Fixed #33037 -- Fixed Trunc() with offset timezones on MySQL, SQLite, Oracle.

[adzshaf]

Fixed ticket: 33037

Changes:

• Fixing split_tzname_delta by adding :00 to offset if the offset does not have minute format.
  Example, offset that will be passed into this function is 10, but MySQL and Oracle cannot parse it since it didn't follow correct format, e.g: 10:00.
  This fix will also fix bug on SQLite. Previously, the tuple unpacking returned error since the offset will be "10" and its split result become ["10"].
• Add fallback during parse datetime on SQLite. tzname become empty from split_tzname_delta result, so it would be error. Hence, use conn_tzname for fallback value.

I added test which is nearly identical of this part of test, with its difference on which timezone that is used. I think we can extract test functions from these parts.

[github-actions]

Hello! Thank you for your contribution 💪

As it's your first contribution be sure to check out the patch review checklist.

If you're fixing a ticket from Trac make sure to set the "Has patch" flag and include a link to this PR in the ticket!

If you have any design or process questions then you can ask in the Django forum.

Welcome aboard ⛵️!

[LilyFoote]

Awesome work!

I've included some suggestions that I think would improve the tests a bit.

[LilyFoote]

buildbot, test on oracle.

[felixxm]

@adzshaf Thanks 👍 Please take a look at @LilyFoote's test suggestions.

[LilyFoote]

Hey @adzshaf! I think you misunderstood what I was suggesting with subTest - I meant that you could loop over the different truncation types (year, quarter, etc) and call the assert helpers in that loop.

Something like:

     truncations = ("year", "quarter", "month", "week", "day", "hour", "minute", "second")
     for truncation in truncations:
         with subTest(truncation=truncation):
             assert_datetime_to_time_kind(truncation)
             assert_datetime_kind(truncation)

[adzshaf]

@LilyFoote my bad for not confirming it. OK, I understand, but should we also do the same for test_trunc_func_with_timezone? Or we just left as it is, and do it only on my test?

[LilyFoote]

Let's leave the other test as it was for now. It might be worth updating it too in a separate PR/commit later, but lets get the main change ready first.

[LilyFoote]

@LilyFoote my bad for not confirming it.

It's OK - I realise I wasn't at all clear with what I meant!

[felixxm]

Let's leave the other test as it was for now. It might be worth updating it too in a separate PR/commit later, but lets get the main change ready first.

Yes, we don't mix cleanups with bugfixes.

[felixxm]

@adzshaf Thanks 👍 Welcome aboard ⛵

I moved cleanups to a separate commit, made small edits, and used subTest() as Lily suggested.

[adzshaf]

@felixxm thanks a lot! Sorry for the inconvenience, haven't got time to fix it yet 🙇