660554 - Opacity through rgba doesn't work for first-child in nested transparent elements [Intel ddx]

Status: UNCONFIRMED

(Core :: Graphics, defect)

Description

[Dmitry Petrov]

User-Agent:       Mozilla/5.0 (X11; Linux i686; rv:2.0.1) Gecko/20100101 Firefox/4.0.1
Build Identifier: Mozilla/5.0 (X11; Linux i686; rv:2.0.1) Gecko/20100101 Firefox/4.0.1

If parent element sets background color through rgba value and also sets and opacity value in it, and it's children also have opaque background through rgba value, than this value doesn't work for the first child element of container.

Reproducible: Always

Steps to Reproduce:
Set background-color: rgba( 255, 255, 255, 0.4 ) for the container div and all it's children.

Actual Results:  
First child of the container will have no background at all.

Expected Results:  
All child elements of the container must have background, as it's set in the css.

As is shown on the example page, bug disappears if there is some visible text in the container before children with the background-color property, or there is some visible text before the container.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Would you mind posting a screenshot of what that testcase looks like for you?  As far as I can tell the rendering I'm seeing on that test page is correct...

Comment 2

[Dmitry Petrov]

Attachment: Bug testcase screenshot

All the words "test" are wrapped in spans's with  background-color: rgba(255, 255, 255, 0.4);

As you can see on the screenshot, the first word on the first row doesn't have any background.

Comment 3

[(mostly gone) XtC4UaLL [:xtc4uall]]

WFM on 4.0.1 & Trunk on WinXP.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Yeah, I don't see the issue in that screenshot in either my trunk Linux build or the 4.0.1 binary I just downloaded from mozilla.

Dmitry, are you using a distro build or one of the mozilla ones?

Comment 5

[Dmitry Petrov]

I'm using the package from the official ubuntu repositories - 4.0.1+build1+nobinonly-0ubuntu0.11.04.3

I don't see the issue in Firefox 4.0.1 for windows

Comment 6

[Boris Zbarsky [:bzbarsky]]

Can you try a mozilla.org Linux build, please?

Comment 7

[Dmitry Petrov]

I still see the bug in the mozilla.org build.

How I tried:
1)Downloaded tar.bz archive from mozilla org
2)Extracted archive contents
3)Entered the program folder and then run program with ./firefox

I don't know if it interferes with distro package.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Did you quit the distro Firefox first?

Comment 9

[Dmitry Petrov]

Yes, I did. Also I've checked if there is any running firefox process

Comment 10

[Boris Zbarsky [:bzbarsky]]

OK.  What Xorg version is this?  This is starting to sound like a bug in the interaction of cairo and your X server...

Comment 11

[Dmitry Petrov]

My system has 7.6+4ubuntu3.1 package installed.

Comment 12

[George Carstoiu]

(In reply to comment #11)
> My system has 7.6+4ubuntu3.1 package installed.

Is this related to the version of Xorg that Dmitry is using?

Comment 13

[Karl Tomlinson (:karlt)]

I think that's xorg-server 1.7.6, which is fairly old now, but it's probably more likely related to device-dependant drivers or maybe even pixman.

Strange though that it would be specific to first-child of first-child.
I don't know what we'd be doing different in that particular case.
I can't rule out a problem on our-side, but I can't reproduce.

Comment 14

[Dmitry Petrov]

I've updated firefox to the latest version 5 and the bug is still there.

Comment 15

[Sjoerd Mulder]

I can actually confirm this issue on Mozilla/5.0 (Ubuntu; X11; Linux x86_64; rv:9.0.1) Gecko/20100101 Firefox/9.0.1

Strangly this only happens when the parent and child rgba values are the same... if they differ it doesnt happen.

I cannot reproduce on Fx3.6.12 WinXP or Fx9.0.1 Win7

See attached screenshot + testcase

Comment 16

[Sjoerd Mulder]

Attachment: screenshot

Comment 17

[Sjoerd Mulder]

Attachment: testcase

Comment 18

[Sjoerd Mulder]

im using the latest Ubuntu 11.10 (64bit) with gnome-panel

Comment 19

[Karl Tomlinson (:karlt)]

Sjoerd and Dmitry, can you describe the video drivers that X is using, please?

Look for "Video Driver" in /var/log/Xorg.0.log. e.g.
[    22.307] (II) Module radeon: vendor="X.Org Foundation"
[    22.307]    compiled for 1.11.2, module version = 6.14.3
[    22.307]    Module class: X.Org Video Driver
[    22.307]    ABI class: X.Org Video Driver, version 11.0

Comment 20

[Sjoerd Mulder]

[     8.026] (II) intel: Driver for Intel Integrated Graphics Chipsets: i810,
[     8.028] 	ABI class: X.Org Video Driver, version 10.0

Its a Sandy Bridge (Core i7 2630QM) laptop with integrated graphics

Video playback / Animations / Gnome Shell / Scrolling is generally smooth, no real performance issues in day-to-day programming :)

Comment 21

[Dmitry Petrov]

I use Lenovo t420 with integrated graphics now and can still reproduce the error. I had another laptop in the time when the bug was reported, but it also had intel graphics.

  [    21.479] (II) LoadModule: "intel"
  [    21.479] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
  [    21.579] (II) Module intel: vendor="X.Org Foundation"
  [    21.579] >--compiled for 1.10.4, module version = 2.17.0
  [    21.579] >--Module class: X.Org Video Driver
  [    21.579] >--ABI class: X.Org Video Driver, version 10.0
...
  [    21.656] (II) intel: Driver for Intel Integrated Graphics Chipsets: i810,
  >---i810-dc100, i810e, i815, i830M, 845G, 854, 852GM/855GM, 865G, 915G,
  >---E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM, Pineview G,
  >---965G, G35, 965Q, 946GZ, 965GM, 965GME/GLE, G33, Q35, Q33, GM45,
  >---4 Series, G45/G43, Q45/Q43, G41, B43, B43, Clarkdale, Arrandale,
  >---Sandybridge Desktop (GT1), Sandybridge Desktop (GT2),
  >---Sandybridge Desktop (GT2+), Sandybridge Mobile (GT1),
  >---Sandybridge Mobile (GT2), Sandybridge Mobile (GT2+),
  >---Sandybridge Server, Ivybridge Mobile (GT1), Ivybridge Mobile (GT2),
  >---Ivybridge Desktop (GT1), Ivybridge Desktop (GT2), Ivybridge Server

Comment 22

[Karl Tomlinson (:karlt)]

Thanks.  Looks like it could be worth filing a bug at freedesktop for the Intel driver.  (I can't reproduce with radeon, nor with Xvnc, which I assume uses the sw framebuffer and pixman.)