Bug 248203 - REGRESSION(256812@main) [GTK] WebKitWebView API tests are flaky getting stuck
Summary: REGRESSION(256812@main) [GTK] WebKitWebView API tests are flaky getting stuck
Status: NEW
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebKitGTK (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-11-21 21:04 PST by Lauro Moura
Modified: 2022-11-22 06:44 PST (History)
3 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Lauro Moura 2022-11-21 21:04:28 PST
WebKitBuild/GTK/Release/bin/TestWebKitAPI/WebKitGTK/TestWebKitWebView

To reproduce:
./Tools/Scripts/run-gtk-tests --flatpak-verbose --release --display-server=xvfb WebKitBuild/GTK/Release/bin/TestWebKitAPI/WebKitGTK/TestWebKitWebView

Expected: Test executable runs and exits normally

Actual: The test executable becomes zombie while one of the spawned WebProcess is kept alive. The EWS bots are timing out stuck (over the 1200s limit).

Bisecting shows that 256812@main introduced the issue. Also, commenting out the test `/webkit/WebKitWebView/fullscreen` makes the issue go away.
Comment 1 Lauro Moura 2022-11-21 21:15:13 PST
Fullscreen test skipped in https://github.com/WebKit/WebKit/pull/6722 to unstuck the bots.
Comment 2 Michael Catanzaro 2022-11-22 06:44:55 PST
(In reply to Lauro Moura from comment #0)
> Bisecting shows that 256812@main introduced the issue.

This is "[iOS] PiP-then-Fullscreen results in broken controls in fullscreen mode." For a web process to stay alive and not notice that the UI process is gone, I guess it has lost control entirely, probably in an infinite loop? If it were to return control to the main loop, then it should notice and die.