[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

STINNER Victor added the comment: Failure on the Ubuntu job run on my PR 24353: https://github.com/python/cpython/pull/24353/checks?check_run_id=1776422914 == FAIL: test_identify (tkinter.test.test_ttk.test_widgets.ComboboxTe

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Ronald Oussoren added the comment: I might be able to debug during the weekend, but my main focus will be on back porting the changes in #41100 to the 3.8 branch (or at least parts of it). -- ___ Python tracker

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Serhiy Storchaka added the comment: All these failures point not on some details of appearance, but that the widget is not rendered at all. -- ___ Python tracker ___

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Ned Deily added the comment: OK, thanks for the clarification, Terry. Considering that our tkinter tests make assumptions about the underlying Tk widgets behavior, assumptions not necessarily shared by the Tk developers :), these kind of version-dependent failures are a risk of and shortcomi

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Terry J. Reedy added the comment: The test may be correctly detecting a change in widget behavior. This is less surprising to me than some keypresses acting like two keypresses a millesecond apart. -- ___ Python tracker

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Ned Deily added the comment: Terry: > It might be the tk/tkinter build, not the tests, that is at fault. Can you elaborate? -- ___ Python tracker ___

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Terry J. Reedy added the comment: It might be the tk/tkinter build, not the tests, that is at fault. That is my current conclusion for #42508 (regarding IDLE as a test). -- nosy: +terry.reedy ___ Python tracker

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Serhiy Storchaka added the comment: Can I get a remote access to machine with macOS? -- ___ Python tracker ___ ___ Python-bugs-list

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Serhiy Storchaka added the comment: I have no ideas. Try to add more calls of update() in these tests. -- ___ Python tracker ___ __

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Ned Deily added the comment: I tested on 3.9 with the backport of GH-23474 merged. The six failures are still there plus there is now an additional failure: == FAIL: test_horizontal_range (tkinter.test.test_ttk.test_extensio

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Serhiy Storchaka added the comment: Try to test with 6cc2c419f6cf5ed336609ba01055e77d7c553e6d (PR-23474). All of failing tests were modified to make them passing on Windows, so I expect that this can help on macOS too (in most cases). If it does not help I will propose other changes for test

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

Change by Ned Deily : -- nosy: +serhiy.storchaka ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail

[issue42507] test_ttk_guionly test failures on macOS with Tcl/Tk 8.6.10

New submission from Ned Deily : As of 3.9.1rc1 (and coming soon to 3.10.x and 3.8.x), we now provide a new additional macOS universal2 installer variant that, for various reasons, provides a built-in version of Tcl/Tk 8.6.10, instead of 8.6.8 as has been provided with the 10.9 installer varia