I tried the patch today without the gem objects and it does not crash, so its some interaction between gem and pd. :(

B.

----- Original Message ----- From: "chris clepper" cgc@humboldtblvd.com To: ben@ekran.org; gem-dev@iem.at Sent: Monday, January 19, 2004 11:56 PM Subject: Re: [GEM-dev] Gem crash pains

This looks like a pd crash as I don't see any GEM function calls in the backtrace. forward this to miller and see what he has to say. maybe GEM is doing something wrong with the pd scheduler?

cgc

At 7:14 PM -0800 1/19/04, ben@ekran.org wrote:

...

Hello all,

I'm now on an OSX machine and having gem segfaults. (attached is my gem crashlog from today) On a hunch I thought I would try my linux crash-test patch on OSX, and low and behold, left control crashes 0.37-0 on OSX the same way as it did in linux:

Program received signal EXC_BAD_ACCESS, Could not access memory. 0x90000e70 in strcmp () (gdb) where #0 0x90000e70 in strcmp () #1 0x00033048 in dogensym () #2 0x00035aa4 in binbuf_text () #3 0x0003b4d0 in socketreceiver_doread () #4 0x0003b810 in socketreceiver_read () #5 0x0003b320 in sys_domicrosleep () #6 0x0003c184 in sys_pollgui () #7 0x00039c04 in m_scheduler () #8 0x000020cc in _start (argc=466652, argv=0x0, envp=0x1) at /SourceCache/Csu/Csu-47/crt.c:267 #9 0x8fe1a700 in __dyld__dyld_start ()

Well after trying it a couple times I found you may have to leave it for

a

...

while counting and then hit it with crtl-l, at least that did it for me. Please try the attached test patch and see what happens.

As for the crashes in the attached log file, these all happend at the

same

...

time as the linux crashes, when sending 1/0 messages to the gemwin while the window is created.

Thanks Ben

Attachment converted: Macintosh HD:crash_gem.log (eApp/CSOm) (0012C0C6) Attachment converted: Macintosh HD:01.redSquare-crash.pd (eApp/CSOm) (0012C0C7)