work with `nix-shell --run`, as Nix has yet to evaluate
our Nix expressions to see which utilities are available.
```
```bash
$ nix-shell --run postgrest-style
# Note that you need to quote any arguments that you would like to pass to
@@ -78,20 +129,28 @@ $ nix-shell --run "postgrest-foo --bar"
A third option is to install utilities that you use very often locally:
```
$ nix-env -f default.nix -iA style
```bash
$ nix-env -f default.nix -iA devTools
# `postgrest-style` can now be run directly:
$ postgrest-style
```
Note that this does not yet work for all utilities (e.g. `postgrest-test-spec` currently
needs to be run within the Nix shell environment).
If you use `nix-shell` very often, you might like to use
https://github.com/xzfc/cached-nix-shell, which skips evaluating all our Nix
expressions if nothing changed, reducing startup time for the shell
considerably.
If you use `nix-shell` very often, you might like to use https://github.com/xzfc/cached-nix-shell,
which skips evaluating all our Nix expressions if nothing changed, reducing startup time for the
shell considerably.
Note: Once inside nix-shell, the utilities work from any directory inside
the PostgREST repo. Paths are resolved relative to the repo root:
```bash
$ cd src
# Even though the current directory is ./src, the config path must still start
# from the repo root:
$ postgrest-run test/io/configs/simple.conf
```
## Testing
@@ -104,11 +163,25 @@ temporary test databases:
$ nix-shell --run postgrest-test-spec
# Run the tests against all supported versions of PostgreSQL:
$ nix-shell --run postgrest-test-spec-all
$ nix-shell --run "postgrest-with-all postgrest-test-spec"
# Run the tests against a specific version of PostgreSQL (use tab-completion in
# nix-shell to see all available versions):
$ nix-shell --run postgrest-test-spec-postgresql-9.5.21
$ nix-shell --run "postgrest-with-postgresql-13 postgrest-test-spec"
```
The io-test that test PostgREST as a black box with inputs and outputs can be
run with `postgrest-test-io`. The test runner under the hood is
[pytest](https://docs.pytest.org/) and you can pass it the usual options:
```bash
# Filter the tests to run by name, including all that contain 'config':
postgrest-test-io -k config
# Run tests in parallel using xdist, specifying the number of processes:
postgrest-test-io -n auto
postgrest-test-io -n 8
```
@@ -126,6 +199,25 @@ $ nix-shell --run postgrest-style
```
There is also `postgrest-style-check` that exits with a non-zero exit code if
the check resulted in any uncommited changes. It's mostly useful for CI.
## General development tools
Tools like `postgrest-build`, `postgrest-run` etc. are simple wrappers around
`cabal` and should do what you expect. `postgrest-check` runs most checks that will
also run in CI, with the exception of the IO and Memory checks that need to be run
separately.
`postgrest-with-postgresql-*` take a command as an argument and will run it
with a temporary database. `postgrest-with-all` will run the command against
all supported PostgreSQL versions. Tests run without `postgrest-with-*` are
run against the latest PostgreSQL version by default.
`postgrest-watch` takes a command as an argument that it will re-run if any source
file is changed. For example, `postgrest-watch postgrest-with-all postgrest-test-spec`
will re-run the full spec test suite against all PostgreSQL versions on every change.
## Tour
The following is not required for working on PostgREST with Nix, but it will
@@ -133,7 +225,7 @@ give you some more background and details on how it works.
### `default.nix`
[`default.nix`](../default.nix) is our 'respository expression' that pulls all
[`default.nix`](../default.nix) is our 'repository expression' that pulls all
the pieces that we define with Nix together. It returns a set (like a dict in
other programming languages), where each attribute is a derivation that Nix
knows how to build, like the `postgrest` attribute from earlier.
@@ -161,3 +253,8 @@ attribute with useful utilities that will be put on the PATH in `nix-shell`.
Our overlays to the Nix package set are defined here. They allow us to tweak our
`pkgs` in `default.nix` by adding new packages or overriding existing ones.
## Upgrading dependencies
See the [upgrading checklist](UPGRADE.md) for how to upgrade the PostgREST
dependencies.
+91
View File
@@ -0,0 +1,91 @@
# Checklist for upgrading Nix dependencies
The Nix dependencies of PostgREST should be updated regularly, in most cases it
should be a very simple operation.
```bash
# Update pinned version of Nixpkgs
nix-shell --run postgrest-nixpkgs-upgrade
# Verify that everything builds
nix-build
```
The following checklist guides you through the complete process in more detail.
## Upgrade the pinned version of `nixpkgs`
The pinned version of [`nixpkgs`](https://github.com/NixOS/nixpkgs) is defined
in [`nix/nixpkgs-version.nix`](nixpkgs-version.nix). The pin refers directly to
a GitHub tarball for the given revision, which is more efficient than pulling
the complete Git repository. To upgrade it to the current `main` of
`nixpkgs`, you can use a small utility script defined in
[`nix/nixpkgs-update.nix`](nixpkgs-update.nix):
```bash
# From the root of the repository, enter nix-shell
nix-shell
# Run the utility script to pin the latest revision in main
postgrest-nixpkgs-upgrade
# Exit the nix-shell with Ctrl-d
```
## Update pinned version of `static-haskell-nix`
We pin [`static-haskell-nix`](https://github.com/nh2/static-haskell-nix) in
[`nix/static-haskell-package.nix`](static-haskell-package.nix). Upgrade the
pinned revision and the tarball hash if necessary. See
[`nix/nixpkgs-upgrade.nix`](nixpkgs-upgrade.nix) for how to get the correct
tarball hash, or just change the hash to an arbitrary value of correct length,
run `nix-build` and use the expected value from the resulting error message.
## Review overlays
Check whether the individual [overlays](overlays) are still required.
## Check if patches are still required and update them as needed
We track a number of PostgREST-specific patches in [`nix/patches`](patches).
Check whether the pull-requests/issues linked in the
[`default.nix`](patches/default.nix) have progressed and remove/modify the
patches if they did. If conflicting changes occurred, you might have to rebase
the respective patches.
## Build everything
Using the PostgREST binary Nix cache is recommended. Install
[Cachix](https://cachix.org/) and run `cachix use postgrest`.
Run `nix-build` in the root directory of the project to build all PostgREST
artifacts. This might take a long time, e.g. when our static GHC version needs
to be rebuilt due to changes to some underlying package. If there are any
errors, this is probably due to one of our patches. Try to fix them and re-run
`nix-build` until everything builds.
## Update the PostgREST binary cache
If you have access to the PostgREST cachix signing key, you can push the
artifacts that you built locally to the binary cache. This will accelerate the
CI builds and tests, sometimes dramatically. This might sometimes even be
required to avoid build timeouts in CI.
You'll need to set the `CACHIX_SIGNING_KEY` before proceeding, e.g. by creating
a file containing `export CACHIX_SIGNING_KEY=...` and sourcing that file, which
avoids having the secret in you shell history.
To push all new artifacts to Cachix, run:
```
nix-store -qR --include-outputs $$(nix-instantiate) | cachix push postgrest
# Or, equivalently
nix-shell --run postgrest-push-cachix
```
The `nix-store` command will query the nix-store to list all dependencies and
build artifacts of PostgREST. The `cachix` command will efficiently push
everything that is not yet cached to the binary cache.
+12
View File
@@ -0,0 +1,12 @@
version: '3'
services:
nix:
container_name: postgrest-nix
build: .
volumes:
- ../:/postgrest
- nix:/nix
volumes:
nix:
-63
View File
@@ -1,63 +0,0 @@
{ postgrest, dockerTools, writeShellScriptBin }:
let
image =
tag:
dockerTools.buildImage {
inherit tag;
name = "postgrest/postgrest";
contents = postgrest;
# Set the current time as the image creation date. This makes the build
# non-reproducible, but that should not be an issue for us.
created = "now";
extraCommands =
''
mkdir etc
cp ${./postgrest.conf} etc/postgrest.conf
'';
config = {
Cmd = [ "/bin/postgrest" "/etc/postgrest.conf" ];
Env = [
"PGRST_DB_URI=postgresql://?user=postgres"
"PGRST_DB_SCHEMA=public"
"PGRST_DB_ANON_ROLE="
"PGRST_DB_POOL=100"
"PGRST_DB_EXTRA_SEARCH_PATH=public"
"PGRST_SERVER_HOST=*4"
"PGRST_SERVER_PORT=3000"
"PGRST_OPENAPI_SERVER_PROXY_URI="
"PGRST_JWT_SECRET="
"PGRST_SECRET_IS_BASE64=false"
"PGRST_JWT_AUD="
"PGRST_MAX_ROWS="
"PGRST_PRE_REQUEST="
"PGRST_ROLE_CLAIM_KEY=.role"
"PGRST_ROOT_SPEC="
"PGRST_RAW_MEDIA_TYPES="
];
User = "1000";
ExposedPorts = {
"3000/tcp" = {};
};
};
};
in
rec {
imageLatest =
image "latest";
imageWithVersion =
image "v${postgrest.version}";
load =
writeShellScriptBin "postgrest-docker-load"
''
set -euo pipefail
docker load -i ${imageLatest}
docker load -i ${imageWithVersion}
'';
}
-19
View File
@@ -1,19 +0,0 @@
db-uri = "$(PGRST_DB_URI)"
db-schema = "$(PGRST_DB_SCHEMA)"
db-anon-role = "$(PGRST_DB_ANON_ROLE)"
db-pool = "$(PGRST_DB_POOL)"
db-extra-search-path = "$(PGRST_DB_EXTRA_SEARCH_PATH)"
server-host = "$(PGRST_SERVER_HOST)"
server-port = "$(PGRST_SERVER_PORT)"
openapi-server-proxy-uri = "$(PGRST_OPENAPI_SERVER_PROXY_URI)"
jwt-secret = "$(PGRST_JWT_SECRET)"
secret-is-base64 = "$(PGRST_SECRET_IS_BASE64)"
jwt-aud = "$(PGRST_JWT_AUD)"
role-claim-key = "$(PGRST_ROLE_CLAIM_KEY)"
max-rows = "$(PGRST_MAX_ROWS)"
pre-request = "$(PGRST_PRE_REQUEST)"
root-spec = "$(PGRST_ROOT_SPEC)"
raw-media-types = "$(PGRST_RAW_MEDIA_TYPES)"
+361
View File
@@ -0,0 +1,361 @@
{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE NamedFieldPuns #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE RecordWildCards #-}
{-# LANGUAGE TupleSections #-}
{-# LANGUAGE TypeFamilies #-}
-- | Haskell Imports and Exports tool
--
-- This tool parses imports and exports from Haskell source files and provides
-- analysis on these imports. For example, you can check whether consistent
-- import aliases are used across your codebase.
module Main (main) where
import qualified Data.Aeson as JSON
import qualified Data.ByteString.Lazy.Char8 as LBS8
import qualified Data.Csv as Csv
import qualified Data.List as List
import qualified Data.Map as Map
import qualified Data.Set as Set
import qualified Data.Text as T
import qualified Data.Text.IO as T
import qualified Dot
import qualified GHC
import qualified Language.Haskell.GHC.ExactPrint.Parsers as ExactPrint
import qualified Options.Applicative as O
import qualified System.FilePath as FP
import Bag (bagToList)
import Data.Aeson.Encode.Pretty (encodePretty)
import Data.Function ((&))
import Data.List (intercalate)
import Data.Maybe (catMaybes, mapMaybe)
import Data.Text (Text)
import GHC.Generics (Generic)
import GHC.Hs.Extension (GhcPs)
import Module (moduleNameString)
import OccName (occNameString)
import RdrName (rdrNameOcc)
import System.Directory.Recursive (getFilesRecursive)
import System.Exit (exitFailure)
-- TYPES
data Options =
Options
{ command :: Command
, sources :: [FilePath]
}
data Command
= Dump OutputFormat
| GraphSymbols
| GraphModules
| CheckAliases
| CheckWildcards [Text]
data OutputFormat = OutputCsv | OutputJson
data ImportedSymbol =
ImportedSymbol
{ impFromModule :: Text
, impModule :: Text
, impQualified :: ImportQualified
, impAlias :: Maybe Text
, impType :: ImportType
, impSymbol :: Maybe Text
, impInternal :: ModuleInternal
, impSource :: FilePath
, impFile :: FilePath
}
deriving (Generic, Csv.ToNamedRecord, Csv.DefaultOrdered, JSON.ToJSON)
data ImportQualified
= Qualified
| NotQualified
deriving (Eq, Generic, JSON.ToJSON)
instance Csv.ToField ImportQualified where
toField Qualified = "qualified"
toField NotQualified = "not qualified"
data ModuleInternal
= Internal
| External
deriving (Eq, Generic, JSON.ToJSON)
instance Csv.ToField ModuleInternal where
toField Internal = "internal"
toField External = "external"
data ImportType
= Wildcard
| Hiding
| Explicit
deriving (Eq, Generic, JSON.ToJSON)
instance Csv.ToField ImportType where
toField Wildcard = "wildcard"
toField Hiding = "hiding"
toField Explicit = "explicit"
-- | Mapping of modules to their aliases and to the files they are found in
type ModuleAliases = [(Text, [(Text, [FilePath])])]
-- | Mapping of modules to files
type WildcardImports = [(FilePath, [Text])]
-- MAIN
main :: IO ()
main =
run =<< O.customExecParser prefs infoOpts
where
prefs = O.prefs $ O.subparserInline <> O.showHelpOnEmpty
infoOpts =
O.info (O.helper <*> opts) $
O.fullDesc
<> O.header "hsie - Swiss army knife for HaSkell Imports and Exports"
<> O.progDesc "Parse Haskell code to analyze imports and exports"
opts =
Options <$> commandOption <*> O.some srcOption
srcOption =
O.argument O.str $
O.metavar "SRCDIR"
<> O.help "Haskell source directory"
<> O.action "directory"
commandOption =
O.subparser $
command "dump-imports" "Dump imported symbols as CSV or JSON"
(Dump <$> jsonOutputFlag)
<> command "graph-modules" "Print dot graph of module imports"
(pure GraphModules)
<> command "graph-symbols" "Print dot graph of symbol imports"
(pure GraphSymbols)
<> command "check-aliases"
"Check that aliases of imported modules are consistent"
(pure CheckAliases)
<> command "check-wildcards"
"Check that no modules are imported as unqualified wildcards"
(CheckWildcards <$> O.many okModuleOption)
command name desc options =
O.command name . O.info (O.helper <*> options) $ O.progDesc desc
jsonOutputFlag =
O.flag OutputCsv OutputJson $
O.long "json" <> O.short 'j' <> O.help "Output JSON"
okModuleOption =
O.strOption $
O.long "ok"
<> O.short 'o'
<> O.metavar "OKMODULE"
<> O.help "Module that is ok to import as unqualified wildcard"
run :: Options -> IO ()
run Options{command, sources} =
runCommand command . markInternal . concat =<< mapM sourceSymbols sources
where
runCommand :: Command -> [ImportedSymbol] -> IO ()
runCommand (Dump format) = LBS8.putStr . dump format
runCommand GraphSymbols = T.putStr . symbolsGraph
runCommand GraphModules = T.putStr . Dot.encode . modulesGraph
runCommand CheckAliases = runInconsistentAliases . inconsistentAliases
runCommand (CheckWildcards okModules) = runWildcards . wildcards okModules
runInconsistentAliases :: ModuleAliases -> IO ()
runInconsistentAliases [] = T.putStrLn "No inconsistent module aliases found."
runInconsistentAliases xs = T.putStr (formatInconsistentAliases xs) >> exitFailure
runWildcards :: WildcardImports -> IO ()
runWildcards [] = T.putStrLn "No unwanted wildcard imports found."
runWildcards xs = T.putStr (formatWildcards xs) >> exitFailure
-- | Mark imports from modules that are among the analyzed ones as internal.
markInternal :: [ImportedSymbol] -> [ImportedSymbol]
markInternal symbols =
fmap mark symbols
where
mark s = s { impInternal = if isInternal s then Internal else External }
isInternal = flip Set.member internalModules . impModule
internalModules = Set.fromList $ fmap impFromModule symbols
-- SYMBOLS
-- | Parse all imported symbols from a source of Haskell source files
sourceSymbols :: FilePath -> IO [ImportedSymbol]
sourceSymbols source = do
files <- filterExts [".hs", ".imports"] <$> getFilesRecursive source
concat <$> mapM moduleSymbols files
where
filterExts exts = filter $ flip elem exts . FP.takeExtension
moduleSymbols filepath = do
GHC.HsModule{..} <- parseModule filepath
return $ concatMap (importSymbols source filepath . GHC.unLoc) hsmodImports
-- | Parse a Haskell module
parseModule :: String -> IO (GHC.HsModule GhcPs)
parseModule filepath = do
result <- ExactPrint.parseModule filepath
case result of
Right (_, hsmod) ->
return $ GHC.unLoc hsmod
Left errs ->
fail $ "Errors with " <> show filepath <> ":\n "
<> List.intercalate "\n " (show <$> bagToList errs)
-- | Symbols imported in an import declaration.
--
-- If the import is a wildcard, i.e. no symbols are selected for import, then
-- only one item is returned.
importSymbols :: FilePath -> FilePath -> GHC.ImportDecl GhcPs -> [ImportedSymbol]
importSymbols _ _ (GHC.XImportDecl _) = mempty
importSymbols source filepath GHC.ImportDecl{..} =
case ideclHiding of
Just (hiding, syms) ->
symbol (if hiding then Hiding else Explicit) . Just . GHC.unLoc <$> GHC.unLoc syms
Nothing ->
[ symbol Wildcard Nothing ]
where
symbol hiding sym =
ImportedSymbol
{ impFile = relativePath filepath
, impSource = source
, impFromModule = T.pack $ moduleFromPath filepath
, impModule = T.pack . moduleNameString . GHC.unLoc $ ideclName
, impQualified = if ideclQualified /= GHC.NotQualified then Qualified else NotQualified
, impAlias = T.pack . moduleNameString . GHC.unLoc <$> ideclAs
, impInternal = External
, impType = hiding
, impSymbol = T.pack . occNameString . rdrNameOcc . GHC.ieName <$> sym
}
moduleFromPath =
intercalate "." . FP.splitDirectories . FP.dropExtension . relativePath
relativePath = FP.makeRelative source
-- DUMP
-- | Dump list of symbols as CSV or JSON
dump :: OutputFormat -> [ImportedSymbol] -> LBS8.ByteString
dump OutputCsv = Csv.encodeDefaultOrderedByName
dump OutputJson = encodePretty
-- ALIASES
-- | Find modules that are imported under different aliases
inconsistentAliases :: [ImportedSymbol] -> ModuleAliases
inconsistentAliases symbols =
fmap moduleAlias symbols
& foldr insertSetMapMap Map.empty
& Map.map (aliases . Map.toList)
& Map.filter ((<) 1 . length)
& Map.toList
where
moduleAlias ImportedSymbol{..} =
(impModule, impAlias, FP.joinPath [impSource, impFile])
insertSetMapMap (k1, k2, v) =
Map.insertWith (Map.unionWith Set.union) k1
(Map.singleton k2 $ Set.singleton v)
aliases :: [(Maybe Text, Set.Set FilePath)] -> [(Text, [FilePath])]
aliases = mapMaybe (\(k, v) -> fmap (, Set.toList v) k)
formatInconsistentAliases :: ModuleAliases -> Text
formatInconsistentAliases modules =
"The following imports have inconsistent aliases:\n\n"
<> T.concat (fmap formatModule modules)
where
formatModule (modName, aliases) =
"Module '"
<> modName
<> "' has the aliases:\n"
<> T.concat (fmap formatAlias aliases)
<> "\n"
formatAlias (alias, sourceFiles) =
" '"
<> alias
<> "' in file"
<> (if length sourceFiles > 2 then "s" else "")
<> ":\n"
<> T.concat (fmap formatFile sourceFiles)
formatFile sourceFile =
" " <> T.pack sourceFile <> "\n"
-- WILDCARDS
-- | Find modules that are imported as wildcards, excluding whitelisted modules.
--
-- Wildcard imports are ones that are not qualified and do not specify which
-- symbols should be imported.
wildcards :: [Text] -> [ImportedSymbol] -> WildcardImports
wildcards okModules =
groupByFile . filter isWildcard . filter (not . isOkModule)
where
isWildcard ImportedSymbol{..} =
impQualified == NotQualified && impType /= Explicit
isOkModule = flip Set.member (Set.fromList okModules) . impModule
groupByFile = Map.toList . fmap Set.toList . foldr insertMap Map.empty
insertMap ImportedSymbol{..} =
Map.insertWith Set.union impFile (Set.singleton impModule)
formatWildcards :: WildcardImports -> Text
formatWildcards files =
"Modules in the following files were imported as wildcards:\n\n"
<> T.concat (fmap formatFile files)
where
formatFile (filepath, modules) =
"In " <> T.pack filepath <> ":\n" <> T.concat (fmap formatModule modules) <> "\n"
formatModule moduleName = " " <> moduleName <> "\n"
-- GRAPHS
modulesGraph :: [ImportedSymbol] -> Dot.DotGraph
modulesGraph symbols =
Dot.DotGraph Dot.Strict Dot.Directed (Just "Modules") $ fmap edge edges
where
edge (from, to) =
Dot.StatementEdge $ Dot.EdgeStatement
(Dot.ListTwo (edgeNode from) (edgeNode to) mempty) mempty
edgeNode t = Dot.EdgeNode $ Dot.NodeId (Dot.Id t) Nothing
edges = unique . fmap edgeTuple . filter ((==) Internal . impInternal) $ symbols
edgeTuple ImportedSymbol{..} = (impFromModule, impModule)
unique = Set.toList . Set.fromList
-- Building Text directly as the Dot package currently doesn't support subgraphs.
symbolsGraph :: [ImportedSymbol] -> Text
symbolsGraph symbols =
"digraph Symbols {\n"
<> " rankdir=LR\n"
<> " ranksep=5\n"
<> T.concat (fmap edge edges)
<> T.concat (fmap cluster symbolsByModule)
<> "}\n"
where
edge (from, to, symbol) =
" "
<> quoted from
<> " -> "
<> quoted (to <> maybe "" ("." <>) symbol)
<> "\n"
cluster (moduleName, clusterSymbols) =
" subgraph "
<> quoted ("cluster_" <> moduleName)
<> " {\n"
<> " " <> quoted moduleName <> "\n"
<> T.concat (fmap (clusterNode moduleName) clusterSymbols)
<> " }\n"
clusterNode moduleName symbol =
" " <> quoted (moduleName <> "." <> symbol) <> "\n"
quoted t = "\"" <> t <> "\""
edges = unique . fmap edgeTuple . filter ((==) Internal . impInternal) $ symbols
edgeTuple ImportedSymbol{..} = (impFromModule, impModule, impSymbol)
unique = Set.toList . Set.fromList
symbolsByModule =
Map.toList . Map.map (catMaybes . Set.toList) . foldr insertMap Map.empty $ edges
insertMap (_, to, symbol) = Map.insertWith Set.union to $ Set.singleton symbol
+67
View File
@@ -0,0 +1,67 @@
# hsie - Swiss army knife for HaSkell Imports and Exports
This tool parses Haskell source code to analyse the imports and exports in a
project. It's available in PostgREST's `nix-shell` by default.
## Dumping imports
Given source code in the directories `src` and `main`, for example, you can run:
```
hsie dump-imports src main
```
This dumps all imports of the modules in the given directory to a CSV file,
printed on `stdout`.
To dump to a JSON file (e.g., to further process with `jq`), add the `--json`
flag:
```
hsie dump-imports --json src main
```
## Graphing imports
The tool can generate `graphviz` graphs of module and symbol imports by printing
a file to `stdout` that can directly be rendered with `dot`:
```
hsie graph-modules src main | dot -Tpng -o modules.png
```
The command `graph-modules` prints a graph of which modules insert which other
modules. `graph-symbols` shows which symbols are imported from which modules.
## Checking imports
To check whether modules are imported under consistent aliases in your project,
run:
```
hsie check-aliases main src
```
This will exit with a non-zero exit code if any inconsistent aliases are found.
The following command checks whether any modules are imported as wildcards, i.e.
not qualified and without specifying symbols.
```
hsie check-wildcards main src
```
To whitelist certain modules to be imported as wildcards, use `--ok`:
```
hsie check-wildcards main src --ok Protolude --ok Test.Module
```
## Current limitations
This tool uses the GHC parser to parse Haskell source code. Language extensions
required to parse each file are detected based on the `{-# LANGUAGE ... #-}`
pragmas. If they are not available (e.g., as they are listed as default
extensions in the `.cabal` file), parses may fail. We can fix this by using
an extended set of non-conflicting extensions by default, as `hlint` does for
example.
+30
View File
@@ -0,0 +1,30 @@
{ ghcWithPackages
, runCommand
}:
let
name = "hsie";
src = ./Main.hs;
modules = ps: [
ps.aeson
ps.aeson-pretty
ps.cassava
ps.dir-traverse
ps.dot
ps.ghc-exactprint
ps.optparse-applicative
];
ghc = ghcWithPackages modules;
hsie =
runCommand "haskellimports" { inherit name src; }
"${ghc}/bin/ghc -O -Werror -Wall -package ghc $src -o $out";
bin =
runCommand name { inherit hsie name; }
''
mkdir -p $out/bin
ln -s $hsie $out/bin/$name
'';
bashCompletion =
runCommand "${name}-bash-completion" { inherit bin name; }
"$bin/bin/$name --bash-completion-script $bin/bin/$name > $out";
in
hsie // { inherit bashCompletion bin; }
-12
View File
@@ -1,12 +0,0 @@
{ writeShellScriptBin, git, silver-searcher, hlint }:
writeShellScriptBin "postgrest-lint"
''
set -euo pipefail
rootdir="$(${git}/bin/git rev-parse --show-toplevel)"
# Lint Haskell files
${silver-searcher}/bin/ag -l -g '\.l?hs$' "$rootdir" \
| xargs ${hlint}/bin/hlint -X QuasiQuotes -X NoPatternSynonyms
''
-41
View File
@@ -1,41 +0,0 @@
{ curl
, jq
, writeShellScriptBin
, nix
}:
# Utility script for pinning the latest unstable version of Nixpkgs.
# Instead of pinning Nixpkgs based on the huge Git repository, we reference a
# specific tarball that only contains the source of the revision that we want
# to pin.
let
name =
"nixpkgs-upgrade";
refUrl =
https://api.github.com/repos/nixos/nixpkgs/git/ref/heads/nixpkgs-unstable;
githubV3Header =
"Accept: application/vnd.github.v3+json";
tarballUrlBase =
https://github.com/nixos/nixpkgs/archive/;
in
writeShellScriptBin name
''
set -euo pipefail
commitHash="$(${curl}/bin/curl "${refUrl}" -H "${githubV3Header}" | ${jq}/bin/jq -r .object.sha)"
tarballUrl="${tarballUrlBase}$commitHash.tar.gz"
tarballHash="$(${nix}/bin/nix-prefetch-url --unpack "$tarballUrl")"
currentDate="$(date --iso)"
cat << EOF
# Pinned version of Nixpkgs, generated with ${name}.
{
date = "$currentDate";
rev = "$commitHash";
tarballHash = "$tarballHash";
}
EOF
''
+4 -4
View File
@@ -1,6 +1,6 @@
# Pinned version of Nixpkgs, generated with nixpkgs-upgrade.
# Pinned version of Nixpkgs, generated with postgrest-nixpkgs-upgrade.
{
date = "2020-04-19";
rev = "10100a97c8964e82b30f180fda41ade8e6f69e41";
tarballHash = "011f36kr3c1ria7rag7px26bh73d1b0xpqadd149bysf4hg17rln";
date = "2021-11-02";
rev = "7053541084bf5ce2921ef307e5585d39d7ba8b3f";
tarballHash = "1flhh5d4zy43x6060hvzjb5hi5cmc51ivc0nwmija9n8d35kcc4x";
}
@@ -0,0 +1,16 @@
# Creates an environment that exposes bashCompletion arguments from all checkedShellScripts
{ buildEnv }:
{ name
, tools
, extra ? { }
}:
let
bashCompletion = builtins.map (tool: tool.bashCompletion) tools;
env = buildEnv {
inherit name;
paths = builtins.map (tool: tool.bin) tools;
};
in
env // { inherit bashCompletion; } // extra
+5
View File
@@ -0,0 +1,5 @@
self: super:
# Overlay that adds `buildToolbox`, an enhanced version of `buildEnv`
{
buildToolbox = super.callPackage ./build-toolbox.nix { };
}
@@ -0,0 +1,141 @@
# Create a bash script that is checked with shellcheck. You can either use it
# directly, or use the .bin attribute to get the script in a bin/ directory,
# to be used in a path for example.
{ argbash
, bash_5
, coreutils
, git
, lib
, runCommand
, shellcheck
, stdenv
, writeTextFile
}:
{ name
, docs
, args ? [ ]
, positionalCompletion ? ""
, inRootDir ? false
, redirectTixFiles ? true
, withEnv ? null
, withPath ? [ ]
, withTmpDir ? false
}: text:
let
# square brackets are a pain to escape - if even possible. just don't use them...
escape = builtins.replaceStrings [ "\n" ] [ " \\n" ];
argsTemplate =
writeTextFile {
inherit name;
destination = "/${name}.m4"; # destination is needed to have the proper basename for completion
text =
''
# BASH_ARGV0 sets $0 - which is used in parser.sh for usage information
# stripping the /nix/store/... path for nicer display
BASH_ARGV0="$(basename "$0")"
# ARG_HELP([${name}], [${escape docs}])
${lib.strings.concatMapStrings (arg: "# " + arg) args}
# ARG_POSITIONAL_DOUBLEDASH()
# ARG_DEFAULTS_POS()
# ARGBASH_GO
'';
};
argsParser =
runCommand "${name}-parser" { }
''
${argbash}/bin/argbash ${argsTemplate}/${name}.m4 > $out
# This forces optional arguments to go *before* positional arguments,
# which allows leftovers to pass optional arguments to sub-commands.
# Example: This way `postgrest-watch -h` will return the help output for watch, while
# `postgrest-watch postgrest-test-spec -h` will return the help output for test-spec.
# Taken from: https://github.com/matejak/argbash/issues/114#issuecomment-557108274
sed '/_positionals_count + 1/a\\t\t\t\tset -- "''${@:1:1}" "--" "''${@:2}"' -i $out
'';
bashCompletion =
runCommand "${name}-completion" { } (
''
${argbash}/bin/argbash --type completion --strip all ${argsTemplate}/${name}.m4 > $out
''
+ lib.optionalString (positionalCompletion != "") ''
sed 's#COMPREPLY.*compgen -o bashdefault .*$#${escape positionalCompletion}#' -i $out
''
);
bin =
writeTextFile {
inherit name;
executable = true;
destination = "/bin/${name}";
text =
''
#!${bash_5}/bin/bash
source ${argsParser}
set -euo pipefail
''
+ lib.optionalString redirectTixFiles ''
# storing tix files in a temporary throw away directory avoids mix/tix conflicts after changes
hpctixdir=$(${coreutils}/bin/mktemp -d)
export HPCTIXFILE="$hpctixdir"/postgrest.tix
trap 'rm -rf $hpctixdir' EXIT
''
+ lib.optionalString inRootDir ''
cd "$(${git}/bin/git rev-parse --show-toplevel)"
if test ! -f postgrest.cabal; then
>&2 echo "Couldn't find postgrest.cabal. Please make sure to" \
"run this command somewhere in the PostgREST repo."
exit 1
fi
''
+ lib.optionalString withTmpDir ''
mkdir -p "''${TMPDIR:-/tmp}/postgrest"
tmpdir="$(${coreutils}/bin/mktemp -d --tmpdir postgrest/${name}-XXX)"
# we keep the tmpdir when an error occurs for debugging
trap 'echo Temporary directory kept at: $tmpdir' ERR
# remove the tmpdir when cancelled (postgrest-watch)
trap 'rm -rf "$tmpdir"' SIGINT SIGTERM
''
+ lib.optionalString (withEnv != null) ''
env="$(cat ${withEnv})"
export PATH="$env/bin:$PATH"
''
+ lib.optionalString (lib.length withPath > 0) ''
export PATH="${lib.concatMapStrings (p: p + "/bin:") withPath}$PATH"
''
+ "(${text})"
+ lib.optionalString withTmpDir ''
rm -rf "$tmpdir"
'';
checkPhase =
''
# check syntax
${stdenv.shell} -n $out/bin/${name}
# check for shellcheck recommendations
${shellcheck}/bin/shellcheck -x $out/bin/${name}
'';
};
script =
runCommand name { inherit bin name; } "ln -s $bin/bin/$name $out";
in
script // { inherit bin bashCompletion; }
@@ -0,0 +1,6 @@
self: super:
# Overlay that adds `checkedShellScript`, an enhanced version of
# writeShellScript and writeShellScriptBin
{
checkedShellScript = super.callPackage ./checked-shell-script.nix { };
}
+9
View File
@@ -0,0 +1,9 @@
{
build-toolbox = import ./build-toolbox;
checked-shell-script = import ./checked-shell-script;
gitignore = import ./gitignore.nix;
haskell-packages = import ./haskell-packages.nix;
postgresql-default = import ./postgresql-default.nix;
postgresql-legacy = import ./postgresql-legacy.nix;
postgresql-future = import ./postgresql-future.nix;
}
+4 -3
View File
@@ -12,8 +12,9 @@ self: super:
gitignoreSrc = super.fetchFromGitHub {
owner = "hercules-ci";
repo = "gitignore";
rev = "2ced4519f865341adcb143c5d668f955a2cb997f";
sha256 = "sha256:0fc5bgv9syfcblp23y05kkfnpgh3gssz6vn24frs8dzw39algk2z";
rev = "211907489e9f198594c0eb0ca9256a1949c9d412";
sha256 = "06j7wpvj54khw0z10fjyi31kpafkr6hi1k0di13k1xp8kywvfyx8";
};
in (super.callPackage gitignoreSrc {}).gitignoreSource;
in
(super.callPackage gitignoreSrc { }).gitignoreSource;
}
+89
View File
@@ -0,0 +1,89 @@
{ compiler, extraOverrides ? (final: prev: { }) }:
self: super:
let
inherit (self.haskell) lib;
overrides =
final: prev:
rec {
# To pin custom versions of Haskell packages:
# protolude =
# prev.callHackageDirect
# {
# pkg = "protolude";
# ver = "0.3.0";
# sha256 = "0iwh4wsjhb7pms88lw1afhdal9f86nrrkkvv65f9wxbd1b159n72";
# }
# { };
#
# To get the sha256:
# nix-prefetch-url --unpack https://hackage.haskell.org/package/protolude-0.3.0/protolude-0.3.0.tar.gz
# To temporarily pin unreleased versions from GitHub:
# <name> =
# prev.callCabal2nixWithOptions "<name>" (super.fetchFromGitHub {
# owner = "<owner>";
# repo = "<repo>";
# rev = "<commit>";
# sha256 = "<sha256>";
# }) "--subpath=<subpath>" {};
#
# To get the sha256:
# nix-prefetch-url --unpack https://github.com/<owner>/<repo>/archive/<commit>.tar.gz
protolude =
prev.callHackageDirect
{
pkg = "protolude";
ver = "0.3.1";
sha256 = "0gf0mn1ycllr69kdq1p07qf7935s10jz0nnhynwqy3d6nmycxr5j";
}
{ };
wai-extra =
prev.callHackageDirect
{
pkg = "wai-extra";
ver = "3.1.8";
sha256 = "1ha8sxc2ii7k7xs5nm06wfwqmf4f1p2acp4ya0jnx6yn6551qps4";
}
{ };
wai-logger =
prev.callHackageDirect
{
pkg = "wai-logger";
ver = "2.3.7";
sha256 = "1d23fdbwbahr3y1vdyn57m1qhljy22pm5cpgb20dy6mlxzdb30xd";
}
{ };
warp =
lib.dontCheck (prev.callHackageDirect
{
pkg = "warp";
ver = "3.3.19";
sha256 = "0y3jj4bhviss6ff9lwxki0zbdcl1rb398bk4s80zvfpnpy7p94cx";
}
{ });
hasql-dynamic-statements =
lib.dontCheck (lib.unmarkBroken prev.hasql-dynamic-statements);
hasql-implicits =
lib.dontCheck (lib.unmarkBroken prev.hasql-implicits);
ptr =
lib.dontCheck (lib.unmarkBroken prev.ptr);
} // extraOverrides final prev;
in
{
haskell =
super.haskell // {
packages = super.haskell.packages // {
"${compiler}" =
super.haskell.packages."${compiler}".override { inherit overrides; };
};
};
}
-10
View File
@@ -1,10 +0,0 @@
The `.nix` files in this directory pin specific versions of Haskell packages
from Hackage. They were generated with `cabal2nix`, e.g.:
```bash
cabal2nix cabal://protolude > protolude.nix
```
Those overrides might become obsolete as the pinned versions are adopted into
our pinned version of nixpkgs.
-11
View File
@@ -1,11 +0,0 @@
self: super:
{
haskellPackages =
super.haskellPackages.override {
overrides =
final: prev: rec {
protolude = prev.callPackage ./protolude.nix {};
};
};
}
@@ -1,43 +0,0 @@
{ mkDerivation
, array
, async
, base
, bytestring
, containers
, deepseq
, ghc-prim
, hashable
, mtl
, mtl-compat
, stdenv
, stm
, text
, transformers
, transformers-compat
}:
mkDerivation {
pname = "protolude";
version = "0.3.0";
sha256 = "4083385a9e03fab9201f63ce198b9ced3fbc1c50d6d42574db5e36c757bedcac";
isLibrary = true;
isExecutable = true;
libraryHaskellDepends = [
array
async
base
bytestring
containers
deepseq
ghc-prim
hashable
mtl
mtl-compat
stm
text
transformers
transformers-compat
];
homepage = "https://github.com/sdiehl/protolude";
description = "A small prelude";
license = stdenv.lib.licenses.mit;
}
+5
View File
@@ -0,0 +1,5 @@
self: super:
# Overlay that sets the default version of PostgreSQL.
{
postgresql = super.postgresql_14;
}
+19
View File
@@ -0,0 +1,19 @@
self: super:
# Overlay that adds future versions of PostgreSQL that are supported by
# PostgREST.
{
## Example for including a postgresql version from a specific nixpks commit:
##
# postgresql_14 =
# let
# rev = "76b1e16c6659ccef7187ca69b287525fea133244";
# tarballHash = "1vsahpcx80k2bgslspb0sa6j4bmhdx77sw6la455drqcrqhdqj6a";
#
# pinnedPkgs =
# builtins.fetchTarball {
# url = "https://github.com/nixos/nixpkgs/archive/${rev}.tar.gz";
# sha256 = tarballHash;
# };
# in
# (import pinnedPkgs { }).pkgs.postgresql_14;
}
+15 -15
View File
@@ -2,19 +2,19 @@ self: super:
# Overlay that adds legacy versions of PostgreSQL that are supported by
# PostgREST.
{
# PostgreSQL 9.4 was removed from Nixpkgs with
# https://github.com/NixOS/nixpkgs/commit/8e2fc57a80d761c46702c3250e61c1bffe021e25
# We pin its parent commit 3b5b9a7 to get the last version that was available.
postgresql_9_4 =
let
rev = "3b5b9a73f59ff93d50156e250203410bdd07f4e0";
tarballHash = "0kr25xqbv1ldp72jbmml21pc5hl7xcfqhclv5qxa5f860jddjznk";
pinnedPkgs =
builtins.fetchTarball {
url = "https://github.com/nixos/nixpkgs/archive/${rev}.tar.gz";
sha256 = tarballHash;
};
in
(import pinnedPkgs {}).pkgs.postgresql_9_4;
# PostgreSQL 9.5 was removed from Nixpkgs with
# https://github.com/NixOS/nixpkgs/commit/72ab382fb6b729b0d654f2c03f5eb25b39f11fbb
# We pin its parent commit to get the last version that was available.
# postgresql_9_5 =
# let
# rev = "55ac7d4580c9ab67848c98cb9519317a1cc399c8";
# tarballHash = "02ffj9f8s1hwhmxj85nx04sv64qb6jm7w0122a1dz9n32fymgklj";
#
# pinnedPkgs =
# builtins.fetchTarball {
# url = "https://github.com/nixos/nixpkgs/archive/${rev}.tar.gz";
# sha256 = tarballHash;
# };
# in
# (import pinnedPkgs { }).pkgs.postgresql_9_5;
}
+16 -17
View File
@@ -1,25 +1,24 @@
{ runCommand }:
{
# Apply patches to a directory.
applyPatches =
name: path: patches:
runCommand name { inherit patches; }
''
set -eou pipefail
name: src: patches:
runCommand
name
{ inherit src patches; }
''
set -eou pipefail
cp -r ${path} $out
chmod -R u+w $out
cp -r $src $out
chmod -R u+w $out
for patch in $patches; do
echo "Applying patch $patch"
patch -d "$out" -p1 < "$patch"
done
'';
for patch in $patches; do
echo "Applying patch $patch"
patch -d "$out" -p1 < "$patch"
done
'';
static-haskell-nix-postgrest-libpq =
./static-haskell-nix-postgrest-libpq.patch;
nixpkgs-revert-ghc-bootstrap =
./nixpkgs-revert-ghc-bootstrap.patch;
# See: https://github.com/NixOS/nixpkgs/pull/87879
nixpkgs-openssl-split-runtime-dependencies-of-static-builds =
./nixpkgs-openssl-split-runtime-dependencies-of-static-builds.patch;
}
@@ -0,0 +1,76 @@
diff --git a/pkgs/development/libraries/openssl/default.nix b/pkgs/development/libraries/openssl/default.nix
index d4be8cc2428..3979698711f 100644
--- a/pkgs/development/libraries/openssl/default.nix
+++ b/pkgs/development/libraries/openssl/default.nix
@@ -50,9 +50,21 @@ let
substituteInPlace crypto/async/arch/async_posix.h \
--replace '!defined(__ANDROID__) && !defined(__OpenBSD__)' \
'!defined(__ANDROID__) && !defined(__OpenBSD__) && 0'
+ '' + optionalString static
+ # On static builds, the ENGINESDIR will be empty, but its path will be
+ # compiled into the library. In order to minimize the runtime dependencies
+ # of packages that statically link openssl, we move it into the OPENSSLDIR,
+ # which will be separated into the 'etc' output.
+ ''
+ substituteInPlace Configurations/unix-Makefile.tmpl \
+ --replace 'ENGINESDIR=$(libdir)/engines-{- $sover_dirname -}' \
+ 'ENGINESDIR=$(OPENSSLDIR)/engines-{- $sover_dirname -}'
'';
- outputs = [ "bin" "dev" "out" "man" ] ++ optional withDocs "doc";
+ outputs = [ "bin" "dev" "out" "man" ]
+ ++ optional withDocs "doc"
+ # Separate output for the runtime dependencies of the static build.
+ ++ optional static "etc";
setOutputFlags = false;
separateDebugInfo =
!stdenv.hostPlatform.isDarwin &&
@@ -101,7 +113,17 @@ let
configureFlags = [
"shared" # "shared" builds both shared and static libraries
"--libdir=lib"
- "--openssldir=etc/ssl"
+ (if !static then
+ "--openssldir=etc/ssl"
+ else
+ # Separate the OPENSSLDIR into its own output, as its path will be
+ # compiled into 'libcrypto.a'. This makes it a runtime dependency of
+ # any package that statically links openssl, so we want to keep that
+ # output minimal. We need to prepend '/.' to the path in order to make
+ # it appear absolute before variable expansion, the 'prefix' would be
+ # prepended to it otherwise.
+ "--openssldir=/.$(etc)/etc/ssl"
+ )
] ++ lib.optionals withCryptodev [
"-DHAVE_CRYPTODEV"
"-DUSE_CRYPTODEV_DIGESTS"
@@ -131,6 +153,9 @@ let
if [ -n "$(echo $out/lib/*.so $out/lib/*.dylib $out/lib/*.dll)" ]; then
rm "$out/lib/"*.a
fi
+
+ # 'etc' is a separate output on static builds only.
+ etc=$out
'' + lib.optionalString (!stdenv.hostPlatform.isWindows)
# Fix bin/c_rehash's perl interpreter line
#
@@ -152,14 +177,15 @@ let
mv $out/include $dev/
# remove dependency on Perl at runtime
- rm -r $out/etc/ssl/misc
+ rm -r $etc/etc/ssl/misc
- rmdir $out/etc/ssl/{certs,private}
+ rmdir $etc/etc/ssl/{certs,private}
'';
postFixup = lib.optionalString (!stdenv.hostPlatform.isWindows) ''
- # Check to make sure the main output doesn't depend on perl
- if grep -r '${buildPackages.perl}' $out; then
+ # Check to make sure the main output and the static runtime dependencies
+ # don't depend on perl
+ if grep -r '${buildPackages.perl}' $out $etc; then
echo "Found an erroneous dependency on perl ^^^" >&2
exit 1
fi
@@ -1,125 +0,0 @@
From 93cc144b215654ca9cb0d2bcb58bb8922895ccab Mon Sep 17 00:00:00 2001
From: Remo <remo>
Date: Sat, 25 Apr 2020 17:31:55 +0200
Subject: [PATCH] Revert "ghc: 8.6.3-binary -> 8.6.5-binary"
This reverts commit 3c7ef6bcd85e3e133ac6f2ab5e238934d56d902e.
---
.../{8.6.5-binary.nix => 8.6.3-binary.nix} | 13 ++++++------
pkgs/top-level/haskell-packages.nix | 20 +++++++++----------
2 files changed, 16 insertions(+), 17 deletions(-)
rename pkgs/development/compilers/ghc/{8.6.5-binary.nix => 8.6.3-binary.nix} (95%)
diff --git a/pkgs/development/compilers/ghc/8.6.5-binary.nix b/pkgs/development/compilers/ghc/8.6.3-binary.nix
similarity index 95%
rename from pkgs/development/compilers/ghc/8.6.5-binary.nix
rename to pkgs/development/compilers/ghc/8.6.3-binary.nix
index 97793d912895..9e0fd12b96cf 100644
--- a/pkgs/development/compilers/ghc/8.6.5-binary.nix
+++ b/pkgs/development/compilers/ghc/8.6.3-binary.nix
@@ -27,19 +27,18 @@ let
in
stdenv.mkDerivation rec {
- version = "8.6.5";
+ version = "8.6.3";
name = "ghc-${version}-binary";
- # https://downloads.haskell.org/~ghc/8.6.5/
src = fetchurl ({
i686-linux = {
- url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-i386-deb9-linux.tar.xz";
- sha256 = "1p2h29qghql19ajk755xa0yxkn85slbds8m9n5196ris743vkp8w";
+ url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-i386-deb8-linux.tar.xz";
+ sha256 = "0bw8a7fxcbskf93rb4m542ff66vrmx5i5kj77qx37cbhijx70w5m";
};
x86_64-linux = {
- url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-x86_64-deb9-linux.tar.xz";
- sha256 = "1pqlx6rdjs2110g0y1i9f8x18lmdizibjqd15f5xahcz39hgaxdw";
+ url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-x86_64-deb8-linux.tar.xz";
+ sha256 = "1m9gaga2pzi2cx5gvasg0rx1dlvr68gmi20l67652kag6xjsa719";
};
aarch64-linux = {
url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-aarch64-ubuntu18.04-linux.tar.xz";
@@ -47,7 +46,7 @@ stdenv.mkDerivation rec {
};
x86_64-darwin = {
url = "http://haskell.org/ghc/dist/${version}/ghc-${version}-x86_64-apple-darwin.tar.xz";
- sha256 = "0s9188vhhgf23q3rjarwhbr524z6h2qga5xaaa2pma03sfqvvhfz";
+ sha256 = "1hbzk57v45176kxcx848p5jn5p1xbp2129ramkbzsk6plyhnkl3r";
};
}.${stdenv.hostPlatform.system}
or (throw "cannot bootstrap GHC on this platform"));
diff --git a/pkgs/top-level/haskell-packages.nix b/pkgs/top-level/haskell-packages.nix
index aa7ef02a3886..712a98e147e0 100644
--- a/pkgs/top-level/haskell-packages.nix
+++ b/pkgs/top-level/haskell-packages.nix
@@ -4,7 +4,7 @@ let
# These are attributes in compiler and packages that don't support integer-simple.
integerSimpleExcludes = [
"ghc822Binary"
- "ghc865Binary"
+ "ghc863Binary"
"ghc844"
"ghcjs"
"ghcjs86"
@@ -42,7 +42,7 @@ in {
ghc822Binary = callPackage ../development/compilers/ghc/8.2.2-binary.nix { };
- ghc865Binary = callPackage ../development/compilers/ghc/8.6.5-binary.nix { };
+ ghc863Binary = callPackage ../development/compilers/ghc/8.6.3-binary.nix { };
ghc844 = callPackage ../development/compilers/ghc/8.4.4.nix {
bootPkgs = packages.ghc822Binary;
@@ -57,31 +57,31 @@ in {
llvmPackages = pkgs.llvmPackages_6;
};
ghc881 = callPackage ../development/compilers/ghc/8.8.1.nix {
- bootPkgs = packages.ghc865Binary;
+ bootPkgs = packages.ghc863Binary;
inherit (buildPackages.python3Packages) sphinx;
buildLlvmPackages = buildPackages.llvmPackages_7;
llvmPackages = pkgs.llvmPackages_7;
};
ghc882 = callPackage ../development/compilers/ghc/8.8.2.nix {
- bootPkgs = packages.ghc865Binary;
+ bootPkgs = packages.ghc863Binary;
inherit (buildPackages.python3Packages) sphinx;
buildLlvmPackages = buildPackages.llvmPackages_7;
llvmPackages = pkgs.llvmPackages_7;
};
ghc883 = callPackage ../development/compilers/ghc/8.8.3.nix {
- bootPkgs = packages.ghc865Binary;
+ bootPkgs = packages.ghc863Binary;
inherit (buildPackages.python3Packages) sphinx;
buildLlvmPackages = buildPackages.llvmPackages_7;
llvmPackages = pkgs.llvmPackages_7;
};
ghc8101 = callPackage ../development/compilers/ghc/8.10.1.nix {
- bootPkgs = packages.ghc865Binary;
+ bootPkgs = packages.ghc863Binary;
inherit (buildPackages.python3Packages) sphinx;
buildLlvmPackages = buildPackages.llvmPackages_9;
llvmPackages = pkgs.llvmPackages_9;
};
ghcHEAD = callPackage ../development/compilers/ghc/head.nix {
- bootPkgs = packages.ghc883; # no binary yet
+ bootPkgs = packages.ghc863Binary;
inherit (buildPackages.python3Packages) sphinx;
buildLlvmPackages = buildPackages.llvmPackages_9;
llvmPackages = pkgs.llvmPackages_9;
@@ -118,9 +118,9 @@ in {
compilerConfig = callPackage ../development/haskell-modules/configuration-ghc-8.2.x.nix { };
packageSetConfig = bootstrapPackageSet;
};
- ghc865Binary = callPackage ../development/haskell-modules {
- buildHaskellPackages = bh.packages.ghc865Binary;
- ghc = bh.compiler.ghc865Binary;
+ ghc863Binary = callPackage ../development/haskell-modules {
+ buildHaskellPackages = bh.packages.ghc863Binary;
+ ghc = bh.compiler.ghc863Binary;
compilerConfig = callPackage ../development/haskell-modules/configuration-ghc-8.6.x.nix { };
packageSetConfig = bootstrapPackageSet;
};
@@ -1,25 +0,0 @@
From bb4c1e27e391eff01591fe60830ff68a9ada41ef Mon Sep 17 00:00:00 2001
From: Remo <remo>
Date: Wed, 22 Apr 2020 11:58:07 +0200
Subject: [PATCH] Add postgrest libpq linking fix
---
survey/default.nix | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/survey/default.nix b/survey/default.nix
index b28606c..fd7cde7 100644
--- a/survey/default.nix
+++ b/survey/default.nix
@@ -1052,6 +1052,11 @@ let
super.squeal-postgresql
[ final.openssl ]
"--libs openssl";
+ postgrest =
+ addStaticLinkerFlagsWithPkgconfig
+ super.postgrest
+ [ final.openssl ]
+ "--libs openssl";
xml-to-json =
addStaticLinkerFlagsWithPkgconfig
+67
View File
@@ -0,0 +1,67 @@
# Derive a fully static Haskell package based on musl instead of glibc.
{ nixpkgs, compiler, patches, allOverlays }:
name: src:
let
# The nh2/static-haskell-nix project does all the hard work for us.
static-haskell-nix =
let
rev = "bd66b86b72cff4479e1c76d5916a853c38d09837";
in
builtins.fetchTarball {
url = "https://github.com/nh2/static-haskell-nix/archive/${rev}.tar.gz";
sha256 = "0rnsxaw7v27znsg9lgqk1i4007ydqrc8gfgimrmhf24lv6galbjh";
};
patched-static-haskell-nix =
patches.applyPatches "patched-static-haskell-nix"
static-haskell-nix
[
# No patches currently required.
];
patchedNixpkgs =
patches.applyPatches "patched-nixpkgs"
nixpkgs
[
patches.nixpkgs-openssl-split-runtime-dependencies-of-static-builds
];
extraOverrides =
final: prev:
rec {
# We need to add our package needs to the package set that we pass to
# static-haskell-nix. Using callCabal2nix on the haskellPackages that
# it returns would result in a dynamic build based on musl, and not the
# fully static build that we want.
"${name}" = prev.callCabal2nix name src { };
};
overlays =
[
allOverlays.postgresql-future
allOverlays.postgresql-default
(allOverlays.haskell-packages { inherit compiler extraOverrides; })
# Disable failing tests for postgresql on musl that should have no impact
# on the libpq that we need (collate.icu.utf8 and foreign regression
# tests)
(self: super:
{ postgresql = super.postgresql.overrideAttrs (_: { doCheck = false; }); }
)
];
# Apply our overlay to the given pkgs.
normalPkgs =
import patchedNixpkgs { inherit overlays; };
defaultCabalPackageVersionComingWithGhc =
{
ghc8107 = "Cabal_3_2_1_0";
}."${compiler}";
# The static-haskell-nix 'survey' derives a full static set of Haskell
# packages, applying fixes where necessary.
survey =
import "${patched-static-haskell-nix}/survey" { inherit normalPkgs compiler defaultCabalPackageVersionComingWithGhc; };
in
survey.haskellPackages."${name}"
-90
View File
@@ -1,90 +0,0 @@
{ mkDerivation
, array
, base
, base-compat
, base-orphans
, binary
, bytestring
, containers
, deepseq
, Diff
, directory
, filepath
, integer-logarithms
, mtl
, optparse-applicative
, parsec
, pretty
, process
, QuickCheck
, stdenv
, stm
, tagged
, tar
, tasty
, tasty-golden
, tasty-hunit
, tasty-quickcheck
, temporary
, text
, time
, transformers
, tree-diff
, unix
}:
mkDerivation {
pname = "Cabal";
version = "3.0.0.0";
sha256 = "5143ec26d740c1a508c93a8860e64407e7546c29b9817db20ff1595c1968d287";
setupHaskellDepends = [ mtl parsec ];
libraryHaskellDepends = [
array
base
binary
bytestring
containers
deepseq
directory
filepath
mtl
parsec
pretty
process
text
time
transformers
unix
];
testHaskellDepends = [
array
base
base-compat
base-orphans
binary
bytestring
containers
deepseq
Diff
directory
filepath
integer-logarithms
optparse-applicative
pretty
process
QuickCheck
stm
tagged
tar
tasty
tasty-golden
tasty-hunit
tasty-quickcheck
temporary
text
tree-diff
];
doCheck = false;
homepage = "http://www.haskell.org/cabal/";
description = "A framework for packaging Haskell software";
license = stdenv.lib.licenses.bsd3;
}
-89
View File
@@ -1,89 +0,0 @@
# Turn a Haskell source package (given its name and source) into a derivation
# for a fully static executable.
{ nixpkgs, compiler, name, src, patches }:
let
# The nh2/static-haskell-nix project does all the hard work for us for
# building static Haskell executables. We apply a patch for PostgREST below
# until the respective pull request is merged. See:
# https://github.com/nh2/static-haskell-nix/pull/91
# statix-haskell-nix builds everything based on Musl instead of glibc, so
# there will be a _lot_ to rebuild if you don't use a binary cache.
static-haskell-nix =
let
rev = "749707fc90b781c3e653e67917a7d571fe82ae7b";
in
builtins.fetchTarball {
url = "https://github.com/nh2/static-haskell-nix/archive/${rev}.tar.gz";
sha256 = "155spda2lww378bhx68w6dxwqd5y6s9kin3qbgl2m23r3vmk3m3w";
};
patched-static-haskell-nix =
patches.applyPatches "patched-static-haskell-nix" static-haskell-nix
[
patches.static-haskell-nix-postgrest-libpq
];
haskellPackagesOverrides =
lib: final: prev:
{
# Add our source package.
"${name}" = prev.callCabal2nix name src {};
# cabal2nix depends on Cabal 3.0.*, while our pinned version of Nixpkgs
# only provides 2.4 or 3.2. So we pinned 3.0.0.0 in ./Cabal.nix
cabal2nix =
prev.cabal2nix.overrideScope
(self: super: { Cabal = self.callPackage ./Cabal.nix {}; });
# Pinned version of protolude.
protolude =
prev.callPackage ../overlays/haskell-packages/protolude.nix {};
# The tests for the packages below took a long time on static
# builds, so we disable them for now - to be investigated.
happy = lib.dontCheck prev.happy;
};
# This overlay adds our source package and applies adjustments to the
# derivation of other packages that it depends on. The overlay applies to
# Haskell packages for the given compiler, which we will later use
# with the static-haskell-nix survey.
overlay =
self: super:
# Override the set of Haskell packages at
# pkgs.haskell.packages."${compiler}".
{
haskell = super.haskell // {
packages = super.haskell.packages // {
"${compiler}" =
super.haskell.packages."${compiler}".override
{ overrides = haskellPackagesOverrides self.haskell.lib; };
};
};
};
patchedNixpkgs =
patches.applyPatches "patched-nixpkgs" nixpkgs
[
patches.nixpkgs-revert-ghc-bootstrap
];
# Apply our overlay to the given pkgs.
normalPkgs =
(import patchedNixpkgs {}).appendOverlays [ overlay ];
# Each version of GHC needs a specific version of Cabal.
defaultCabalPackageVersionComingWithGhc =
{
ghc883 = "Cabal_3_2_0_0";
}."${compiler}";
# Let the static-haskell-nix project do the hard work of deriving a set of
# fully static Haskell executables, including one for the our source package
# that we added through the overlay.
survey =
import "${patched-static-haskell-nix}/survey"
{ inherit normalPkgs compiler defaultCabalPackageVersionComingWithGhc; };
in
# Return the fully static derivation of our source package.
survey.haskellPackages."${name}"
-37
View File
@@ -1,37 +0,0 @@
{ writeShellScriptBin
, git
, nixpkgs-fmt
, silver-searcher
, stylish-haskell
, buildEnv
}:
let
style =
writeShellScriptBin "postgrest-style"
''
set -euo pipefail
rootdir="$(${git}/bin/git rev-parse --show-toplevel)"
# Format Nix files
${nixpkgs-fmt}/bin/nixpkgs-fmt "$rootdir" > /dev/null 2> /dev/null
# Format Haskell files
${silver-searcher}/bin/ag -l -g '\.l?hs$' "$rootdir" \
| xargs ${stylish-haskell}/bin/stylish-haskell -i
'';
check =
writeShellScriptBin "postgrest-style-check"
''
set -euo pipefail
${style}/bin/${style.name}
${git}/bin/git diff-index --exit-code HEAD -- '*.hs' '*.lhs' '*.nix'
'';
in
buildEnv {
name = "postgrest-style";
paths = [ style check ];
}
-90
View File
@@ -1,90 +0,0 @@
# Utilities for running the PostgREST test suite
{ buildEnv
, cabal-install
, git
, lib
, postgrestBuildEnv
, postgresqlVersions
, runtimeShell
, writeShellScript
, writeShellScriptBin
}:
let
# Wrap the `test/with_tmp_db` script with the required dependencies from Nix.
withTmpDb =
postgresql:
writeShellScript "postgrest-test-${postgresql.name}"
''
set -euo pipefail
export PATH=${postgresql}/bin:${git}/bin:${runtimeShell}/bin:"$PATH"
exec ${../test/with_tmp_db} "$@"
'';
# Script to run the Haskell test suite against a specific version of
# PostgreSQL.
testSpec =
name: postgresql:
writeShellScriptBin name
''
set -euo pipefail
cat << EOF
Running spec against ${postgresql.name}...
EOF
# TODO: Make this work outside nix-shell when installed with nix-env.
# Probably using postgrestBuildEnv somehow?
${withTmpDb postgresql} ${cabal-install}/bin/cabal v2-test \
--test-show-detail=direct
cat << EOF
Done running spec against ${postgresql.name}.
EOF
'';
# The PostgreSQL version that we run the tests against by default.
defaultPostgresql =
builtins.head postgresqlVersions;
defaultTestSpec =
testSpec "postgrest-test-spec" defaultPostgresql;
# Create a `testSpec` for each PostgreSQL version that we want to test
# against.
testSpecVersions =
map
(postgresql: testSpec "postgrest-test-spec-${postgresql.name}" postgresql)
postgresqlVersions;
# Helper script for running the tests against all PostgreSQL versions.
testSpecAllVersions =
let
testRunners =
map (test: "${test}/bin/${test.name}") testSpecVersions;
in
writeShellScriptBin "postgrest-test-spec-all"
''
set -euo pipefail
${lib.concatStringsSep "\n" testRunners}
'';
in
# Create an environment that contains all the utility scripts for running tests
# that we defined above.
buildEnv {
name =
"postgrest-tests";
paths =
[
defaultTestSpec
testSpecAllVersions
] ++ testSpecVersions;
}
+57
View File
@@ -0,0 +1,57 @@
{ buildToolbox
, cabal-install
, checkedShellScript
, devCabalOptions
, postgrest
}:
let
build =
checkedShellScript
{
name = "postgrest-build";
docs = "Build PostgREST interactively using cabal-install.";
args = [ "ARG_LEFTOVERS([Cabal arguments])" ];
inRootDir = true;
withEnv = postgrest.env;
}
''
exec ${cabal-install}/bin/cabal v2-build ${devCabalOptions} "''${_arg_leftovers[@]}"
'';
clean =
checkedShellScript
{
name = "postgrest-clean";
docs = "Clean the PostgREST project, including all cabal-install artifacts.";
inRootDir = true;
}
''
# clean old coverage data, too
rm -rf .hpc coverage
exec ${cabal-install}/bin/cabal v2-clean
'';
run =
checkedShellScript
{
name = "postgrest-run";
docs = "Run PostgREST after buidling it interactively with cabal-install";
args = [ "ARG_LEFTOVERS([PostgREST arguments])" ];
inRootDir = true;
withEnv = postgrest.env;
}
''
exec ${cabal-install}/bin/cabal v2-run ${devCabalOptions} --verbose=0 -- \
postgrest "''${_arg_leftovers[@]}"
'';
in
buildToolbox
{
name = "postgrest-cabal";
tools = [
build
clean
run
];
}
+296
View File
@@ -0,0 +1,296 @@
{ buildToolbox
, cabal-install
, cachix
, checkedShellScript
, devCabalOptions
, entr
, git
, gnugrep
, graphviz
, haskellPackages
, hsie
, nix
, silver-searcher
, style
, tests
, withTools
}:
let
watch =
checkedShellScript
{
name = "postgrest-watch";
docs =
''
Watch the project for changes and reinvoke the given command.
Example:
postgrest-watch postgrest-test-io
'';
args =
[
"ARG_POSITIONAL_SINGLE([command], [Command to run])"
"ARG_LEFTOVERS([command arguments])"
];
positionalCompletion = "_command";
redirectTixFiles = false; # will be done by sub-command
inRootDir = true;
}
''
while true; do
(! ${silver-searcher}/bin/ag -l . | ${entr}/bin/entr -dr "$_arg_command" "''${_arg_leftovers[@]}")
done
'';
pushCachix =
checkedShellScript
{
name = "postgrest-push-cachix";
docs = ''
Push all build artifacts to cachix.
Requires authentication with `cachix authtoken ...`.
'';
inRootDir = true;
}
''
${nix}/bin/nix-instantiate \
| while read -r drv; do
${nix}/bin/nix-store -qR --include-outputs "$drv"
done \
| ${cachix}/bin/cachix push postgrest
'';
check =
checkedShellScript
{
name = "postgrest-check";
docs =
''
Run most checks that will also run on CI, but only against the
latest PostgreSQL version.
This currently excludes the memory and spec-idempotence tests,
as those are particularly expensive.
'';
inRootDir = true;
}
''
${tests}/bin/postgrest-test-spec
${tests}/bin/postgrest-test-querycost
${tests}/bin/postgrest-test-doctests
${tests}/bin/postgrest-test-io
${style}/bin/postgrest-lint
${style}/bin/postgrest-style-check
'';
gitHooks =
let
name = "postgrest-git-hooks";
in
checkedShellScript
{
inherit name;
docs =
''
Enable or disable git pre-commit and pre-push hooks.
Basic is faster and will only run:
- pre-commit: postgrest-style
- pre-push: postgrest-lint
Full takes a lot more time and will run:
- pre-commit: postgrest-style && postgrest-lint
- pre-push: postgrest-check
Changes made by postgrest-style will be staged automatically.
Example usage:
postgrest-git-hooks disable
postgrest-git-hooks enable basic
postgrest-git-hooks enable full
The "run" operation and "--hook" argument are only used internally.
'';
args =
[
"ARG_POSITIONAL_SINGLE([operation], [Operation])"
"ARG_TYPE_GROUP_SET([OPERATION], [OPERATION], [operation], [disable,enable,run])"
"ARG_POSITIONAL_SINGLE([mode], [Mode], [basic])"
"ARG_TYPE_GROUP_SET([MODE], [MODE], [mode], [basic,full])"
"ARG_OPTIONAL_SINGLE([hook], , [Hook], [pre-commit])"
"ARG_TYPE_GROUP_SET([HOOK], [HOOK], [hook], [pre-commit,pre-push])"
];
positionalCompletion =
''
if test "$prev" == "${name}"; then
COMPREPLY=( $(compgen -W "enable disable" -- "$cur") )
elif test "$prev" == "enable" || test "$prev" == "disable"; then
COMPREPLY=( $(compgen -W "basic full" -- "$cur") )
fi
'';
inRootDir = true;
}
''
if [ run != "$_arg_operation" ]; then
# Remove all hooks first and ignore failures because the file might be missing.
# This assumes that we're only adding lines that include "postgrest-git-hooks"
# to the hook file.
sed -i -e '/postgrest-git-hooks/d' .git/hooks/pre-{commit,push} 2> /dev/null || true
if [ disabled != "$_arg_mode" ]; then
# The nix-shell && + nix-shell || pattern makes sure we can run the hook
# in a pure nix-shell, where nix-shell itself is not available, too.
# The $(nix-shell --run "command -v ...") pattern ensures we only need to enable
# the hooks once and still run the latest of our hook scripts, even when we
# update them in the repo.
echo 'command -v nix-shell > /dev/null || postgrest-git-hooks --hook=pre-commit run' "$_arg_mode" \
>> .git/hooks/pre-commit
# shellcheck disable=SC2016
echo 'command -v nix-shell > /dev/null && $(nix-shell --quiet -Q --run "command -v postgrest-git-hooks") --hook=pre-commit run' "$_arg_mode" \
>> .git/hooks/pre-commit
chmod +x .git/hooks/pre-commit
echo 'command -v nix-shell > /dev/null || postgrest-git-hooks --hook=pre-push run' "$_arg_mode" \
>> .git/hooks/pre-push
# shellcheck disable=SC2016
echo 'command -v nix-shell > /dev/null && $(nix-shell --quiet -Q --run "command -v postgrest-git-hooks") --hook=pre-push run' "$_arg_mode" \
>> .git/hooks/pre-push
chmod +x .git/hooks/pre-push
fi
else
# When run from a git hook, the GIT_ environment variables conflict with our withGit helper.
# The following unsets all GIT_ variables.
unset "''${!GIT_@}"
case "$_arg_mode" in
basic)
case "$_arg_hook" in
pre-commit)
# To be able to automatically add only changes from postgrest-style to the staging area,
# we need to run postgrest-style twice. Otherwise we'd risk merge conflicts when popping
# the stash afterwards.
${style}/bin/postgrest-style
stash="postgrest-git-hooks-$RANDOM"
${git}/bin/git stash push --include-untracked --keep-index -m "$stash"
if [ "$(git stash list --grep $stash)" ]; then
# Only create the stash pop trap, if we actually created a stash.
# Otherwise stash pop will cause havoc.
trap '${git}/bin/git stash pop $(git stash list --format=format:%gD --grep "$stash" -n1)' EXIT
fi
${style}/bin/postgrest-style
${git}/bin/git add .
;;
pre-push)
# Create a clean working tree without any uncomitted changes.
${withTools.withGit} HEAD ${style}/bin/postgrest-lint
;;
esac
;;
full)
case "$_arg_hook" in
pre-commit)
# To be able to automatically add only changes from postgrest-style to the staging area,
# we need to run postgrest-style twice. Otherwise we'd risk merge conflicts when popping
# the stash afterwards.
${style}/bin/postgrest-style
stash="postgrest-git-hooks-$RANDOM"
${git}/bin/git stash push --include-untracked --keep-index -m "$stash"
if [ "$(git stash list --grep $stash)" ]; then
# Only create the stash pop trap, if we actually created a stash.
# Otherwise stash pop will cause havoc.
trap '${git}/bin/git stash pop $(git stash list --format=format:%gD --grep "$stash" -n1)' EXIT
fi
${style}/bin/postgrest-style
${git}/bin/git add .
${style}/bin/postgrest-lint
;;
pre-push)
# Create a clean working tree without any uncomitted changes.
${withTools.withGit} HEAD ${check}
;;
esac
;;
esac
fi
'';
dumpMinimalImports =
checkedShellScript
{
name = "postgrest-dump-minimal-imports";
docs = "Dump minimal imports into given directory.";
args = [ "ARG_POSITIONAL_SINGLE([dumpdir], [Output directory])" ];
inRootDir = true;
withTmpDir = true;
}
''
mkdir -p "$_arg_dumpdir"
${cabal-install}/bin/cabal v2-build ${devCabalOptions} \
--builddir="$tmpdir" \
--ghc-option=-ddump-minimal-imports \
--ghc-option=-dumpdir="$_arg_dumpdir" \
1>&2
# Fix OverloadedRecordFields imports
# shellcheck disable=SC2016
sed -E 's/\$sel:.*://g' -i "$_arg_dumpdir"/*
'';
hsieMinimalImports =
checkedShellScript
{
name = "postgrest-hsie-minimal-imports";
docs = "Run hsie with a provided dump of minimal imports.";
args = [ "ARG_LEFTOVERS([hsie arguments])" ];
withTmpDir = true;
}
''
${dumpMinimalImports} "$tmpdir"
${hsie} "$tmpdir" "''${_arg_leftovers[@]}"
'';
hsieGraphModules =
checkedShellScript
{
name = "postgrest-hsie-graph-modules";
docs = "Create a PNG graph of modules imported within the codebase.";
args = [ "ARG_POSITIONAL_SINGLE([outfile], [Output filename])" ];
}
''
${hsie} graph-modules main src | ${graphviz}/bin/dot -Tpng -o "$_arg_outfile"
'';
hsieGraphSymbols =
checkedShellScript
{
name = "postgrest-hsie-graph-symbols";
docs = "Create a PNG graph of symbols imported within the codebase.";
args = [ "ARG_POSITIONAL_SINGLE([outfile], [Output filename])" ];
}
''
${hsieMinimalImports} graph-symbols | ${graphviz}/bin/dot -Tpng -o "$_arg_outfile"
'';
in
buildToolbox
{
name = "postgrest-dev";
tools = [
watch
pushCachix
check
gitHooks
dumpMinimalImports
hsieMinimalImports
hsieGraphModules
hsieGraphSymbols
];
}
+95
View File
@@ -0,0 +1,95 @@
# Docker image built with Nix
In order to build an optimal PostgREST Docker image, we create the image from
scratch (i.e., without a parent image like `debian` or `alpine`), and only
include the file that is essential for running PostgREST: the static
PostgREST binary.
This is similar to what you would get with the following `Dockerfile`:
```Dockerfile
# `scratch` is a minimal, reserved image in Docker, see
# https://docs.docker.com/develop/develop-images/baseimages/ . It essentially
# means "don't use a parent image and start with an empty one".
FROM scratch
# The static PostgREST executable has no runtime dependencies, so it's all we
# need to include for running the application.
ADD /absolute/path/to/postgrest /bin/postgrest
EXPOSE 3000
# This is the user id that Docker will run our image under by default. Note
# that we don't actually add the user to `/etc/passwd` or `/etc/shadow`. This
# means that tools like whoami would not work properly, but we don't include
# those in the image anyway. Not adding the user has the benefit that the image
# can be run under any user you specify.
USER 1000
CMD [ "/bin/postgrest" ]
```
# Building the Docker image with Nix
As we are building the static PostgREST executable with Nix and that's the main
input to the Docker file, we can also create the Docker image directly with Nix
using the [`dockerTools`
utilities](https://nixos.org/nixpkgs/manual/#sec-pkgs-dockerTools). Those
utilities don't actually use `Dockerfiles` or Docker to build Docker images,
but create them directly by putting together the required `json` and `tar`
files that make up an image. This is more efficient, does not rely on Docker or
root permissions and results in fully reproducible builds. See
[`nix/docker/default.nix`](./default.nix) for details how the image is built.
# Building and loading the image
The Nix expression provides a helper script `postgrest-docker-load` that loads
the optimized image into your local Docker instance (using `docker load -i
<image file>` under the hood). You can use it by running:
```
# Running from the root directory of the repository:
# Build the `docker` attribute from `default.nix`, the result will be symlinked
# to `result`:
nix-build -A docker
# Run the loading script:
result/bin/postgrest-docker-load
```
The Docker image built with Nix always has the name "postgrest:latest" when
loaded.
# Inspecting the optimized image
The image does not come with the usual utilities like `bash` and `ls`.
You can, however, explore the `tar` file of the image by saving it with `docker
save postgrest:latest > image.tar`.
[Dive](https://github.com/wagoodman/dive) is also useful for looking at the
contents of the image:
```
┃ ● Layers ┣━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ │ Current Layer Contents ├────────────────────────────────────────────────────────────────────────────────
Cmp Size Command Permission UID:GID Size Filetree
14 MB FROM 20ee65c811575d2 dr-xr-xr-x 0:0 14 MB ├── bin
-r-xr-xr-x 0:0 14 MB │ └── postgrest
│ Layer Details ├───────────────────────────────────────────────────────────────────────────────────────── drwxr-xr-x 0:0 783 B ├── etc
-r--r--r-- 0:0 783 B │ └── postgrest.conf
Tags: (unavailable) dr-xr-xr-x 0:0 23 kB └── nix
Id: 20ee65c811575d206eb673e1887e7f7e6b7ccde902a63ccb924c5faa50b32cee dr-xr-xr-x 0:0 23 kB └── store
Digest: sha256:ece77302b83fd38fb54395dabc10c2eba06fc1d1933801d36cc2c4732d9c8f38 dr-xr-xr-x 0:0 23 kB └── s440jbrn94wmpzy7f8yfsp6jr2shllw5-openssl-1.1.1g-etc
Command: dr-xr-xr-x 0:0 23 kB └── etc
dr-xr-xr-x 0:0 23 kB └── ssl
-r--r--r-- 0:0 412 B ├── ct_log_list.cnf
│ Image Details ├───────────────────────────────────────────────────────────────────────────────────────── -r--r--r-- 0:0 412 B ├── ct_log_list.cnf.dist
dr-xr-xr-x 0:0 0 B ├── engines-1.1
-r--r--r-- 0:0 11 kB ├── openssl.cnf
Total Image size: 14 MB -r--r--r-- 0:0 11 kB └── openssl.cnf.dist
Potential wasted space: 0 B
Image efficiency score: 100 %
Count Total Space Path
```
+49
View File
@@ -0,0 +1,49 @@
{ buildToolbox
, postgrest
, dockerTools
, checkedShellScript
}:
let
image =
dockerTools.buildImage {
name = "postgrest";
tag = "latest";
contents = postgrest;
# Set the current time as the image creation date. This makes the build
# non-reproducible, but that should not be an issue for us.
created = "now";
extraCommands =
''
rmdir share
'';
config = {
Cmd = [ "/bin/postgrest" ];
User = "1000";
ExposedPorts = {
"3000/tcp" = { };
};
};
};
load =
checkedShellScript
{
name = "postgrest-docker-load";
docs = "Load the PostgREST image into Docker.";
}
''
docker load -i ${image}
'';
in
buildToolbox
{
name = "postgrest-docker";
tools = [ load ];
extra = {
inherit image;
};
}
+174
View File
@@ -0,0 +1,174 @@
{ buildToolbox
, checkedShellScript
, jq
, python3Packages
, vegeta
, withTools
, writers
}:
let
runner =
checkedShellScript
{
name = "postgrest-loadtest-runner";
docs = "Run vegeta. Assume PostgREST to be running.";
args = [
"ARG_LEFTOVERS([additional vegeta arguments])"
"ARG_USE_ENV([PGRST_SERVER_UNIX_SOCKET], [], [Unix socket to connect to running PostgREST instance])"
];
}
''
# ARG_USE_ENV only adds defaults or docs for environment variables
# We manually implement a required check here
# See also: https://github.com/matejak/argbash/issues/80
: "''${PGRST_SERVER_UNIX_SOCKET:?PGRST_SERVER_UNIX_SOCKET is required}"
${vegeta}/bin/vegeta -cpus 1 attack \
-unix-socket "$PGRST_SERVER_UNIX_SOCKET" \
-max-workers 1 \
-workers 1 \
-rate 0 \
-duration 60s \
"''${_arg_leftovers[@]}"
'';
loadtest =
checkedShellScript
{
name = "postgrest-loadtest";
docs = "Run the vegeta loadtests with PostgREST.";
args = [
"ARG_OPTIONAL_SINGLE([output], [o], [Filename to dump json output to], [./loadtest/result.bin])"
"ARG_OPTIONAL_SINGLE([testdir], [t], [Directory to load tests and fixtures from], [./test/load])"
"ARG_LEFTOVERS([additional vegeta arguments])"
];
inRootDir = true;
}
''
export PGRST_DB_CONFIG="false"
export PGRST_DB_POOL="1"
export PGRST_DB_TX_END="rollback-allow-override"
export PGRST_LOG_LEVEL="crit"
mkdir -p "$(dirname "$_arg_output")"
# shellcheck disable=SC2145
${withTools.withPg} --fixtures "$_arg_testdir"/fixtures.sql \
${withTools.withPgrst} \
sh -c "cd \"$_arg_testdir\" && ${runner} -targets targets.http \"''${_arg_leftovers[@]}\"" \
| tee "$_arg_output" \
| ${vegeta}/bin/vegeta report -type=text
'';
loadtestAgainst =
let
name = "postgrest-loadtest-against";
in
checkedShellScript
{
inherit name;
docs =
''
Run the vegeta loadtest twice:
- once on the <target> branch
- once in the current worktree
'';
args = [
"ARG_POSITIONAL_SINGLE([target], [Commit-ish reference to compare with])"
"ARG_LEFTOVERS([additional vegeta arguments])"
];
positionalCompletion =
''
if test "$prev" == "${name}"; then
__gitcomp_nl "$(__git_refs)"
fi
'';
inRootDir = true;
}
''
cat << EOF
Running loadtest on "$_arg_target"...
EOF
# Runs the test files from the current working tree
# to make sure both tests are run with the same files.
# Save the results in the current working tree, too,
# otherwise they'd be lost in the temporary working tree
# created by withTools.withGit.
${withTools.withGit} "$_arg_target" ${loadtest} --output "$PWD/loadtest/$_arg_target.bin" --testdir "$PWD/test/load" "''${_arg_leftovers[@]}"
cat << EOF
Done running on "$_arg_target".
EOF
cat << EOF
Running loadtest on HEAD...
EOF
${loadtest} --output "$PWD/loadtest/head.bin" --testdir "$PWD/test/load" "''${_arg_leftovers[@]}"
cat << EOF
Done running on HEAD.
EOF
'';
reporter =
checkedShellScript
{
name = "postgrest-loadtest-reporter";
docs = "Create a named json report for a single result file.";
args = [
"ARG_POSITIONAL_SINGLE([file], [Filename of result to create report for])"
"ARG_LEFTOVERS([additional vegeta arguments])"
];
inRootDir = true;
}
''
${vegeta}/bin/vegeta report -type=json "$_arg_file" \
| ${jq}/bin/jq --arg branch "$(basename "$_arg_file" .bin)" '. + {branch: $branch}'
'';
toMarkdown =
writers.writePython3 "postgrest-loadtest-to-markdown"
{
libraries = [ python3Packages.pandas python3Packages.tabulate ];
}
''
import sys
import pandas as pd
pd.read_json(sys.stdin) \
.set_index('param') \
.drop(['branch', 'earliest', 'end', 'latest']) \
.convert_dtypes() \
.to_markdown(sys.stdout, floatfmt='.0f')
'';
report =
checkedShellScript
{
name = "postgrest-loadtest-report";
docs = "Create a report of all loadtest reports as markdown.";
inRootDir = true;
}
''
find loadtest -type f -iname '*.bin' -exec ${reporter} {} \; \
| ${jq}/bin/jq '[leaf_paths as $path | {param: $path | join("."), (.branch): getpath($path)}]' \
| ${jq}/bin/jq --slurp 'flatten | group_by(.param) | map(add)' \
| ${toMarkdown}
'';
in
buildToolbox {
name = "postgrest-loadtest";
tools = [ loadtest loadtestAgainst report ];
}
+28
View File
@@ -0,0 +1,28 @@
# The memory tests have large dependencies (a profiled build of PostgREST)
# and are run less often than the spec tests, so we don't include them in
# the default test environment. We make them available through a separate module.
{ buildToolbox
, checkedShellScript
, curl
, postgrestProfiled
, withTools
}:
let
test =
checkedShellScript
{
name = "postgrest-test-memory";
docs = "Run the memory tests.";
inRootDir = true;
withPath = [ postgrestProfiled curl ];
}
''
${withTools.withPg} test/memory/memory-tests.sh
'';
in
buildToolbox
{
name = "postgrest-memory";
tools = [ test ];
}
+53
View File
@@ -0,0 +1,53 @@
{ buildToolbox
, checkedShellScript
, curl
, jq
, nix
}:
# Utility script for pinning the latest unstable version of Nixpkgs.
# Instead of pinning Nixpkgs based on the huge Git repository, we reference a
# specific tarball that only contains the source of the revision that we want
# to pin.
let
name =
"postgrest-nixpkgs-upgrade";
refUrl =
"https://api.github.com/repos/nixos/nixpkgs/git/ref/heads/nixpkgs-unstable";
githubV3Header =
"Accept: application/vnd.github.v3+json";
tarballUrlBase =
"https://github.com/nixos/nixpkgs/archive/";
upgrade =
checkedShellScript
{
inherit name;
docs = "Pin the newest unstable version of Nixpkgs.";
inRootDir = true;
}
''
commitHash="$(${curl}/bin/curl "${refUrl}" -H "${githubV3Header}" | ${jq}/bin/jq -r .object.sha)"
tarballUrl="${tarballUrlBase}$commitHash.tar.gz"
tarballHash="$(${nix}/bin/nix-prefetch-url --unpack "$tarballUrl")"
currentDate="$(date --iso)"
cat > nix/nixpkgs-version.nix << EOF
# Pinned version of Nixpkgs, generated with ${name}.
{
date = "$currentDate";
rev = "$commitHash";
tarballHash = "$tarballHash";
}
EOF
'';
in
buildToolbox
{
name = "postgrest-nixpkgs";
tools = [ upgrade ];
}
+54
View File
@@ -0,0 +1,54 @@
{ buildToolbox
, checkedShellScript
, curl
, jq
}:
let
dockerHubDescription =
let
description =
./docker-hub-description.md;
fullDescription =
./docker-hub-full-description.md;
in
checkedShellScript
{
name = "postgrest-release-dockerhub-description";
docs = "Update the repository description on Docker Hub.";
args = [
"ARG_USE_ENV([DOCKER_USER], [], [DockerHub user name])"
"ARG_USE_ENV([DOCKER_PASS], [], [DockerHub password])"
"ARG_USE_ENV([DOCKER_REPO], [], [DockerHub repository])"
];
}
''
# ARG_USE_ENV only adds defaults or docs for environment variables
# We manually implement a required check here
# See also: https://github.com/matejak/argbash/issues/80
: "''${DOCKER_USER:?DOCKER_USER is required}"
: "''${DOCKER_PASS:?DOCKER_PASS is required}"
: "''${DOCKER_REPO:?DOCKER_REPO is required}"
echo "Logging in to Docker Hub to get an auth token..."
token="$(
${curl}/bin/curl --fail -s \
--data-urlencode "username=$DOCKER_USER" \
--data-urlencode "password=$DOCKER_PASS" \
"https://hub.docker.com/v2/users/login/" \
| ${jq}/bin/jq -r .token
)"
repo_url="https://hub.docker.com/v2/repositories/$DOCKER_REPO/postgrest/"
echo "Patching both descriptions at $repo_url ..."
${curl}/bin/curl --fail -X PATCH "$repo_url" \
-H "Authorization: JWT $token" \
--data-urlencode description@${description} \
--data-urlencode full_description@${fullDescription}
'';
in
buildToolbox
{
name = "postgrest-release";
tools = [ dockerHubDescription ];
}
@@ -0,0 +1 @@
REST API for any Postgres database
@@ -0,0 +1,70 @@
# PostgREST
[![Join the chat at https://gitter.im/begriffs/postgrest](https://img.shields.io/badge/gitter-join%20chat%20%E2%86%92-brightgreen.svg)](https://gitter.im/begriffs/postgrest)
[![Donate](https://img.shields.io/badge/Donate-Patreon-orange.svg?colorB=F96854)](https://www.patreon.com/postgrest)
[![Donate](https://img.shields.io/badge/Donate-PayPal-green.svg)](https://www.paypal.me/postgrest)
[![Docs](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://postgrest.org)
[![Build Status](https://github.com/postgrest/postgrest/actions/workflows/ci.yaml/badge.svg?branch=main)](https://github.com/PostgREST/postgrest/actions?query=branch%3Amain)
PostgREST serves a fully RESTful API from any existing PostgreSQL database. It
provides a cleaner, more standards-compliant, faster API than you are likely to
write from scratch.
## Sponsors
<table>
<tbody>
<tr>
<td align="center" valign="middle">
<a href="https://www.cybertec-postgresql.com/en/?utm_source=postgrest.org&utm_medium=referral&utm_campaign=postgrest" target="_blank">
<img width="222px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/cybertec-new.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://www.2ndquadrant.com/en/?utm_campaign=External%20Websites&utm_source=PostgREST&utm_medium=Logo" target="_blank">
<img width="296px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/2ndquadrant.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://tryretool.com/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/retool.png">
</a>
</td>
</tr>
<tr></tr>
<tr>
<td align="center" valign="middle">
<a href="https://gnuhost.eu/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/gnuhost.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://supabase.io?utm_source=postgrest%20backers&utm_medium=open%20source%20partner&utm_campaign=postgrest%20backers%20github&utm_term=homepage" target="_blank">
<img width="296px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/supabase.png">
</a>
</td>
<td align="center" valign="middle">
<a href="https://oblivious.ai/?utm_source=sponsor&utm_campaign=postgrest" target="_blank">
<img width="296px" src="https://raw.githubusercontent.com/PostgREST/postgrest/main/static/oblivious.jpg">
</a>
</td>
</tr>
</tbody>
</table>
# Usage
To learn how to use this container, see the [PostgREST Docker
documentation](https://postgrest.com/en/stable/install.html#docker).
You can configure the PostgREST image by setting
[enviroment variables](https://postgrest.org/en/stable/configuration.html).
# How this image is built
The image is built from scratch using
[Nix](https://nixos.org/nixpkgs/manual/#sec-pkgs-dockerTools) instead of a
`Dockerfile`, which yields a higly secure and optimized image. This is also why
no commands are listed in the image history. See the [PostgREST
respository](https://github.com/PostgREST/postgrest/tree/main/nix/docker) for
details on the build process and how to inspect the image.
+84
View File
@@ -0,0 +1,84 @@
{ actionlint
, black
, buildToolbox
, checkedShellScript
, git
, hlint
, hsie
, nixpkgs-fmt
, shellcheck
, silver-searcher
, statix
, stylish-haskell
}:
let
style =
checkedShellScript
{
name = "postgrest-style";
docs = "Automatically format Haskell, Nix and Python files.";
inRootDir = true;
}
''
# Format Nix files
${statix}/bin/statix fix
${nixpkgs-fmt}/bin/nixpkgs-fmt . > /dev/null 2> /dev/null
# Format Haskell files
# --vimgrep fixes a bug in ag: https://github.com/ggreer/the_silver_searcher/issues/753
${silver-searcher}/bin/ag -l --vimgrep -g '\.l?hs$' . \
| xargs ${stylish-haskell}/bin/stylish-haskell -i
# Format Python files
${black}/bin/black . 2> /dev/null
'';
# Script to check whether any uncommited changes result from postgrest-style
styleCheck =
checkedShellScript
{
name = "postgrest-style-check";
docs = "Check whether postgrest-style results in any uncommited changes.";
inRootDir = true;
}
''
${style}
trap "echo postgrest-style-check failed. Run postgrest-style to fix issues automatically." ERR
${git}/bin/git diff-index --exit-code HEAD -- '*.hs' '*.lhs' '*.nix'
'';
lint =
checkedShellScript
{
name = "postgrest-lint";
docs = "Lint all Haskell files, bash scripts and github workflows.";
inRootDir = true;
}
''
echo "Checking consistency of import aliases in Haskell code..."
${hsie} check-aliases main src
echo "Linting Haskell files..."
# --vimgrep fixes a bug in ag: https://github.com/ggreer/the_silver_searcher/issues/753
${silver-searcher}/bin/ag -l --vimgrep -g '\.l?hs$' . \
| xargs ${hlint}/bin/hlint -X QuasiQuotes -X NoPatternSynonyms
echo "Linting bash scripts..."
${shellcheck}/bin/shellcheck \
.github/get_cirrusci_freebsd \
.github/release \
test/with_tmp_db
echo "Linting workflows..."
${actionlint}/bin/actionlint
'';
in
buildToolbox
{
name = "postgrest-style";
tools = [ style styleCheck lint ];
}
+212
View File
@@ -0,0 +1,212 @@
{ buildToolbox
, cabal-install
, checkedShellScript
, devCabalOptions
, ghc
, glibcLocales
, gnugrep
, haskellPackages
, hpc-codecov
, jq
, postgrest
, python3
, runtimeShell
, withTools
, yq
}:
let
testSpec =
checkedShellScript
{
name = "postgrest-test-spec";
docs = "Run the Haskell test suite";
inRootDir = true;
withEnv = postgrest.env;
}
''
${withTools.withPg} ${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:spec
'';
testQuerycost =
checkedShellScript
{
name = "postgrest-test-querycost";
docs = "Run the Haskell test suite for query costs";
inRootDir = true;
withEnv = postgrest.env;
}
''
${withTools.withPg} ${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:querycost
'';
testDoctests =
checkedShellScript
{
name = "postgrest-test-doctests";
docs = "Run the Haskell doctest test suite";
inRootDir = true;
withEnv = postgrest.env;
}
''
# For unknown reasons, doctests uses the wrong GHC package database outside
# nix-shell and fails, so we set the package path explicitly
#ghcWithPackages="$(cat ${postgrest.env})"
#ghcVersion="$(ls "$ghcWithPackages/lib")"
#export GHC_PACKAGE_PATH="$ghcWithPackages/lib/$ghcVersion/package.conf.d/"
${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:doctests
'';
testSpecIdempotence =
checkedShellScript
{
name = "postgrest-test-spec-idempotence";
docs = "Check that the Haskell tests can be run multiple times against the same db.";
inRootDir = true;
withEnv = postgrest.env;
}
''
${withTools.withPg} ${runtimeShell} -c " \
${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:spec && \
${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:spec"
'';
ioTestPython =
python3.withPackages (ps: [
ps.pyjwt
ps.pytest
ps.pytest_xdist
ps.pyyaml
ps.requests
ps.requests-unixsocket
]);
testIO =
checkedShellScript
{
name = "postgrest-test-io";
docs = "Run the pytest-based IO tests.";
args = [ "ARG_LEFTOVERS([pytest arguments])" ];
inRootDir = true;
withEnv = postgrest.env;
}
''
${cabal-install}/bin/cabal v2-build ${devCabalOptions}
${cabal-install}/bin/cabal v2-exec -- ${withTools.withPg} -f test/io/fixtures.sql \
${ioTestPython}/bin/pytest -v test/io "''${_arg_leftovers[@]}"
'';
dumpSchema =
checkedShellScript
{
name = "postgrest-dump-schema";
docs = "Dump the loaded schema's DbStructure as a yaml file.";
inRootDir = true;
withEnv = postgrest.env;
withPath = [ jq ];
}
''
${withTools.withPg} \
${cabal-install}/bin/cabal v2-run ${devCabalOptions} --verbose=0 -- \
postgrest --dump-schema \
| ${yq}/bin/yq -y .
'';
coverage =
checkedShellScript
{
name = "postgrest-coverage";
docs = "Run spec and io tests while collecting hpc coverage data. First runs weeder to detect dead code.";
args = [ "ARG_LEFTOVERS([hpc report arguments])" ];
inRootDir = true;
redirectTixFiles = false;
withEnv = postgrest.env;
withTmpDir = true;
}
''
export LOCALE_ARCHIVE="${glibcLocales}/lib/locale/locale-archive"
# clean up previous coverage reports
mkdir -p coverage
rm -rf coverage/*
# build once before running all the tests
${cabal-install}/bin/cabal v2-build ${devCabalOptions} exe:postgrest lib:postgrest test:spec test:querycost
${haskellPackages.weeder}/bin/weeder --config=./test/weeder.dhall || echo Found dead code: Check file list above.
# collect all tests
HPCTIXFILE="$tmpdir"/io.tix \
${withTools.withPg} -f test/io/fixtures.sql ${cabal-install}/bin/cabal v2-exec ${devCabalOptions} -- \
${ioTestPython}/bin/pytest -v test/io
HPCTIXFILE="$tmpdir"/spec.tix \
${withTools.withPg} ${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:spec
HPCTIXFILE="$tmpdir"/querycost.tix \
${withTools.withPg} ${cabal-install}/bin/cabal v2-run ${devCabalOptions} test:querycost
# Note: No coverage for doctests, as doctests leverage GHCi and GHCi does not support hpc
# collect all the tix files
${ghc}/bin/hpc sum --union --exclude=Paths_postgrest --output="$tmpdir"/tests.tix \
"$tmpdir"/io*.tix "$tmpdir"/spec.tix "$tmpdir"/querycost.tix
# prepare the overlay
${ghc}/bin/hpc overlay --output="$tmpdir"/overlay.tix test/coverage.overlay
${ghc}/bin/hpc sum --union --output="$tmpdir"/tests-overlay.tix "$tmpdir"/tests.tix "$tmpdir"/overlay.tix
# check nothing in the overlay is actually tested
${ghc}/bin/hpc map --function=inv --output="$tmpdir"/inverted.tix "$tmpdir"/tests.tix
${ghc}/bin/hpc combine --function=sub \
--output="$tmpdir"/check.tix "$tmpdir"/overlay.tix "$tmpdir"/inverted.tix
# returns zero exit code if any count="<non-zero>" lines are found, i.e.
# something is covered by both the overlay and the tests
if ${ghc}/bin/hpc report --xml "$tmpdir"/check.tix | ${gnugrep}/bin/grep -qP 'count="[^0]'
then
${ghc}/bin/hpc markup --highlight-covered --destdir=coverage/overlay "$tmpdir"/overlay.tix || true
${ghc}/bin/hpc markup --highlight-covered --destdir=coverage/check "$tmpdir"/check.tix || true
echo "ERROR: Something is covered by both the tests and the overlay:"
echo "file://$(pwd)/coverage/check/hpc_index.html"
exit 1
else
# copy the result .tix file to the coverage/ dir to make it available to postgrest-coverage-draft-overlay, too
cp "$tmpdir"/tests-overlay.tix coverage/postgrest.tix
# prepare codecov json report
${hpc-codecov}/bin/hpc-codecov --mix=.hpc --out=coverage/codecov.json coverage/postgrest.tix
# create html and stdout reports
${ghc}/bin/hpc markup --destdir=coverage coverage/postgrest.tix
echo "file://$(pwd)/coverage/hpc_index.html"
${ghc}/bin/hpc report coverage/postgrest.tix "''${_arg_leftovers[@]}"
fi
'';
coverageDraftOverlay =
checkedShellScript
{
name = "postgrest-coverage-draft-overlay";
docs = "Create a draft overlay from current coverage report.";
inRootDir = true;
}
''
${ghc}/bin/hpc draft --output=test/coverage.overlay coverage/postgrest.tix
sed -i 's|^module \(.*\):|module \1/|g' test/coverage.overlay
'';
in
buildToolbox
{
name = "postgrest-tests";
tools =
[
testSpec
testQuerycost
testDoctests
testSpecIdempotence
testIO
dumpSchema
coverage
coverageDraftOverlay
];
}
+274
View File
@@ -0,0 +1,274 @@
{ bashCompletion
, buildToolbox
, cabal-install
, checkedShellScript
, curl
, devCabalOptions
, git
, lib
, postgresqlVersions
, postgrest
, writeText
}:
let
withTmpDb =
{ name, postgresql }:
checkedShellScript
{
name = "postgrest-with-${name}";
docs = "Run the given command in a temporary database with ${name}";
args =
[
"ARG_OPTIONAL_SINGLE([fixtures], [f], [SQL file to load fixtures from], [test/spec/fixtures/load.sql])"
"ARG_POSITIONAL_SINGLE([command], [Command to run])"
"ARG_LEFTOVERS([command arguments])"
"ARG_USE_ENV([PGUSER], [postgrest_test_authenticator], [Authenticator PG role])"
"ARG_USE_ENV([PGDATABASE], [postgres], [PG database name])"
"ARG_USE_ENV([PGRST_DB_SCHEMAS], [test], [Schema to expose])"
"ARG_USE_ENV([PGRST_DB_ANON_ROLE], [postgrest_test_anonymous], [Anonymous PG role])"
];
positionalCompletion = "_command";
inRootDir = true;
redirectTixFiles = false;
withPath = [ postgresql ];
withTmpDir = true;
}
''
# avoid starting multiple layers of withTmpDb
if test -v PGRST_DB_URI; then
exec "$_arg_command" "''${_arg_leftovers[@]}"
fi
setuplog="$tmpdir/setup.log"
log () {
echo "$1" >> "$setuplog"
}
mkdir -p "$tmpdir"/{db,socket}
# remove data dir, even if we keep tmpdir - no need to upload it to artifacts
trap 'rm -rf $tmpdir/db' EXIT
export PGDATA="$tmpdir/db"
export PGHOST="$tmpdir/socket"
export PGUSER
export PGDATABASE
export PGRST_DB_URI="postgresql:///$PGDATABASE?host=$PGHOST&user=$PGUSER"
export PGRST_DB_SCHEMAS
export PGRST_DB_ANON_ROLE
log "Initializing database cluster..."
# We try to make the database cluster as independent as possible from the host
# by specifying the timezone, locale and encoding.
PGTZ=UTC initdb --no-locale --encoding=UTF8 --nosync -U "$PGUSER" --auth=trust \
>> "$setuplog"
log "Starting the database cluster..."
# Instead of listening on a local port, we will listen on a unix domain socket.
pg_ctl -l "$tmpdir/db.log" -w start -o "-F -c listen_addresses=\"\" -k $PGHOST" \
>> "$setuplog"
stop () {
log "Stopping the database cluster..."
pg_ctl stop -m i >> "$setuplog"
rm -rf "$tmpdir/db"
}
trap stop EXIT
log "Loading fixtures..."
psql -v ON_ERROR_STOP=1 -f "$_arg_fixtures" >> "$setuplog"
log "Done. Running command..."
("$_arg_command" "''${_arg_leftovers[@]}")
'';
# Helper script for running a command against all PostgreSQL versions.
withPgAll =
let
runners =
builtins.map
(version:
''
cat << EOF
Running against ${version.name}...
EOF
trap 'echo "Failed on ${version.name}"' exit
(${withTmpDb version} "$_arg_command" "''${_arg_leftovers[@]}")
trap "" exit
cat << EOF
Done running against ${version.name}.
EOF
'')
postgresqlVersions;
in
checkedShellScript
{
name = "postgrest-with-all";
docs = "Run command against all supported PostgreSQL versions.";
args =
[
"ARG_POSITIONAL_SINGLE([command], [Command to run])"
"ARG_LEFTOVERS([command arguments])"
];
positionalCompletion = "_command";
inRootDir = true;
}
(lib.concatStringsSep "\n\n" runners);
# Create a `postgrest-with-postgresql-` for each PostgreSQL version
withPgVersions = builtins.map withTmpDb postgresqlVersions;
withPg = builtins.head withPgVersions;
withGit =
let
name = "postgrest-with-git";
in
checkedShellScript
{
inherit name;
docs =
''
Create a new worktree of the postgrest repo in a temporary directory and
check out <commit>, then run <command> with arguments inside the temporary folder.
'';
args =
[
"ARG_POSITIONAL_SINGLE([commit], [Commit-ish reference to run command with])"
"ARG_POSITIONAL_SINGLE([command], [Command to run])"
"ARG_LEFTOVERS([command arguments])"
];
positionalCompletion =
''
if test "$prev" == "${name}"; then
__gitcomp_nl "$(__git_refs)"
else
_command_offset 2
fi
'';
inRootDir = true;
}
''
# not using withTmpDir here, because we don't want to keep the directory on error
tmpdir="$(mktemp -d)"
trap 'rm -rf "$tmpdir"' EXIT
${git}/bin/git worktree add -f "$tmpdir" "$_arg_commit" > /dev/null
cd "$tmpdir"
("$_arg_command" "''${_arg_leftovers[@]}")
${git}/bin/git worktree remove -f "$tmpdir" > /dev/null
'';
legacyConfig =
writeText "legacy.conf"
''
# Using this config file to support older postgrest versions for `postgrest-loadtest-against`
db-uri="$(PGRST_DB_URI)"
db-schema="$(PGRST_DB_SCHEMAS)"
db-anon-role="$(PGRST_DB_ANON_ROLE)"
db-pool="$(PGRST_DB_POOL)"
server-unix-socket="$(PGRST_SERVER_UNIX_SOCKET)"
log-level="$(PGRST_LOG_LEVEL)"
'';
waitForPgrstPid =
checkedShellScript
{
name = "postgrest-wait-for-pgrst-pid";
docs = "Wait for PostgREST to be running. Needs to be a separate command for timeout to work below.";
args = [
"ARG_USE_ENV([PGRST_SERVER_UNIX_SOCKET], [], [Unix socket to check for running PostgREST instance])"
];
}
''
# ARG_USE_ENV only adds defaults or docs for environment variables
# We manually implement a required check here
# See also: https://github.com/matejak/argbash/issues/80
: "''${PGRST_SERVER_UNIX_SOCKET:?PGRST_SERVER_UNIX_SOCKET is required}"
until [ -S "$PGRST_SERVER_UNIX_SOCKET" ]
do
sleep 0.1
done
# return pid of postgrest process
lsof -t -c '/^postgrest$/' "$PGRST_SERVER_UNIX_SOCKET"
'';
waitForPgrstReady =
checkedShellScript
{
name = "postgrest-wait-for-pgrst-ready";
docs = "Wait for PostgREST to be ready to serve requests. Needs to be a separate command for timeout to work below.";
args = [
"ARG_USE_ENV([PGRST_SERVER_UNIX_SOCKET], [], [Unix socket to check for running PostgREST instance])"
];
}
''
# ARG_USE_ENV only adds defaults or docs for environment variables
# We manually implement a required check here
# See also: https://github.com/matejak/argbash/issues/80
: "''${PGRST_SERVER_UNIX_SOCKET:?PGRST_SERVER_UNIX_SOCKET is required}"
function check_status () {
${curl}/bin/curl -s -o /dev/null -w "%{http_code}" --unix-socket "$PGRST_SERVER_UNIX_SOCKET" http://localhost/
}
while [[ "$(check_status)" != "200" ]];
do sleep 0.1;
done
'';
withPgrst =
checkedShellScript
{
name = "postgrest-with-pgrst";
docs = "Build and run PostgREST and run <command> with PGRST_SERVER_UNIX_SOCKET set.";
args =
[
"ARG_POSITIONAL_SINGLE([command], [Command to run])"
"ARG_LEFTOVERS([command arguments])"
];
positionalCompletion = "_command";
inRootDir = true;
withEnv = postgrest.env;
withTmpDir = true;
}
''
export PGRST_SERVER_UNIX_SOCKET="$tmpdir"/postgrest.socket
${cabal-install}/bin/cabal v2-build ${devCabalOptions} > "$tmpdir"/build.log 2>&1
${cabal-install}/bin/cabal v2-run ${devCabalOptions} --verbose=0 -- \
postgrest ${legacyConfig} > "$tmpdir"/run.log 2>&1 &
# to get the pid of the postgrest process, we need to jump through some hoops
# $! will return the pid of cabal - but killing this, will not propagate to postgrest
pid=$(timeout -s TERM 1 ${waitForPgrstPid})
cleanup() {
kill "$pid" || true
}
trap cleanup EXIT
timeout -s TERM 5 ${waitForPgrstReady}
("$_arg_command" "''${_arg_leftovers[@]}")
'';
in
buildToolbox
{
name = "postgrest-with";
tools = [ withPgAll withGit withPgrst ] ++ withPgVersions;
# make withTools available for other nix files
extra = { inherit withGit withPg withPgAll withPgrst; };
}
+168 -119
View File
@@ -1,8 +1,8 @@
name: postgrest
version: 7.0.1
version: 9.0.1
synopsis: REST API for any Postgres database
description: Reads the schema of a PostgreSQL database and creates RESTful routes
for the tables and views, supporting all HTTP verbs that security
for tables, views, and functions, supporting all HTTP verbs that security
permits.
license: MIT
license-file: LICENSE
@@ -19,61 +19,88 @@ source-repository head
type: git
location: git://github.com/PostgREST/postgrest.git
flag ci
flag dev
default: False
manual: True
description: No warnings allowed in continuous integration
description: Development flags
flag hpc
default: True
manual: True
description: Enable HPC (dev only)
library
exposed-modules: PostgREST.ApiRequest
PostgREST.App
default-language: Haskell2010
default-extensions: OverloadedStrings
NoImplicitPrelude
hs-source-dirs: src
exposed-modules: PostgREST.App
PostgREST.AppState
PostgREST.Auth
PostgREST.CLI
PostgREST.Config
PostgREST.DbRequestBuilder
PostgREST.Config.Database
PostgREST.Config.JSPath
PostgREST.Config.PgVersion
PostgREST.Config.Proxy
PostgREST.ContentType
PostgREST.Cors
PostgREST.DbStructure
PostgREST.DbStructure.Identifiers
PostgREST.DbStructure.Proc
PostgREST.DbStructure.Relationship
PostgREST.DbStructure.Table
PostgREST.Error
PostgREST.GucHeader
PostgREST.Logger
PostgREST.Middleware
PostgREST.OpenAPI
PostgREST.Parsers
PostgREST.QueryBuilder
PostgREST.Statements
PostgREST.Query.QueryBuilder
PostgREST.Query.SqlFragment
PostgREST.Query.Statements
PostgREST.RangeQuery
PostgREST.Types
PostgREST.Request.ApiRequest
PostgREST.Request.DbRequestBuilder
PostgREST.Request.Parsers
PostgREST.Request.Preferences
PostgREST.Request.Types
PostgREST.Version
PostgREST.Workers
other-modules: Paths_postgrest
PostgREST.Private.Common
PostgREST.Private.QueryFragment
hs-source-dirs: src
build-depends: base >= 4.9 && < 4.15
build-depends: base >= 4.9 && < 4.16
, HTTP >= 4000.3.7 && < 4000.4
, Ranged-sets >= 0.3 && < 0.5
, aeson >= 1.4.7 && < 1.5
, ansi-wl-pprint >= 0.6.7 && < 0.7
, base64-bytestring >= 1 && < 1.2
, aeson >= 1.4.7 && < 1.6
, auto-update >= 0.1.4 && < 0.2
, base64-bytestring >= 1 && < 1.3
, bytestring >= 0.10.8 && < 0.11
, case-insensitive >= 1.2 && < 1.3
, cassava >= 0.4.5 && < 0.6
, configurator-pg >= 0.2 && < 0.3
, containers >= 0.5.7 && < 0.7
, contravariant >= 1.4 && < 1.6
, contravariant-extras >= 0.3.3 && < 0.4
, cookie >= 0.4.2 && < 0.5
, either >= 4.4.1 && < 5.1
, gitrev >= 1.2 && < 1.4
, hasql >= 1.4 && < 1.5
, hasql-dynamic-statements == 0.3.1
, hasql-notifications >= 0.1 && < 0.3
, hasql-pool >= 0.5 && < 0.6
, hasql-transaction >= 0.7.2 && < 1.1
, hasql-transaction >= 1.0.1 && < 1.1
, heredoc >= 0.2 && < 0.3
, http-types >= 0.12.2 && < 0.13
, insert-ordered-containers >= 0.2.2 && < 0.3
, interpolatedstring-perl6 >= 1 && < 1.1
, jose >= 0.8.1 && < 0.9
, lens >= 4.14 && < 4.20
, lens >= 4.14 && < 5.1
, lens-aeson >= 1.0.1 && < 1.2
, mtl >= 2.2.2 && < 2.3
, network-uri >= 2.6.1 && < 2.8
, optparse-applicative >= 0.13 && < 0.16
, optparse-applicative >= 0.13 && < 0.17
, parsec >= 3.1.11 && < 3.2
, protolude >= 0.3 && < 0.4
, protolude >= 0.3.1 && < 0.4
, regex-tdfa >= 1.2.2 && < 1.4
, retry >= 0.7.4 && < 0.10
, scientific >= 0.3.4 && < 0.4
, swagger2 >= 2.4 && < 2.7
, text >= 1.2.2 && < 1.3
@@ -82,54 +109,61 @@ library
, vector >= 0.11 && < 0.13
, wai >= 3.2.1 && < 3.3
, wai-cors >= 0.2.5 && < 0.3
, wai-extra >= 3.0.19 && < 3.1
, wai-middleware-static >= 0.8.1 && < 0.9
default-language: Haskell2010
default-extensions: OverloadedStrings
QuasiQuotes
NoImplicitPrelude
ghc-options: -O2 -Werror -Wall -fwarn-identities
, wai-extra >= 3.1.8 && < 3.2
, warp >= 3.3.19 && < 3.4
-- -fno-spec-constr may help keep compile time memory use in check,
-- see https://gitlab.haskell.org/ghc/ghc/issues/16017#note_219304
-- -optP-Wno-nonportable-include-path
-- prevents build failures on case-insensitive filesystems (macos),
-- see https://github.com/commercialhaskell/stack/issues/3918
ghc-options: -Werror -Wall -fwarn-identities
-fno-spec-constr -optP-Wno-nonportable-include-path
-- -fno-spec-constr may help keep compile time memory use in check,
-- see https://gitlab.haskell.org/ghc/ghc/issues/16017#note_219304
-- -optP-Wno-nonportable-include-path
-- prevents build failures on case-insensitive filesystems (macos),
-- see https://github.com/commercialhaskell/stack/issues/3918
if flag(dev)
ghc-options: -O0 -fwrite-ide-info
if flag(hpc)
ghc-options: -fhpc -hpcdir .hpc
else
ghc-options: -O2
if !os(windows)
build-depends:
unix
, directory >= 1.2.6 && < 1.4
, network >= 2.6 && < 3.2
exposed-modules:
PostgREST.Unix
executable postgrest
main-is: Main.hs
hs-source-dirs: main
build-depends: base >= 4.9 && < 4.15
, auto-update >= 0.1.4 && < 0.2
, base64-bytestring >= 1 && < 1.2
, bytestring >= 0.10.8 && < 0.11
, directory >= 1.2.6 && < 1.4
, either >= 4.4.1 && < 5.1
, hasql >= 1.4 && < 1.5
, hasql-pool >= 0.5 && < 0.6
, hasql-transaction >= 0.7.2 && < 1.1
, network < 3.2
, postgrest
, protolude >= 0.3 && < 0.4
, retry >= 0.7.4 && < 0.9
, text >= 1.2.2 && < 1.3
, time >= 1.6 && < 1.11
, wai >= 3.2.1 && < 3.3
, warp >= 3.2.12 && < 3.4
default-language: Haskell2010
default-extensions: OverloadedStrings
QuasiQuotes
NoImplicitPrelude
ghc-options: -threaded -rtsopts "-with-rtsopts=-N -I2"
hs-source-dirs: main
main-is: Main.hs
build-depends: base >= 4.9 && < 4.16
, containers >= 0.5.7 && < 0.7
, postgrest
, protolude >= 0.3.1 && < 0.4
ghc-options: -threaded -rtsopts "-with-rtsopts=-N -I0 -qg"
-O2 -Werror -Wall -fwarn-identities
-fno-spec-constr -optP-Wno-nonportable-include-path
if !os(windows)
build-depends: unix
other-modules: UnixSocket
if flag(dev)
ghc-options: -O0 -fwrite-ide-info
-- https://github.com/PostgREST/postgrest/issues/387
-with-rtsopts=-K1K
if flag(hpc)
ghc-options: -fhpc -hpcdir .hpc
else
ghc-options: -O2
test-suite spec
type: exitcode-stdio-1.0
default-language: Haskell2010
default-extensions: OverloadedStrings
QuasiQuotes
NoImplicitPrelude
hs-source-dirs: test/spec
main-is: Main.hs
other-modules: Feature.AndOrParamsSpec
Feature.AsymmetricJwtSpec
@@ -139,102 +173,117 @@ test-suite spec
Feature.ConcurrentSpec
Feature.CorsSpec
Feature.DeleteSpec
Feature.DisabledOpenApiSpec
Feature.EmbedDisambiguationSpec
Feature.EmbedInnerJoinSpec
Feature.ExtraSearchPathSpec
Feature.HtmlRawOutputSpec
Feature.InsertSpec
Feature.IgnorePrivOpenApiSpec
Feature.JsonOperatorSpec
Feature.LegacyGucsSpec
Feature.MultipleSchemaSpec
Feature.NoJwtSpec
Feature.NonexistentSchemaSpec
Feature.PgVersion95Spec
Feature.PgVersion96Spec
Feature.OpenApiSpec
Feature.OptionsSpec
Feature.ProxySpec
Feature.QueryLimitedSpec
Feature.QuerySpec
Feature.RangeSpec
Feature.RawOutputTypesSpec
Feature.RollbackSpec
Feature.RootSpec
Feature.RpcPreRequestGucsSpec
Feature.RpcSpec
Feature.SingularSpec
Feature.StructureSpec
Feature.UnicodeSpec
Feature.UpdateSpec
Feature.UpsertSpec
Feature.RawOutputTypesSpec
Feature.HtmlRawOutputSpec
Feature.MultipleSchemaSpec
SpecHelper
TestTypes
hs-source-dirs: test
build-depends: base >= 4.9 && < 4.15
, aeson >= 1.4.7 && < 1.5
build-depends: base >= 4.9 && < 4.16
, aeson >= 1.4.7 && < 1.6
, aeson-qq >= 0.8.1 && < 0.9
, async >= 2.1.1 && < 2.3
, auto-update >= 0.1.4 && < 0.2
, base64-bytestring >= 1 && < 1.2
, base64-bytestring >= 1 && < 1.3
, bytestring >= 0.10.8 && < 0.11
, case-insensitive >= 1.2 && < 1.3
, cassava >= 0.4.5 && < 0.6
, containers >= 0.5.7 && < 0.7
, contravariant >= 1.4 && < 1.6
, hasql >= 1.4 && < 1.5
, hasql-pool >= 0.5 && < 0.6
, hasql-transaction >= 0.7.2 && < 1.1
, hasql-transaction >= 1.0.1 && < 1.1
, heredoc >= 0.2 && < 0.3
, hspec >= 2.3 && < 2.8
, hspec-wai >= 0.10 && < 0.11
, hspec-wai-json >= 0.10 && < 0.11
, hspec >= 2.3 && < 2.9
, hspec-wai >= 0.10 && < 0.12
, hspec-wai-json >= 0.10 && < 0.12
, http-types >= 0.12.3 && < 0.13
, lens >= 4.14 && < 4.20
, lens >= 4.14 && < 5.1
, lens-aeson >= 1.0.1 && < 1.2
, monad-control >= 1.0.1 && < 1.1
, postgrest
, process >= 1.4.2 && < 1.7
, protolude >= 0.3 && < 0.4
, protolude >= 0.3.1 && < 0.4
, regex-tdfa >= 1.2.2 && < 1.4
, text >= 1.2.2 && < 1.3
, time >= 1.6 && < 1.11
, transformers-base >= 0.4.4 && < 0.5
, wai >= 3.2.1 && < 3.3
, wai-extra >= 3.0.19 && < 3.1
, wai-extra >= 3.0.19 && < 3.2
ghc-options: -O0 -Werror -Wall -fwarn-identities
-fno-spec-constr -optP-Wno-nonportable-include-path
-fno-warn-missing-signatures
-fwrite-ide-info
-- https://github.com/PostgREST/postgrest/issues/387
-with-rtsopts=-K33K
test-suite querycost
type: exitcode-stdio-1.0
default-language: Haskell2010
default-extensions: OverloadedStrings
QuasiQuotes
NoImplicitPrelude
ghc-options: -threaded -rtsopts -with-rtsopts=-N
hs-source-dirs: test/spec
main-is: QueryCost.hs
other-modules: SpecHelper
build-depends: base >= 4.9 && < 4.16
, aeson >= 1.4.7 && < 1.6
, base64-bytestring >= 1 && < 1.3
, bytestring >= 0.10.8 && < 0.11
, case-insensitive >= 1.2 && < 1.3
, containers >= 0.5.7 && < 0.7
, contravariant >= 1.4 && < 1.6
, hasql >= 1.4 && < 1.5
, hasql-dynamic-statements == 0.3.1
, hasql-pool >= 0.5 && < 0.6
, hasql-transaction >= 1.0.1 && < 1.1
, heredoc >= 0.2 && < 0.3
, hspec >= 2.3 && < 2.9
, hspec-wai >= 0.10 && < 0.12
, http-types >= 0.12.3 && < 0.13
, lens >= 4.14 && < 5.1
, lens-aeson >= 1.0.1 && < 1.2
, postgrest
, process >= 1.4.2 && < 1.7
, protolude >= 0.3.1 && < 0.4
, regex-tdfa >= 1.2.2 && < 1.4
, wai-extra >= 3.0.19 && < 3.2
ghc-options: -O0 -Werror -Wall -fwarn-identities
-fno-spec-constr -optP-Wno-nonportable-include-path
-fwrite-ide-info
-- https://github.com/PostgREST/postgrest/issues/387
-with-rtsopts=-K1K
Test-Suite spec-querycost
Type: exitcode-stdio-1.0
Default-Language: Haskell2010
default-extensions: OverloadedStrings, QuasiQuotes, NoImplicitPrelude
Hs-Source-Dirs: test
Main-Is: QueryCost.hs
Other-Modules: SpecHelper
Build-Depends: base >= 4.9 && < 4.15
, aeson >= 1.4.7 && < 1.5
, aeson-qq >= 0.8.1 && < 0.9
, async >= 2.1.1 && < 2.3
, auto-update >= 0.1.4 && < 0.2
, base64-bytestring >= 1 && < 1.2
, bytestring >= 0.10.8 && < 0.11
, case-insensitive >= 1.2 && < 1.3
, cassava >= 0.4.5 && < 0.6
, containers >= 0.5.7 && < 0.7
, contravariant >= 1.4 && < 1.6
, hasql >= 1.4 && < 1.5
, hasql-pool >= 0.5 && < 0.6
, hasql-transaction >= 0.7.2 && < 1.1
, heredoc >= 0.2 && < 0.3
, hspec >= 2.3 && < 2.8
, hspec-wai >= 0.10 && < 0.11
, hspec-wai-json >= 0.10 && < 0.11
, http-types >= 0.12.3 && < 0.13
, lens >= 4.14 && < 4.20
, lens-aeson >= 1.0.1 && < 1.2
, monad-control >= 1.0.1 && < 1.1
, postgrest
, process >= 1.4.2 && < 1.7
, protolude >= 0.3 && < 0.4
, regex-tdfa >= 1.2.2 && < 1.4
, text >= 1.2.2 && < 1.3
, time >= 1.6 && < 1.11
, transformers-base >= 0.4.4 && < 0.5
, wai >= 3.2.1 && < 3.3
, wai-extra >= 3.0.19 && < 3.1
test-suite doctests
type: exitcode-stdio-1.0
default-language: Haskell2010
default-extensions: OverloadedStrings
NoImplicitPrelude
hs-source-dirs: test/doc
main-is: Main.hs
build-depends: base >= 4.9 && < 4.16
, doctest >= 0.8
, postgrest
, pretty-simple
, protolude >= 0.3.1 && < 0.4
ghc-options: -threaded -O0 -Werror -Wall -fwarn-identities
-fno-spec-constr -optP-Wno-nonportable-include-path
+49 -12
View File
@@ -1,23 +1,60 @@
with (import ./default.nix);
pkgs.lib.overrideDerivation env (
# The additional modules below have large dependencies and are therefore
# disabled by default. You can activate them by passing arguments to nix-shell,
# e.g.:
#
# nix-shell --arg docker true
#
# We highly recommend that use the PostgREST binary cache by installing cachix
# (https://app.cachix.org/) and running `cachix use postgrest`.
{ docker ? false
, memory ? false
}:
let
postgrest =
import ./default.nix;
inherit (postgrest) pkgs;
inherit (pkgs) lib;
toolboxes =
[
postgrest.cabalTools
postgrest.devTools
postgrest.loadtest
postgrest.nixpkgsTools
postgrest.style
postgrest.tests
postgrest.withTools
postgrest.release
]
++ lib.optional docker postgrest.docker
++ lib.optional memory postgrest.memory;
in
lib.overrideDerivation postgrest.env (
base: {
buildInputs =
base.buildInputs ++ [
pkgs.cabal-install
pkgs.stack
pkgs.cabal2nix
pkgs.git
pkgs.postgresql
nixpkgsUpgrade
tests
style
lint
];
postgrest.hsie.bin
]
++ toolboxes;
shellHook =
''
# Set our pinned version of Nixpkgs in the NIX_PATH so that
# `stack --nix` also uses that version.
NIX_PATH="nixpkgs=${nixpkgs}"
'';
source ${pkgs.bashCompletion}/etc/profile.d/bash_completion.sh
source ${pkgs.git}/share/git/contrib/completion/git-completion.bash
source ${postgrest.hsie.bashCompletion}
''
+ builtins.concatStringsSep "\n" (
builtins.map (bashCompletion: "source ${bashCompletion}") (
builtins.concatLists (builtins.map (toolbox: toolbox.bashCompletion) toolboxes)
)
);
}
)
-346
View File
@@ -1,346 +0,0 @@
{-|
Module : PostgREST.ApiRequest
Description : PostgREST functions to translate HTTP request to a domain type called ApiRequest.
-}
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE MultiWayIf #-}
module PostgREST.ApiRequest (
ApiRequest(..)
, InvokeMethod(..)
, ContentType(..)
, Action(..)
, Target(..)
, mutuallyAgreeable
, userApiRequest
) where
import qualified Data.Aeson as JSON
import qualified Data.ByteString as BS
import qualified Data.ByteString.Lazy as BL
import qualified Data.CaseInsensitive as CI
import qualified Data.Csv as CSV
import qualified Data.HashMap.Strict as M
import qualified Data.List as L
import qualified Data.Set as S
import qualified Data.Text as T
import qualified Data.Vector as V
import Control.Arrow ((***))
import Data.Aeson.Types (emptyArray, emptyObject)
import Data.List (last, lookup, partition)
import Data.List.NonEmpty (head)
import Data.Maybe (fromJust)
import Data.Ranged.Ranges (Range (..), emptyRange,
rangeIntersection)
import Network.HTTP.Base (urlEncodeVars)
import Network.HTTP.Types.Header (hAuthorization, hCookie)
import Network.HTTP.Types.URI (parseQueryReplacePlus,
parseSimpleQuery)
import Network.Wai (Request (..))
import Network.Wai.Parse (parseHttpAccept)
import Web.Cookie (parseCookiesText)
import Data.Ranged.Boundaries
import PostgREST.Error (ApiRequestError (..))
import PostgREST.RangeQuery (NonnegRange, allRange, rangeGeq,
rangeLimit, rangeOffset, rangeRequested,
restrictRange)
import PostgREST.Types
import Protolude hiding (head, toS)
import Protolude.Conv (toS)
type RequestBody = BL.ByteString
data InvokeMethod = InvHead | InvGet | InvPost deriving Eq
-- | Types of things a user wants to do to tables/views/procs
data Action = ActionCreate | ActionRead{isHead :: Bool}
| ActionUpdate | ActionDelete
| ActionSingleUpsert | ActionInvoke InvokeMethod
| ActionInfo | ActionInspect{isHead :: Bool}
deriving Eq
-- | The target db object of a user action
data Target = TargetIdent QualifiedIdentifier
| TargetProc{tpQi :: QualifiedIdentifier, tpIsRootSpec :: Bool}
| TargetDefaultSpec{tdsSchema :: Schema} -- The default spec offered at root "/"
| TargetUnknown [Text]
deriving Eq
{-|
Describes what the user wants to do. This data type is a
translation of the raw elements of an HTTP request into domain
specific language. There is no guarantee that the intent is
sensible, it is up to a later stage of processing to determine
if it is an action we are able to perform.
-}
data ApiRequest = ApiRequest {
iAction :: Action -- ^ Similar but not identical to HTTP verb, e.g. Create/Invoke both POST
, iRange :: M.HashMap ByteString NonnegRange -- ^ Requested range of rows within response
, iTopLevelRange :: NonnegRange -- ^ Requested range of rows from the top level
, iTarget :: Target -- ^ The target, be it calling a proc or accessing a table
, iAccepts :: [ContentType] -- ^ Content types the client will accept, [CTAny] if no Accept header
, iPayload :: Maybe PayloadJSON -- ^ Data sent by client and used for mutation actions
, iPreferRepresentation :: PreferRepresentation -- ^ If client wants created items echoed back
, iPreferParameters :: Maybe PreferParameters -- ^ How to pass parameters to a stored procedure
, iPreferCount :: Maybe PreferCount -- ^ Whether the client wants a result count
, iPreferResolution :: Maybe PreferResolution -- ^ Whether the client wants to UPSERT or ignore records on PK conflict
, iFilters :: [(Text, Text)] -- ^ Filters on the result ("id", "eq.10")
, iLogic :: [(Text, Text)] -- ^ &and and &or parameters used for complex boolean logic
, iSelect :: Maybe Text -- ^ &select parameter used to shape the response
, iOnConflict :: Maybe Text -- ^ &on_conflict parameter used to upsert on specific unique keys
, iColumns :: Maybe Text -- ^ &columns parameter used to shape the payload
, iOrder :: [(Text, Text)] -- ^ &order parameters for each level
, iCanonicalQS :: ByteString -- ^ Alphabetized (canonical) request query string for response URLs
, iJWT :: Text -- ^ JSON Web Token
, iHeaders :: [(Text, Text)] -- ^ HTTP request headers
, iCookies :: [(Text, Text)] -- ^ Request Cookies
, iPath :: ByteString -- ^ Raw request path
, iMethod :: ByteString -- ^ Raw request method
, iProfile :: Maybe Schema -- ^ The request profile for enabling use of multiple schemas. Follows the spec in hhttps://www.w3.org/TR/dx-prof-conneg/ttps://www.w3.org/TR/dx-prof-conneg/.
, iSchema :: Schema -- ^ The request schema. Can vary depending on iProfile.
}
-- | Examines HTTP request and translates it into user intent.
userApiRequest :: NonEmpty Schema -> Maybe Text -> Request -> RequestBody -> Either ApiRequestError ApiRequest
userApiRequest confSchemas rootSpec req reqBody
| isJust profile && fromJust profile `notElem` confSchemas = Left $ UnacceptableSchema $ toList confSchemas
| isTargetingProc && method `notElem` ["HEAD", "GET", "POST"] = Left ActionInappropriate
| topLevelRange == emptyRange = Left InvalidRange
| shouldParsePayload && isLeft payload = either (Left . InvalidBody . toS) witness payload
| otherwise = Right ApiRequest {
iAction = action
, iTarget = target
, iRange = ranges
, iTopLevelRange = topLevelRange
, iAccepts = maybe [CTAny] (map decodeContentType . parseHttpAccept) $ lookupHeader "accept"
, iPayload = relevantPayload
, iPreferRepresentation = representation
, iPreferParameters = if | hasPrefer (show SingleObject) -> Just SingleObject
| hasPrefer (show MultipleObjects) -> Just MultipleObjects
| otherwise -> Nothing
, iPreferCount = if | hasPrefer (show ExactCount) -> Just ExactCount
| hasPrefer (show PlannedCount) -> Just PlannedCount
| hasPrefer (show EstimatedCount) -> Just EstimatedCount
| otherwise -> Nothing
, iPreferResolution = if | hasPrefer (show MergeDuplicates) -> Just MergeDuplicates
| hasPrefer (show IgnoreDuplicates) -> Just IgnoreDuplicates
| otherwise -> Nothing
, iFilters = filters
, iLogic = [(toS k, toS $ fromJust v) | (k,v) <- qParams, isJust v, endingIn ["and", "or"] k ]
, iSelect = toS <$> join (lookup "select" qParams)
, iOnConflict = toS <$> join (lookup "on_conflict" qParams)
, iColumns = columns
, iOrder = [(toS k, toS $ fromJust v) | (k,v) <- qParams, isJust v, endingIn ["order"] k ]
, iCanonicalQS = toS $ urlEncodeVars
. L.sortOn fst
. map (join (***) toS . second (fromMaybe BS.empty))
$ qString
, iJWT = tokenStr
, iHeaders = [ (toS $ CI.foldedCase k, toS v) | (k,v) <- hdrs, k /= hCookie]
, iCookies = maybe [] parseCookiesText $ lookupHeader "Cookie"
, iPath = rawPathInfo req
, iMethod = method
, iProfile = profile
, iSchema = schema
}
where
-- queryString with '+' converted to ' '(space)
qString = parseQueryReplacePlus True $ rawQueryString req
-- rpcQParams = Rpc query params e.g. /rpc/name?param1=val1, similar to filter but with no operator(eq, lt..)
(filters, rpcQParams) =
case action of
ActionInvoke InvGet -> partitionFlts
ActionInvoke InvHead -> partitionFlts
_ -> (flts, [])
partitionFlts = partition (liftM2 (||) (isEmbedPath . fst) (hasOperator . snd)) flts
flts =
[ (toS k, toS $ fromJust v) |
(k,v) <- qParams, isJust v,
k `notElem` ["select", "columns"],
not (endingIn ["order", "limit", "offset", "and", "or"] k) ]
hasOperator val = any (`T.isPrefixOf` val) $
((<> ".") <$> "not":M.keys operators) ++
((<> "(") <$> M.keys ftsOperators)
isEmbedPath = T.isInfixOf "."
isTargetingProc = case target of
TargetProc _ _ -> True
_ -> False
isTargetingDefaultSpec = case target of
TargetDefaultSpec _ -> True
_ -> False
contentType = decodeContentType . fromMaybe "application/json" $ lookupHeader "content-type"
columns
| action `elem` [ActionCreate, ActionUpdate, ActionInvoke InvPost] = toS <$> join (lookup "columns" qParams)
| otherwise = Nothing
payload =
case (contentType, action) of
(_, ActionInvoke InvGet) -> Right rpcPrmsToJson
(_, ActionInvoke InvHead) -> Right rpcPrmsToJson
(CTApplicationJSON, _) ->
if isJust columns
then Right $ RawJSON reqBody
else note "All object keys must match" . payloadAttributes reqBody
=<< if BL.null reqBody && isTargetingProc
then Right emptyObject
else JSON.eitherDecode reqBody
(CTTextCSV, _) -> do
json <- csvToJson <$> CSV.decodeByName reqBody
note "All lines must have same number of fields" $ payloadAttributes (JSON.encode json) json
(CTOther "application/x-www-form-urlencoded", _) ->
let json = M.fromList . map (toS *** JSON.String . toS) . parseSimpleQuery $ toS reqBody
keys = S.fromList $ M.keys json in
Right $ ProcessedJSON (JSON.encode json) keys
(ct, _) ->
Left $ toS $ "Content-Type not acceptable: " <> toMime ct
rpcPrmsToJson = ProcessedJSON (JSON.encode $ M.fromList $ second JSON.toJSON <$> rpcQParams) (S.fromList $ fst <$> rpcQParams)
topLevelRange = fromMaybe allRange $ M.lookup "limit" ranges -- if no limit is specified, get all the request rows
action =
case method of
-- The HEAD method is identical to GET except that the server MUST NOT return a message-body in the response
-- From https://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html#sec9.4
"HEAD" | isTargetingDefaultSpec -> ActionInspect{isHead=True}
| isTargetingProc -> ActionInvoke InvHead
| otherwise -> ActionRead{isHead=True}
"GET" | isTargetingDefaultSpec -> ActionInspect{isHead=False}
| isTargetingProc -> ActionInvoke InvGet
| otherwise -> ActionRead{isHead=False}
"POST" -> if isTargetingProc
then ActionInvoke InvPost
else ActionCreate
"PATCH" -> ActionUpdate
"PUT" -> ActionSingleUpsert
"DELETE" -> ActionDelete
"OPTIONS" -> ActionInfo
_ -> ActionInspect{isHead=False}
defaultSchema = head confSchemas
profile
| length confSchemas <= 1 -- only enable content negotiation by profile when there are multiple schemas specified in the config
= Nothing
| action `elem` [ActionCreate, ActionUpdate, ActionSingleUpsert, ActionDelete, ActionInvoke InvPost] -- POST/PATCH/PUT/DELETE don't use the same header as per the spec
= Just $ maybe defaultSchema toS $ lookupHeader "Content-Profile"
| action `elem` [ActionRead True, ActionRead False, ActionInvoke InvGet, ActionInvoke InvHead,
ActionInspect False, ActionInspect True, ActionInfo]
= Just $ maybe defaultSchema toS $ lookupHeader "Accept-Profile"
| otherwise = Nothing
schema = fromMaybe defaultSchema profile
target = case path of
[] -> case rootSpec of
Just pName -> TargetProc (QualifiedIdentifier schema pName) True
Nothing -> TargetDefaultSpec schema
[table] -> TargetIdent $ QualifiedIdentifier schema table
["rpc", proc] -> TargetProc (QualifiedIdentifier schema proc) False
other -> TargetUnknown other
shouldParsePayload =
action `elem`
[ActionCreate, ActionUpdate, ActionSingleUpsert,
ActionInvoke InvPost,
-- Though ActionInvoke{isGet=True}(a GET /rpc/..) doesn't really have a payload, we use the payload variable as a way
-- to store the query string arguments to the function.
ActionInvoke InvGet,
ActionInvoke InvHead]
relevantPayload | shouldParsePayload = rightToMaybe payload
| otherwise = Nothing
path = pathInfo req
method = requestMethod req
hdrs = requestHeaders req
qParams = [(toS k, v)|(k,v) <- qString]
lookupHeader = flip lookup hdrs
hasPrefer :: Text -> Bool
hasPrefer val = any (\(h,v) -> h == "Prefer" && val `elem` split v) hdrs
where
split :: BS.ByteString -> [Text]
split = map T.strip . T.split (==',') . toS
representation
| hasPrefer (show Full) = Full
| hasPrefer (show None) = None
| otherwise = if action == ActionCreate
then HeadersOnly -- Assume the user wants the Location header(for POST) by default
else None
auth = fromMaybe "" $ lookupHeader hAuthorization
tokenStr = case T.split (== ' ') (toS auth) of
("Bearer" : t : _) -> t
_ -> ""
endingIn:: [Text] -> Text -> Bool
endingIn xx key = lastWord `elem` xx
where lastWord = last $ T.split (=='.') key
headerRange = rangeRequested hdrs
replaceLast x s = T.intercalate "." $ L.init (T.split (=='.') s) ++ [x]
limitParams :: M.HashMap ByteString NonnegRange
limitParams = M.fromList [(toS (replaceLast "limit" k), restrictRange (readMaybe =<< (toS <$> v)) allRange) | (k,v) <- qParams, isJust v, endingIn ["limit"] k]
offsetParams :: M.HashMap ByteString NonnegRange
offsetParams = M.fromList [(toS (replaceLast "limit" k), maybe allRange rangeGeq (readMaybe =<< (toS <$> v))) | (k,v) <- qParams, isJust v, endingIn ["offset"] k]
urlRange = M.unionWith f limitParams offsetParams
where
f rl ro = Range (BoundaryBelow o) (BoundaryAbove $ o + l - 1)
where
l = fromMaybe 0 $ rangeLimit rl
o = rangeOffset ro
ranges = M.insert "limit" (rangeIntersection headerRange (fromMaybe allRange (M.lookup "limit" urlRange))) urlRange
{-|
Find the best match from a list of content types accepted by the
client in order of decreasing preference and a list of types
producible by the server. If there is no match but the client
accepts */* then return the top server pick.
-}
mutuallyAgreeable :: [ContentType] -> [ContentType] -> Maybe ContentType
mutuallyAgreeable sProduces cAccepts =
let exact = listToMaybe $ L.intersect cAccepts sProduces in
if isNothing exact && CTAny `elem` cAccepts
then listToMaybe sProduces
else exact
type CsvData = V.Vector (M.HashMap Text BL.ByteString)
{-|
Converts CSV like
a,b
1,hi
2,bye
into a JSON array like
[ {"a": "1", "b": "hi"}, {"a": 2, "b": "bye"} ]
The reason for its odd signature is so that it can compose
directly with CSV.decodeByName
-}
csvToJson :: (CSV.Header, CsvData) -> JSON.Value
csvToJson (_, vals) =
JSON.Array $ V.map rowToJsonObj vals
where
rowToJsonObj = JSON.Object .
M.map (\str ->
if str == "NULL"
then JSON.Null
else JSON.String $ toS str
)
payloadAttributes :: RequestBody -> JSON.Value -> Maybe PayloadJSON
payloadAttributes raw json =
-- Test that Array contains only Objects having the same keys
case json of
JSON.Array arr ->
case arr V.!? 0 of
Just (JSON.Object o) ->
let canonicalKeys = S.fromList $ M.keys o
areKeysUniform = all (\case
JSON.Object x -> S.fromList (M.keys x) == canonicalKeys
_ -> False) arr in
if areKeysUniform
then Just $ ProcessedJSON raw canonicalKeys
else Nothing
Just _ -> Nothing
Nothing -> Just emptyPJArray
JSON.Object o -> Just $ ProcessedJSON raw (S.fromList $ M.keys o)
-- truncate everything else to an empty array.
_ -> Just emptyPJArray
where
emptyPJArray = ProcessedJSON (JSON.encode emptyArray) S.empty
+589 -394
View File
File diff suppressed because it is too large Load Diff
+155
View File
@@ -0,0 +1,155 @@
{-# LANGUAGE RecordWildCards #-}
module PostgREST.AppState
( AppState
, getConfig
, getDbStructure
, getIsWorkerOn
, getJsonDbS
, getMainThreadId
, getPgVersion
, getPool
, getTime
, getRetryNextIn
, init
, initWithPool
, logWithZTime
, putConfig
, putDbStructure
, putIsWorkerOn
, putJsonDbS
, putPgVersion
, putRetryNextIn
, releasePool
, signalListener
, waitListener
) where
import qualified Hasql.Pool as SQL
import Control.AutoUpdate (defaultUpdateSettings, mkAutoUpdate,
updateAction)
import Data.IORef (IORef, atomicWriteIORef, newIORef,
readIORef)
import Data.Time (ZonedTime, defaultTimeLocale, formatTime,
getZonedTime)
import Data.Time.Clock (UTCTime, getCurrentTime)
import PostgREST.Config (AppConfig (..))
import PostgREST.Config.PgVersion (PgVersion (..), minimumPgVersion)
import PostgREST.DbStructure (DbStructure)
import Protolude
data AppState = AppState
{ statePool :: SQL.Pool -- | Connection pool, either a 'Connection' or a 'ConnectionError'
, statePgVersion :: IORef PgVersion
-- | No schema cache at the start. Will be filled in by the connectionWorker
, stateDbStructure :: IORef (Maybe DbStructure)
-- | Cached DbStructure in json
, stateJsonDbS :: IORef ByteString
-- | Helper ref to make sure just one connectionWorker can run at a time
, stateIsWorkerOn :: IORef Bool
-- | Binary semaphore used to sync the listener(NOTIFY reload) with the connectionWorker.
, stateListener :: MVar ()
-- | Config that can change at runtime
, stateConf :: IORef AppConfig
-- | Time used for verifying JWT expiration
, stateGetTime :: IO UTCTime
-- | Time with time zone used for worker logs
, stateGetZTime :: IO ZonedTime
-- | Used for killing the main thread in case a subthread fails
, stateMainThreadId :: ThreadId
-- | Keeps track of when the next retry for connecting to database is scheduled
, stateRetryNextIn :: IORef Int
}
init :: AppConfig -> IO AppState
init conf = do
newPool <- initPool conf
initWithPool newPool conf
initWithPool :: SQL.Pool -> AppConfig -> IO AppState
initWithPool newPool conf =
AppState newPool
<$> newIORef minimumPgVersion -- assume we're in a supported version when starting, this will be corrected on a later step
<*> newIORef Nothing
<*> newIORef mempty
<*> newIORef False
<*> newEmptyMVar
<*> newIORef conf
<*> mkAutoUpdate defaultUpdateSettings { updateAction = getCurrentTime }
<*> mkAutoUpdate defaultUpdateSettings { updateAction = getZonedTime }
<*> myThreadId
<*> newIORef 0
initPool :: AppConfig -> IO SQL.Pool
initPool AppConfig{..} =
SQL.acquire (configDbPoolSize, configDbPoolTimeout, toUtf8 configDbUri)
getPool :: AppState -> SQL.Pool
getPool = statePool
releasePool :: AppState -> IO ()
releasePool AppState{..} = SQL.release statePool >> throwTo stateMainThreadId UserInterrupt
getPgVersion :: AppState -> IO PgVersion
getPgVersion = readIORef . statePgVersion
putPgVersion :: AppState -> PgVersion -> IO ()
putPgVersion = atomicWriteIORef . statePgVersion
getDbStructure :: AppState -> IO (Maybe DbStructure)
getDbStructure = readIORef . stateDbStructure
putDbStructure :: AppState -> DbStructure -> IO ()
putDbStructure appState structure =
atomicWriteIORef (stateDbStructure appState) $ Just structure
getJsonDbS :: AppState -> IO ByteString
getJsonDbS = readIORef . stateJsonDbS
putJsonDbS :: AppState -> ByteString -> IO ()
putJsonDbS appState = atomicWriteIORef (stateJsonDbS appState)
getIsWorkerOn :: AppState -> IO Bool
getIsWorkerOn = readIORef . stateIsWorkerOn
putIsWorkerOn :: AppState -> Bool -> IO ()
putIsWorkerOn = atomicWriteIORef . stateIsWorkerOn
getRetryNextIn :: AppState -> IO Int
getRetryNextIn = readIORef . stateRetryNextIn
putRetryNextIn :: AppState -> Int -> IO ()
putRetryNextIn = atomicWriteIORef . stateRetryNextIn
getConfig :: AppState -> IO AppConfig
getConfig = readIORef . stateConf
putConfig :: AppState -> AppConfig -> IO ()
putConfig = atomicWriteIORef . stateConf
getTime :: AppState -> IO UTCTime
getTime = stateGetTime
-- | Log to stderr with local time
logWithZTime :: AppState -> Text -> IO ()
logWithZTime appState txt = do
zTime <- stateGetZTime appState
hPutStrLn stderr $ toS (formatTime defaultTimeLocale "%d/%b/%Y:%T %z: " zTime) <> txt
getMainThreadId :: AppState -> ThreadId
getMainThreadId = stateMainThreadId
-- | As this IO action uses `takeMVar` internally, it will only return once
-- `stateListener` has been set using `signalListener`. This is currently used
-- to syncronize workers.
waitListener :: AppState -> IO ()
waitListener = takeMVar . stateListener
-- tryPutMVar doesn't lock the thread. It should always succeed since
-- the connectionWorker is the only mvar producer.
signalListener :: AppState -> IO ()
signalListener appState = void $ tryPutMVar (stateListener appState) ()
+56 -88
View File
@@ -1,5 +1,3 @@
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE LambdaCase #-}
{-|
Module : PostgREST.Auth
Description : PostgREST authorization functions.
@@ -12,103 +10,73 @@ Authentication should always be implemented in an external service.
In the test suite there is an example of simple login function that can be used for a
very simple authentication system inside the PostgreSQL database.
-}
module PostgREST.Auth (
containsRole
{-# LANGUAGE RecordWildCards #-}
module PostgREST.Auth
( containsRole
, jwtClaims
, JWTAttempt(..)
, parseSecret
, JWTClaims
) where
import qualified Crypto.JOSE.Types as JOSE.Types
import qualified Crypto.JWT as JWT
import qualified Data.Aeson as JSON
import qualified Data.HashMap.Strict as M
import Data.Vector as V
import qualified Data.Vector as V
import Control.Lens (set)
import Data.Time.Clock (UTCTime)
import Control.Lens (set)
import Control.Monad.Except (liftEither)
import Data.Either.Combinators (mapLeft)
import Data.Time.Clock (UTCTime)
import Control.Lens.Operators
import Crypto.JWT
import PostgREST.Config (AppConfig (..), JSPath, JSPathExp (..))
import PostgREST.Error (Error (..))
import PostgREST.Types
import Protolude hiding (toS)
import Protolude.Conv (toS)
import Protolude
{-|
Possible situations encountered with client JWTs
-}
data JWTAttempt = JWTInvalid JWTError
| JWTMissingSecret
| JWTClaims (M.HashMap Text JSON.Value)
deriving (Eq, Show)
{-|
Receives the JWT secret and audience (from config) and a JWT and returns a map
of JWT claims.
-}
jwtClaims :: Maybe JWKSet -> Maybe StringOrURI -> LByteString -> UTCTime -> Maybe JSPath -> IO JWTAttempt
jwtClaims _ _ "" _ _ = return $ JWTClaims M.empty
jwtClaims secret audience payload time jspath =
case secret of
Nothing -> return JWTMissingSecret
Just s -> do
let validation = set allowedSkew 1 $ defaultJWTValidationSettings (maybe (const True) (==) audience)
eJwt <- runExceptT $ do
jwt <- decodeCompact payload
verifyClaimsAt validation s time jwt
return $ case eJwt of
Left e -> JWTInvalid e
Right jwt -> JWTClaims $ claims2map jwt jspath
type JWTClaims = M.HashMap Text JSON.Value
{-|
Turn JWT ClaimSet into something easier to work with,
also here the jspath is applied to put the "role" in the map
-}
claims2map :: ClaimsSet -> Maybe JSPath -> M.HashMap Text JSON.Value
claims2map claims jspath = (\case
-- | Receives the JWT secret and audience (from config) and a JWT and returns a
-- map of JWT claims.
jwtClaims :: Monad m =>
AppConfig -> LByteString -> UTCTime -> ExceptT Error m JWTClaims
jwtClaims _ "" _ = return M.empty
jwtClaims AppConfig{..} payload time = do
secret <- liftEither . maybeToRight JwtTokenMissing $ configJWKS
eitherClaims <-
lift . runExceptT $
JWT.verifyClaimsAt validation secret time =<< JWT.decodeCompact payload
liftEither . mapLeft jwtClaimsError $ claimsMap configJwtRoleClaimKey <$> eitherClaims
where
validation =
JWT.defaultJWTValidationSettings audienceCheck & set JWT.allowedSkew 1
audienceCheck :: JWT.StringOrURI -> Bool
audienceCheck = maybe (const True) (==) configJwtAudience
jwtClaimsError :: JWT.JWTError -> Error
jwtClaimsError JWT.JWTExpired = JwtTokenInvalid "JWT expired"
jwtClaimsError e = JwtTokenInvalid $ show e
-- | Turn JWT ClaimSet into something easier to work with.
--
-- Also, here the jspath is applied to put the "role" in the map.
claimsMap :: JSPath -> JWT.ClaimsSet -> JWTClaims
claimsMap jspath claims =
case JSON.toJSON claims of
val@(JSON.Object o) ->
let role = maybe M.empty (M.singleton "role") $
walkJSPath (Just val) =<< jspath in
M.delete "role" o `M.union` role -- mutating the map
_ -> M.empty
) $ JSON.toJSON claims
M.delete "role" o `M.union` role val
_ ->
M.empty
where
role value =
maybe M.empty (M.singleton "role") $ walkJSPath (Just value) jspath
walkJSPath :: Maybe JSON.Value -> JSPath -> Maybe JSON.Value
walkJSPath x [] = x
walkJSPath (Just (JSON.Object o)) (JSPKey key:rest) = walkJSPath (M.lookup key o) rest
walkJSPath (Just (JSON.Array ar)) (JSPIdx idx:rest) = walkJSPath (ar V.!? idx) rest
walkJSPath _ _ = Nothing
walkJSPath :: Maybe JSON.Value -> JSPath -> Maybe JSON.Value
walkJSPath x [] = x
walkJSPath (Just (JSON.Object o)) (JSPKey key:rest) = walkJSPath (M.lookup key o) rest
walkJSPath (Just (JSON.Array ar)) (JSPIdx idx:rest) = walkJSPath (ar V.!? idx) rest
walkJSPath _ _ = Nothing
{-|
Whether a response from jwtClaims contains a role claim
-}
containsRole :: JWTAttempt -> Bool
containsRole (JWTClaims claims) = M.member "role" claims
containsRole _ = False
{-|
Parse `jwt-secret` configuration option and turn into a JWKSet.
There are three ways to specify `jwt-secret`: text secret, JSON Web Key
(JWK), or JSON Web Key Set (JWKS). The first two are converted into a JWKSet
with one key and the last is converted as is.
-}
parseSecret :: ByteString -> JWKSet
parseSecret str =
fromMaybe (maybe secret (\jwk' -> JWKSet [jwk']) maybeJWK)
maybeJWKSet
where
maybeJWKSet = JSON.decode (toS str) :: Maybe JWKSet
maybeJWK = JSON.decode (toS str) :: Maybe JWK
secret = JWKSet [jwkFromSecret str]
{-|
Internal helper to generate a symmetric HMAC-SHA256 JWK from a text secret.
-}
jwkFromSecret :: ByteString -> JWK
jwkFromSecret key =
fromKeyMaterial km
& jwkUse ?~ Sig
& jwkAlg ?~ JWSAlg HS256
where
km = OctKeyMaterial (OctKeyParameters (JOSE.Types.Base64Octets key))
-- | Whether a response from jwtClaims contains a role claim
containsRole :: JWTClaims -> Bool
containsRole = M.member "role"
+221
View File
@@ -0,0 +1,221 @@
{-# LANGUAGE NamedFieldPuns #-}
{-# LANGUAGE QuasiQuotes #-}
{-# LANGUAGE RecordWildCards #-}
module PostgREST.CLI
( main
, CLI (..)
, Command (..)
, readCLIShowHelp
) where
import qualified Data.Aeson as JSON
import qualified Data.ByteString.Char8 as BS
import qualified Data.ByteString.Lazy as LBS
import qualified Hasql.Pool as SQL
import qualified Hasql.Transaction.Sessions as SQL
import qualified Options.Applicative as O
import Data.Text.IO (hPutStrLn)
import Text.Heredoc (str)
import PostgREST.AppState (AppState)
import PostgREST.Config (AppConfig (..))
import PostgREST.DbStructure (queryDbStructure)
import PostgREST.Version (prettyVersion)
import PostgREST.Workers (reReadConfig)
import qualified PostgREST.App as App
import qualified PostgREST.AppState as AppState
import qualified PostgREST.Config as Config
import Protolude hiding (hPutStrLn)
main :: App.SignalHandlerInstaller -> Maybe App.SocketRunner -> CLI -> IO ()
main installSignalHandlers runAppWithSocket CLI{cliCommand, cliPath} = do
conf@AppConfig{..} <-
either panic identity <$> Config.readAppConfig mempty cliPath Nothing
appState <- AppState.init conf
-- Override the config with config options from the db
-- TODO: the same operation is repeated on connectionWorker, ideally this
-- would be done only once, but dump CmdDumpConfig needs it for tests.
when configDbConfig $ reReadConfig True appState
exec cliCommand appState
where
exec :: Command -> AppState -> IO ()
exec CmdDumpConfig appState = putStr . Config.toText =<< AppState.getConfig appState
exec CmdDumpSchema appState = putStrLn =<< dumpSchema appState
exec CmdRun appState = App.run installSignalHandlers runAppWithSocket appState
-- | Dump DbStructure schema to JSON
dumpSchema :: AppState -> IO LBS.ByteString
dumpSchema appState = do
AppConfig{..} <- AppState.getConfig appState
result <-
let transaction = if configDbPreparedStatements then SQL.transaction else SQL.unpreparedTransaction in
SQL.use (AppState.getPool appState) $
transaction SQL.ReadCommitted SQL.Read $
queryDbStructure
(toList configDbSchemas)
configDbExtraSearchPath
configDbPreparedStatements
SQL.release $ AppState.getPool appState
case result of
Left e -> do
hPutStrLn stderr $ "An error ocurred when loading the schema cache:\n" <> show e
exitFailure
Right dbStructure -> return $ JSON.encode dbStructure
-- | Command line interface options
data CLI = CLI
{ cliCommand :: Command
, cliPath :: Maybe FilePath
}
data Command
= CmdRun
| CmdDumpConfig
| CmdDumpSchema
-- | Read command line interface options. Also prints help.
readCLIShowHelp :: Bool -> IO CLI
readCLIShowHelp hasEnvironment =
O.customExecParser prefs opts
where
prefs = O.prefs $ O.showHelpOnError <> O.showHelpOnEmpty
opts = O.info parser $ O.fullDesc <> progDesc <> footer
parser = O.helper <*> exampleParser <*> cliParser
progDesc =
O.progDesc $
"PostgREST "
<> BS.unpack prettyVersion
<> " / create a REST API to an existing Postgres database"
footer =
O.footer $
"To run PostgREST, please pass the FILENAME argument"
<> " or set PGRST_ environment variables."
exampleParser =
O.infoOption exampleConfigFile $
O.long "example"
<> O.short 'e'
<> O.help "Show an example configuration file"
cliParser :: O.Parser CLI
cliParser =
CLI
<$> (dumpConfigFlag <|> dumpSchemaFlag)
<*> optionalIf hasEnvironment configFileOption
configFileOption =
O.strArgument $
O.metavar "FILENAME"
<> O.help "Path to configuration file (optional with PGRST_ environment variables)"
dumpConfigFlag =
O.flag CmdRun CmdDumpConfig $
O.long "dump-config"
<> O.help "Dump loaded configuration and exit"
dumpSchemaFlag =
O.flag CmdRun CmdDumpSchema $
O.long "dump-schema"
<> O.help "Dump loaded schema as JSON and exit (for debugging, output structure is unstable)"
optionalIf :: Alternative f => Bool -> f a -> f (Maybe a)
optionalIf True = O.optional
optionalIf False = fmap Just
exampleConfigFile :: [Char]
exampleConfigFile =
[str|### REQUIRED:
|db-uri = "postgres://user:pass@localhost:5432/dbname"
|db-schema = "public"
|db-anon-role = "postgres"
|
|### OPTIONAL:
|## number of open connections in the pool
|db-pool = 10
|
|## Time to live, in seconds, for an idle database pool connection.
|db-pool-timeout = 10
|
|## extra schemas to add to the search_path of every request
|db-extra-search-path = "public"
|
|## limit rows in response
|# db-max-rows = 1000
|
|## stored proc to exec immediately after auth
|# db-pre-request = "stored_proc_name"
|
|## stored proc that overrides the root "/" spec
|## it must be inside the db-schema
|# db-root-spec = "stored_proc_name"
|
|## Notification channel for reloading the schema cache
|db-channel = "pgrst"
|
|## Enable or disable the notification channel
|db-channel-enabled = true
|
|## Enable in-database configuration
|db-config = true
|
|## Determine if GUC request settings for headers, cookies and jwt claims use the legacy names (string with dashes, invalid starting from PostgreSQL v14) with text values instead of the new names (string without dashes, valid on all PostgreSQL versions) with json values.
|## For PostgreSQL v14 and up, this setting will be ignored.
|db-use-legacy-gucs = true
|
|## how to terminate database transactions
|## possible values are:
|## commit (default)
|## transaction is always committed, this can not be overriden
|## commit-allow-override
|## transaction is committed, but can be overriden with Prefer tx=rollback header
|## rollback
|## transaction is always rolled back, this can not be overriden
|## rollback-allow-override
|## transaction is rolled back, but can be overriden with Prefer tx=commit header
|db-tx-end = "commit"
|
|## enable or disable prepared statements. disabling is only necessary when behind a connection pooler.
|## when disabled, statements will be parametrized but won't be prepared.
|db-prepared-statements = true
|
|server-host = "!4"
|server-port = 3000
|
|## unix socket location
|## if specified it takes precedence over server-port
|# server-unix-socket = "/tmp/pgrst.sock"
|
|## unix socket file mode
|## when none is provided, 660 is applied by default
|# server-unix-socket-mode = "660"
|
|## determine if the OpenAPI output should follow or ignore role privileges or be disabled entirely
|## admitted values: follow-privileges, ignore-privileges, disabled
|openapi-mode = "follow-privileges"
|
|## base url for the OpenAPI output
|openapi-server-proxy-uri = ""
|
|## choose a secret, JSON Web Key (or set) to enable JWT auth
|## (use "@filename" to load from separate file)
|# jwt-secret = "secret_with_at_least_32_characters"
|# jwt-aud = "your_audience_claim"
|jwt-secret-is-base64 = false
|
|## jspath to the role claim key
|jwt-role-claim-key = ".role"
|
|## content types to produce raw output
|# raw-media-types="image/png, image/jpg"
|
|## logging level, the admitted values are: crit, error, warn and info.
|log-level = "error"
|]
+384 -242
View File
@@ -1,217 +1,367 @@
{-|
Module : PostgREST.Config
Description : Manages PostgREST configuration options.
Description : Manages PostgREST configuration type and parser.
This module provides a helper function to read the command line
arguments using the optparse-applicative and the AppConfig type to store
them. It also can be used to define other middleware configuration that
may be delegated to some sort of external configuration.
It currently includes a hardcoded CORS policy but this could easly be
turned in configurable behaviour if needed.
Other hardcoded options such as the minimum version number also belong here.
-}
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TemplateHaskell #-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE RecordWildCards #-}
{-# OPTIONS_GHC -fno-warn-type-defaults #-}
module PostgREST.Config ( prettyVersion
, docsVersion
, readOptions
, corsPolicy
, AppConfig (..)
, configPoolTimeout'
)
where
module PostgREST.Config
( AppConfig (..)
, Environment
, JSPath
, JSPathExp(..)
, LogLevel(..)
, OpenAPIMode(..)
, Proxy(..)
, toText
, isMalformedProxyUri
, readAppConfig
, readPGRSTEnvironment
, toURI
, parseSecret
) where
import qualified Data.ByteString as B
import qualified Data.ByteString.Char8 as BS
import qualified Data.CaseInsensitive as CI
import qualified Data.Configurator as C
import qualified Text.PrettyPrint.ANSI.Leijen as L
import qualified Crypto.JOSE.Types as JOSE
import qualified Crypto.JWT as JWT
import qualified Data.Aeson as JSON
import qualified Data.ByteString as BS
import qualified Data.ByteString.Base64 as B64
import qualified Data.ByteString.Lazy as LBS
import qualified Data.Configurator as C
import qualified Data.Map.Strict as M
import qualified Data.Text as T
import qualified Data.Text.Encoding as T
import Control.Lens (preview)
import Control.Monad (fail)
import Crypto.JWT (StringOrURI, stringOrUri)
import Data.List (lookup)
import Data.List.NonEmpty (fromList)
import Data.Scientific (floatingOrInteger)
import Data.Text (dropEnd, dropWhileEnd,
intercalate, splitOn, strip, take,
unpack)
import Data.Text.IO (hPutStrLn)
import Data.Version (versionBranch)
import Development.GitRev (gitHash)
import Network.Wai.Middleware.Cors (CorsResourcePolicy (..))
import Numeric (readOct)
import Paths_postgrest (version)
import System.IO.Error (IOError)
import System.Posix.Types (FileMode)
import Control.Lens (preview)
import Control.Monad (fail)
import Crypto.JWT (JWK, JWKSet, StringOrURI, stringOrUri)
import Data.Aeson (toJSON)
import Data.Either.Combinators (mapLeft)
import Data.List (lookup)
import Data.List.NonEmpty (fromList, toList)
import Data.Maybe (fromJust)
import Data.Scientific (floatingOrInteger)
import Data.Time.Clock (NominalDiffTime)
import Numeric (readOct, showOct)
import System.Environment (getEnvironment)
import System.Posix.Types (FileMode)
import Control.Applicative
import Data.Monoid
import Network.Wai
import Options.Applicative hiding (str)
import Text.Heredoc
import Text.PrettyPrint.ANSI.Leijen hiding ((<$>), (<>))
import PostgREST.Config.JSPath (JSPath, JSPathExp (..),
dumpJSPath, pRoleClaimKey)
import PostgREST.Config.Proxy (Proxy (..),
isMalformedProxyUri, toURI)
import PostgREST.DbStructure.Identifiers (QualifiedIdentifier, dumpQi,
toQi)
import PostgREST.Error (ApiRequestError (..))
import PostgREST.Parsers (pRoleClaimKey)
import PostgREST.Types (JSPath, JSPathExp (..))
import Protolude hiding (concat, hPutStrLn, intercalate, null,
take, toS, (<>))
import Protolude.Conv (toS)
import Protolude hiding (Proxy, toList)
-- | Config file settings for the server
data AppConfig = AppConfig {
configDatabase :: Text
, configAnonRole :: Text
, configOpenAPIProxyUri :: Maybe Text
, configSchemas :: NonEmpty Text
, configHost :: Text
, configPort :: Int
, configSocket :: Maybe FilePath
, configSocketMode :: Either Text FileMode
, configJwtSecret :: Maybe B.ByteString
, configJwtSecretIsBase64 :: Bool
, configJwtAudience :: Maybe StringOrURI
, configPool :: Int
, configPoolTimeout :: Int
, configMaxRows :: Maybe Integer
, configReqCheck :: Maybe Text
, configQuiet :: Bool
, configSettings :: [(Text, Text)]
, configRoleClaimKey :: Either ApiRequestError JSPath
, configExtraSearchPath :: [Text]
, configRootSpec :: Maybe Text
, configRawMediaTypes :: [B.ByteString]
data AppConfig = AppConfig
{ configAppSettings :: [(Text, Text)]
, configDbAnonRole :: Text
, configDbChannel :: Text
, configDbChannelEnabled :: Bool
, configDbExtraSearchPath :: [Text]
, configDbMaxRows :: Maybe Integer
, configDbPoolSize :: Int
, configDbPoolTimeout :: NominalDiffTime
, configDbPreRequest :: Maybe QualifiedIdentifier
, configDbPreparedStatements :: Bool
, configDbRootSpec :: Maybe QualifiedIdentifier
, configDbSchemas :: NonEmpty Text
, configDbConfig :: Bool
, configDbTxAllowOverride :: Bool
, configDbTxRollbackAll :: Bool
, configDbUri :: Text
, configDbUseLegacyGucs :: Bool
, configFilePath :: Maybe FilePath
, configJWKS :: Maybe JWKSet
, configJwtAudience :: Maybe StringOrURI
, configJwtRoleClaimKey :: JSPath
, configJwtSecret :: Maybe BS.ByteString
, configJwtSecretIsBase64 :: Bool
, configLogLevel :: LogLevel
, configOpenApiMode :: OpenAPIMode
, configOpenApiServerProxyUri :: Maybe Text
, configRawMediaTypes :: [BS.ByteString]
, configServerHost :: Text
, configServerPort :: Int
, configServerUnixSocket :: Maybe FilePath
, configServerUnixSocketMode :: FileMode
}
configPoolTimeout' :: (Fractional a) => AppConfig -> a
configPoolTimeout' =
fromRational . toRational . configPoolTimeout
data LogLevel = LogCrit | LogError | LogWarn | LogInfo
dumpLogLevel :: LogLevel -> Text
dumpLogLevel = \case
LogCrit -> "crit"
LogError -> "error"
LogWarn -> "warn"
LogInfo -> "info"
defaultCorsPolicy :: CorsResourcePolicy
defaultCorsPolicy = CorsResourcePolicy Nothing
["GET", "POST", "PATCH", "PUT", "DELETE", "OPTIONS"] ["Authorization"] Nothing
(Just $ 60*60*24) False False True
data OpenAPIMode = OAFollowPriv | OAIgnorePriv | OADisabled
deriving Eq
-- | CORS policy to be used in by Wai Cors middleware
corsPolicy :: Request -> Maybe CorsResourcePolicy
corsPolicy req = case lookup "origin" headers of
Just origin -> Just defaultCorsPolicy {
corsOrigins = Just ([origin], True)
, corsRequestHeaders = "Authentication":accHeaders
, corsExposedHeaders = Just [
"Content-Encoding", "Content-Location", "Content-Range", "Content-Type"
, "Date", "Location", "Server", "Transfer-Encoding", "Range-Unit"
dumpOpenApiMode :: OpenAPIMode -> Text
dumpOpenApiMode = \case
OAFollowPriv -> "follow-privileges"
OAIgnorePriv -> "ignore-privileges"
OADisabled -> "disabled"
-- | Dump the config
toText :: AppConfig -> Text
toText conf =
unlines $ (\(k, v) -> k <> " = " <> v) <$> pgrstSettings ++ appSettings
where
-- apply conf to all pgrst settings
pgrstSettings = (\(k, v) -> (k, v conf)) <$>
[("db-anon-role", q . configDbAnonRole)
,("db-channel", q . configDbChannel)
,("db-channel-enabled", T.toLower . show . configDbChannelEnabled)
,("db-extra-search-path", q . T.intercalate "," . configDbExtraSearchPath)
,("db-max-rows", maybe "\"\"" show . configDbMaxRows)
,("db-pool", show . configDbPoolSize)
,("db-pool-timeout", show . floor . configDbPoolTimeout)
,("db-pre-request", q . maybe mempty dumpQi . configDbPreRequest)
,("db-prepared-statements", T.toLower . show . configDbPreparedStatements)
,("db-root-spec", q . maybe mempty dumpQi . configDbRootSpec)
,("db-schemas", q . T.intercalate "," . toList . configDbSchemas)
,("db-config", T.toLower . show . configDbConfig)
,("db-tx-end", q . showTxEnd)
,("db-uri", q . configDbUri)
,("db-use-legacy-gucs", T.toLower . show . configDbUseLegacyGucs)
,("jwt-aud", T.decodeUtf8 . LBS.toStrict . JSON.encode . maybe "" toJSON . configJwtAudience)
,("jwt-role-claim-key", q . T.intercalate mempty . fmap dumpJSPath . configJwtRoleClaimKey)
,("jwt-secret", q . T.decodeUtf8 . showJwtSecret)
,("jwt-secret-is-base64", T.toLower . show . configJwtSecretIsBase64)
,("log-level", q . dumpLogLevel . configLogLevel)
,("openapi-mode", q . dumpOpenApiMode . configOpenApiMode)
,("openapi-server-proxy-uri", q . fromMaybe mempty . configOpenApiServerProxyUri)
,("raw-media-types", q . T.decodeUtf8 . BS.intercalate "," . configRawMediaTypes)
,("server-host", q . configServerHost)
,("server-port", show . configServerPort)
,("server-unix-socket", q . maybe mempty T.pack . configServerUnixSocket)
,("server-unix-socket-mode", q . T.pack . showSocketMode)
]
}
Nothing -> Nothing
where
headers = requestHeaders req
accHeaders = case lookup "access-control-request-headers" headers of
Just hdrs -> map (CI.mk . toS . strip . toS) $ BS.split ',' hdrs
Nothing -> []
-- | User friendly version number
prettyVersion :: Text
prettyVersion =
intercalate "." (map show $ versionBranch version)
<> " (" <> take 7 $(gitHash) <> ")"
-- quote all app.settings
appSettings = second q <$> configAppSettings conf
-- | Version number used in docs
docsVersion :: Text
docsVersion = "v" <> dropEnd 1 (dropWhileEnd (/= '.') prettyVersion)
-- quote strings and replace " with \"
q s = "\"" <> T.replace "\"" "\\\"" s <> "\""
-- | Function to read and parse options from the command line
readOptions :: IO AppConfig
readOptions = do
-- First read the config file path from command line
cfgPath <- customExecParser parserPrefs opts
-- Now read the actual config file
conf <- catches (C.load cfgPath)
[ Handler (\(ex :: IOError) -> exitErr $ "Cannot open config file:\n\t" <> show ex)
, Handler (\(C.ParseError err) -> exitErr $ "Error parsing config file:\n" <> err)
]
showTxEnd c = case (configDbTxRollbackAll c, configDbTxAllowOverride c) of
( False, False ) -> "commit"
( False, True ) -> "commit-allow-override"
( True , False ) -> "rollback"
( True , True ) -> "rollback-allow-override"
showJwtSecret c
| configJwtSecretIsBase64 c = B64.encode secret
| otherwise = secret
where
secret = fromMaybe mempty $ configJwtSecret c
showSocketMode c = showOct (configServerUnixSocketMode c) mempty
case C.runParser parseConfig conf of
-- This class is needed for the polymorphism of overrideFromDbOrEnvironment
-- because C.required and C.optional have different signatures
class JustIfMaybe a b where
justIfMaybe :: a -> b
instance JustIfMaybe a a where
justIfMaybe a = a
instance JustIfMaybe a (Maybe a) where
justIfMaybe a = Just a
-- | Reads and parses the config and overrides its parameters from env vars,
-- files or db settings.
readAppConfig :: [(Text, Text)] -> Maybe FilePath -> Maybe Text -> IO (Either Text AppConfig)
readAppConfig dbSettings optPath prevDbUri = do
env <- readPGRSTEnvironment
-- if no filename provided, start with an empty map to read config from environment
conf <- maybe (return $ Right M.empty) loadConfig optPath
case C.runParser (parser optPath env dbSettings) =<< mapLeft show conf of
Left err ->
exitErr $ "Error parsing config file:\n\t" <> err
Right appConf ->
return appConf
return . Left $ "Error in config " <> err
Right parsedConfig ->
Right <$> decodeLoadFiles parsedConfig
where
parseConfig =
AppConfig
<$> reqString "db-uri"
<*> reqString "db-anon-role"
<*> optString "openapi-server-proxy-uri"
<*> (fromList . splitOnCommas <$> reqValue "db-schema")
<*> (fromMaybe "!4" <$> optString "server-host")
<*> (fromMaybe 3000 <$> optInt "server-port")
<*> (fmap unpack <$> optString "server-unix-socket")
<*> parseSocketFileMode "server-unix-socket-mode"
<*> (fmap encodeUtf8 <$> optString "jwt-secret")
<*> (fromMaybe False <$> optBool "secret-is-base64")
<*> parseJwtAudience "jwt-aud"
<*> (fromMaybe 10 <$> optInt "db-pool")
<*> (fromMaybe 10 <$> optInt "db-pool-timeout")
<*> optInt "max-rows"
<*> optString "pre-request"
<*> pure False
<*> (fmap (fmap coerceText) <$> C.subassocs "app.settings" C.value)
<*> (maybe (Right [JSPKey "role"]) parseRoleClaimKey <$> optValue "role-claim-key")
<*> (maybe ["public"] splitOnCommas <$> optValue "db-extra-search-path")
<*> optString "root-spec"
<*> (maybe [] (fmap encodeUtf8 . splitOnCommas) <$> optValue "raw-media-types")
-- Both C.ParseError and IOError are shown here
loadConfig :: FilePath -> IO (Either SomeException C.Config)
loadConfig = try . C.load
parseSocketFileMode :: C.Key -> C.Parser C.Config (Either Text FileMode)
decodeLoadFiles :: AppConfig -> IO AppConfig
decodeLoadFiles parsedConfig =
decodeJWKS <$>
(decodeSecret =<< readSecretFile =<< readDbUriFile prevDbUri parsedConfig)
parser :: Maybe FilePath -> Environment -> [(Text, Text)] -> C.Parser C.Config AppConfig
parser optPath env dbSettings =
AppConfig
<$> parseAppSettings "app.settings"
<*> reqString "db-anon-role"
<*> (fromMaybe "pgrst" <$> optString "db-channel")
<*> (fromMaybe True <$> optBool "db-channel-enabled")
<*> (maybe ["public"] splitOnCommas <$> optValue "db-extra-search-path")
<*> optWithAlias (optInt "db-max-rows")
(optInt "max-rows")
<*> (fromMaybe 10 <$> optInt "db-pool")
<*> (fromIntegral . fromMaybe 10 <$> optInt "db-pool-timeout")
<*> (fmap toQi <$> optWithAlias (optString "db-pre-request")
(optString "pre-request"))
<*> (fromMaybe True <$> optBool "db-prepared-statements")
<*> (fmap toQi <$> optWithAlias (optString "db-root-spec")
(optString "root-spec"))
<*> (fromList . splitOnCommas <$> reqWithAlias (optValue "db-schemas")
(optValue "db-schema")
"missing key: either db-schemas or db-schema must be set")
<*> (fromMaybe True <$> optBool "db-config")
<*> parseTxEnd "db-tx-end" snd
<*> parseTxEnd "db-tx-end" fst
<*> reqString "db-uri"
<*> (fromMaybe True <$> optBool "db-use-legacy-gucs")
<*> pure optPath
<*> pure Nothing
<*> parseJwtAudience "jwt-aud"
<*> parseRoleClaimKey "jwt-role-claim-key" "role-claim-key"
<*> (fmap encodeUtf8 <$> optString "jwt-secret")
<*> (fromMaybe False <$> optWithAlias
(optBool "jwt-secret-is-base64")
(optBool "secret-is-base64"))
<*> parseLogLevel "log-level"
<*> parseOpenAPIMode "openapi-mode"
<*> parseOpenAPIServerProxyURI "openapi-server-proxy-uri"
<*> (maybe [] (fmap encodeUtf8 . splitOnCommas) <$> optValue "raw-media-types")
<*> (fromMaybe "!4" <$> optString "server-host")
<*> (fromMaybe 3000 <$> optInt "server-port")
<*> (fmap T.unpack <$> optString "server-unix-socket")
<*> parseSocketFileMode "server-unix-socket-mode"
where
parseAppSettings :: C.Key -> C.Parser C.Config [(Text, Text)]
parseAppSettings key = addFromEnv . fmap (fmap coerceText) <$> C.subassocs key C.value
where
addFromEnv f = M.toList $ M.union fromEnv $ M.fromList f
fromEnv = M.mapKeys fromJust $ M.filterWithKey (\k _ -> isJust k) $ M.mapKeys normalize env
normalize k = ("app.settings." <>) <$> T.stripPrefix "PGRST_APP_SETTINGS_" (toS k)
parseSocketFileMode :: C.Key -> C.Parser C.Config FileMode
parseSocketFileMode k =
C.optional k C.string >>= \case
Nothing -> pure $ Right 432 -- return default 660 mode if no value was provided
optString k >>= \case
Nothing -> pure 432 -- return default 660 mode if no value was provided
Just fileModeText ->
case (readOct . unpack) fileModeText of
case readOct $ T.unpack fileModeText of
[] ->
pure $ Left "Invalid server-unix-socket-mode: not an octal"
fail "Invalid server-unix-socket-mode: not an octal"
(fileMode, _):_ ->
if fileMode < 384 || fileMode > 511
then pure $ Left "Invalid server-unix-socket-mode: needs to be between 600 and 777"
else pure $ Right fileMode
then fail "Invalid server-unix-socket-mode: needs to be between 600 and 777"
else pure fileMode
parseOpenAPIMode :: C.Key -> C.Parser C.Config OpenAPIMode
parseOpenAPIMode k =
optString k >>= \case
Nothing -> pure OAFollowPriv
Just "follow-privileges" -> pure OAFollowPriv
Just "ignore-privileges" -> pure OAIgnorePriv
Just "disabled" -> pure OADisabled
Just _ -> fail "Invalid openapi-mode. Check your configuration."
parseOpenAPIServerProxyURI :: C.Key -> C.Parser C.Config (Maybe Text)
parseOpenAPIServerProxyURI k =
optString k >>= \case
Nothing -> pure Nothing
Just val | isMalformedProxyUri val -> fail "Malformed proxy uri, a correct example: https://example.com:8443/basePath"
| otherwise -> pure $ Just val
parseJwtAudience :: C.Key -> C.Parser C.Config (Maybe StringOrURI)
parseJwtAudience k =
C.optional k C.string >>= \case
optString k >>= \case
Nothing -> pure Nothing -- no audience in config file
Just aud -> case preview stringOrUri (unpack aud) of
Just aud -> case preview stringOrUri (T.unpack aud) of
Nothing -> fail "Invalid Jwt audience. Check your configuration."
(Just "") -> pure Nothing
aud' -> pure aud'
reqString :: C.Key -> C.Parser C.Config Text
reqString k = C.required k C.string
parseLogLevel :: C.Key -> C.Parser C.Config LogLevel
parseLogLevel k =
optString k >>= \case
Nothing -> pure LogError
Just "crit" -> pure LogCrit
Just "error" -> pure LogError
Just "warn" -> pure LogWarn
Just "info" -> pure LogInfo
Just _ -> fail "Invalid logging level. Check your configuration."
reqValue :: C.Key -> C.Parser C.Config C.Value
reqValue k = C.required k C.value
parseTxEnd :: C.Key -> ((Bool, Bool) -> Bool) -> C.Parser C.Config Bool
parseTxEnd k f =
optString k >>= \case
-- RollbackAll AllowOverride
Nothing -> pure $ f (False, False)
Just "commit" -> pure $ f (False, False)
Just "commit-allow-override" -> pure $ f (False, True)
Just "rollback" -> pure $ f (True, False)
Just "rollback-allow-override" -> pure $ f (True, True)
Just _ -> fail "Invalid transaction termination. Check your configuration."
parseRoleClaimKey :: C.Key -> C.Key -> C.Parser C.Config JSPath
parseRoleClaimKey k al =
optWithAlias (optString k) (optString al) >>= \case
Nothing -> pure [JSPKey "role"]
Just rck -> either (fail . show) pure $ pRoleClaimKey rck
reqWithAlias :: C.Parser C.Config (Maybe a) -> C.Parser C.Config (Maybe a) -> [Char] -> C.Parser C.Config a
reqWithAlias orig alias err =
orig >>= \case
Just v -> pure v
Nothing ->
alias >>= \case
Just v -> pure v
Nothing -> fail err
optWithAlias :: C.Parser C.Config (Maybe a) -> C.Parser C.Config (Maybe a) -> C.Parser C.Config (Maybe a)
optWithAlias orig alias =
orig >>= \case
Just v -> pure $ Just v
Nothing -> alias
reqString :: C.Key -> C.Parser C.Config Text
reqString k = overrideFromDbOrEnvironment C.required k coerceText
optString :: C.Key -> C.Parser C.Config (Maybe Text)
optString k = mfilter (/= "") <$> C.optional k C.string
optString k = mfilter (/= "") <$> overrideFromDbOrEnvironment C.optional k coerceText
optValue :: C.Key -> C.Parser C.Config (Maybe C.Value)
optValue k = C.optional k C.value
optValue k = overrideFromDbOrEnvironment C.optional k identity
optInt :: (Read i, Integral i) => C.Key -> C.Parser C.Config (Maybe i)
optInt k = join <$> C.optional k (coerceInt <$> C.value)
optInt k = join <$> overrideFromDbOrEnvironment C.optional k coerceInt
optBool :: C.Key -> C.Parser C.Config (Maybe Bool)
optBool k = join <$> C.optional k (coerceBool <$> C.value)
optBool k = join <$> overrideFromDbOrEnvironment C.optional k coerceBool
overrideFromDbOrEnvironment :: JustIfMaybe a b =>
(C.Key -> C.Parser C.Value a -> C.Parser C.Config b) ->
C.Key -> (C.Value -> a) -> C.Parser C.Config b
overrideFromDbOrEnvironment necessity key coercion =
case reloadableDbSetting <|> M.lookup envVarName env of
Just dbOrEnvVal -> pure $ justIfMaybe $ coercion $ C.String dbOrEnvVal
Nothing -> necessity key (coercion <$> C.value)
where
dashToUnderscore '-' = '_'
dashToUnderscore c = c
envVarName = "PGRST_" <> (toUpper . dashToUnderscore <$> toS key)
reloadableDbSetting =
let dbSettingName = T.pack $ dashToUnderscore <$> toS key in
if dbSettingName `notElem` [
"server_host", "server_port", "server_unix_socket", "server_unix_socket_mode", "log_level",
"db_anon_role", "db_uri", "db_channel_enabled", "db_channel", "db_pool", "db_pool_timeout", "db_config"]
then lookup dbSettingName dbSettings
else Nothing
coerceText :: C.Value -> Text
coerceText (C.String s) = s
@@ -219,90 +369,82 @@ readOptions = do
coerceInt :: (Read i, Integral i) => C.Value -> Maybe i
coerceInt (C.Number x) = rightToMaybe $ floatingOrInteger x
coerceInt (C.String x) = readMaybe $ toS x
coerceInt (C.String x) = readMaybe x
coerceInt _ = Nothing
coerceBool :: C.Value -> Maybe Bool
coerceBool (C.Bool b) = Just b
coerceBool (C.String b) = readMaybe $ toS b
coerceBool (C.String s) =
-- parse all kinds of text: True, true, TRUE, "true", ...
case readMaybe $ T.toTitle $ T.filter isAlpha $ toS s of
Just b -> Just b
-- numeric instead?
Nothing -> (> 0) <$> (readMaybe s :: Maybe Integer)
coerceBool _ = Nothing
parseRoleClaimKey :: C.Value -> Either ApiRequestError JSPath
parseRoleClaimKey (C.String s) = pRoleClaimKey s
parseRoleClaimKey v = pRoleClaimKey $ show v
splitOnCommas :: C.Value -> [Text]
splitOnCommas (C.String s) = strip <$> splitOn "," s
splitOnCommas (C.String s) = T.strip <$> T.splitOn "," s
splitOnCommas _ = []
opts = info (helper <*> pathParser) $
fullDesc
<> progDesc (
"PostgREST "
<> toS prettyVersion
<> " / create a REST API to an existing Postgres database"
)
<> footerDoc (Just $
text "Example Config File:"
L.<> nest 2 (hardline L.<> exampleCfg)
)
-- | Read the JWT secret from a file if configJwtSecret is actually a
-- filepath(has @ as its prefix). To check if the JWT secret is provided is
-- in fact a file path, it must be decoded as 'Text' to be processed.
readSecretFile :: AppConfig -> IO AppConfig
readSecretFile conf =
maybe (return conf) readSecret maybeFilename
where
maybeFilename = T.stripPrefix "@" . decodeUtf8 =<< configJwtSecret conf
readSecret filename = do
jwtSecret <- chomp <$> BS.readFile (toS filename)
return $ conf { configJwtSecret = Just jwtSecret }
chomp bs = fromMaybe bs (BS.stripSuffix "\n" bs)
parserPrefs = prefs showHelpOnError
decodeSecret :: AppConfig -> IO AppConfig
decodeSecret conf@AppConfig{..} =
case (configJwtSecretIsBase64, configJwtSecret) of
(True, Just secret) ->
either fail (return . updateSecret) $ decodeB64 secret
_ -> return conf
where
updateSecret bs = conf { configJwtSecret = Just bs }
decodeB64 = B64.decode . encodeUtf8 . T.strip . replaceUrlChars . decodeUtf8
replaceUrlChars = T.replace "_" "/" . T.replace "-" "+" . T.replace "." "="
exitErr :: Text -> IO a
exitErr err = do
hPutStrLn stderr err
exitFailure
-- | Parse `jwt-secret` configuration option and turn into a JWKSet.
--
-- There are three ways to specify `jwt-secret`: text secret, JSON Web Key
-- (JWK), or JSON Web Key Set (JWKS). The first two are converted into a JWKSet
-- with one key and the last is converted as is.
decodeJWKS :: AppConfig -> AppConfig
decodeJWKS conf =
conf { configJWKS = parseSecret <$> configJwtSecret conf }
exampleCfg :: Doc
exampleCfg = vsep . map (text . toS) . lines $
[str|db-uri = "postgres://user:pass@localhost:5432/dbname"
|db-schema = "public" # this schema gets added to the search_path of every request
|db-anon-role = "postgres"
|db-pool = 10
|db-pool-timeout = 10
|
|server-host = "!4"
|server-port = 3000
|
|## unix socket location
|## if specified it takes precedence over server-port
|# server-unix-socket = "/tmp/pgrst.sock"
|## unix socket file mode
|## when none is provided, 660 is applied by default
|# server-unix-socket-mode = "660"
|
|## base url for swagger output
|# openapi-server-proxy-uri = ""
|
|## choose a secret, JSON Web Key (or set) to enable JWT auth
|## (use "@filename" to load from separate file)
|# jwt-secret = "secret_with_at_least_32_characters"
|# secret-is-base64 = false
|# jwt-aud = "your_audience_claim"
|
|## limit rows in response
|# max-rows = 1000
|
|## stored proc to exec immediately after auth
|# pre-request = "stored_proc_name"
|
|## jspath to the role claim key
|# role-claim-key = ".role"
|
|## extra schemas to add to the search_path of every request
|# db-extra-search-path = "extensions, util"
|
|## stored proc that overrides the root "/" spec
|## it must be inside the db-schema
|# root-spec = "stored_proc_name"
|
|## content types to produce raw output
|# raw-media-types="image/png, image/jpg"
|]
parseSecret :: ByteString -> JWKSet
parseSecret bytes =
fromMaybe (maybe secret (\jwk' -> JWT.JWKSet [jwk']) maybeJWK)
maybeJWKSet
where
maybeJWKSet = JSON.decodeStrict bytes :: Maybe JWKSet
maybeJWK = JSON.decodeStrict bytes :: Maybe JWK
secret = JWT.JWKSet [JWT.fromKeyMaterial keyMaterial]
keyMaterial = JWT.OctKeyMaterial . JWT.OctKeyParameters $ JOSE.Base64Octets bytes
pathParser :: Parser FilePath
pathParser =
strArgument $
metavar "FILENAME" <>
help "Path to configuration file"
-- | Read database uri from a separate file if `db-uri` is a filepath.
readDbUriFile :: Maybe Text -> AppConfig -> IO AppConfig
readDbUriFile maybeDbUri conf =
case maybeDbUri of
Just prevDbUri ->
pure $ conf { configDbUri = prevDbUri }
Nothing ->
case T.stripPrefix "@" $ configDbUri conf of
Nothing -> return conf
Just filename -> do
dbUri <- T.strip <$> readFile (toS filename)
return $ conf { configDbUri = dbUri }
type Environment = M.Map [Char] Text
-- | Read environment variables that start with PGRST_
readPGRSTEnvironment :: IO Environment
readPGRSTEnvironment =
M.map T.pack . M.fromList . filter (isPrefixOf "PGRST_" . fst) <$> getEnvironment
+67
View File
@@ -0,0 +1,67 @@
{-# LANGUAGE QuasiQuotes #-}
module PostgREST.Config.Database
( pgVersionStatement
, queryDbSettings
, queryPgVersion
) where
import PostgREST.Config.PgVersion (PgVersion (..))
import qualified Hasql.Decoders as HD
import qualified Hasql.Encoders as HE
import qualified Hasql.Pool as SQL
import Hasql.Session (Session, statement)
import qualified Hasql.Statement as SQL
import qualified Hasql.Transaction as SQL
import qualified Hasql.Transaction.Sessions as SQL
import Text.InterpolatedString.Perl6 (q)
import Protolude
queryPgVersion :: Session PgVersion
queryPgVersion = statement mempty pgVersionStatement
pgVersionStatement :: SQL.Statement () PgVersion
pgVersionStatement = SQL.Statement sql HE.noParams versionRow False
where
sql = "SELECT current_setting('server_version_num')::integer, current_setting('server_version')"
versionRow = HD.singleRow $ PgVersion <$> column HD.int4 <*> column HD.text
queryDbSettings :: SQL.Pool -> Bool -> IO (Either SQL.UsageError [(Text, Text)])
queryDbSettings pool prepared =
let transaction = if prepared then SQL.transaction else SQL.unpreparedTransaction in
SQL.use pool . transaction SQL.ReadCommitted SQL.Read $
SQL.statement mempty dbSettingsStatement
-- | Get db settings from the connection role. Global settings will be overridden by database specific settings.
dbSettingsStatement :: SQL.Statement () [(Text, Text)]
dbSettingsStatement = SQL.Statement sql HE.noParams decodeSettings False
where
sql = [q|
WITH
role_setting (database, setting) AS (
SELECT setdatabase,
unnest(setconfig)
FROM pg_catalog.pg_db_role_setting
WHERE setrole = CURRENT_USER::regrole::oid
AND setdatabase IN (0, (SELECT oid FROM pg_catalog.pg_database WHERE datname = CURRENT_CATALOG))
),
kv_settings (database, k, v) AS (
SELECT database,
substr(setting, 1, strpos(setting, '=') - 1),
substr(setting, strpos(setting, '=') + 1)
FROM role_setting
WHERE setting LIKE 'pgrst.%'
)
SELECT DISTINCT ON (key)
replace(k, 'pgrst.', '') AS key,
v AS value
FROM kv_settings
ORDER BY key, database DESC;
|]
decodeSettings = HD.rowList $ (,) <$> column HD.text <*> column HD.text
column :: HD.Value a -> HD.Row a
column = HD.column . HD.nonNullable
+51
View File
@@ -0,0 +1,51 @@
module PostgREST.Config.JSPath
( JSPath
, JSPathExp(..)
, dumpJSPath
, pRoleClaimKey
) where
import qualified Text.ParserCombinators.Parsec as P
import Data.Either.Combinators (mapLeft)
import Text.ParserCombinators.Parsec ((<?>))
import Text.Read (read)
import Protolude
-- | full jspath, e.g. .property[0].attr.detail
type JSPath = [JSPathExp]
-- | jspath expression, e.g. .property, .property[0] or ."property-dash"
data JSPathExp
= JSPKey Text
| JSPIdx Int
dumpJSPath :: JSPathExp -> Text
-- TODO: this needs to be quoted properly for special chars
dumpJSPath (JSPKey k) = "." <> show k
dumpJSPath (JSPIdx i) = "[" <> show i <> "]"
-- Used for the config value "role-claim-key"
pRoleClaimKey :: Text -> Either Text JSPath
pRoleClaimKey selStr =
mapLeft show $ P.parse pJSPath ("failed to parse role-claim-key value (" <> toS selStr <> ")") (toS selStr)
pJSPath :: P.Parser JSPath
pJSPath = toJSPath <$> (period *> pPath `P.sepBy` period <* P.eof)
where
toJSPath :: [(Text, Maybe Int)] -> JSPath
toJSPath = concatMap (\(key, idx) -> JSPKey key : maybeToList (JSPIdx <$> idx))
period = P.char '.' <?> "period (.)"
pPath :: P.Parser (Text, Maybe Int)
pPath = (,) <$> pJSPKey <*> P.optionMaybe pJSPIdx
pJSPKey :: P.Parser Text
pJSPKey = toS <$> P.many1 (P.alphaNum <|> P.oneOf "_$@") <|> pQuotedValue <?> "attribute name [a..z0..9_$@])"
pJSPIdx :: P.Parser Int
pJSPIdx = P.char '[' *> (read <$> P.many1 P.digit) <* P.char ']' <?> "array index [0..n]"
pQuotedValue :: P.Parser Text
pQuotedValue = toS <$> (P.char '"' *> P.many (P.noneOf "\"") <* P.char '"')
+60
View File
@@ -0,0 +1,60 @@
{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DeriveGeneric #-}
module PostgREST.Config.PgVersion
( PgVersion(..)
, minimumPgVersion
, pgVersion96
, pgVersion100
, pgVersion109
, pgVersion110
, pgVersion112
, pgVersion114
, pgVersion121
, pgVersion130
, pgVersion140
) where
import qualified Data.Aeson as JSON
import Protolude
data PgVersion = PgVersion
{ pgvNum :: Int32
, pgvName :: Text
}
deriving (Eq, Generic, JSON.ToJSON)
instance Ord PgVersion where
(PgVersion v1 _) `compare` (PgVersion v2 _) = v1 `compare` v2
-- | Tells the minimum PostgreSQL version required by this version of PostgREST
minimumPgVersion :: PgVersion
minimumPgVersion = pgVersion96
pgVersion96 :: PgVersion
pgVersion96 = PgVersion 90600 "9.6"
pgVersion100 :: PgVersion
pgVersion100 = PgVersion 100000 "10"
pgVersion109 :: PgVersion
pgVersion109 = PgVersion 100009 "10.9"
pgVersion110 :: PgVersion
pgVersion110 = PgVersion 110000 "11.0"
pgVersion112 :: PgVersion
pgVersion112 = PgVersion 110002 "11.2"
pgVersion114 :: PgVersion
pgVersion114 = PgVersion 110004 "11.4"
pgVersion121 :: PgVersion
pgVersion121 = PgVersion 120001 "12.1"
pgVersion130 :: PgVersion
pgVersion130 = PgVersion 130000 "13.0"
pgVersion140 :: PgVersion
pgVersion140 = PgVersion 140000 "14.0"
+77
View File
@@ -0,0 +1,77 @@
{-|
Module : PostgREST.Private.ProxyUri
Description : Proxy Uri validator
-}
module PostgREST.Config.Proxy
( Proxy(..)
, isMalformedProxyUri
, toURI
) where
import qualified Data.Text as T
import Data.Maybe (fromJust)
import Network.URI (URI (..), URIAuth (..), isAbsoluteURI, parseURI)
import Protolude hiding (Proxy)
data Proxy = Proxy
{ proxyScheme :: Text
, proxyHost :: Text
, proxyPort :: Integer
, proxyPath :: Text
}
{-|
Test whether a proxy uri is malformed or not.
A valid proxy uri should be an absolute uri without query and user info,
only http(s) schemes are valid, port number range is 1-65535.
For example
http://postgrest.com/openapi.json
https://postgrest.com:8080/openapi.json
-}
isMalformedProxyUri :: Text -> Bool
isMalformedProxyUri uri
| isAbsoluteURI (toS uri) = not $ isUriValid $ toURI uri
| otherwise = True
toURI :: Text -> URI
toURI uri = fromJust $ parseURI (toS uri)
isUriValid:: URI -> Bool
isUriValid = fAnd [isSchemeValid, isQueryValid, isAuthorityValid]
fAnd :: [a -> Bool] -> a -> Bool
fAnd fs x = all ($ x) fs
isSchemeValid :: URI -> Bool
isSchemeValid URI {uriScheme = s}
| T.toLower (T.pack s) == "https:" = True
| T.toLower (T.pack s) == "http:" = True
| otherwise = False
isQueryValid :: URI -> Bool
isQueryValid URI {uriQuery = ""} = True
isQueryValid _ = False
isAuthorityValid :: URI -> Bool
isAuthorityValid URI {uriAuthority = a}
| isJust a = fAnd [isUserInfoValid, isHostValid, isPortValid] $ fromJust a
| otherwise = False
isUserInfoValid :: URIAuth -> Bool
isUserInfoValid URIAuth {uriUserInfo = ""} = True
isUserInfoValid _ = False
isHostValid :: URIAuth -> Bool
isHostValid URIAuth {uriRegName = ""} = False
isHostValid _ = True
isPortValid :: URIAuth -> Bool
isPortValid URIAuth {uriPort = ""} = True
isPortValid URIAuth {uriPort = (':':p)} =
case readMaybe p of
Just i -> i > (0 :: Integer) && i < 65536
Nothing -> False
isPortValid _ = False
+64
View File
@@ -0,0 +1,64 @@
{-# LANGUAGE DuplicateRecordFields #-}
module PostgREST.ContentType
( ContentType(..)
, toHeader
, toMime
, decodeContentType
) where
import qualified Data.ByteString as BS
import qualified Data.ByteString.Internal as BS (c2w)
import Network.HTTP.Types.Header (Header, hContentType)
import Protolude
-- | Enumeration of currently supported response content types
data ContentType
= CTApplicationJSON
| CTSingularJSON
| CTTextCSV
| CTTextPlain
| CTOpenAPI
| CTUrlEncoded
| CTOctetStream
| CTAny
| CTOther ByteString
deriving (Eq)
-- | Convert from ContentType to a full HTTP Header
toHeader :: ContentType -> Header
toHeader ct = (hContentType, toMime ct <> charset)
where
charset = case ct of
CTOctetStream -> mempty
CTOther _ -> mempty
_ -> "; charset=utf-8"
-- | Convert from ContentType to a ByteString representing the mime type
toMime :: ContentType -> ByteString
toMime CTApplicationJSON = "application/json"
toMime CTTextCSV = "text/csv"
toMime CTTextPlain = "text/plain"
toMime CTOpenAPI = "application/openapi+json"
toMime CTSingularJSON = "application/vnd.pgrst.object+json"
toMime CTUrlEncoded = "application/x-www-form-urlencoded"
toMime CTOctetStream = "application/octet-stream"
toMime CTAny = "*/*"
toMime (CTOther ct) = ct
-- | Convert from ByteString to ContentType. Warning: discards MIME parameters
decodeContentType :: BS.ByteString -> ContentType
decodeContentType ct =
case BS.takeWhile (/= BS.c2w ';') ct of
"application/json" -> CTApplicationJSON
"text/csv" -> CTTextCSV
"text/plain" -> CTTextPlain
"application/openapi+json" -> CTOpenAPI
"application/vnd.pgrst.object+json" -> CTSingularJSON
"application/vnd.pgrst.object" -> CTSingularJSON
"application/x-www-form-urlencoded" -> CTUrlEncoded
"application/octet-stream" -> CTOctetStream
"*/*" -> CTAny
ct' -> CTOther ct'
+42
View File
@@ -0,0 +1,42 @@
{-|
Module : PostgREST.Cors
Description : Wai Middleware to set cors policy.
-}
module PostgREST.Cors (middleware) where
import qualified Data.ByteString.Char8 as BS
import qualified Data.CaseInsensitive as CI
import qualified Network.Wai as Wai
import qualified Network.Wai.Middleware.Cors as Wai
import Data.List (lookup)
import Protolude
middleware :: Wai.Middleware
middleware = Wai.cors corsPolicy
-- | CORS policy to be used in by Wai Cors middleware
corsPolicy :: Wai.Request -> Maybe Wai.CorsResourcePolicy
corsPolicy req = case lookup "origin" headers of
Just origin ->
Just Wai.CorsResourcePolicy
{ Wai.corsOrigins = Just ([origin], True)
, Wai.corsMethods = ["GET", "POST", "PATCH", "PUT", "DELETE", "OPTIONS"]
, Wai.corsRequestHeaders = "Authorization" : accHeaders
, Wai.corsExposedHeaders = Just
[ "Content-Encoding", "Content-Location", "Content-Range", "Content-Type"
, "Date", "Location", "Server", "Transfer-Encoding", "Range-Unit"]
, Wai.corsMaxAge = Just $ 60*60*24
, Wai.corsVaryOrigin = False
, Wai.corsRequireOrigin = False
, Wai.corsIgnoreFailures = True
}
Nothing -> Nothing
where
headers = Wai.requestHeaders req
accHeaders = case lookup "access-control-request-headers" headers of
Just hdrs -> map (CI.mk . BS.strip) $ BS.split ',' hdrs
-- Impossible case, Middleware.Cors will not evaluate this when
-- the Access-Control-Request-Headers header is not set.
Nothing -> []
-351
View File
@@ -1,351 +0,0 @@
{-|
Module : PostgREST.DbRequestBuilder
Description : PostgREST database request builder
This module is in charge of building an intermediate representation(ReadRequest, MutateRequest) between the HTTP request and the final resulting SQL query.
A query tree is built in case of resource embedding. By inferring the relationship between tables, join conditions are added for every embedded resource.
-}
{-# LANGUAGE DuplicateRecordFields #-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE NamedFieldPuns #-}
module PostgREST.DbRequestBuilder (
readRequest
, mutateRequest
, returningCols
) where
import qualified Data.HashMap.Strict as M
import qualified Data.Set as S
import Control.Arrow ((***))
import Data.Either.Combinators (mapLeft)
import Data.Foldable (foldr1)
import Data.List (delete)
import Data.Text (isInfixOf)
import Control.Applicative
import Data.Tree
import Network.Wai
import PostgREST.ApiRequest (Action (..), ApiRequest (..))
import PostgREST.Error (ApiRequestError (..), errorResponseFor)
import PostgREST.Parsers
import PostgREST.RangeQuery (NonnegRange, allRange, restrictRange)
import PostgREST.Types
import Protolude hiding (from)
readRequest :: Schema -> TableName -> Maybe Integer -> [Relation] -> ApiRequest -> Either Response ReadRequest
readRequest schema rootTableName maxRows allRels apiRequest =
mapLeft errorResponseFor $
treeRestrictRange maxRows =<<
augmentRequestWithJoin schema rootRels =<<
addFiltersOrdersRanges apiRequest <*>
(initReadRequest rootName <$> pRequestSelect sel)
where
sel = fromMaybe "*" $ iSelect apiRequest -- default to all columns requested (SELECT *) for a non existent ?select querystring param
(rootName, rootRels) = rootWithRels schema rootTableName allRels (iAction apiRequest)
-- Get the root table name with its relationships according to the Action type.
-- This is done because of the shape of the final SQL Query. The mutation cases are wrapped in a WITH {sourceCTEName}(see Statements.hs).
-- So we need a FROM {sourceCTEName} instead of FROM {tableName}.
rootWithRels :: Schema -> TableName -> [Relation] -> Action -> (QualifiedIdentifier, [Relation])
rootWithRels schema rootTableName allRels action = case action of
ActionRead _ -> (QualifiedIdentifier schema rootTableName, allRels) -- normal read case
_ -> (QualifiedIdentifier mempty sourceCTEName, mapMaybe toSourceRel allRels ++ allRels) -- mutation cases and calling proc
where
-- To enable embedding in the sourceCTEName cases we need to replace the foreign key tableName in the Relation
-- with {sourceCTEName}. This way findRel can find relationships with sourceCTEName.
toSourceRel :: Relation -> Maybe Relation
toSourceRel r@Relation{relTable=t}
| rootTableName == tableName t = Just $ r {relTable=t {tableName=sourceCTEName}}
| otherwise = Nothing
-- Build the initial tree with a Depth attribute so when a self join occurs we can differentiate the parent and child tables by having
-- an alias like "table_depth", this is related to http://github.com/PostgREST/postgrest/issues/987.
initReadRequest :: QualifiedIdentifier -> [Tree SelectItem] -> ReadRequest
initReadRequest rootQi =
foldr (treeEntry rootDepth) initial
where
rootDepth = 0
rootSchema = qiSchema rootQi
rootName = qiName rootQi
initial = Node (Select [] rootQi Nothing [] [] [] [] allRange, (rootName, Nothing, Nothing, Nothing, rootDepth)) []
treeEntry :: Depth -> Tree SelectItem -> ReadRequest -> ReadRequest
treeEntry depth (Node fld@((fn, _),_,alias, embedHint) fldForest) (Node (q, i) rForest) =
let nxtDepth = succ depth in
case fldForest of
[] -> Node (q {select=fld:select q}, i) rForest
_ -> Node (q, i) $
foldr (treeEntry nxtDepth)
(Node (Select [] (QualifiedIdentifier rootSchema fn) Nothing [] [] [] [] allRange,
(fn, Nothing, alias, embedHint, nxtDepth)) [])
fldForest:rForest
treeRestrictRange :: Maybe Integer -> ReadRequest -> Either ApiRequestError ReadRequest
treeRestrictRange maxRows request = pure $ nodeRestrictRange maxRows <$> request
where
nodeRestrictRange :: Maybe Integer -> ReadNode -> ReadNode
nodeRestrictRange m (q@Select {range_=r}, i) = (q{range_=restrictRange m r }, i)
augmentRequestWithJoin :: Schema -> [Relation] -> ReadRequest -> Either ApiRequestError ReadRequest
augmentRequestWithJoin schema allRels request =
addRels schema allRels Nothing request
>>= addJoinConditions Nothing
addRels :: Schema -> [Relation] -> Maybe ReadRequest -> ReadRequest -> Either ApiRequestError ReadRequest
addRels schema allRels parentNode (Node (query@Select{from=tbl}, (nodeName, _, alias, hint, depth)) forest) =
case parentNode of
Just (Node (Select{from=parentNodeQi}, _) _) ->
let newFrom r = if qiName tbl == nodeName then tableQi (relFTable r) else tbl
newReadNode = (\r -> (query{from=newFrom r}, (nodeName, Just r, alias, Nothing, depth))) <$> rel
rel = findRel schema allRels (qiName parentNodeQi) nodeName hint
in
Node <$> newReadNode <*> (updateForest . hush $ Node <$> newReadNode <*> pure forest)
_ ->
let rn = (query, (nodeName, Nothing, alias, Nothing, depth)) in
Node rn <$> updateForest (Just $ Node rn forest)
where
updateForest :: Maybe ReadRequest -> Either ApiRequestError [ReadRequest]
updateForest rq = mapM (addRels schema allRels rq) forest
-- Finds a relationship between an origin and a target in the request: /origin?select=target(*)
-- If more than one relationship is found then the request is ambiguous and we return an error.
-- In that case the request can be disambiguated by adding precision to the target or by using a hint: /origin?select=target!hint(*)
-- The elements will be matched according to these rules:
-- origin = table / view
-- target = table / view / constraint / column-from-origin
-- hint = table / view / constraint / column-from-origin / column-from-target
-- (hint can take table / view values to aid in finding the junction in an m2m relationship)
findRel :: Schema -> [Relation] -> NodeName -> NodeName -> Maybe EmbedHint -> Either ApiRequestError Relation
findRel schema allRels origin target hint =
case rel of
[] -> Left $ NoRelBetween origin target
[r] -> Right r
rs ->
-- Return error if more than one relationship is found, unless we're in a self reference case.
--
-- Here we handle a self reference relationship to not cause a breaking change:
-- In a self reference we get two relationships with the same foreign key and relTable/relFtable but with different cardinalities(m2o/o2m)
-- We output the O2M rel, the M2O rel can be obtained by using the origin column as an embed hint.
let [rel0, rel1] = take 2 rs in
if length rs == 2 && relConstraint rel0 == relConstraint rel1 && relTable rel0 == relTable rel1 && relFTable rel0 == relFTable rel1
then note (NoRelBetween origin target) (find (\r -> relType r == O2M) rs)
else Left $ AmbiguousRelBetween origin target rs
where
matchFKSingleCol hint_ cols = length cols == 1 && hint_ == (colName <$> head cols)
rel = filter (
\Relation{relTable, relColumns, relConstraint, relFTable, relFColumns, relType, relJunction} ->
-- Both relationship ends need to be on the exposed schema
schema == tableSchema relTable && schema == tableSchema relFTable &&
(
-- /projects?select=clients(*)
origin == tableName relTable && -- projects
target == tableName relFTable || -- clients
-- /projects?select=projects_client_id_fkey(*)
(
origin == tableName relTable && -- projects
Just target == relConstraint -- projects_client_id_fkey
) ||
-- /projects?select=client_id(*)
(
origin == tableName relTable && -- projects
matchFKSingleCol (Just target) relColumns -- client_id
)
) && (
isNothing hint || -- hint is optional
-- /projects?select=clients!projects_client_id_fkey(*)
hint == relConstraint || -- projects_client_id_fkey
-- /projects?select=clients!client_id(*) or /projects?select=clients!id(*)
matchFKSingleCol hint relColumns || -- client_id
matchFKSingleCol hint relFColumns || -- id
-- /users?select=tasks!users_tasks(*)
(
relType == M2M && -- many-to-many between users and tasks
hint == (tableName . junTable <$> relJunction) -- users_tasks
)
)
) allRels
-- previousAlias is only used for the case of self joins
addJoinConditions :: Maybe Alias -> ReadRequest -> Either ApiRequestError ReadRequest
addJoinConditions previousAlias (Node node@(query@Select{from=tbl}, nodeProps@(_, rel, _, _, depth)) forest) =
case rel of
Just r@Relation{relType=O2M} -> Node (augmentQuery r, nodeProps) <$> updatedForest
Just r@Relation{relType=M2O} -> Node (augmentQuery r, nodeProps) <$> updatedForest
Just r@Relation{relType=M2M, relJunction=junction} ->
case junction of
Just Junction{junTable} ->
let rq = augmentQuery r in
Node (rq{implicitJoins=tableQi junTable:implicitJoins rq}, nodeProps) <$> updatedForest
Nothing ->
Left UnknownRelation
Nothing -> Node node <$> updatedForest
where
newAlias = case isSelfReference <$> rel of
Just True
| depth /= 0 -> Just (qiName tbl <> "_" <> show depth) -- root node doesn't get aliased
| otherwise -> Nothing
_ -> Nothing
augmentQuery r =
foldr
(\jc rq@Select{joinConditions=jcs} -> rq{joinConditions=jc:jcs})
query{fromAlias=newAlias}
(getJoinConditions previousAlias newAlias r)
updatedForest = mapM (addJoinConditions newAlias) forest
-- previousAlias and newAlias are used in the case of self joins
getJoinConditions :: Maybe Alias -> Maybe Alias -> Relation -> [JoinCondition]
getJoinConditions previousAlias newAlias (Relation Table{tableSchema=tSchema, tableName=tN} cols _ Table{tableName=ftN} fCols typ jun) =
case typ of
O2M ->
zipWith (toJoinCondition tN ftN) cols fCols
M2O ->
zipWith (toJoinCondition tN ftN) cols fCols
M2M -> case jun of
Just (Junction jt _ jc1 _ jc2) ->
let jtn = tableName jt in
zipWith (toJoinCondition tN jtn) cols jc1 ++ zipWith (toJoinCondition ftN jtn) fCols jc2
Nothing -> []
where
toJoinCondition :: Text -> Text -> Column -> Column -> JoinCondition
toJoinCondition tb ftb c fc =
let qi1 = removeSourceCTESchema tSchema tb
qi2 = removeSourceCTESchema tSchema ftb in
JoinCondition (maybe qi1 (QualifiedIdentifier mempty) previousAlias, colName c)
(maybe qi2 (QualifiedIdentifier mempty) newAlias, colName fc)
-- On mutation and calling proc cases we wrap the target table in a WITH {sourceCTEName}
-- if this happens remove the schema `FROM "schema"."{sourceCTEName}"` and use only the
-- `FROM "{sourceCTEName}"`. If the schema remains the FROM would be invalid.
removeSourceCTESchema :: Schema -> TableName -> QualifiedIdentifier
removeSourceCTESchema schema tbl = QualifiedIdentifier (if tbl == sourceCTEName then mempty else schema) tbl
addFiltersOrdersRanges :: ApiRequest -> Either ApiRequestError (ReadRequest -> ReadRequest)
addFiltersOrdersRanges apiRequest = foldr1 (liftA2 (.)) [
flip (foldr addFilter) <$> filters,
flip (foldr addOrder) <$> orders,
flip (foldr addRange) <$> ranges,
flip (foldr addLogicTree) <$> logicForest
]
{-
The esence of what is going on above is that we are composing tree functions
of type (ReadRequest->ReadRequest) that are in (Either ApiRequestError a) context
-}
where
filters :: Either ApiRequestError [(EmbedPath, Filter)]
filters = mapM pRequestFilter flts
logicForest :: Either ApiRequestError [(EmbedPath, LogicTree)]
logicForest = mapM pRequestLogicTree logFrst
action = iAction apiRequest
-- there can be no filters on the root table when we are doing insert/update/delete
(flts, logFrst) =
case action of
ActionInvoke _ -> (iFilters apiRequest, iLogic apiRequest)
ActionRead _ -> (iFilters apiRequest, iLogic apiRequest)
_ -> join (***) (filter (( "." `isInfixOf` ) . fst)) (iFilters apiRequest, iLogic apiRequest)
orders :: Either ApiRequestError [(EmbedPath, [OrderTerm])]
orders = mapM pRequestOrder $ iOrder apiRequest
ranges :: Either ApiRequestError [(EmbedPath, NonnegRange)]
ranges = mapM pRequestRange $ M.toList $ iRange apiRequest
addFilterToNode :: Filter -> ReadRequest -> ReadRequest
addFilterToNode flt (Node (q@Select {where_=lf}, i) f) = Node (q{where_=addFilterToLogicForest flt lf}::ReadQuery, i) f
addFilter :: (EmbedPath, Filter) -> ReadRequest -> ReadRequest
addFilter = addProperty addFilterToNode
addOrderToNode :: [OrderTerm] -> ReadRequest -> ReadRequest
addOrderToNode o (Node (q,i) f) = Node (q{order=o}, i) f
addOrder :: (EmbedPath, [OrderTerm]) -> ReadRequest -> ReadRequest
addOrder = addProperty addOrderToNode
addRangeToNode :: NonnegRange -> ReadRequest -> ReadRequest
addRangeToNode r (Node (q,i) f) = Node (q{range_=r}, i) f
addRange :: (EmbedPath, NonnegRange) -> ReadRequest -> ReadRequest
addRange = addProperty addRangeToNode
addLogicTreeToNode :: LogicTree -> ReadRequest -> ReadRequest
addLogicTreeToNode t (Node (q@Select{where_=lf},i) f) = Node (q{where_=t:lf}::ReadQuery, i) f
addLogicTree :: (EmbedPath, LogicTree) -> ReadRequest -> ReadRequest
addLogicTree = addProperty addLogicTreeToNode
addProperty :: (a -> ReadRequest -> ReadRequest) -> (EmbedPath, a) -> ReadRequest -> ReadRequest
addProperty f ([], a) rr = f a rr
addProperty f (targetNodeName:remainingPath, a) (Node rn forest) =
case pathNode of
Nothing -> Node rn forest -- the property is silenty dropped in the Request does not contain the required path
Just tn -> Node rn (addProperty f (remainingPath, a) tn:delete tn forest)
where
pathNode = find (\(Node (_,(nodeName,_,alias,_,_)) _) -> nodeName == targetNodeName || alias == Just targetNodeName) forest
mutateRequest :: Schema -> TableName -> ApiRequest -> S.Set FieldName -> [FieldName] -> ReadRequest -> Either Response MutateRequest
mutateRequest schema tName apiRequest cols pkCols readReq = mapLeft errorResponseFor $
case action of
ActionCreate -> do
confCols <- case iOnConflict apiRequest of
Nothing -> pure pkCols
Just param -> pRequestOnConflict param
pure $ Insert qi cols ((,) <$> iPreferResolution apiRequest <*> Just confCols) [] returnings
ActionUpdate -> Update qi cols <$> combinedLogic <*> pure returnings
ActionSingleUpsert ->
(\flts ->
if null (iLogic apiRequest) &&
S.fromList (fst <$> iFilters apiRequest) == S.fromList pkCols &&
not (null (S.fromList pkCols)) &&
all (\case
Filter _ (OpExpr False (Op "eq" _)) -> True
_ -> False) flts
then Insert qi cols (Just (MergeDuplicates, pkCols)) <$> combinedLogic <*> pure returnings
else
Left InvalidFilters) =<< filters
ActionDelete -> Delete qi <$> combinedLogic <*> pure returnings
_ -> Left UnsupportedVerb
where
qi = QualifiedIdentifier schema tName
action = iAction apiRequest
returnings =
if iPreferRepresentation apiRequest == None
then []
else returningCols readReq
filters = map snd <$> mapM pRequestFilter mutateFilters
logic = map snd <$> mapM pRequestLogicTree logicFilters
combinedLogic = foldr addFilterToLogicForest <$> logic <*> filters
-- update/delete filters can be only on the root table
(mutateFilters, logicFilters) = join (***) onlyRoot (iFilters apiRequest, iLogic apiRequest)
onlyRoot = filter (not . ( "." `isInfixOf` ) . fst)
returningCols :: ReadRequest -> [FieldName]
returningCols rr@(Node _ forest) = returnings
where
fldNames = fstFieldNames rr
-- Without fkCols, when a mutateRequest to /projects?select=name,clients(name) occurs, the RETURNING SQL part would be
-- `RETURNING name`(see QueryBuilder).
-- This would make the embedding fail because the following JOIN would need the "client_id" column from projects.
-- So this adds the foreign key columns to ensure the embedding succeeds, result would be `RETURNING name, client_id`.
-- This also works for the other relType's.
fkCols = concat $ mapMaybe (\case
Node (_, (_, Just Relation{relColumns=cols, relType=relTyp}, _, _, _)) _ -> case relTyp of
O2M -> Just cols
M2O -> Just cols
M2M -> Just cols
_ -> Nothing
) forest
-- However if the "client_id" is present, e.g. mutateRequest to /projects?select=client_id,name,clients(name)
-- we would get `RETURNING client_id, name, client_id` and then we would produce the "column reference \"client_id\" is ambiguous"
-- error from PostgreSQL. So we deduplicate with Set:
returnings = S.toList . S.fromList $ fldNames ++ (colName <$> fkCols)
-- Traditional filters(e.g. id=eq.1) are added as root nodes of the LogicTree
-- they are later concatenated with AND in the QueryBuilder
addFilterToLogicForest :: Filter -> [LogicTree] -> [LogicTree]
addFilterToLogicForest flt lf = Stmnt flt : lf
File diff suppressed because it is too large Load Diff
+42
View File
@@ -0,0 +1,42 @@
{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DeriveGeneric #-}
module PostgREST.DbStructure.Identifiers
( QualifiedIdentifier(..)
, Schema
, TableName
, FieldName
, dumpQi
, toQi
) where
import qualified Data.Aeson as JSON
import qualified Data.Text as T
import Protolude
-- | Represents a pg identifier with a prepended schema name "schema.table".
-- When qiSchema is "", the schema is defined by the pg search_path.
data QualifiedIdentifier = QualifiedIdentifier
{ qiSchema :: Schema
, qiName :: TableName
}
deriving (Eq, Ord, Generic, JSON.ToJSON, JSON.ToJSONKey)
instance Hashable QualifiedIdentifier
dumpQi :: QualifiedIdentifier -> Text
dumpQi (QualifiedIdentifier s i) =
(if T.null s then mempty else s <> ".") <> i
-- TODO: Handle a case where the QI comes like this: "my.fav.schema"."my.identifier"
-- Right now it only handles the schema.identifier case
toQi :: Text -> QualifiedIdentifier
toQi txt = case T.drop 1 <$> T.breakOn "." txt of
(i, "") -> QualifiedIdentifier mempty i
(s, i) -> QualifiedIdentifier s i
type Schema = Text
type TableName = Text
type FieldName = Text
+85
View File
@@ -0,0 +1,85 @@
{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DeriveGeneric #-}
module PostgREST.DbStructure.Proc
( PgType(..)
, ProcDescription(..)
, ProcParam(..)
, ProcVolatility(..)
, ProcsMap
, RetType(..)
, procReturnsScalar
, procReturnsSingle
, procTableName
) where
import qualified Data.Aeson as JSON
import qualified Data.HashMap.Strict as M
import PostgREST.DbStructure.Identifiers (QualifiedIdentifier (..),
Schema, TableName)
import Protolude
data PgType
= Scalar
| Composite QualifiedIdentifier
deriving (Eq, Ord, Generic, JSON.ToJSON)
data RetType
= Single PgType
| SetOf PgType
deriving (Eq, Ord, Generic, JSON.ToJSON)
data ProcVolatility
= Volatile
| Stable
| Immutable
deriving (Eq, Ord, Generic, JSON.ToJSON)
data ProcDescription = ProcDescription
{ pdSchema :: Schema
, pdName :: Text
, pdDescription :: Maybe Text
, pdParams :: [ProcParam]
, pdReturnType :: RetType
, pdVolatility :: ProcVolatility
, pdHasVariadic :: Bool
}
deriving (Eq, Generic, JSON.ToJSON)
data ProcParam = ProcParam
{ ppName :: Text
, ppType :: Text
, ppReq :: Bool
, ppVar :: Bool
}
deriving (Eq, Ord, Generic, JSON.ToJSON)
-- Order by least number of params in the case of overloaded functions
instance Ord ProcDescription where
ProcDescription schema1 name1 des1 prms1 rt1 vol1 hasVar1 `compare` ProcDescription schema2 name2 des2 prms2 rt2 vol2 hasVar2
| schema1 == schema2 && name1 == name2 && length prms1 < length prms2 = LT
| schema2 == schema2 && name1 == name2 && length prms1 > length prms2 = GT
| otherwise = (schema1, name1, des1, prms1, rt1, vol1, hasVar1) `compare` (schema2, name2, des2, prms2, rt2, vol2, hasVar2)
-- | A map of all procs, all of which can be overloaded(one entry will have more than one ProcDescription).
-- | It uses a HashMap for a faster lookup.
type ProcsMap = M.HashMap QualifiedIdentifier [ProcDescription]
procReturnsScalar :: ProcDescription -> Bool
procReturnsScalar proc = case proc of
ProcDescription{pdReturnType = (Single Scalar)} -> True
ProcDescription{pdReturnType = (SetOf Scalar)} -> True
_ -> False
procReturnsSingle :: ProcDescription -> Bool
procReturnsSingle proc = case proc of
ProcDescription{pdReturnType = (Single _)} -> True
_ -> False
procTableName :: ProcDescription -> Maybe TableName
procTableName proc = case pdReturnType proc of
SetOf (Composite qi) -> Just $ qiName qi
Single (Composite qi) -> Just $ qiName qi
_ -> Nothing

Some files were not shown because too many files have changed in this diff Show More