gh-128014: fix handling of a default empty string passed to tkinter.Wm.wm_iconbitmap

[Xiaokang2022]

• Issue: tkinter.Wm.wm_iconbitmap has no effect when passing an empty string to the parameter default #128014

[Xiaokang2022]

I think this might need a "skip news" label.

[tomasr8]

This is a user-visible change, right? If so, this will need a news entry.

[terryjreedy]

Currently, widget.wm_iconbitmap treats default='' the same as default=None. This seems like a bug.

A possible blurb: When passing default='' to tkinter's widget.wm_iconbitmap, treat '' as an icon path instead of as None.

@serhiy-storchaka This seems liike a trivial fix. Apply? Backport?

[serhiy-storchaka]

It is not the only site where all empty values are treated the same way as None. There was even a discussion about this somewhere. We should fix this issue all at once (if this is an issue) instead of one line per PR.

[serhiy-storchaka]

On other hand, this looks like a bug that blocks a useful use case. So we can fix only this bug in this PR. But we need a test for this.

[Xiaokang2022]

Sorry, I don't know how to test this.

[serhiy-storchaka]

No problem, I added a test.

[terryjreedy]

Macos-13 had 2 failures.

Seem like should be spurious, in test_multiprocessing, but failed on retest, along with test_tkinter.

2024-12-22T13:24:41.7713420Z FAIL: test_repr_rlock (test.test_multiprocessing_spawn.test_processes.WithProcessesTestLock.test_repr_rlock)
2024-12-22T13:24:41.7714380Z ----------------------------------------------------------------------
2024-12-22T13:24:41.7714900Z Traceback (most recent call last):
2024-12-22T13:24:41.7715760Z File "/Users/runner/work/cpython/cpython/Lib/test/_test_multiprocessing.py", line 1530, in test_repr_rlock
2024-12-22T13:24:41.7717320Z self.assertEqual('<RLock(SomeOtherThread, nonzero)>', repr(lock))
2024-12-22T13:24:41.7717990Z ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2024-12-22T13:24:41.7718650Z AssertionError: '<RLock(SomeOtherThread, nonzero)>' != '<RLock(None, 0)>'
2024-12-22T13:24:41.7719310Z - <RLock(SomeOtherThread, nonzero)>
2024-12-22T13:24:41.7719850Z + <RLock(None, 0)>

Real, in new test iconbitmap in test_tkinter.
2024-12-22T13:27:50.6347190Z FAIL: test_wm_iconbitmap (test.test_tkinter.test_misc.WmTest.test_wm_iconbitmap)
2024-12-22T13:27:50.6347930Z ----------------------------------------------------------------------
2024-12-22T13:27:50.6348460Z Traceback (most recent call last):
2024-12-22T13:27:50.6349360Z File "/Users/runner/work/cpython/cpython/Lib/test/test_tkinter/test_misc.py", line 561, in test_wm_iconbitmap
2024-12-22T13:27:50.6350360Z self.assertEqual(t.wm_iconbitmap(), 'hourglass')
2024-12-22T13:27:50.6350860Z ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2024-12-22T13:27:50.6351350Z AssertionError: '' != 'hourglass'
2024-12-22T13:27:50.6351730Z + hourglass

[Xiaokang2022]

Even before this PR change, tkinter would have produced such results on macOS. This doesn't seem to be related to this PR, and a separate issue may need to be opened to discuss it.

[serhiy-storchaka]

Oh, there is a bug in Tk (https://core.tcl-lang.org/tk/tktview/13ac26b35dc55f7c37f70b39d59d7ef3e63017c8).

[serhiy-storchaka]

It is better to wait until this bug was fixed in Tk, so we could skip the check conditionally.

[miss-islington-app]

Thanks @Xiaokang2022 for the PR, and @serhiy-storchaka for merging it 🌮🎉.. I'm working now to backport this PR to: 3.12, 3.13.
🐍🍒⛏🤖

[miss-islington-app]

Sorry, @Xiaokang2022 and @serhiy-storchaka, I could not cleanly backport this to 3.12 due to a conflict.
Please backport using cherry_picker on command line.

cherry_picker 58e9f95c4aa970db32a94b9152b51ede22f823bd 3.12

[bedevere-app]

GH-128418 is a backport of this pull request to the 3.13 branch.

[bedevere-app]

GH-128420 is a backport of this pull request to the 3.12 branch.

[vstinner]

The new test_wm_iconbitmap() test added by this change fails randomly on macos-26-intel with a timeout: see issue gh-146531.