-
Notifications
You must be signed in to change notification settings - Fork 183
Make test valid for backspace/delete #2549
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make test valid for backspace/delete #2549
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
I think this test would have been disabled years ago if it hadn't false positive'd at the time that Display.post tests were all disabled in Bug 553754 commit b846e85 |
The tests failed locally for me, but they pass on the build machine for all the targets currently tested. |
Do you run on Wayland or X11? Note that XWayland inherits all security changes of Wayland so one has to use plain X11 for real test. Test machines are running on X11 only AFAIK. |
Yes, I think that is probably the issue. I will set up an x11 only run to see what happens. |
This test taking a very long time (10 seconds) on Wayland backend. The test as written was invalid because the loop had a timeout without the exit condition being true. This commit updates the test to: - resolve the false positive by checking condition after timeout loop - using common code, SwtTestUtil.processEvents - asserts on the display.post so the underlying error is flagged as the error - adds SwtTestUtil.processEvents at the beginning of the function to ensure window is focused so that the current windown can be posted to - for styled text changes from invokeAction to posting key events, this is to ensure the original intention of the test worked to ensure we don't have a regression in the future on sending BS event - see https://bugs.eclipse.org/565164. `test_invokeActionI` also has a test for `invokeAction(ST.DELETE_PREVIOUS)` so nothing is lost here
9cc2334
to
adbd0ca
Compare
These tests will now show as a failure on gtk4 - which is fine because it was previously showing as a false pass (for me at least). Thanks @akurtakov - you were right, setting up a real x11 backend shows the expected behaviour. I have update the top comment with the new commit message as I cleaned up theses two tests, and learned a bunch of new stuff too. |
I documented my setup in https://kichwacoders.com/2025/09/30/testing-and-developing-swt-on-gtk/ |
The updated tests pass on GitHub actions, all platforms. But failed on Windows I-build https://download.eclipse.org/eclipse/downloads/drops4/I20250930-1800/testResults.php - so more work (and more for me to learn about different platforms) |
This test taking a very long time (10 seconds) on Wayland backend. The test as written was invalid because the loop had a timeout without the exit condition being true.
This commit updates the test to:
test_invokeActionI
also has a test forinvokeAction(ST.DELETE_PREVIOUS)
so nothing is lost here