1008610 - White border around button

Status: VERIFIED FIXED

(Core :: Graphics, defect)

Description

[Tom S [:evilpie]]

Attachment: button.png

Latest Firefox Nightly on Linux Mint.

Comment 1

[:Gijs (he/him)]

Got a regression range?

Comment 2

[Tom S [:evilpie]]

Last good revision: 2a03b34c8953 (2014-05-08)
First bad revision: cf89b5d018f8 (2014-05-09)
Pushlog:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=2a03b34c8953&tochange=cf89b5d018f8

I tried bisecting revisions with mozregression, but that doesn't seem to work.

Comment 3

[Cork]

This affects the "allow" button for extension installation and firebug2s toolbar button too

Last good revision: 5bbc85136202
First bad revision: d1155f994a3a
Pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=5bbc85136202&tochange=d1155f994a3a

Comment 4

[:Gijs (he/him)]

These regression ranges overlap, but one isn't a subset of the other.

That's very confusing. :-(

Comment 5

[Cork]

Just to be clear mozregression marks those as a subset from inside of 08/05 - 09/05. I don't have a tree at work so can't do a build test until i get home.

Comment 6

[:Gijs (he/him)]

(In reply to Cork from comment #5)
> Just to be clear mozregression marks those as a subset from inside of 08/05
> - 09/05. I don't have a tree at work so can't do a build test until i get
> home.

Huh, interesting. d1155f994a3a is in the range, but 5bbc85136202 is not. I guess it's because of merging, so I suspect (but am not 100% sure) that the real range is the intersection of the two, which is quite small:

https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=8a5a9a06f59a&tochange=d1155f994a3a

Comment 7

[Tom S [:evilpie]]

Maybe this has something to do with Bug 1003707? I could imagine if for some reason we calculate a little bit to large element size and thus end up with a white background.

Comment 8

[Cork]

The first bad revision is:
changeset:   182069:3b5fb4abaa3f
user:        Matt Woodrow <[e-mail]>
date:        Thu May 08 11:23:44 2014 +1200
summary:     Bug 1003707 - Pass surfaces sizes in to CreateSourceSurfaceFromNativeSurface instead of trying to extract it from cairo. r=Bas

Comment 9

[Masatoshi Kimura [:emk]]

Why bug 1003707 was landed on m-c and m-a at all? Firefox 31+ had no problem.

Comment 10

[Matt Woodrow (:mattwoodrow)]

Attachment: Fix copy-paste error

Comment 11

[Matt Woodrow (:mattwoodrow)]

We should track this since we're uplifting bug 1003707.

Comment 12

[Matt Woodrow (:mattwoodrow)]

 https://hg.mozilla.org/integration/mozilla-inbound/rev/b22b1b69249

Comment 13

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/b22b1b69249b

Comment 14

[Lukas Blakk [:lsblakk] use ?needinfo]

Please nominate for uplift

Comment 15

[Matt Woodrow (:mattwoodrow)]

Comment on attachment 8421441 [details] [diff] [review]
Fix copy-paste error

[Approval Request Comment]
Bug caused by (feature/regressing bug #): Bug 1003707
User impact if declined: Broken native theme rendering on linux
Testing completed (on m-c, etc.): Been on m-c, manually tested
Risk to taking this patch (and alternatives if risky): Very low risk, just fixes a typo.
String or IDL/UUID changes made by this patch: None.

Comment 16

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/b261b1657477
https://hg.mozilla.org/releases/mozilla-beta/rev/1c2c267a019f

Comment 17

[Francesca Ciceri [:madamezou]]

Hi, 

I was able to reproduce it on Mozilla/5.0 (X11; Linux x86_64; rv:32.0) Gecko/20100101 Firefox/32.0 ID:20140510030204 CSet: 7586c29906b5

I can confirm that the issue is no longer present in the latest Nightly Mozilla/5.0 (X11; Linux x86_64; rv:32.0) Gecko/20100101 Firefox/32.0 ID:20140514030203 CSet: 2f8af55d6e9a

Cheers,
Francesca

Comment 18

[Francesca Ciceri [:madamezou]]

And now verified on latest Aurora and latest Beta as well.

Cheers,
Francesca

Comment 19

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-b2g30_v1_4/rev/1c2c267a019f