User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030314 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030314 I was doing a google search and this was the top hit. I clicked on the url and mozilla crashed. Tested it a few more times, each time I try to open the url in question mozilla crashes. Reproducible: Always Steps to Reproduce: 1. Open http://www.movabletype.org/docs/mtinstall.html Actual Results: mozilla crashed Expected Results: displayed the page Linux 1.3 release doesn't appear to have talkback, sorry.

WFM 1.3 OS X

WFM Linux 2003031908 and Last night CVS build with GTK2/XFT

WFM 2003031804 PC/WinXP

WFM Linux (RH 7.3) Build 2003032105

WFM, current trunk CVS, Linux. (RH8, "all" plugins) Reporter: Did you install from rpm? If so: Was it the RH8 RPM with Gtk2?

Reporter mailed: "Yes, I installed from rpm. Specifically, the RH8 rpms with Xft support." Over to Blizzard.

worksforme with linux trunk/CVS/gtk2/xft if you start Mozilla from a terminal, does it print a message when it crashes?

Looking at the css on that page I'm going to make a wild guess that you have some .fon files in your font path. Can you move them out of your font path temporarily to test if they are the problem? *** This bug has been marked as a duplicate of 180309 ***

No, not a dup.

So. Does anything print to the console when Mozilla crashes?

> a 'locate .fon' might be enough. I'm guessing that you had one of the > .fon files in your path. No, locate doesn't turn up any files ending in ".fon". > bash$ fc-list "" family file This spews out a huge list of fonts, but none of them are .fon -- most are .pfa, .pfc, .pfb, .font or .fontmap.

> So. Does anything print to the console when Mozilla crashes? Not a thing! Is there any way to get verbose debugging information? This install is from the RH 8.0 RPMs which don't have any sort of talkback.

% mozilla --sync & % gdb /usr/lib/mozilla-1.3/mozilla-bin _pid_of_mozilla (gdb) b abort (gdb) b exit (gdb) cont [make Mozilla crash] (gdb) bt [create an attachment for output of bt]

Mozilla hung as soon as I attempted to attach gdb session & set breakpoints. I don't know why the gdb session was suspended halfway through reading symbols, I didn't do that explicitly. I also don't know what caused the SIG32, as I never had a chance to enter the offending url from the original bug report.

Did you install the RH8.0 glibc update before you made this backtrace?

If that glibc update was a security issue, then I believe so -- the machine is kept current with up2date. RH has so many updates it's hard to keep track. $ rpm --query glibc glibc-2.3.2-4.80

Strange observation: I had two different instances of mozilla 1.3 running, one local and one tunneled over ssh from another RH 8.0 machine. I hit the url that causes the crash, on the network instance of mozilla, and immediately BOTH instances of mozilla crashed.

slaton: can you try gdb again? the SIG32 weirdness is because of the glibc update. when it does SIG32, just do "cont" again. you might have to do it a few times. Once it stops giving you SIG32s, make Mozilla crash for real, and then do "bt". unless you're careful, the RPM build will not bring up a second instance of Mozilla (it opens a new window instead), even when starting Mozilla from different machines.

> Strange observation: I had two different instances of mozilla 1.3 running, one > local and one tunneled over ssh from another RH 8.0 machine. I hit the url that > causes the crash, on the network instance of mozilla, and immediately BOTH > instances of mozilla crashed. I'll bet that due to -remote magic that they are actually the same instance. :)

Bug Day: 1-13-2004 Tested With: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6b) Gecko/20031210 Are you still having this problem with a more recent build? The page loads just fine for me.

Bug day 04.02.03 No comment from reporter in over 10 months. WFM in 1.4 xft release RPM on Mandrake 9.2 and as indicated in other comments.