Bug 202594 - REGRESSION(r250540): GitLab CI status is broken
Summary: REGRESSION(r250540): GitLab CI status is broken
Status: RESOLVED DUPLICATE of bug 202827
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: WebKit Nightly Build
Hardware: PC Linux
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks: 201994
  Show dependency treegraph
 
Reported: 2019-10-04 12:32 PDT by Michael Catanzaro
Modified: 2019-10-14 23:10 PDT (History)
7 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Michael Catanzaro 2019-10-04 12:32:17 PDT
Since 2.27.1, GitLab CI status pages (e.g. https://gitlab.gnome.org/GNOME/gnome-boxes/-/jobs/455823) are broken. The terminal window does not appear at all.
Comment 1 Michael Catanzaro 2019-10-04 16:36:25 PDT
This broke in r250540: "Inline caching is wrong for custom accessors and custom values"
Comment 2 Radar WebKit Bug Importer 2019-10-06 17:40:41 PDT
<rdar://problem/56022868>
Comment 3 Carlos Garcia Campos 2019-10-09 02:12:46 PDT
I can't reproduce it in mac. With GTK port I get an error in the js console:

[Error] TypeError: o is not a function. (In 'o(!1)', 'o' is an instance of Object)
	(anonymous function) (main.f8ac0dda.chunk.js:130:201836)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:130:110589)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:136:126956)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:130:201645)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:93:7337)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:136:126850)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:136:126778)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:156:115284)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (main.f8ac0dda.chunk.js:58:15103)
	(anonymous function) (main.f8ac0dda.chunk.js:58:19612)
	c (runtime.15f5b13a.bundle.js:1:523)
	(anonymous function) (pages.projects.merge_requests.show.bc041ae3.chunk.js:1:167621)
	c (runtime.15f5b13a.bundle.js:1:523)
	t (runtime.15f5b13a.bundle.js:1:389)
	r (runtime.15f5b13a.bundle.js:1:252)
	Global Code (pages.projects.merge_requests.show.bc041ae3.chunk.js:1)
Comment 4 Carlos Garcia Campos 2019-10-09 02:14:22 PDT
(In reply to Carlos Garcia Campos from comment #3)
> I can't reproduce it in mac. With GTK port I get an error in the js console:
> 
> [Error] TypeError: o is not a function. (In 'o(!1)', 'o' is an instance of
> Object)
> 	(anonymous function) (main.f8ac0dda.chunk.js:130:201836)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:130:110589)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:136:126956)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:130:201645)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:93:7337)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:136:126850)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:136:126778)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:156:115284)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function) (main.f8ac0dda.chunk.js:58:15103)
> 	(anonymous function) (main.f8ac0dda.chunk.js:58:19612)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	(anonymous function)
> (pages.projects.merge_requests.show.bc041ae3.chunk.js:1:167621)
> 	c (runtime.15f5b13a.bundle.js:1:523)
> 	t (runtime.15f5b13a.bundle.js:1:389)
> 	r (runtime.15f5b13a.bundle.js:1:252)
> 	Global Code (pages.projects.merge_requests.show.bc041ae3.chunk.js:1)

This is not CI status page is just a normal merge request page, the comments are never loaded.
Comment 5 Carlos Garcia Campos 2019-10-14 03:14:19 PDT
I've noticed that reloading the page sometimes fixes the issue, so it seems to be a race condition. Is there anything else I can do to help to fix this (with my limited jsc knowledge)?
Comment 6 Michael Catanzaro 2019-10-14 06:46:53 PDT
Rollout...?
Comment 7 Saam Barati 2019-10-14 08:32:31 PDT
I have a fix in:
https://bugs.webkit.org/show_bug.cgi?id=202827

That I will land today. Please confirm it fixes your bug
Comment 8 Michael Catanzaro 2019-10-14 11:00:37 PDT
Thanks Saam!

(In reply to Carlos Garcia Campos from comment #5)
> I've noticed that reloading the page sometimes fixes the issue, so it seems
> to be a race condition. Is there anything else I can do to help to fix this
> (with my limited jsc knowledge)?

If you could test Saam's fix, please, then I won't need to build WebKit on my travel laptop? :)
Comment 9 Saam Barati 2019-10-14 11:48:52 PDT
(In reply to Carlos Garcia Campos from comment #5)
> I've noticed that reloading the page sometimes fixes the issue, so it seems
> to be a race condition. Is there anything else I can do to help to fix this
> (with my limited jsc knowledge)?

I don't think it's a typical race. You're probably just seeing effects of concurrent JIT. If you turn of concurrent JIT, it should reproduce every time in my experience with other web pages that broke.
Comment 10 Saam Barati 2019-10-14 11:49:22 PDT
(In reply to Saam Barati from comment #9)
> (In reply to Carlos Garcia Campos from comment #5)
> > I've noticed that reloading the page sometimes fixes the issue, so it seems
> > to be a race condition. Is there anything else I can do to help to fix this
> > (with my limited jsc knowledge)?
> 
> I don't think it's a typical race. You're probably just seeing effects of
> concurrent JIT. If you turn of concurrent JIT, it should reproduce every
> time in my experience with other web pages that broke.

And if my patch doesn't fix the issue, it'd be super good to know :-)
Comment 11 Carlos Garcia Campos 2019-10-14 23:10:07 PDT
(In reply to Saam Barati from comment #10)
> (In reply to Saam Barati from comment #9)
> > (In reply to Carlos Garcia Campos from comment #5)
> > > I've noticed that reloading the page sometimes fixes the issue, so it seems
> > > to be a race condition. Is there anything else I can do to help to fix this
> > > (with my limited jsc knowledge)?
> > 
> > I don't think it's a typical race. You're probably just seeing effects of
> > concurrent JIT. If you turn of concurrent JIT, it should reproduce every
> > time in my experience with other web pages that broke.
> 
> And if my patch doesn't fix the issue, it'd be super good to know :-)

It fixed it, thanks!
Comment 12 Carlos Garcia Campos 2019-10-14 23:10:19 PDT

*** This bug has been marked as a duplicate of bug 202827 ***