Closed
Bug 882990
Opened 11 years ago
Closed 9 years ago
firefox crashes when viewing MJPEG stream from mjpg-streamer
Categories
(Core :: Graphics: ImageLib, defect)
Tracking
()
RESOLVED
WORKSFORME
People
(Reporter: illumilore, Unassigned)
References
Details
(Keywords: crash, stackwanted)
Attachments
(1 file)
39.66 KB,
application/x-gzip
|
Details |
User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:21.0) Gecko/20100101 Firefox/21.0 (Beta/Release) Build ID: 20130512193848 Steps to reproduce: Viewing mjpg streaming website Actual results: firefox crashes Expected results: firefox should not crash
Comment 1•11 years ago
|
||
Please provide the crash ID from about:crashes. Which is the URL?
The website is any server set up with http://sourceforge.net/projects/mjpg-streamer/ A crash ID doesn't show up because the computer usually locks up before that happens. This happens in linux and windows with any FF 21 installed.
Flags: needinfo?(illumilore)
Comment 3•11 years ago
|
||
(In reply to ill from comment #2) > The website is any server set up with > http://sourceforge.net/projects/mjpg-streamer/ Can you provide an URL?
Flags: needinfo?(illumilore)
A URL to what? It differs based on what the server is.
Flags: needinfo?(illumilore)
Updated•11 years ago
|
Status: UNCONFIRMED → RESOLVED
Closed: 11 years ago
Resolution: --- → INCOMPLETE
![]() |
||
Comment 5•11 years ago
|
||
To a server with such a stream, of course.
Summary: firefox crashes when viewing webpage → firefox crashes when viewing MJPEG stream from mjpg-streamer
192.168.0.11:8080/?action=stream
Status: RESOLVED → UNCONFIRMED
Resolution: INCOMPLETE → ---
![]() |
||
Comment 8•11 years ago
|
||
According to the reporter (bug 861545), it started with Firefox 20.
Version: 21 Branch → 20 Branch
Comment 9•11 years ago
|
||
(In reply to ill from comment #6) > 192.168.0.11:8080/?action=stream this is a local address not available to any testers, correct? if so, it's not useful
Status: UNCONFIRMED → RESOLVED
Closed: 11 years ago → 11 years ago
Resolution: --- → INCOMPLETE
Reporter | ||
Comment 11•11 years ago
|
||
"if so, it's not useful" So to reproduce, create a streaming mjpeg server and then browse to whatever that url is with firefox.
Status: RESOLVED → UNCONFIRMED
Resolution: INCOMPLETE → ---
Comment 12•11 years ago
|
||
I'm having similar issues viewing Motion JPG stream from an AXIS IP camera. Update rate is very slow with a lot of hard drive access. Viewing with the camera's viewer is fine.
Comment 13•10 years ago
|
||
Difficult to reproduce, but seems worth keeping the bug open in case anyone wants to try or has a similar setup.
Component: Untriaged → Video/Audio
Product: Firefox → Core
Comment 14•10 years ago
|
||
MJPEG is handled outside of the content/media code, not sure if the correct component is ImageLib or something under GFX.
Component: Video/Audio → ImageLib
Comment 16•10 years ago
|
||
It seems the problem is not crashing, but leaking memory, see bug 1082783 and bug 858615.
Depends on: 858615
Comment 17•10 years ago
|
||
I found these 2 examples: http://192.73.220.130/livevid.html?ds=8 http://141.89.114.58/demo/edu320x240v.html Is it reproducible with FF33 on Linux?
Flags: needinfo?(illumilor.e)
Comment 18•10 years ago
|
||
It is reproducible on 32.0.2 on opensuse. I don't think the repos have 33 up yet.
Flags: needinfo?(illumilor.e)
Comment 19•10 years ago
|
||
I see this on Firefox 33.1.1 (32 bit) on Windows 7 x64. The problem occurs when opening the MJPG file directly (e.g. entering the URL in the address bar or by right clicking the image on a page it's embedded in and selecting "View image"). The problem does NOT occur when viewing it embedded in an HTML page. I cannot provide a public test case, but I can perform tests on a few different platforms and nightlies, if needed.
Comment 20•10 years ago
|
||
As we said, without public testcase, it's not easy to test and debug. Are you able to crash Firefox with your STR by using both URLs I gave in comment #17?
Flags: needinfo?(tomas.creemers)
Comment 21•10 years ago
|
||
I am able to recreate the crash from the first link you gave, when I right click the stream, then click view image.
Comment 22•10 years ago
|
||
Nice! Can you type about:crashes in the location bar and paste some crash URLs (bp-...) related to this crash, it would help.
Comment 23•10 years ago
|
||
When the crash does happen, the whole OS locks up and I have to hard reboot because the hard drive can't keep up with what firefox is trying to write to it once it runs out of memory. That is really inconvenient, so I have been trying to avoid that. Does the hard drive not immediately start thrashing when you try it?
Comment 24•10 years ago
|
||
If you crash due to running out of memory, the crash report is not going to be useful, you want bug 858615. The only reason I didn't mark this bug as a duplicate was because it wasn't clearly indicated that this bug is also about running out of memory (followed by a crash). FWIW I can't reproduce this using http://192.73.220.130/livevid.html?ds=8 (neither with the page, nor after clicking view image) on Nightly (OS X) and on Firefox 31 that comes with Ubuntu 14.04 (in a VM), so there must be something that triggers this behavior.
Comment 25•10 years ago
|
||
This is indeed caused by running out of memory. I have let Firefox run out of memory with http://192.73.220.130/now.jpg?snap=spush?ds=8?dummy=1416274037212 (direct link to the MJPEG stream in the first link in comment #17). I saved about:memory when it was using over 3GB of memory (see attachment). It did not generate a crash report at all. I got a message box with title "Mozilla crash reporter" and as message (translated from Dutch) "[...]Unfortunately, the crash reporter cannot send a crash report. Details: The application did not specify a crash reporting server."
Flags: needinfo?(tomas.creemers)
![]() |
||
Comment 26•10 years ago
|
||
Yeah, let's move the memory leak discussion to bug 858615.
Comment 27•9 years ago
|
||
Tomas, does this issue still reproduce now that bug 858615 is resolved?
Flags: needinfo?(tomas.creemers)
Comment 28•9 years ago
|
||
(In reply to Anthony Hughes, QA Mentor (:ashughes) from comment #27) > Tomas, does this issue still reproduce now that bug 858615 is resolved? I have not re-tested, but in bug 858615 comment 24 I wrote that it was not reproducible anymore. Thank you for taking the time to go through the backlog of issues :-)
Flags: needinfo?(tomas.creemers)
Comment 29•9 years ago
|
||
Thanks for following up, Tomas. I'm resolving this bug as WORKSFORME based on your comment. Please reopen this bug report if the issue returns.
Status: UNCONFIRMED → RESOLVED
Closed: 11 years ago → 9 years ago
Resolution: --- → WORKSFORME
You need to log in
before you can comment on or make changes to this bug.
Description
•