* Batch transaction boilerplate
* Refactoring transaction boilerplate
* Independent sessions test
* Transactions - partial
* Missing only one test
* All tests passing for mongo db
* Tests on Travis
* Transactions on postgres
* Fix travis to restart mongodb
* Remove mongodb service and keep only mongodb runner
* MongoDB service back
* Initialize replicaset
* Remove mongodb runner again
* Again only with mongodb-runner and removing cache
* Trying with pretest and posttest
* WiredTiger
* Pretest and posttest again
* Removing inexistent scripts
* wiredTiger
* One more attempt
* Trying another way to run mongodb-runner
* Fixing tests
* Include batch transaction on direct access
* Add tests to direct access
* Changed count to be approximate. Should help with postgres slowness
* refactored last commit to only fall back to estimate if no complex query
* handlign variables correctly
* Trying again because it was casting to lowercase table names which doesnt work for us/
* syntax error
* Adding quotations to pg query
* hopefully final pg fix
* Postgres will now use an approximate count unless there is a more complex query specified
* handling edge case
* Fix for count being very slow on large Parse Classes' collections in Postgres. Replicating fix for Mongo in issue 5264
* Fixed silly spelling error resulting from copying over notes
* Lint fixes
* limiting results to 1 on approximation
* suppress test that we can no longer run for postgres
* removed tests from Postgres that no longer apply
* made changes requested by dplewis
* fixed count errors
* updated package.json
* removed test exclude for pg
* removed object types from method
* test disabled for postgres
* returned type
* add estimate count test
* fix mongo test
* Auth module refactoring in order to be reusable
* Ensure cache controller is properly forwarded from helpers
* Nits
* Adds support for static validation
* Adds support for CLP in Live query (no support for roles yet)
* Adds e2e test to validate liveQuery hooks is properly called
* Adds tests over LiveQueryController to ensure data is correctly transmitted
* nits
* Fixes for flow types
* Removes usage of Parse.Promise
* Use the Auth module for authentication and caches
* Cleaner implementation of getting auth
* Adds authCache that stores auth promises
* Proper testing of the caching
* nits
* Various improvements in test name / de-duplications
* Reverts to class by class deletion, introduced fast mode that just delete data for mongo
- Speeds up are incredible Executed 1695 of 1713 specs INCOMPLETE (18 PENDING) in 4 mins 19 secs.
* Adds documentation about the deleteEverything
* Fixes an issue that would let the beforeDelete be called when user has no access to the object
* Ensure we properly lock user
- Improves find method so we can attempt to read for a write poking the right ACL instead of using masterKey
- This ensure we do not run beforeDelete/beforeFind/beforeSave in the wrong scenarios
* nits
* Caps insufficient
* add failing test to demonstrate that you can't sort on a
field in an object.
* Only validate the base of the field name.
* fix test name
* Only test sort for mongo.
* pg order by nested object
* level 2 test
* Factor out operation to get a field's base name. Add comment.
* tweak comment wording so it wont make my grammar teacher angry.
* adding nor to specialQuerykeys
* adding nor suport
* adding test
* CRLF
* adding postgres NOR
* adding nor validation
* adding NOR test
* adding test amd fixing NOR cases
* revert the nor logic
* Introduces flow types for storage
* Better typing of QueryOptions
* Adds flow types to SchemaCOntroller,
- runs flow on pre tests
- fixes flow
* Adds ClassLevelPermissions type
* Moves Controller types into a single file
* Changes import styles
* Changes import styles
* fixing method setIndexesWithSchemaFormat (#4454)
Fixing invalid database code in method `setIndexesWithSchemaFormat`:
* It must be a transaction, not a task, as it executes multiple database changes
* It should contain the initial queries inside the transaction, providing the context, not outside it;
* Replaced the code with the ES6 Generator notation
* Removing the use of batch, as the value of the result promise is irrelevant, only success/failure that matters
* nits
* Fixes tests, improves flow typing
* Add Indexes to Schema API
* error handling
* ci errors
* postgres support
* full text compound indexes
* pg clean up
* get indexes on startup
* test compound index on startup
* add default _id to index, full Text index on startup
* lint
* fix test
* Support for Aggregate Queries
* improve pg and coverage
* Mongo 3.4 aggregates and tests
* replace _id with objectId
* improve tests for objectId
* project with group query
* typo
* Refactors pushStatusHandler to use HTTP interface so we can bind CloudCode hooks
* Handle correctly nested dot atomic operations
* Better handling of restricted class names, add support for afterSave _PushStatus
* Adds simple testing for afterSave(PushStatus)
* Reverts jobStatusHandler
* Addresses fixes
* adds delays to all methods
* Makes sure we don't override roles
* Reduces the query size whith pointer permissions
- Does not return as $and if not needed
- Returns just the query with the additional constraint
* Do not use $in if include is just of length 1
* Full Text Support
* invalid input test
* Support for sort
* index exist test
* clean up
* better error messaging
* postgres support
* error instructions for $diacritic and $case sensitivity
* nit
* nit
* nit
* separate test for full text
Mongo has a hard limit on 1 $near operation per query. Restructuring to
avoid SERVER-13732 should not invalidate a query by creating multiple
$near operations.
Additionally, queries with multiple $ors are now recursively handled,
whereas before, ops at the top level would only have been pushed one
level deeper.
https://github.com/parse-community/parse-server/issues/3767
* Adds index on _Role name property
In order to avoid having different _Role objects with the same name, adding an index on the name property of _Role is necessary.
Fixes#3579
* Uses throw instead of Promise.reject when enforcing unique indexes
* Fixes wrong sorting of results in schemas tests
In a prior commit, improvements were made to the addition of `_rperm`
in the case of `$or` queries, to avoid MongoDB bug SERVER-13732.
As the vast majority of $or queries previously hit this bug due to the
presence of `_rperm` on most Parse queries), the present solution
avoids the bug and improves query performance in most cases.
However, it's still possible for clients to supply their own queries
which hit that bug, such as those with `_created_at` or `_updated_at`
filters, or their own properties from their data model.
This commit makes the logic currently present for `_rperm` available
to all top level filters that exist alongside an $or query, meaning
SERVER-13732 should be avoided in all cases where keys at the top and
inner levels do not have name clashes.
- https://github.com/ParsePlatform/parse-server/pull/3476
- https://jira.mongodb.org/browse/SERVER-13732
MongoDB has an unfixed bug in all supported versions 2.6-3.4 which
results in suboptimal index usage for `$or` queries when the query has
implicit `$and`s at the query root.
When adding `_rperm` to `$or` queries, Parse accidentally creates
queries which hit this bug.
The fix in this commit applies the suggested workaround of putting the
`_rperm` property inside all `$or` subdocuments, moving it from the top
level and leaving `$or` as the only top-level operator.
MongoDB Bug Link: https://jira.mongodb.org/browse/SERVER-13732
