Opened 9 years ago
Last modified 3 months ago
#25889 assigned Cleanup/optimization
Organize tests in tests/queries
Reported by: | Claude Paroz | Owned by: | Wassef Ben Ahmed |
---|---|---|---|
Component: | Testing framework | Version: | dev |
Severity: | Normal | Keywords: | |
Cc: | Anssi Kääriäinen, Ülgen Sarıkavak | Triage Stage: | Accepted |
Has patch: | yes | Needs documentation: | no |
Needs tests: | no | Patch needs improvement: | yes |
Easy pickings: | no | UI/UX: | no |
Description
There are a bunch of test_ticket_<number>
tests in tests/queries/tests.py
or classes named Ticket<number>Tests
, sometimes without any docstring at all. I think that this is very poor naming, and make it difficult to browse the file researching for specific tests or test domains.
I'm sure other test apps also suffer from this, but I was struck by this specific file and this could be a starting point for a larger cleanup.
Change History (14)
follow-up: 5 comment:1 by , 9 years ago
Cc: | added |
---|---|
Summary: | Rename tests in queries tests → Organize tests in queries tests |
Triage Stage: | Unreviewed → Accepted |
comment:2 by , 9 years ago
Summary: | Organize tests in queries tests → Organize tests in tests/queries |
---|
comment:3 by , 8 years ago
Owner: | changed from | to
---|---|
Status: | new → assigned |
comment:4 by , 8 years ago
After cleaning most of the method names and adding some context in comments. There are some methods that have multiple tests. Eg. test_ticketno_1 then we have test_ticketno_2. What should be done with such tests. We could group them into a single class maybe and provide some additional information in the comments? Or we could have like test_context_of_test_1?
comment:5 by , 8 years ago
Replying to Tim Graham:
If we can break up the single
tests.py
file into multiple tests files (perhaps based on the models they use) andmodels.py
into a package similar totests/auth_tests/models
, I think that would be a useful task. If we can do some renaming along the way, that seems okay but it would be easiest to review in a separate commit from one that moves things around.
I think this sort of thing won't be easy. As most of the models aren't reused. If the patch required something there was a high chance that the contributor made the model similar to that given in the ticket. According to me the maximum we can do is to rename the tests such that they provide the context of what the test is for without loosing the information provided by the original author of the test.
comment:7 by , 4 years ago
Owner: | removed |
---|---|
Status: | assigned → new |
comment:8 by , 8 months ago
Cc: | added |
---|
comment:9 by , 4 months ago
Needs documentation: | set |
---|
comment:10 by , 4 months ago
Needs documentation: | unset |
---|---|
Owner: | set to |
Status: | new → assigned |
comment:11 by , 4 months ago
Patch needs improvement: | unset |
---|
comment:12 by , 4 months ago
Patch needs improvement: | set |
---|
comment:13 by , 4 months ago
Patch needs improvement: | unset |
---|
comment:14 by , 3 months ago
Patch needs improvement: | set |
---|
I think the issue is that for queries tests, they are not easily categorized or described. Many of the tickets are "here is query that regressed." If we can break up the single
tests.py
file into multiple tests files (perhaps based on the models they use) andmodels.py
into a package similar totests/auth_tests/models
, I think that would be a useful task. If we can do some renaming along the way, that seems okay but it would be easiest to review in a separate commit from one that moves things around.