692350 - Table background stops repeating after ~1600 lines

Status: RESOLVED DUPLICATE of bug 671302

(Core :: Graphics, defect)

Description

[Tim (fmdeveloper)]

Attachment: Zip file containing {reduced} test case

The table background in the attached test case stops repeating after about 1600 lines.

This looks to be a recent regression - if someone can confirm that this should work I can do a regression window

See Original question - https://support.mozilla.com/en-US/questions/882563

Confirmed working on Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.23) Gecko/20110920 Firefox/3.6.23 ID:20110920075126

Confirmed failing on Mozilla/5.0 (Windows NT 5.1; rv:10.0a1) Gecko/20111005 Firefox/10.0a1 ID:20111005030932

Comment 1

[Boris Zbarsky [:bzbarsky]]

Tim, this seems to work for me on Mac (2011-10-05 build).

Do you normally get hardware-accelerated rendering?  If so, does turning off hardware acceleration change anything?

And yes, the behavior described in the sumo report sounds wrong; a regression range would be great.

Comment 2

[Tim (fmdeveloper)]

Attachment: Screenshot from nightly on Windows XP

Will work on a regression range

I do not have H/A on this computer. Graphics section below. This is on a fresh profile for nightly.

  Graphics

        Adapter Description
        NVIDIA GeForce4 4200 Go (Dell Mobile)

        Vendor ID
        10de

        Device ID
        0286

        Adapter RAM
        Unknown

        Adapter Drivers
        nv4_disp

        Driver Version
        6.7.4.2

        Driver Date
        10-26-2004

        Vendor ID (GPU #2)
        000e

        Device ID (GPU #2)
        000e

        Adapter RAM (GPU #2)
        Unknown

        Adapter Drivers (GPU #2)
        Unknown

        Driver Version (GPU #2)
        7.1.542.0

        Driver Date (GPU #2)
        5-22-2006

        WebGL Renderer
        Blocked for your graphics driver version. Try updating your graphics driver to version 257.21 or newer.

        GPU Accelerated Windows
        0/1. Blocked for your graphics driver version. Try updating your graphics driver to version 257.21 or newer.

Comment 3

[Scott Johnson (:jwir3)]

This works for me as well, Ubuntu Linux 11.04/x86_64, with 2011-10-05 build of Nightly.

Comment 4

[Tim (fmdeveloper)]

Here's the regression window I came up with using http://harthur.github.com/mozregression/

Last good nightly: 2011-05-25
First bad nightly: 2011-05-26

Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=836aa9658341&tochange=831f8e040f38

Comment 5

[Boris Zbarsky [:bzbarsky]]

Ah, the cairo 1.10 update.  Lovely.  :(

Comment 6

[Bruce A. Wittmeier]

Here is another demo - seems not tied to table rows but perhaps pixels, image iterations, ???

The two tables give different results - variation of about 20 rows.

http://gotnarcosis.com/brucewittmeier/technical/TableRowDemo.html

Comment 7

[Mike Starov]

Confirmed failing on Win7/Vista, Firefox 7.0.1 on VMWare virtual machine. Related bug 697507 opened.

Any estimates when this will be worked on?

Comment 9

[Boris Zbarsky [:bzbarsky]]

Given the cairo connection, a good start is getting this in the right component...

Comment 10

[Bruce A. Wittmeier]

Im not trying to push this, but I did a little more experimenting - dont know if it will help or not.


First experiment single image - 1,280px × 288px
http://www.gotnarcosis.com/brucewittmeier/images/waterspotbg.jpg

They both truncate around 1320 or so.
http://gotnarcosis.com/brucewittmeier/technical/TableRowBGDemoTallerImage.html


Second experiment:
I increased the row size and the image of the second table to a significantly taller image.
They both truncate around 610 or so.  Since both truncate about the same rows and the images are different in height, this suggests the problem is in pixels rather than rows.  The images dont seem to be a factor.

http://gotnarcosis.com/brucewittmeier/technical/TableRowBGDemoTallerImage.html

First image - 1,280px × 288px
http://www.gotnarcosis.com/brucewittmeier/images/waterspotbg.jpg

Second image - 1,024px × 768px
http://www.gotnarcosis.com/brucewittmeier/images/26044.jpg

Comment 11

[Bruce A. Wittmeier]

Cancel the prev post.


I did a little more experimenting - dont know if it will help or not.


First experiment single image - 1,280px × 288px
http://www.gotnarcosis.com/brucewittmeier/images/waterspotbg.jpg

They both truncate around 1320 or so.
http://gotnarcosis.com/brucewittmeier/technical/TableRowDemo.html


Second experiment:
I increased the row size and the image of the second table to a significantly taller image.
They both truncate around 610 or so.  Since both truncate about the same rows and the images are different in height, this suggests the problem is in pixels rather than rows.  The images dont seem to be a factor.

http://gotnarcosis.com/brucewittmeier/technical/TableRowBGDemoTallerImage.html

First image - 1,280px × 288px
http://www.gotnarcosis.com/brucewittmeier/images/waterspotbg.jpg

Second image - 1,024px × 768px
http://www.gotnarcosis.com/brucewittmeier/images/26044.jpg

Comment 12

[Jeff Muizelaar [:jrmuizel]]

This sounds a lot like bug 671302