summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Fix [247d80ddd8]: Chinese translation - tk msgjan.nijtmans2022-02-071-0/+92
|
* Add support for windows-armjan.nijtmans2022-02-073-7/+36
|
* Fix [b54a9193f6]: broken HTTPS URL in ttk::style manual pagejan.nijtmans2022-02-071-1/+1
|
* Unneeded TEXT() macro in tkMain.c. Comment formattingjan.nijtmans2022-02-071-13/+12
|
* Fix [3fefb364e2]: crash in ttk::combobox current with empty combobox.fvogel2022-02-032-1/+16
|\
| * Tune error message slightly again.fvogel2022-02-032-2/+2
| |
| * Change slightly the error message to match what 8.7 returns ([3fefb364e2] is ↵fvogel2022-02-032-2/+2
| | | | | | | | not present in trunk, the problem is only with 8.6).
| * Fix [3fefb364e2]: crash in ttk::combobox current. Test combobox-2.8 now passes.fvogel2022-02-011-1/+11
| |
| * Add (currently) crashing test combobox-2.8fvogel2022-02-011-0/+5
|/
* Fix [1501749fff] - Crash on embedded window deletion bound to <Map> events.fvogel2022-01-313-13/+43
|\
| * Backout temporary commits helping the debug, specifically [33363e00] and ↵fvogel2022-01-312-2/+2
| | | | | | | | [ec5c33bb].
| * Remove debug code inadvertently left in.fvogel2022-01-301-5/+0
| |
| * Re-introduce the fix [83d63659], which was correct after more analysis. One ↵fvogel2022-01-301-13/+15
| | | | | | | | cannot make any use of the 'client' pointer after Tk_MaintainGeometry/Tk_MapWindow have been called because they may run event handlers that in turn may destroy the embedded window. In the process the embedded window pointer will be freed and 'client' will point to already freed memory. This analysis was supported by valgrind wich showed the problem with textWind-18.1.
| * Fix memory problem revealed by valgrind with textWind-18.2. At this point ↵fvogel2022-01-301-0/+7
| | | | | | | | dInfoPtr is already freed when running textWind-18.2.
| * Ditto for macOS with clang.fvogel2022-01-241-1/+1
| |
| * Temporarily set verbose mode for classic tests on Windows with gcc, to see ↵fvogel2022-01-241-1/+1
| | | | | | | | where exactly it crashes in textWind.test.
| * Fix the first part of [1501749fff]. Now textWind-18.1 passes.fvogel2022-01-232-6/+9
| |
| * Don't use return in DisplayText but 'goto end' so that Tcl_Preserve / ↵fvogel2022-01-231-1/+1
| | | | | | | | Tcl_Release calls keep being balanced.
| * More accurate comment. Indeed long ago there used to be ↵fvogel2022-01-231-1/+1
| | | | | | | | Tcl_Preserve/Tcl_Release calls around the call to TkTextPickCurrent but they have been replaced by refcount management when TIP #155 was implemented, see [e58248ce5f8b5af2].
| * Fix the second part of [1501749fff]. Now textWind-18.2 passes.fvogel2022-01-232-1/+8
| |
| * Add some time in textWind-18.1 to make sure it crashes before the test is ↵fvogel2022-01-231-0/+1
| | | | | | | | over. The crash can be seen with TESTFLAGS='-verbose bepst': the log stops with '---- textWind-18.1 start' (i.e. the log shows no end of the test).
| * Add (currently crashing) test textWind-18.2 dealing with the issue ↵fvogel2022-01-231-0/+15
| | | | | | | | demonstrated by the second script in comment dated 2022-01-08 23:30:30 in [1501749fff]. This test is constrained with 'knownBug'.
| * merge core-8-6-branchfvogel2022-01-238-126/+210
| |\
* | | Slightly better description for TK_OPTION_RELIEF (regarding TK_RELIEF_NULL)jan.nijtmans2022-01-261-2/+2
| | |
* | | Add some spare stub entriesjan.nijtmans2022-01-253-5/+45
| |/ |/|
* | Cherrypick [baf43e83] from trunk.fvogel2022-01-171-0/+0
| |
* | Merge 8.5jan.nijtmans2022-01-171-1/+1
|\ \
| * | Add '--disable-xft' on MacOSX Github ACTIONS builds. Thanks, François, for ↵jan.nijtmans2022-01-171-1/+1
| | | | | | | | | | | | the suggestion!
* | | Fix [b1d115fa60]: No delivery of <Enter> event upon destruction of toplevel. ↵fvogel2022-01-164-21/+74
|\ \ \ | | | | | | | | | | | | This fix is entirely in tkInt.h (and is tested in event.test). This fix revealed a latent crash in the test suite (focus.test) with macOS, fix that also in tkFocus.c (see details in the discussion in ticket [b1d115fa60] and in successive commits in PR#9 at GitHub: https://github.com/tcltk/tk/pull/9). Finally, remove dead code from filebox.test (we were there when analyzing the previously mentioned crash in focus.test).
| * \ \ merge 8.6fvogel2022-01-017-75/+105
| |\ \ \
| * | | | Fix crash in focus.test on macOS at Github (only). See problem analysis in ↵fvogel2021-12-311-0/+12
| | | | | | | | | | | | | | | | | | | | PR#9 at GitHub: [https://github.com/tcltk/tk/pull/9]
| * | | | merge 8.6fvogel2021-12-2024-248/+323
| |\ \ \ \
| * | | | | Remove useless statement in filebox.testfvogel2021-12-191-1/+0
| | | | | |
| * | | | | Remove code doing nothing in filebox.testfvogel2021-12-191-3/+0
| | | | | |
| * | | | | Remove useless or repeated code in filebox.testfvogel2021-12-191-15/+0
| | | | | |
| * | | | | Remove superfluous destroy statement.fvogel2021-12-131-2/+1
| | | | | |
| * | | | | Move cleanup statement in the correct (event-9.1) test it belongs to.fvogel2021-12-131-1/+1
| | | | | |
| * | | | | Revert previous commit since it does not help with the crash in focus.test.fvogel2021-12-131-8/+2
| | | | | |
| * | | | | Tests event-9.1 and -9.2: use controlPointerWarpTiming after generation of ↵fvogel2021-12-121-2/+8
| | | | | | | | | | | | | | | | | | | | | | | | the warp events. Also force focus on the considered window before warping, and save and restore focus at test start/end. This could prevent the crash in focus.test seen at GitHub Actions for macOS.
| * | | | | Avoid white space only changes.fvogel2021-12-121-1/+1
| | | | | |
| * | | | | merge 8.6fvogel2021-12-113-10/+46
| |\ \ \ \ \
| * \ \ \ \ \ merge 8.6fvogel2021-12-088-80/+179
| |\ \ \ \ \ \
| * \ \ \ \ \ \ merge 8.6fvogel2021-12-0422-198/+619
| |\ \ \ \ \ \ \
| * \ \ \ \ \ \ \ merge 8.6fvogel2021-11-2912-209/+323
| |\ \ \ \ \ \ \ \
| * | | | | | | | | Put the Tk_TopWinHierarchy check inside TkGetContainer.fvogel2021-11-172-7/+9
| | | | | | | | | |
| * | | | | | | | | Adequate hygiene in the added tests prevents several failures (false ↵fvogel2021-11-171-1/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | positives) when running the complete test suite.
| * | | | | | | | | We're rather testing whether the window is at the top of the hierarchy ↵fvogel2021-11-171-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | rather than if it's a toplevel. See the difference in tk.h
| * | | | | | | | | Fix [b1d115fa60]: No delivery of <Enter> event upon destruction of toplevel. ↵fvogel2021-11-171-1/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix proposed by Erik Leunissen. Test event-9.1 now passes.
| * | | | | | | | | Add tests event-9.1 and -9.2. Test event-9.1 enters '.' when destroying a ↵fvogel2021-11-171-0/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | toplevel covering it. This test currently fails due to the issue reported in [b1d115fa60]. Test event-9.2 enters '.top1', which is a child toplevel of '.', when destroying a toplevel covering it. This test passes without changes.
| * | | | | | | | | Merge 8.6bug-b1d115fa60fvogel2021-11-16277-2252/+3418
| |\ \ \ \ \ \ \ \ \