Commit Graph
925 Commits
Author SHA1 Message Date
Robert Vollmert 2b8ffc8e61 src: update for changed map type in aeson-2
This means that we're now using Data.Map.Strict instead of
Data.HashMap.Strict for JSON objects in general, and specifically
for claims maps and CSV rows.

This addresses certain hash flooding vulnerabilities, but may
have performance downsides.

Compare e.g. https://frasertweedale.github.io/blog-fp/posts/2021-10-12-aeson-hash-flooding-protection.html
2022-06-13 13:25:54 +02:00
Robert Vollmert 5e6987b1d8 src: consistently import HashMap as HM, Map as M
With both HashMap and Map imported as M in different modules,
linter rules prevented ever importing both modules in one place.
2022-06-13 13:16:50 +02:00
Laurence IslaandGitHub f9688c3553 Upgrade protolude to 0.3.1 2022-06-08 11:41:39 -05:00
Franz-Josef FärberandSteve Chavez 807364bd7b RPC POST for function w/single unnamed XML param 2022-06-03 11:52:57 -05:00
steve-chavez f7b173163c fix: ignore views on col/fk as target
Also refactor self relationship findRel logic
2022-05-16 12:22:24 -05:00
steve-chavez 2be63b36d6 refactor: self relationship
* Add test for self relationship in view
2022-05-09 21:31:19 -05:00
steve-chavez c60380b5fc perf: RelationshipsMap with foreign schema as key 2022-05-09 21:31:19 -05:00
steve-chavez d88b16e5ab fix: allow cast on types with underscores/numbers
e.g. select=oid_array::_int4
2022-05-09 13:02:33 -05:00
Franz-Josef FärberandSteve Chavez 5a3fbd7111 Allow returning XML from single-column queries 2022-05-04 10:27:28 -05:00
Laurence IslaandGitHub a2349d90c0 Reorganize error groups
* Grouped Hasql errors into one error and renamed it to InternalError
* Move all the General errors to ApiRequest errors
2022-05-02 21:16:31 -05:00
steve-chavez 50b275d3e2 refactor: move PUT checks to ApiRequest 2022-05-02 11:12:02 -05:00
steve-chavez e90391b7ac limited update/delete requires explicit order
* limited update/delete now works on views with explicit order
* no default order, enforce order presence
* apply row count to ensure limited mutations
* move requiring order to ApiRequest
2022-05-02 11:12:02 -05:00
Laurence IslaandGitHub 5ad8800773 feat: Allow limit=0 in query params to return an empty array 2022-04-28 17:14:29 -05:00
Franz-Josef FärberandSteve Chavez d2aa50be52 Allow returning XML from RPCs 2022-04-26 17:37:59 -05:00
steve-chavez 115dae7484 perf: Relationship list to a hash map 2022-04-25 21:41:51 -05:00
steve-chavez e4006da9bc perf: remove toSourceRel
toSourceRel is an expensive operation since it iterates over all the
relationships detected. We now use an alias on the CTE so embedding can
work without toSourceRel.
2022-04-25 21:41:51 -05:00
steve-chavez a47c59baca fix: restrict generated m2m relationships
Only add m2m relationship based on real junctions:
a table has FKs to two other tables and the FK columns
are part of the PK columns.

Also add some stats to big_schema.sql
2022-04-25 12:35:16 -05:00
steve-chavez 07d619981e perf: don't lookup Table pkCols unnecessarily
handleCreate only needs it on some headers, udpate and delete don't need
pkcols.
2022-04-22 19:04:16 -05:00
steve-chavez 778b8d074d refactor: merge Column list into Table
* Remove Table from Column
* Correct OpenAPI test where the schema was appearing for the enum unnecessarily
* Increase memory-tests consumption by 2M

Seems the Hasql decoder consumes more memory now that the allTables
results also have [Column] inside.
2022-04-22 19:04:16 -05:00
steve-chavez 902e4e8921 refactor: clean allColumns query
* Add OpenAPI test to ensure enum labels are present
* Remove unnecessary fk/pk columns cte
* Remove unnecessary step for parsing enum labels
2022-04-22 11:46:45 -05:00
steve-chavez d2719420f4 refactor: PKcols in table, ViewKeyDependency type
* Get PKcols inside tables - done with SQL for tables and
  with an additional step in Haskell for views.

  This fixes an fk column being considered as a pk column on views
  and corrects the test added on
  https://github.com/PostgREST/postgrest/pull/1875/files/1d549768580310e18aac4ffa6dbd01c5b77934a7#r853674126

* classify view key dependencies in SQL

* remove Column from Relationship

* Merge cols/fcols in Relationship and
  ensure allM2ORels and allViewsKeyDependencies fk columns
  are ordered - done by attnum in SQL

* Cardinality now contains relColumns instead of Relationship -
  this simplifies getJoinConditions.
2022-04-21 14:44:45 -05:00
steve-chavez f4e171aa2b refactor: ignore partitions in allM2ORels
Also clarify foreign key filter
2022-04-19 17:23:12 -05:00
steve-chavez 7f1507b9fb refactor: remove Table from Relationship
Just having the QualifiedIdentifier gets us closer to having
Relationship as a Table attribute since it avoids a cyclic dependency

* remove unnecessary findTable
* modify RootSpec test
2022-04-18 15:48:03 -05:00
steve-chavez cdcc175abf refactor: table list to table map 2022-04-18 15:48:03 -05:00
steve-chavez f995799e8f refactor: remove fkey from allPrimaryKeys query
The query had the unnecessary 'f' in `c.contype in ('p', 'u', 'f')`

Also comment and clarify the query by using explicit JOINs
2022-04-18 15:48:03 -05:00
steve-chavez 44e2b2fbb4 refactor: DRY allTables/accessibleTables query 2022-04-18 15:48:03 -05:00
Laurence IslaandGitHub 7589102c6c Fix misleading disambiguation error where relationship looks like valid syntax
* Add columns for the m2m relationship
2022-04-12 00:02:19 +02:00
Michael Kane JunckerandWolfgang Walther 007f49a8bc feat(query): add basic regular expression operator support - refs #2236 2022-04-11 08:33:16 +02:00
steve-chavez f3f9030c48 disallow limit/offset on views 2022-03-26 15:29:13 +01:00
steve-chavez 73dc2692b1 refactor: add findTable function 2022-03-26 15:29:13 +01:00
steve-chavez 87ffe39746 fix: ignore max-rows on POST/PATCH/PUT/DELETE 2022-03-26 15:29:13 +01:00
steve-chavez 331e88ea39 feat: add limited delete 2022-03-26 15:29:13 +01:00
steve-chavez f4becf99ad feat: add limited update 2022-03-26 15:29:13 +01:00
Ezequiel AlvarezandGitHub 799daa7556 fix: json/jsonb columns should not have type "string" in OpenAPI spec (#2203)
* Switch to no type for json/jsonb
2022-03-26 15:16:35 +01:00
Steve ChavezandGitHub a3c1d9977f Correct live/ready checks to consider special host values (#2182) 2022-03-11 15:16:45 +01:00
Laurence IslaandGitHub d6834e8bf8 Add 'PGRST' error code to differentiate from PostgreSQL errors 2022-03-04 15:57:38 -05:00
Laurence IslaandGitHub 99d0b805df 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
2022-02-28 13:23:55 -05:00
Laurence IslaandGitHub 69070f341a fix: Remove admin-server-port from in-db configuration 2022-02-23 17:26:01 -05:00
Wolfgang WaltherandWolfgang Walther 41a4396147 fix: Remove uncallable functions from schema cache and OpenAPI output
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-02-08 09:08:27 +01:00
Wolfgang WaltherandWolfgang Walther d9f7f6f509 refactor: Remove pg_catalog prefix from schema cache queries
Since the search path is voided at the beginning, there is no need to qualify those.

Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-02-08 09:08:27 +01:00
Wolfgang WaltherandWolfgang Walther bd62008e1c 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>
2022-02-05 17:48:22 +01:00
Wolfgang WaltherandWolfgang Walther ded8981368 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>
2022-02-05 17:48:22 +01:00
Wolfgang WaltherandWolfgang Walther 52d628f1ed fix: Return 204 No Content without Content-Type for RPCs returning VOID
Resolves #2001

BREAKING CHANGE: Previously, those RPCs would return "null" as a body with Content-Type: application/json.
2022-02-05 09:28:28 +01:00
Wolfgang WaltherandWolfgang Walther 9f1b5c0a81 feat: Support accessing array items and fields of composite types through json operators
This is supported in select=, in filters and in order=.

Resolves #1543

Resolves #2075

Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-02-03 08:13:38 +01:00
Wolfgang WaltherandWolfgang Walther 58f76f3d6d fix: Return 405 Method Not Allowed for unsupported verbs
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-01-28 19:26:33 +01:00
Wolfgang WaltherandWolfgang Walther 8980b09419 refactor: Sort error types
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-01-28 19:26:33 +01:00
Wolfgang WaltherandWolfgang Walther 927ff6f1e5 test: Split spec tests into multiple subfolders
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-01-28 19:26:33 +01:00
Wolfgang WaltherandWolfgang Walther 8a722e2cfe refactor: Replace elem with pattern matching to highlight missing code coverage
Signed-off-by: Wolfgang Walther <walther@technowledgy.de>
2022-01-28 19:26:33 +01:00
Wolfgang WaltherandWolfgang Walther fc7113e52e 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.
2022-01-28 16:45:36 +01:00
Steve ChavezandGitHub c7cadcd776 Fix embedded filter silently getting ignored (#2133)
When the embedded resource is not included in the request
2022-01-24 17:52:27 -05:00