1736384 - Regression: FF segfault with GTK_DEBUG=interactive (GTK inspector)

Status: RESOLVED FIXED

(Core :: Widget: Gtk, defect, P2)

Description

[Valentin Villenave]

Launching Firefox with GTK_DEBUG=interactive results in a segfault.

I tracked it down to this commit (not sure exactly why):

Bug 1721347 - Add OS theme info to about:support.

I’m using GTK 3.24.5-5 on a x86_64 Debian 10 system.

See also Thunderbird’s bug 1736369.

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please revert this change in case you think the bot is wrong.

Comment 2

[Martin Stránský [:stransky] (ni? me)]

Can you please backtrace of the crash? You can find it at about:crashes - please submit it.
Thanks.

Comment 3

[Valentin Villenave]

(In reply to Martin Stránský [:stransky] (ni? me) from comment #2)

Can you please backtrace of the crash?

Sure: there you go.

Comment 4

[Martin Stránský [:stransky] (ni? me)]

Will look at it, Thanks.

Comment 6

[Martin Stránský [:stransky] (ni? me)]

This is a recursion in gfxPlatform::Init(), we call gtk_init(nullptr, nullptr) here, then mozilla::widget::IMContextWrapper::OnThemeChanged() which leads back to gfxPlatform::Init().

Comment 7

[Martin Stránský [:stransky] (ni? me)]

Hm, may the new LookAndFeel::GetThemeInfo() call at nsSystemInfo::Init() cause it?

Comment 8

[Martin Stránský [:stransky] (ni? me)]

I can reproduce it but I'm not sure where to put the check we're already in the nested loop. Emilio, any idea?
Thanks.

Comment 9

[Emilio Cobos Álvarez (:emilio)]

I can't repro on a debug build, does this need an opt build or something?

Comment 10

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1736384 - Don't create IMContextWrapper GTK style provider unnecessarily. r=stransky

I don't know whether all this style provider business is still
necessary looking at bug 1461307, but this should help by avoiding
initializing LookAndFeel too early in any case.

Comment 11

[Emilio Cobos Álvarez (:emilio)]

I think that should help, but I couldn't repro the bug so if you could confirm it'd be great.

Comment 12

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1721347

Comment 13

[Martin Stránský [:stransky] (ni? me)]

(In reply to Emilio Cobos Álvarez (:emilio) from comment #9)

I can't repro on a debug build, does this need an opt build or something?

I can reproduce that on nightly binaries produced by Mozilla under Wayland / Fedora 34.

Comment 14

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/6cebf8df6b78
Don't create IMContextWrapper GTK style provider unnecessarily. r=stransky

Comment 15

[Valentin Villenave]

I can reproduce it using daily builds:

$ mozregression -c 'sh -c "GTK_DEBUG=interactive {binary}"'
INFO: Using local file: /tmp/mozregression/persist/2021-11-09--mozilla-central--firefox-96.0a1.en-US.linux-x86_64.tar.bz2
INFO: Running mozilla-central build for 2021-11-09
INFO: application_buildid: 20211109190508
 INFO: application_version: 96.0a1
 INFO: Running test command: `sh -c "GTK_DEBUG=interactive /tmp/tmplm_4d2pn/firefox/firefox"`
ExceptionHandler::GenerateDump cloned child ExceptionHandler::WaitForContinueSignal waiting for continue signal...
10832
ExceptionHandler::SendContinueSignalToChild sent continue signal to child
INFO: Test command result: 11 (build is bad)

This is on two GNU/Linux x86_64 installs, one Debian oldstable and one Fedora rawhide.

Comment 16

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/6cebf8df6b78

Comment 17

[BugBot [:suhaib / :marco/ :calixte]]

The patch landed in nightly and beta is affected.
:emilio, is this bug important enough to require an uplift?
If not please set status_beta to wontfix.

For more information, please visit auto_nag documentation.

Comment 18

[Emilio Cobos Álvarez (:emilio)]

Martin can you confirm the fix? Should we uplift this?

Comment 19

[Martin Stránský [:stransky] (ni? me)]

Yes, I can confirm the fix. I don't thin we need to uplift it as GTK_DEBUG=interactive is a debugging and rather unused feature.