912769 - All the chrome is painted more often on linux Nightlies than on FF23

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Olli Pettay [:smaug][[e-mail]]]

Based on nglayout.debug.paint_flashing=true we end up painting the whole
browser chrome after each page load.
FF23 paints only the throbber, page title in the tab and the changed parts of the
location bar. And looks like we're doing more full paints also in the content area,
at least if one can trust the visual feedback from paint flashing.

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

beta doesn't have this bug, but aurora does.

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

My testcase is to load http://mxr.mozilla.org/mozilla-central/ search for file nsINode, click the
first result and click there the first included file. Chrome shouldn't get fully repainted.

I think this is the right range
2013-07-19-03-02-04-mozilla-central is ok     http://hg.mozilla.org/mozilla-central/rev/af4e3ce8c487
2013-07-20-03-02-14-mozilla-central is not ok http://hg.mozilla.org/mozilla-central/rev/bf73e10f5e54
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=af4e3ce8c487&tochange=bf73e10f5e54

Could be some change to FF chrome or to invalidation, I guess

Comment 3

[Mike Conley (:mconley) (:⚙️)]

Do we have a sense of which platforms are affected? The platform listed in this bug is Linux, but sometimes that's set just because that's the platform the reporter is using.

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

I use only linux. In theory I could test this on Windows too.

Comment 5

[Timothy Nikkel (:tnikkel)]

Maybe narrowing the regression range down would be useful? Or testing on other platforms?

Comment 6

[Robert O'Callahan (:roc) (email my personal email if necessary)]

(In reply to Olli Pettay [:smaug] from comment #2)
> My testcase is to load http://mxr.mozilla.org/mozilla-central/ search for
> file nsINode, click the
> first result and click there the first included file. Chrome shouldn't get
> fully repainted.

These steps don't do any unnecessary repainting for me, on Linux trunk (2d495522a53a).

Comment 7

[Wayne Mery (:wsmwk)]

(In reply to Timothy Nikkel (:tn) from comment #5)
> Maybe narrowing the regression range down would be useful? Or testing on
> other platforms?

smaug, can you wrangle these two points further?

(I probably should not have put linux in bug summary, but will leave there for now)

Comment 8

[Alice0775 White]

I can reproduce on ubuntu12.04 32bit.

STR
1. Enable "Bookmarks Toolbar"
2. Open http://www.mozilla.org/en-US/about/contact.html#map-mountain_view
3. Click "DESKTOP" "MOBILE" "RELEASES"  each
   --- observe toolbar color

Actual Results:
Toolbox("Bookmarks Toolbar" etc) is repainted

Regression window(m-c)
Good:
http://hg.mozilla.org/mozilla-central/rev/76a944fa6b25
Mozilla/5.0 (X11; Linux i686; rv:25.0) Gecko/20130802 Firefox/25.0 ID:20130802001119
Bad:
http://hg.mozilla.org/mozilla-central/rev/d0edf8086809
Mozilla/5.0 (X11; Linux i686; rv:25.0) Gecko/20100101 Firefox/25.0 ID:20130803030205
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=76a944fa6b25&tochange=d0edf8086809


Regression window(m-i)
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/7fe79f773eb4
Mozilla/5.0 (X11; Linux i686; rv:25.0) Gecko/20100101 Firefox/25.0 ID:20130802094418
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/93f36094bf6f
Mozilla/5.0 (X11; Linux i686; rv:25.0) Gecko/20100101 Firefox/25.0 ID:20130802095818
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=7fe79f773eb4&tochange=93f36094bf6f

Regressed by:
3f36094bf6f	Benoit Girard — Bug 894773 - Don't force layers on single opacity change. r=roc


Graphics
--------

Adapter Description: VMware, Inc. -- Gallium 0.4 on SVGA3D; build: RELEASE;
Device ID: Gallium 0.4 on SVGA3D; build: RELEASE;
Driver Version: 2.1 Mesa 8.0.4
GPU Accelerated Windows: 0/1 Basic
Vendor ID: VMware, Inc.
WebGL Renderer: VMware, Inc. -- Gallium 0.4 on SVGA3D; build: RELEASE;
windowLayerManagerRemote: false
AzureCanvasBackend: cairo
AzureContentBackend: cairo
AzureFallbackCanvasBackend: none
AzureSkiaAccelerated: 0

Comment 9

[Benoit Girard (:BenWa)]

The patch from bug 894773 was backout. The bug doesn't really reflect that well.