Can't set the value of a DateTimeField with auto_now or auto_now_add #23

[rodbv]

Describe the change
A clear and concise description of what the change is.

PR Checklist

• Change is covered with tests
• CHANGELOG.md is updated

[rodbv]

I've faced an issue similar to this and gave it a go. One issue with this implementation is that it only works if either _refresh_after_create is set to True, or the user calls refresh_from_db() directly like I did on the test.

I could call instance.refresh_from_db() here but this would override _refresh_after_create

[amureki]

@rodbv oh, many thanks for this attempt to tackle an annoying issue! 👍 This is promising!
I would need to dig more into the implementation, but meanwhile, may I ask you to look into the failing tests? I think, there are edge cases here that are causing errors.

Best,
Rust

[rodbv]

@amureki thanks! I've fixed the tests and the linter error. When running the tests workflow on my repo there's an error on the ruff linter but it seems to be a setup issue.

Edit: Some tests are still failing with timezone issues, I'll get to them

[rodbv]

I've dealt with my test failing depending on the use_tz setting. Ruff linter is still failing, is there anything I can do?

[amureki]

I've dealt with my test failing depending on the use_tz setting. Ruff linter is still failing, is there anything I can do?

Thanks for fixing the tests! I'll take a look into tz stuff before the end of this week. And also will do a proper check on your code.
Ruff part should be resolved if you update your branch on the main.

Sorry for the complications on the contribution way.

[amureki]

@rodbv I finally got some time and figured the issue. This was happening due to django.conf.settings tampering, which is usually not a good practice (but I know, we still have it in a couple of tests).
I updated your test to use a fixture provided by pytest-django, which solved the issue.

I also went ahead and fixed the last ruff problem, so it is good to go now, I am going to merge it and release now!

[amureki]

I released it as 1.17.0

[rodbv]

@rodbv I finally got some time and figured the issue. This was happening due to django.conf.settings tampering, which is usually not a good practice (but I know, we still have it in a couple of tests). I updated your test to use a fixture provided by pytest-django, which solved the issue.

I also went ahead and fixed the last ruff problem, so it is good to go now, I am going to merge it and release now!

Awesome! Thanks a lot for the support and for being so welcoming, I'm looking forward to contribute more, I'll have another look at the list of open issues :)

[amureki]

@rodbv well, thank you for the contribution, patience and kind words, much appreciated! ✨