Closed Bug 343993 Opened 18 years ago Closed 18 years ago

Some backgrounds corrupted with "screenshot" of other content at high resolutions

Categories

(Core :: Graphics, defect)

x86
Linux
defect
Not set
major

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: aguertin+bugzilla, Unassigned)

Details

(Keywords: regression, testcase)

Attachments

(2 files)

2006-07-08 linux trunk nightly (and many earlier nightlies, too)

At large screen sizes (1561px+ ?) on some pages the backgrounds of some elements and/or the entire page are corrupted with external content. This content can be anything on the screen, not just things from firefox.

Example urls are bugzilla (the blue header at top) and http://speex.org/ (the main page background)

This has been happening since the late 20s in june, I think. 

Screenshot coming in a moment. Testcase and regression range coming tomorrow, probably.
Attached image Example screenshot
This happens in builds as far back as  2006-05-20 (and probably more, I didn't check), too, but I know it *didn't* happen originally on that date... so it must be something else that changed on my computer. One candidate is my switch from x.org 6.8 to 7.0

It doesn't happen on the 1.8 branch, so it's probably related to cairo.
Attached file testcase
Of note: The background must be a valid image file, although what image doesn't seem to matter. A .jpg from a different server also showed the problem. Image size might matter, though, I didn't test.

The bug will happen with repeat-x and repeat (or unspecified as that defaults to repeat), but not repeat-y or no-repeat. Again, image size might matter, as the image used in the testcase is very wide but not very tall.
This doesn't seem to happen on http://speex.org anymore, which was what I had the URL as. It might be intermittent, or more likely dependent on other factors. I'm still seeing it on many other sites, including bugzilla. I have seen the bug go away on bugzilla for a short period of time twice, which would possibly back up being dependent on other factors.

(Also note that the screenshot doesn't do justice to how bad this bug is: If it happens on a page's background instead of an element's, I oftentimes have to go into DOMI and delete the background just to read the text.)
Flags: blocking1.9a1?
Keywords: testcase
This no longer happens for me, even in old builds it used to happen in. I don't know what fixed it, although I did recently rebuild my entire system with gcc 4.1.1 (from 3.3.6), as well as upgrade a few X libraries and glibc.

-> WORKSFORME. If anyone else ever has this problem, please reopen.
Status: NEW → RESOLVED
Closed: 18 years ago
Resolution: --- → WORKSFORME
Flags: blocking1.9a1?
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: