summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Split pack-18.1 into two tests and remove the tempNotWin constraint. ↵fvogel2022-11-011-6/+33
| | | | pack-18.1.1 checks that, on Linux or macOS, width/height changes are taken into account while the window is unmapped ; pack-18.1.2 checks that, on Windows, width/height changes are taken into account on window remapping. This way we're at least checking correct sizes on Windows when the window gets remapped, while we weren't testing anything on Windows previously.
* place-8.1, place-8.2: Remove platform-specific updating in these two tests.fvogel2022-11-011-16/+6
|
* pack-18.1, pack-18.2: Remove platform-specific updating in these two tests. ↵fvogel2022-11-011-31/+16
| | | | Remove duplicate code. Replace very old constraint tempNotPc by tempNotWin (which is the modern name for the same constraint).
* Fix mistake in previous commit when defining constraint haveCourier37Font.fvogel2022-10-311-1/+1
|
* font-38.11, font-44.2, unixFont-2.11: Eradicate all remaining ↵fvogel2022-10-314-7/+20
| | | | failsOnUbuntuNoXft constraints, replacing them by more specific constraints.
* Move constraint haveTimes14Bold in constraints.tclfvogel2022-10-312-4/+6
|
* Rename constraint hasTimes14Bold --> haveTimes14Bold.fvogel2022-10-311-2/+2
|
* textTag.test: Replace constraint failsOnUbuntuNoXft by a more specific ↵fvogel2022-10-311-8/+17
| | | | constraint bigFontTwiceLargerThanTextFont checking exactly what is needed. Note: Without the constraint three tests (-16.2, -16.6 and -16.7) fail when run at Github Actions CI with XQuartz because the font actually selected when requesting bigFont as {Helvetica 24} is 'times 9', which has actual size of 9, for an unknown reason.
* textTag.test: Remove useless haveCourier12 constraint, I can't see how this ↵fvogel2022-10-311-442/+155
| | | | constraint can evaluate to false anyway.
* fontchooser-4.4: Replace constraint failsOnUbuntuNoXft by a more specific ↵fvogel2022-10-301-3/+2
| | | | constraint hasTimes14Bold checking exactly what is needed. Note: Without the constraint this test returns '9' instead of '14 bold' when run at Github Actions CI with XQuartz because the font actually selected in the fontchooser when requesting -font {times 14 bold} is 'times 9', for an unknown reason.
* scrollbar-6.35 and scrollbar-6.44: Remove constraint failsOnXQuarz by ↵fvogel2022-10-301-15/+18
| | | | adapting the tests to XQuartz smaller scrollbar arrows without changing what is tested.
* scrollbar-6.12.* and scrollbar-6.29.*: Remove constraint failsOnXQuarz by ↵fvogel2022-10-301-6/+10
| | | | adapting the tests to XQuartz smaller scrollbar arrows without changing what is tested.
* scrollbar-3.36 and scrollbar-3.38: Remove constraint failsOnXQuarz by making ↵fvogel2022-10-301-4/+6
| | | | the tests calculate its expected result.
* spinbox-8.18: Remove constraint failsOnUbuntuNoXft by making the test adapt ↵fvogel2022-10-292-6/+5
| | | | its expected result to the font size.
* entry-8.18: Remove constraint failsOnUbuntuNoXft by making the test adapt ↵fvogel2022-10-291-5/+2
| | | | its expected result to the font size.
* Remove a word in comment, making it more general.fvogel2022-10-291-1/+1
|
* entry-5.8: Remove superfluous constraint failsOnXQuarz, it tests OK without ↵fvogel2022-10-291-2/+1
| | | | it (macOS Ventura, and at the CI runner) with an X11 build with XQuartz.
* clrpick-2.3: And now constraint failsOnXQuarz can be removed. This ↵fvogel2022-10-291-3/+1
| | | | constraint should not have been added, the failure was revealing a real issue (dealt with in the two previous commits).
* clrpick-2.3: Fix test failure (this test only runs on unix notAqua). The ↵fvogel2022-10-291-1/+0
| | | | line deleted in this commit was added in [a093deecc7] but it was superfluous because the WM_DELETE_WINDOW thing already did the job. The problem with binding to <Destroy> was that when closing the dialog with the 'Ok' button, proc OkCmd fires and correctly sets Priv(selectColor) so that this color is remembered when the dialog is reopened, but then proc CancelCmd fired as well due to the binding to <Destroy> and was resetting Priv(selectColor) to ''. When reopening the dialog the previously selected color was then forgotten and clrpick-2.3 (which purpose is to check that rememberance) failed.
* clrpick.test: When [testConstraint defaultPseudocolor8] is false we always ↵fvogel2022-10-291-1/+1
| | | | have leftover colors. Therefore constraint colorsLeftOver should be true so that the correspondingly constrainted tests do run. Setting it to false was a mistake from the beginning ([ec889b9906]).
* Fix unintended fork (again).fvogel2022-10-281-15/+2
|\
| * Fix unintended fork.fvogel2022-10-280-0/+0
| |
| * clrpick.test: Fix typos, and simplify clrpick-2.1fvogel2022-10-281-15/+2
| |
* | clipboard-6.2: Remove constraint failsOnXQuarz since it does not fail on ↵fvogel2022-10-281-3/+1
|/ | | | macOS 13.0 (Ventura) when Tk is built with clang with an X11 windowing system (XQuartz).
* canvText-20.1: Tolerate rounding error of at most 1 pixel on the bounding ↵fvogel2022-10-281-3/+9
| | | | boxes values.
* Remove failsOnUbuntu and failsOnXQuarz constraints on canvText-20.1. Using ↵fvogel2022-10-281-5/+2
| | | | TkDefaultFont instead of Hevetica should let the test pass on all platforms without the rounding errors on the bbox that led this test to failure on Linux.
* merge core-8-6-branchfvogel2022-10-28114-2652/+4959
|\
| * Test hygiene. Remove every file that is created.dgp2022-10-281-2/+4
| |
| * Make sure `make distclean` clears away any tkUuid.h file in build area.dgp2022-10-282-2/+2
| |
| * Fix [5e4e44a93b]: ttk::panedwindow chokes on dragjan.nijtmans2022-10-281-6/+6
| |\
| | * Fix [5e4e44a93b]: ttk::panedwindow chokes on drag.fvogel2022-10-251-6/+6
| | |
| * | Tk_GetScrollInfo() -> Tk_GetScrollInfoObj(). Make invalid TK_SCROLL_???? ↵jan.nijtmans2022-10-274-18/+17
| | | | | | | | | | | | values behave the same as TK_SCROLL_ERROR
| * | Fix [43573999ca]: Problem with tkBind.c since Tk 8.6.10 (various assertions ↵fvogel2022-10-262-58/+103
| |\ \ | | |/ | |/| | | | fail).
| | * Rework comments a bit more.fvogel2022-10-251-3/+3
| | |
| | * Rework comments a bit.fvogel2022-10-251-47/+47
| | |
| | * Fix [43573999ca]: Problem with tkBind.c since Tk 8.6.10 (various assertions ↵fvogel2022-10-241-17/+41
| | | | | | | | | | | | fail). Test bind-37.1 now passes.
| | * Focus is not needed to let bind-37.1 crash.fvogel2022-10-241-1/+0
| | |
| | * Add test bind-37.1 demonstrating bug [43573999ca]. This test currently ↵fvogel2022-10-241-0/+22
| |/ | | | | | | crashes (on all platforms).
| * Add a few more test adjustmentsculler2022-10-172-12/+15
| |\
| | * Add a few other adjustments to unstable_testsculler2022-10-102-20/+50
| | |
| * | Merge unstable_tests branch. This commit stabilizes several more tests that ↵fvogel2022-10-1710-105/+104
| |\ \ | | | | | | | | | | | | could sporadically fail on macOS. No test expected results were changed.
| | * | Fix comment in tkMacOSXTest.c because it did not follow the successive ↵fvogel2022-10-151-2/+2
| | | | | | | | | | | | | | | | changes in how we signal testing events to tkProcessMouseEvent.
| | * | Document why there is an 'update' in proc deleteWindows.fvogel2022-10-151-0/+4
| | | |
| | * | Increase the timeout value in canvImg-4.2 since it was seen triggering once ↵fvogel2022-10-151-1/+1
| | | | | | | | | | | | | | | | on macOS in more than 30 runs of the test suite.
| | * | Merge 8.6. At this point (before the present merge), branch unstable_tests ↵fvogel2022-10-1511-60/+105
| | |\ \ | | | | | | | | | | | | | | | runs the test suite successfully on all plaftorms with completely stable results. Especially on macOS, the test suite showed no failure in 30 CI runs at Github Actions (except one timeout failure in canvImg-4.2).
| | * | | Revert [14f82014], i.e. remove TESTFLAGS='-verbose bepst' on macOS.fvogel2022-10-131-2/+2
| | | | |
| | * | | The latter commit made font-47.2, font-47.3, frame612.3 and textImage-4.2 ↵fvogel2022-10-133-8/+8
| | | | | | | | | | | | | | | | | | | | fail again (but not always). Try more updates.
| | * | | Further try to avoid possible hangs in frame-12.3, textImage-4.2, font-47.2 ↵fvogel2022-10-113-10/+12
| | | |/ | | |/| | | | | | | | | and font-47.3. Hangs were (still) observed with frame-12.3 on macOS. This commit is based on an analysis of what is happening (from the events point of view) when changing a font configuration such as its size.
| | * | Test Marc's finding about failures in unixEmbed.test due to interactions ↵fvogel2022-10-101-0/+1
| | | | | | | | | | | | | | | | between successive tests.
| | * | Further try to avoid possible hangs in frame-12.3 and textImage-4.2. Hangs ↵fvogel2022-10-102-2/+2
| | | | | | | | | | | | | | | | were (rarely) observed with frame-12.3 on macOS.
generated by cgit v0.12 at 2024-09-21 18:02:35 (GMT)