Red Hat Bugzilla – Bug 429413
firefox doesn't show screenshots
Last modified: 2008-04-08 03:48:54 EDT
Description of problem:
I went to read <http://fedoraproject.org/wiki/Interviews/PackageKit>, the
screenshots for that one show just as blank areas.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
The bug does not reproduce for me on the link that Horst provided (I can see
the screenshots), but I encounter it from time to time. It's quite real.
Right now it happens to me with this:
I wonder if this is some kind of caching issue: if the image fails to load
once, it stays black until it expires. It seems to happen with sites which
This is still an issue, at least for me.
Forget about the "cache", it seems to be related to image resizing.
All black images occur when an image is shown with constraints.
I went to the upstream bugzilla and amazingly enough there's no
such bug filed, at least as far as I can see. Perhaps it's Fedora
specific. I'm surprised this bug is not getting any attention,
I see tons of the black rectangles everywhere.
I'm using vanilla upstream with a few compilation fix exceptions. This surely
isn't Fedora specific, but may be affected by one of the newer GNOME libraries
we ship? If this still occurs in the latest packages (20080221 or later) then
we should make sure this gets filed upstream.
Still happening to me.
Cannot reproduce here. Something's weird. What happens when you run firefox with
Matej, -no-remote does exactly nothing for me, I tested.
Notice, however, that the original link that Horst provided does
not fail anymore, because the bug requires constraints placed on
images (forced resizing), and apparently CSS was changed in our
Wiki not to scale images anymore. You can see those screencaps
extending into the right column. Make sure you visited the link
that I provided in comment #1.
Sorry, that was stupid type, I meant -safe-mode (so that we are sure no
plugins/extensions are involved). It just works for me.
Created attachment 300180 [details]
illustration of non-reproduction
(In reply to comment #6)
> You can see those screencaps extending into the right column. Make sure you
> visited the link that I provided in comment #1.
Apparently I cannot reproduce it there as well (is this what was meant to be
seen?). This bug is getting dangerously close to WORKSFORME.
Now (firefox-3.0-0.52.beta5.fc9.i386) the page in my original report and the one
in comment #1 work for me.
[No, that isn't really WORKSFORME, its just that for a while firefox had a new
version out each day...]
Created attachment 301355 [details]
The illustration above is taken with -safe-mode.
I suppose it's ok to close if it works for Horst.
This only happens on one of my Rawhide systems, so obviously it's
something specific to it. I just cannot guess what. Same architecture
(x86_64). The video hardware is different, but why would it affect
Pete, let's give it a month. I will put you in NEEDINFO and if you find some way
for us to reproduce this issue, let us know. Otherwise, if I won't be able to
reproduce it in a month, I will close this bug. OK?
The page in #1 shows OK for me (firefox-3.0-0.52.beta5.fc9.i386, i686, nVidia
Corporation G72M [Quadro NVS 110M/GeForce Go 7300] rev 161), using the nv driver
(the nouveau one has strange issues, e.g.
Horst's comment #13 gave me an idea. I shut down all instances of Firefox
and ran one at a time carefuly, with ssh -Y to display it across. Result:
Firefox local @niphredil: FAIL (the original failure)
Firefox local @simbelmyne: OK
Firefox running @niphrefil, displaying to simbelmyne: OK
Firefox running @simbelmyne, displaying to niphredil: FAIL
So, the bug stays with the X server. Can it be a bug in X?
Niphredil uses "radeon" (supplied by -drv-ati).
I'll try to switch DRI off and check if that helps...
Matej, Horst, what should I do about this bug? Clone?
Never mind, looks like it fixed itself with -16, but I haven't restarted
X server since 220.127.116.111-1.20080307.fc9. I'm happy now. Matej, we can
close this as far as I'm concerned.
OK, so I will close this now, and if this issue turns out to still be
reproducible in the latest update, please reopen this bug with additional
Closing as RAWHIDE.