feat: add --require-unique-paramset-ids option skips pytest internal … #13757

gomri15 · 2025-09-26T18:27:12Z

Added flag to skip attempting to generate parametrization ids and just raise error on duplicate ids during collection

Closes #13737

Zac-HD

Should i documentation like there is for --collect-only?

Yes please - and then I think we'll be ready to merge 🙂

doc/en/reference/reference.rst

changelog/13737.feature.rst

src/_pytest/main.py

src/_pytest/python.py

testing/test_collection.py

gomri15 · 2025-09-27T16:45:20Z

Should i documentation like there is for --collect-only?

Yes please - and then I think we'll be ready to merge 🙂

Is this till needed or did with @nicoddemus say cover this?

src/_pytest/main.py

testing/test_collection.py

gomri15

Done all fixes

testing/test_collection.py

src/_pytest/python.py

nicoddemus · 2025-10-02T12:16:07Z

@Zac-HD just realized something: doesn't make more sense for this to be an ini option rather than a command-line flag? I understand this is something one would like to setup permanently for a test suite, rather than something to be passed on the command line on occasion.

Zac-HD · 2025-10-03T06:25:31Z

doesn't make more sense for this to be an ini option rather than a command-line flag?

Ideally both, I guess?

gomri15 · 2025-10-03T11:41:41Z

doesn't make more sense for this to be an ini option rather than a command-line flag?

Ideally both, I guess?

@nicoddemus, @Zac-HD
added support for the ini flag and tests for it.

Zac-HD

Minor edits to the docs, and then I'll give @nicoddemus a few days to comment, and then merge!

changelog/13737.feature.rst

doc/en/reference/reference.rst

bluetech · 2025-10-12T13:23:52Z

just realized something: doesn't make more sense for this to be an ini option rather than a command-line flag? I understand this is something one would like to setup permanently for a test suite, rather than something to be passed on the command line on occasion.

FWIW I agree with this, I don't see a reason for this to be a CLI flag, only ini. Having fewer flags is generally a good thing.

In the future we could also consider adding it to the planned strict setting. A while ago we deprecated and removed strict with the intention of re-adding it such that it enabled various "strict" options, like this one. Though I do reckon that if we add it it should be as an INI flag not CLI.

gomri15 · 2025-10-12T14:03:05Z

I understand wanting to have fewer flags to reduce complexity and setup pain.

Because we can get the same results from an ini and most likely, this parameter wills be set at the level of the whole execution and not just a few tests being run

Would you like me to remove the cli flag oart of this pr and leave the ini?

bluetech

Some comments from me. I think this can be a nice strictness flag.

changelog/13737.feature.rst

doc/en/reference/reference.rst

src/_pytest/python.py

changelog/13737.feature.rst

testing/test_collection.py

src/_pytest/python.py

bluetech · 2025-10-12T14:05:54Z

Thanks for the input bluetech, would you like me to remove the cli flag oart of this pr and leave the ini?

I'll let @Zac-HD decide, I don't feel strongly about it.

gomri15 · 2025-10-14T10:53:18Z

@bluetech I went over and fixed all your comments, thanks for the input.

Main changes made:

rename flag to - "require-unique-parameterization-ids" i agree it's much more readable and would be easier for pytest users to understand
Added fix suggestion to error message (it might be too explicit, but I was thinking it would be nice for the message to be a one stop shop for solving the problem)
Inlined testing until function which called create testfile
Improved wording of description for clarity.

bluetech

Thanks for the update! Please see my comments below.

testing/test_collection.py

src/_pytest/python.py

changelog/13737.feature.rst

testing/test_collection.py

bluetech · 2025-10-14T16:48:46Z

Thanks for the input bluetech, would you like me to remove the cli flag oart of this pr and leave the ini?

I'll let @Zac-HD decide, I don't feel strongly about it.

BTW, if we go with ini only, I think it would make sense to use the name strict_parametrization_ids or such, similar to strict_xfail, to indicate it's a strictness flag. And if in the future we add the strict ini it'd be clear that it's part of it.

gomri15 · 2025-10-17T13:10:17Z

@bluetech Fixed all the comments is there anything else you think needs doing?

bluetech · 2025-10-18T17:06:46Z

@gomri15 Instead of another round of pedantic comments, I pushed a commit with the changes I thought about. Some comments:

I made an "executive decision" to rename the param to strict_parametrization_ids and remove the CLI flag. Should tie in with --strict deprecation was botched - what to do about it #13823.
Tweaks the comments/messages a bit.
Simplified the test a bit.
Added handling of pytest.HIDDEN_PARAM
As a separate commit, enabled the flag in pytest's own test suite (always good to "dogfood") and fixed the couple of issues it found (which were actual useless duplicates!).

Please take a look at the changes and let me know what you think. If it's good I can merge.

gomri15 · 2025-10-19T12:30:16Z

src/_pytest/python.py

+                    id if id is not HIDDEN_PARAM else "<hidden>" for id in resolved_ids
+                )
+                duplicates = ", ".join(
+                    id if id is not HIDDEN_PARAM else "<hidden>"


not a blocker or a reason not to merge.
this is a pattern that does repeat itself a few times in the code
id if id is not HIDDEN_PARAM else "<hidden>"

I use https://en.wikipedia.org/wiki/Rule_of_three_(computer_programming) for this :)

It's a little repetition but close together and unlikely to get out of sync. So I prefer making the code more readable.

Haha i like it 2 is just under 3 haha

gomri15 · 2025-10-19T12:31:11Z

@bluetech Thank you very much for getting this a cross the line, really appreciate your input and I'll make a point of growing from it.

the changes look good to me.
I just added myself to Authors.

Fix pytest-dev#13737.

bluetech · 2025-10-19T13:41:35Z

Merged, thanks @gomri15 it was a great first PR. I will definitely use this option in my own projects.

gomri15 · 2025-10-19T13:43:56Z

Thank you @bluetech
I am looking forward to contributing more and taking the learning from this review forward.

Pleasure working with you

psf-chronographer bot added the bot:chronographer:provided (automation) changelog entry is part of PR label Sep 26, 2025

Zac-HD reviewed Sep 26, 2025

View reviewed changes

doc/en/reference/reference.rst Outdated Show resolved Hide resolved

nicoddemus requested changes Sep 27, 2025

View reviewed changes

nicoddemus mentioned this pull request Sep 27, 2025

Fix skip bug in which if exception group has all skips it reported error instead of skip #13741

Open

nicoddemus reviewed Sep 27, 2025

View reviewed changes

testing/test_collection.py Outdated Show resolved Hide resolved

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 2 times, most recently from 907cf3e to 8a770c9 Compare September 27, 2025 16:44

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 2 times, most recently from c0ce285 to a07ff8a Compare September 27, 2025 17:01

gomri15 requested review from Zac-HD and nicoddemus September 27, 2025 17:12

Zac-HD reviewed Sep 30, 2025

View reviewed changes

src/_pytest/main.py Outdated Show resolved Hide resolved

testing/test_collection.py Outdated Show resolved Hide resolved

gomri15 force-pushed the 13737-require-unique-paramset-ids branch from 35da7df to 4d6fa15 Compare October 1, 2025 07:09

gomri15 requested a review from Zac-HD October 1, 2025 07:10

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 3 times, most recently from 71152b2 to 9cebbab Compare October 1, 2025 19:45

gomri15 commented Oct 2, 2025

View reviewed changes

nicoddemus reviewed Oct 2, 2025

View reviewed changes

testing/test_collection.py Outdated Show resolved Hide resolved

nicoddemus reviewed Oct 2, 2025

View reviewed changes

src/_pytest/python.py Outdated Show resolved Hide resolved

gomri15 force-pushed the 13737-require-unique-paramset-ids branch from b1bae36 to b825eb4 Compare October 3, 2025 11:41

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 4 times, most recently from 03b31f8 to 358627e Compare October 3, 2025 19:28

gomri15 force-pushed the 13737-require-unique-paramset-ids branch from 358627e to f547106 Compare October 11, 2025 09:08

Zac-HD reviewed Oct 12, 2025

View reviewed changes

changelog/13737.feature.rst Outdated Show resolved Hide resolved

doc/en/reference/reference.rst Outdated Show resolved Hide resolved

doc/en/reference/reference.rst Outdated Show resolved Hide resolved

bluetech reviewed Oct 12, 2025

View reviewed changes

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 3 times, most recently from 57b5f00 to 9a1d479 Compare October 14, 2025 10:48

gomri15 requested review from Zac-HD and bluetech October 14, 2025 12:04

gomri15 force-pushed the 13737-require-unique-paramset-ids branch from 9a1d479 to 4782ea6 Compare October 14, 2025 12:05

bluetech reviewed Oct 14, 2025

View reviewed changes

gomri15 force-pushed the 13737-require-unique-paramset-ids branch 3 times, most recently from 9d7870a to 4eaf20e Compare October 14, 2025 20:53

gomri15 requested a review from bluetech October 15, 2025 18:53

gomri15 force-pushed the 13737-require-unique-paramset-ids branch from 4eaf20e to 767a11f Compare October 17, 2025 13:09

bluetech mentioned this pull request Oct 18, 2025

--strict deprecation was botched - what to do about it #13823

Open

gomri15 commented Oct 19, 2025

View reviewed changes

gomri15 and others added 2 commits October 19, 2025 16:24

python: add strict_parametrization_ids option

526288b

Fix pytest-dev#13737.

testing: enable strict_parametrization_ids and fix problems

5df54c6

bluetech force-pushed the 13737-require-unique-paramset-ids branch from cca6f1c to 5df54c6 Compare October 19, 2025 13:25

bluetech merged commit ebcd836 into pytest-dev:main Oct 19, 2025
33 checks passed

Uh oh!

feat: add --require-unique-paramset-ids option skips pytest internal … #13757

feat: add --require-unique-paramset-ids option skips pytest internal … #13757

Conversation

gomri15 commented Sep 26, 2025 • edited by Zac-HD Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Zac-HD left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

gomri15 commented Sep 27, 2025

Uh oh!

Uh oh!

Uh oh!

gomri15 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

nicoddemus commented Oct 2, 2025

Uh oh!

Zac-HD commented Oct 3, 2025

Uh oh!

gomri15 commented Oct 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Zac-HD left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bluetech commented Oct 12, 2025

Uh oh!

gomri15 commented Oct 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bluetech left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bluetech commented Oct 12, 2025

Uh oh!

gomri15 commented Oct 14, 2025

Uh oh!

bluetech left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bluetech commented Oct 14, 2025

Uh oh!

gomri15 commented Oct 17, 2025

Uh oh!

bluetech commented Oct 18, 2025

Uh oh!

gomri15 commented Sep 26, 2025 •

edited by Zac-HD

Loading

gomri15 commented Oct 3, 2025 •

edited

Loading

gomri15 commented Oct 12, 2025 •

edited

Loading