Using build 2000-03-08-09 commercial build on Mac, I crash Launching Mail. 1. Launch Netscape 3. Select Mail from the Tasks menu Crash- getting Talkback report now.

Not sure this is limited to mail. Shrirang sees this clicking other components. Still investigating.

I am seeing random crashes launching the browser using a new profile or using menus and not specifically using mail. This is what I get as a stack trace: Trigger Type: Program Crash Trigger Reason: Illegal PowerPC instruction Call Stack: (Signature = 0x00000008 fe35fbac) 0x00000008 0x000074c0 nsThreadPoolRunnable::Run() [nsThread.cpp, line 607] nsThread::Main() [nsThread.cpp, line 83] _PR_UserRunThread() [pruthr.c, line 489] _PR_UserDestroyThread() [pruthr.c, line 360]

With this mac build, we get the same crash simply clicking on the Buddy List tab in the Browser sidebar every time. cc amusil, prass, scalkins

Mozilla build crashes too.

Adding beta1 for PDT consideration. Note that bug 29908 exists on a crash when quitting after using mail, but this bug sounds different.

smoketest blocker; cc'ing warren and beard, as they have the most recent experience with mac threading changes.

Any more helpful stack traces?

When I run a commercial, installed build, I crash at (CurStackBase does not seem to apply...dumping 4K.) Calling chain using A6/R1 links Back chain ISA Caller 0D28B380 PPC 1DA1996C PRP_TryLock+00994 0D28B310 PPC 1D9184A0 nsThread::Main(void*)+00028 0D28B2C0 PPC 1D91A328 nsThreadPoolRunnable::Run()+0004C 0D28B270 PPC 1C8F40F0 nsFileTransport::Run()+F40F0 PowerPC unmapped memory exception at 1C8F4AFC nsFileTransport::Process()+F4AFC Closing log I think this bug is a dup of 30477, which describes a very similar problem.

Judging from the stack, it seems different from bug 30477 to me.

warren volunteered. Thanks warren!

Well, both this and 30477 talk about crashes clicking between sidebar tabs, and both seem timing dependent (I don't see this crash in my debug build). I would not put too much faith in the stack from Talkback; my stack has nsFileTransport::Process on the top, which I do believe.

I've been seeing the crash Simon describes for two days upon launching Mac OS commercial & Mozilla builds, rendering them unusable for any testing. Added "dogfood" keyword.

Also note bug #31040, "Win98 crashes soon after launching a profile"

Lack of memory on the Mac maybe? The memory partition was getting too small for debug builds, maybe we are reaching the limit for optimized builds too.

Nevermind -- I read the stack wrong. Yes, this appears to be a dup. *** This bug has been marked as a duplicate of 30477 ***

Bug 30477 has been verified. Mark it verified/dup.

verified dup