901955 - CSS animations not working for SVG opacity

Status: VERIFIED FIXED

(Core :: SVG, defect)

Description

[Sebastian Widmer]

Attachment: example

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:25.0) Gecko/20100101 Firefox/25.0 (Beta/Release)
Build ID: 20130806004002

Steps to reproduce:

Create an SVG element and try to animate its opacity style property with CSS3.


Actual results:

Animation does not work; opacity change can be observed only when the element is clicked. At the same time, similar animation works OK for HTML <div>.

I've been using Firefox 24a2 for a while, and this issue has appeared just recently (round Aug 1 IIRC). Firefox 25 nightly shows the same. In Firefox 23, everything works OK.


Expected results:

SVG element should fade in/out smoothly.

Comment 1

[Robert Longson [:longsonr]]

Would you be willing to use http://mozilla.github.io/mozregression/ to find a regression range?

Comment 2

[Cork]

Last good nightly: 2013-05-25
First bad nightly: 2013-05-26

Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=7a2f7a45819a&tochange=0fed3377c839

Makes me suspect bug 875175.

Comment 3

[Robert Longson [:longsonr]]

Could also be bug 854765 or bug 875037.

Comment 4

[Sebastian Widmer]

(In reply to Cork from comment #2)
> Last good nightly: 2013-05-25
> First bad nightly: 2013-05-26
> 
> Pushlog:
> http://hg.mozilla.org/mozilla-central/
> pushloghtml?fromchange=7a2f7a45819a&tochange=0fed3377c839

Confirming. Don't know why I had an impression that it broke just a week ago :-\

Comment 5

[Alice0775 White]

Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=243533ca7769&tochange=d1b9087dd689

Regressed by Bug 854765

Comment 6

[Jonathan Watt [:jwatt]]

Okay, I know what's going on. Patch coming up shortly.

Comment 7

[Jonathan Watt [:jwatt]]

Attachment: patch

This used to work because nsSVGPathGeometryFrame::DidSetStyleContext used to blindly invalidate and reflow, so we'd always invalidate for an opacity property change. Bug 854765 got rid of that code, and for opacity changes we now rely on DLBI. Unfortunately that breaks down when the SVG opacity optimization kicks in. The optimization is that if 'opacity' is set on an nsSVGPathGeometryFrame but nsSVGUtils::CanOptimizeOpacity() returns true, then we avoid creating a layer (previously, more directly avoided creating and compositing an offscreen surface) by treating the 'opacity' as 'fill-opacity' or 'stroke-opacity'. More specifically, nsIFrame::BuildDisplayListForStackingContext does not create an nsDisplayOpacity display item. As a result, DLBI can't invalidate for the change.

The simplest thing to do in order to keep this safe for beta is to invalidate the frame in DidSetStyleContext in the case that the old and new 'opacity' is different and nsSVGUtils::CanOptimizeOpacity() returns true.

Comment 8

[Jonathan Watt [:jwatt]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/1d64db569ebf

Comment 9

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/1d64db569ebf

Comment 10

[bhavana bajaj [:bajaj]]

:jwatt - ping, is this ready for uplift ?

Comment 11

[Jonathan Watt [:jwatt]]

Comment on attachment 789886 [details] [diff] [review]
patch

Yes.

[Approval Request Comment]
Bug caused by (feature/regressing bug #): bug 854765
User impact if declined: broken SVG
Testing completed (on m-c, etc.): been on m-c for days
Risk to taking this patch (and alternatives if risky): very low
String or IDL/UUID changes made by this patch: none

Comment 12

[Jonathan Watt [:jwatt]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/62ec86acf4c6
https://hg.mozilla.org/releases/mozilla-beta/rev/3dbd75529f99

Comment 13

[Ioana (away)]

Verified as fixed on:
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:24.0) Gecko/20100101 Firefox/24.0 (20130806170643)

Comment 14

[Ioana (away)]

(In reply to Ioana Budnar, QA [:ioana] from comment #13)
> Verified as fixed on:
> Mozilla/5.0 (Windows NT 6.1; WOW64; rv:24.0) Gecko/20100101 Firefox/24.0
> (20130806170643)

Correct build ID for the build with the fix: 20130822154523.

On the one in comment 13 the bug does reproduce.

Comment 15

[Ioana (away)]

Also verified on Firefox 24 beta 5 on Mac OSX 10.7.5 and Ubuntu 13.04 64bit.

Comment 16

[Ryan VanderMeulen [:RyanVM]]

This was backed out from Fx24 in bug 907503.

Comment 17

[Ioana (away)]

Verified as fixed on Firefox 25:
Mac OS X 10.8.4 - 20130912004004
Ubuntu 13.04 - 20130911004006
Windows 7 - 20130912004004

Comment 18

[Bader Zaidan]

Looks like the fix works on Firefox Nightly 27.0:

Trisquel GNU/Linux 6.0 64bit- 20130918030202

Comment 19

[Bogdan Maris, Desktop Test Engineering]

Mozilla/5.0 (X11; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:26.0) Gecko/20100101 Firefox/26.0
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:26.0) Gecko/20100101 Firefox/26.0

Verified as fixed on latest Aurora 26.0a2 (buildID: 20131009004001) and latest Nightly 27.0a1 (buildID: 20131008030202).