Add missing features in unified test format #820
Conversation
| @@ -1452,6 +1535,29 @@ operation_iterate_until_document_or_error (test_t *test, | |||
| return ret; | |||
| } | |||
|
|
|||
| static bool | |||
| operation_close (test_t *test, | |||
There was a problem hiding this comment.
This operation is just for cursor types. Should the function be named operation_cursor_close instead?
Edit: I just realized that isn't typically done for other operation methods, so this is fine as-is.
There was a problem hiding this comment.
Yeah, this surprised me as well, as the operations don't seem to be tied to objects. @kevinAlbs want me to create a ticket for this?
There was a problem hiding this comment.
The name of the operation functions should be 1:1 with the operation names from UTF, e.g. fooBar => operation_foo_bar.
I think the operation names must be unique in UTF. operation_close seems correct.
There is some precedent. Operations that can act on multiple object types, like operation_create_change_stream, currently checks the operation type within the function.
There was a problem hiding this comment.
I think the operation names must be unique in UTF.
FWIW, this is not actually a requirement of the spec even if the names might appear unique in practice (if you're considering the variations of aggregate to be the same). I suppose we can punt on this until it actually becomes an issue.
Since the only tests that are using the loop operation and storeEventsAsEntities are the drivers-atlas-testing tests, we can defer implementation of those two features until the atlas testing project.
Note that implementing the serverless topology check will be done separately
Co-authored-by: Kevin Albertson <[email protected]>
alcaeus
force-pushed
the
unified-test-format-features
branch
from
e06407c to
823c905
Compare
CDRIVER-4066, CDRIVER-4060, CDRIVER-4064
While implementing the command redaction tests, I selectively implemented version 1.5 of the unified test format, ignoring 1.2, 1.3, and 1.4 as they weren't necessary. This PR fixes this.
For schema version 1.2, I've skipped implementing the
loopoperation and thestoreEventsAsEntitiesoption, as they are only required once we resume work on drivers atlas testing. We may decide to implement these earlier, but I didn't want to do so given the tight timeline on the other features.In version 1.3 (load balancer support), the new
load-balancedtopology mode is accepted, but it obviously never matches. This will be added in a separate pull request related to load balancer.