FROM : Steven Frank
DATE : Mon Aug 28 22:14:27 2006
Hello,
I have a project which breaks into the debugger with SIGTRAP when
debugged on an Intel Mac, but not on a PowerPC Mac.
There is some sort of relation with MallocStackLogging -- when it's
set to 1, I get the SIGTRAP. When it's set to 0, I don't. On
PowerPC, it doesn't happen at all.
One of the threads is always in a fairly innocuous-looking call to
usleep() at the time of the SIGTRAP. All of the other threads are in
various states of sitting around not doing much.
Obviously, I can work around it by disabling MallocStackLogging --
but is this symptomatic of a memory issue in my code, or is it simply
a bug in MallocStackLogging? In which case, should it be reported?
(Is MallocStackLogging not threadsafe?)
Thanks for any advice,
Steven
http://panic.com/
http://stevenf.com/
DATE : Mon Aug 28 22:14:27 2006
Hello,
I have a project which breaks into the debugger with SIGTRAP when
debugged on an Intel Mac, but not on a PowerPC Mac.
There is some sort of relation with MallocStackLogging -- when it's
set to 1, I get the SIGTRAP. When it's set to 0, I don't. On
PowerPC, it doesn't happen at all.
One of the threads is always in a fairly innocuous-looking call to
usleep() at the time of the SIGTRAP. All of the other threads are in
various states of sitting around not doing much.
Obviously, I can work around it by disabling MallocStackLogging --
but is this symptomatic of a memory issue in my code, or is it simply
a bug in MallocStackLogging? In which case, should it be reported?
(Is MallocStackLogging not threadsafe?)
Thanks for any advice,
Steven
http://panic.com/
http://stevenf.com/
| Related mails | Author | Date |
|---|---|---|
| No related mails found. | ||






Cocoa mail archive

