Closed Bug 22185 Opened 25 years ago Closed 25 years ago

Multiple redraws on ZDNet

Categories

(Core :: CSS Parsing and Computation, defect, P3)

x86
Linux
defect

Tracking

()

VERIFIED WORKSFORME

People

(Reporter: pontus, Assigned: attinasi)

References

()

Details

(Keywords: helpwanted, perf)

When visiting the URL
http://www.zdnet.com/pcweek/stories/jumps/0,4270,2411217,00.html
Mozilla M11 will exit with a segmentation fault.
I reached the page by visiting www.linuxworld.com and clicking on
"Windows isn't the only game in developer town"  under the heading "Net News".

I use Debian GNU/Linux 2.2 (kernel 2.2.13) and glibc 2.1.2, window manager is
Enlightenment DR0.16.

Mozilla was installed using the Debian package for mozilla, version 19991118.M11-2.
Severity: major → critical
changing assignment from nobody@netscape.com to nobody@mozilla.org
[Correcting assignation from "nobody@netscape.com" to "nobody@mozilla.org".]
Using Win95 1999123008, I do not crash, either by going directly to the zdnet
URL, or by clicking on this anchor in the Linuxworld page (which opens a new
browser window, if that makes a difference):

  <a href="http://www.zdnet.com/pcweek/stories/jumps/0,4270,2411217,00.html"
    target=other>Windows isn't the only game in developer town</a>

However, the page loads very slowly, completely blanking the canvas four times
during loading. [Although I know that there are a number of bugs filed for
URLs at zdnet, pcweek, et al].
pontus, 3jrgm or petersen, can you check this with latest build please.  Lots of 
performance work has been done. And this may be better for you now.  Go get M13:
http://www.mozilla.org/projects/seamonkey/release-notes/
Assignee: nobody → troy
Component: Browser-General → Layout
Keywords: perf
QA Contact: nobody → petersen
Yes, it flashes while loading for me as well. The page has a linked style sheet 
and it appears to be the known bug where when the linked style sheet loads we 
discard the current frames and re-create them
Assignee: troy → pierre
Component: Layout → Style System
QA Contact: petersen → chrisd
It's not a dup of bug 17309. I saved the page as a local file and removed the 
linked style sheet: we still have the flicker. Could someone please look into it 
during the bug contest this week?
Whiteboard: need help
Replacing "need help" in the Status Whiteboard with the 'helpwanted' keyword.
Keywords: helpwanted
Whiteboard: need help
track bug precession. pierre: let me know if i've misunderstood what this bug is
currently about.
Summary: Mozilla M11 segfaults on the given URL → asynchronous stylesheet arrival causes excess repaint
Nope, the summary should not be "asynchronous stylesheet arrival causes excess 
repaint" otherwise it would be a dup of bug 17309, which is not as I wrote on 
2000-02-01 22:23.

Changed the summary to "Multiple redraws on ZDNet" and CCd Vidur who has probably 
seen something like that before.
Summary: asynchronous stylesheet arrival causes excess repaint → Multiple redraws on ZDNet
It could be related to bug 26320 that I reassigned yesterday to Vidur.
I'll take a look at this tomorrow. I did make an attempt at narrowing this
down a month ago, and, if memory serves me (heh!), this page was not simply
the result of asynch stylesheet loading. However, I never was able to get
this down to a simple set of conditions, but maybe another pass at this will
produce a better characterization. I'll also look at bug 17309 and bug 26320,
and dup if that is the case. 
I just tested this site with a 3/1 build and I see no jumping or multiple 
painting. Like bug 26320 that Pierre suggested might be related, this appears to 
be fixed now. If anybody else is still seeing this with recent builds please 
make it known, otherwise marking WORKSFORME.

Taking over to help reduce Pierre's bug count.
Assignee: pierre → attinasi
yes. I don't see this anymore either. Marking WORKSFORME. 

(However, I did notice a small problem with style not being reapplied
when the page is reloaded -- opened bug #30308 on that point). 
Status: NEW → RESOLVED
Closed: 25 years ago
Resolution: --- → WORKSFORME
Verified Worksforme on build 2000-04-24-09 (Red Hat 6.2, kernel 2.2.14).
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.