Closed Bug 152429 Opened 22 years ago Closed 22 years ago

flash in an iframe crashes with Gdk-ERROR **: BadDrawable

Categories

(Core Graveyard :: Plug-ins, defect)

x86
Linux
defect
Not set
critical

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: reidr, Assigned: srgchrpv)

References

Details

(Keywords: crash, regression, testcase)

Attachments

(4 files)

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1a) Gecko/20020610, build 2002061108 A mostly stock RedHat 7.3 system, though I've upgraded some things trying to fix this, like XFree86 Over the last several days, I've seen frequent crashes with the following error message: Gdk-ERROR **: BadDrawable (invalid Pixmap or Window parameter)-1.1a 2:53PM serial 2919 error_code 9 request_code 146 minor_code 3 [1] + 19258 exit 1 ./run-mozilla.sh I've gone back and tried various nightly versions that used to work fine for days and weeks at a time, including 1.1alpha, but they all seem to crash now. I think the culprit may be the talkback pages associated with news articles at http://www.zdnet.com/news, though at this point that's just a hunch. I've restarted and am trying to see how long things last without visiting that site. Previously, it'd crash after 10-15 minutes of browsing. I don't think it's anything to do with plugins; flash works fine, for example. I don't know what else might have changed on my system to cause this, but I'd believe it's not a mozilla-specific problem. (I don't use many other gnome applications, but gnomecal for one worked...)
Keywords: crash, stackwanted
I tried running mozilla under gdb, but unfortunately it seems to exit cleanly when this happens, so I'm not sure how to get a stack trace. I'm also just running binaries, not building from source. This definitely seems to be triggered by viewing some of the talkback posts at http://forums.com.com/group/zd.News.Talkback/zdnn/tb.tpt/@thread@47106@forward@1@D-,D@ALL/@article@47106?EXP=ALL&VWM=hr&ROS=1&PAGETP=2100&SHOST=zdnet.com.com&NODEID=1104 It doesn't happen right away, but usually only after looking at 2-3 of them.
WFM with linux build 20020618 (tried at least 10, no problems) gdk exits, so there is no crash and gdb won't kick in. to get gdb to catch the problem, see bug 104550 comment 32. also, if you're using a mozilla binary (from tar.gz), the stack won't be incredibly useful. If this doesn't happen consistently on any particular page, does it at least always happen on a page if you reload enough times? If you can get that to occur, developing a testcase would be very helpful.
Well, I can no longer get it to crash. Same version of mozilla. I can't think of anything that changed. Maybe it was an offensive ad that they removed from their rotation.... I'll keep an eye out for this in the future and see if I can nail it down better if it happens again.
-> wfm please reopn if you see this again
Status: UNCONFIRMED → RESOLVED
Closed: 22 years ago
Resolution: --- → WORKSFORME
Well, it's been happening to me again. I think I just had a lucky period where I wasn't visiting an offending site for awhile. This most often happens while visiting zdnet, as I mentioned in my original comment, but that's a reflection of my browsing habits. I also had mozilla crash while visiting this link: http://download.com.com/redir?pid=2555655&merid=50632&mfgid=50632&ltype=dl_dlnow&lop=link&edId=3&siteId=4&oId=3000-2032-2555655&ontId=2032&destUrl=%2F3001-2032-2555655.html There's a redirect to an intermediate page, which then goes to a page at www.real.com (discovered by running opera). The crash seemed to occur at the intermediate page; it flashed up briefly before dying. It looks like it's a simple page, though, so maybe it really was the real.com page that caused the problem. The above link crashed my browser twice in a row, so it's repeatable here; I hope it's not just some strange problem with my system. I'm starting to think the problem may be related to flash, but if so, it's not consistent. Some embedded flashes work fine for me, that's for sure. Using Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1b) Gecko/20020725 at the moment, but it's been happening with a variety of nightlies.
Status: RESOLVED → UNCONFIRMED
Resolution: WORKSFORME → ---
Attached file testcase (deleted) —
the brings Mozilla down with the Gdk error, build 20020724. load the testcase, then go to another page (or just reload).
the testcase crashes with linux build 2002072408, but not 2002072221. also loading the iframe's target works fine I'll attach a stack here shortly so we can get a component (could be frames or plugins)...
Component: Browser-General → HTMLFrames
Keywords: regression
Attached file stacktrace (deleted) —
a wonderfully useful stacktrace
marking NEW ==> Plugins
Assignee: Matti → beppe
Status: UNCONFIRMED → NEW
Component: HTMLFrames → Plug-ins
Ever confirmed: true
QA Contact: imajes-qa → shrir
Attached file a slightly better stack (deleted) —
did a bit more testing... all that is necessary is a flash plugin in an iframe. updating summary
Keywords: stackwanted
Summary: Crash with Gdk-ERROR **: BadDrawable → flash in an iframe crashes with Gdk-ERROR **: BadDrawable
hmm, the last trunk build which WFM is 20020722... Andrew, thanks for test case.
Assignee: beppe → serge
backing out bug 139294 fixes the trunk.
Depends on: 139294
This looks to be a dupe of bug 159155. Both may have been caused by bug 139294. No talkback available when this crash happens, so we've no idea how much of a topcrash this may be. Happened to me when using accel-w to close the second to last tab.
*** Bug 159155 has been marked as a duplicate of this bug. ***
Keywords: testcase
Attached patch patch v1 (deleted) — Splinter Review
set the content viewer as non-sticky for nsObjectFrame
Comment on attachment 92953 [details] [diff] [review] patch v1 r=bzbarsky (since this is just restoring the staus quo. ;) ).
Attachment #92953 - Flags: review+
Comment on attachment 92953 [details] [diff] [review] patch v1 sr=jst
Attachment #92953 - Flags: superreview+
Comment on attachment 92953 [details] [diff] [review] patch v1 a=asa (on behalf of drivers) for checkin to 1.1
Attachment #92953 - Flags: approval+
checked in on the trunk /cvsroot/mozilla/layout/html/base/src/nsObjectFrame.cpp,v <-- nsObjectFrame.cpp new revision: 1.354; previous revision: 1.353 thanks all.
Status: NEW → RESOLVED
Closed: 22 years ago22 years ago
Resolution: --- → FIXED
*** Bug 159208 has been marked as a duplicate of this bug. ***
*** Bug 159428 has been marked as a duplicate of this bug. ***
*** Bug 159748 has been marked as a duplicate of this bug. ***
*** Bug 159943 has been marked as a duplicate of this bug. ***
*** Bug 160087 has been marked as a duplicate of this bug. ***
tried tests from duped bugs, look good. verified on 0812 trunk.
Status: RESOLVED → VERIFIED
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: