@alejsdev could you please take a look? This is relatively trivial and contains a test as requested by@svlandeg.

Hello! This is ready for review.

@tamird there's no need to tag individual people nor to keep posting comments.

Your aggressive and demeaning attitude and comments towards others here are not well received and are not appreciated.

This is open source, MIT licensed, free. The result of any work from others here is received by you and/or your employer for free. Neither you nor your employer (it seems Google, from your GitHub) are paying anything to receive this open source project or our work for free. You actually have no particular right to demand reviews, merges, timelines, etc.

I request that you abstain from adding unrequested comments, in particular tagging others or with the only intention of pressing us to give a higher priority to reviewing your PRs, handling your needs, or reporting a status on the process of handling them. In particular, I request you to abstain from making any comment that could be in any way interpreted as demeaning towards others.

@tiangolo your points are well taken. The last comment here was 5 weeks ago.

Looks like it's not so easy now as we run tests on Python 3.12 and Pydantic V1 is not supported there..

Looks like it's not so easy now as we run tests on Python 3.12 and Pydantic V1 is not supported there..

Just needed a little copy & paste.

@tamird, thanks for resolving the merge conflict!

If I recall correctly, previously we switched to using existingErrorDict anddisplay_errors (import them frompydantic.v1) for simplicity. But now it turned out we can't just use them and have to duplicate that code.
Would it simplify the code if we switch back to just using our own implementations?

@tamird, thanks for resolving the merge conflict!

If I recall correctly, previously we switched to using existingErrorDict anddisplay_errors (import them frompydantic.v1) for simplicity. But now it turned out we can't just use them and have to duplicate that code. Would it simplify the code if we switch back to just using our own implementations?

Sure, it would probably be simpler. However this PR has now been open for over a year, and there's no indication that there's a desire to merge it. I hope you understand that I'm reluctant to invest further effort at this time without some indication that it will be accepted (or rejected).

This pull request has a merge conflict that needs to be resolved.

This pull request has a merge conflict that needs to be resolved.