[plt-scheme] Seg fault in 371.3

From: Abdulaziz Ghuloum (aghuloum at cs.indiana.edu)
Date: Sat Jan 5 08:08:24 EST 2008

On Jan 5, 2008, at 7:34 AM, Matthew Flatt wrote:

> This isn't a crash. MzScheme's GC will handle the signal to  
> implement a
> write barrier.
> Use the commands
>   handle SIGSEGV nostop noprint
>   c
> to continue in gdb.


Interesting that you mention this.  I used to do this SEGV trap when
Ikarus was running on the PowerPC (back in the day).  One day after
some I don't know how many hours of chasing after a GC bug, I decided
to enter gdb.  I used your trick above to let gdb continue after the
signal and GDB ITSELF SEGFAULTED!  That was the end of that hack for
me.  Now maybe I should've started another gdb to debug the first gdb
that was debugging ikarus.  THAT would've been quality time.

Just wanted to share that "story from the trenches of another camp".

Now back to your normal programming.


Posted on the users mailing list.