2056926 - Paper over partial update issues causing startup crashes when loading UA stylesheets.

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect)

Comment 1

[Emilio Cobos Álvarez [:emilio]]

Attachment: Bug 2056926 - Paper over bug 1681745 / bug 1941972. r=#style!

See the comment for reasoning, and bug 1681745 for context.

Not a fan in many ways, since this is likely to cause wrong rendering at
best, and random crashes later down the line at worst. But at least I
add the crash reporter annotation message, if we then crash later down
the line...

Tested manually by removing select.css from my objdir and using the
non-nightly codepath.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/2a9f3f2c1c17
https://hg.mozilla.org/integration/autoland/rev/1c8f51b10d8c
Paper over bug 1681745 / bug 1941972. r=firefox-style-system-reviewers,dshin

Comment 3

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/1c8f51b10d8c

Comment 4

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/1c8f51b10d8c

Comment 5

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/1c8f51b10d8c

Comment 6

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/1c8f51b10d8c

Comment 7

[Daniel Holbert [:dholbert]]

ni=emilio to request uplift, since I think we're trying to get this into a dot-release

(maybe you were wanting to add a bit more before doing so -- but maybe this is still independently worth uplifting on its own?)

Comment 8

[Daniel Holbert [:dholbert]]

(In reply to Emilio Cobos Álvarez [:emilio] from comment #1)

Not a fan in many ways, since this is likely to cause wrong rendering at
best, and random crashes later down the line at worst.

To get an idea of how bad this might end up for users who would otherwise have crashed: I tested the same franken-build situation from earlier releases (v152.0 with v151 final-dot-release omni.ja, and v151.0 with v150 final-dot-release omni.ja), and indeed, both technically launch but are terribly unusable.

However, the v151.0-with-v150-omni.ja franken-build was at least able to self-update (and help|about successfully kicked that off, even if the rest of the UI was pretty broken). So the potential for that self-update lifeline is a vote in favor of taking this and avoiding the startup-crash, in the hopes that users who end up in this situation in the future will have a way that their Firefox installations might get back to working (via later self-updates)

Comment 9

[Emilio Cobos Álvarez [:emilio]]

Attachment: Bug 2056926 - Paper over bug 1681745 / bug 1941972.

See the comment for reasoning, and bug 1681745 + bug 1941972 for
context.

Not a fan in many ways, since this is likely to cause wrong rendering at
best, and random crashes later down the line at worst. But at least I
add the crash reporter annotation message, if we then crash later down
the line...

Tested manually by removing select.css from my objdir and using the
non-nightly codepath.

Original Revision: https://phabricator.services.mozilla.com/D313486

Comment 10

[Phabricator Automation]

firefox-beta Uplift Approval Request

• User impact if declined/Reason for urgency: Wallpaper for bug 1681745.
• Code covered by automated testing?: no
• Fix verified in Nightly?: no
• Needs manual QE testing?: no
• Steps to reproduce for manual QE testing:
• Risk associated with taking this patch: low
• Explanation of risk level: Avoids assuming that the omnijar contains the file we expect.
• String changes made/needed?: none
• Is Android affected?: yes

Comment 11

[Emilio Cobos Álvarez [:emilio]]

Attachment: Bug 2056926 - Paper over bug 1681745 / bug 1941972.

See the comment for reasoning, and bug 1681745 + bug 1941972 for
context.

Not a fan in many ways, since this is likely to cause wrong rendering at
best, and random crashes later down the line at worst. But at least I
add the crash reporter annotation message, if we then crash later down
the line...

Tested manually by removing select.css from my objdir and using the
non-nightly codepath.

Original Revision: https://phabricator.services.mozilla.com/D313486

Comment 12

[Phabricator Automation]

firefox-release Uplift Approval Request

• User impact if declined/Reason for urgency: Wallpaper for bug 1681745.
• Code covered by automated testing?: no
• Fix verified in Nightly?: no
• Needs manual QE testing?: no
• Steps to reproduce for manual QE testing:
• Risk associated with taking this patch: low
• Explanation of risk level: Avoids assuming that the omnijar contains the file we expect.
• String changes made/needed?: none
• Is Android affected?: yes

Comment 13

[Emilio Cobos Álvarez [:emilio]]

Yeah I wanted to do bug 2057731 as well as per Gijs' request.

Comment 14

[Pulsebot]

https://github.com/mozilla-firefox/firefox/commit/7d361d2e4116
https://hg.mozilla.org/releases/mozilla-beta/rev/32c4d1fb04e2

Comment 15

[Pulsebot]

https://github.com/mozilla-firefox/firefox/commit/161813000a4a
https://hg.mozilla.org/releases/mozilla-release/rev/f3cc5f70cb4d

Comment 16

[Pulsebot]

https://hg.mozilla.org/releases/mozilla-esr153/rev/0c8fc61d8bb8