[darcs-users] Segfaults on Solaris 8/SPARC darcs 1.0.2

Jason Dagit dagit at cs.orst.edu
Mon Feb 14 19:22:14 UTC 2005


On Mon, 14 Feb 2005, David Roundy wrote:

> On Sun, Feb 13, 2005 at 10:46:43PM -0800, Jason Dagit wrote:
>> Hello,
>>
>> I managed to build darcs on solaris 8,  but
>> if I do the following I get a segfault:
>> mkdir test
>> cd test
>> darcs init
>> touch foo
>> darcs add foo   <-- This is the line that segfaults.
>>
>> I think it's crashing in pthreads, but I don't really know how to be
>> sure of that.  I tried compiling with profiling enabled, but all that
>> changed was the way darcs crashes (no extra output or anything).  If
>> profiling is enabled I get Illegal Instruction instead of segfault.
>
> I don't think add should be using the pthreads code, since it doesn't need
> to write or read a compressed patch.  On the other hand, I may be wrong.
> Can you run something like strace or valgrind to see what's happening
> before it crashes?

Unfortunately no.  I'll demonstrate:

$ strace ls
ERROR: unable to open /dev/log

strace appears to be broken on that machine and valgrind is not installed. 
I would have to compile it myself, which is perhaps what I should do.

I can run it from gdb, in which case I have the following backtrace:
Program received signal SIGSEGV, Segmentation fault.
0xfefdff20 in ?? ()
(gdb) back
#0  0xfefdff20 in ?? ()
#1  0x3f8b9c in schedule ()
#2  0x3f9558 in waitThread_ ()
#3  0x3f92d4 in scheduleWaitThread ()
#4  0x3f7d88 in rts_evalLazyIO ()
#5  0x3f76f8 in main ()

Thanks,
Jason




More information about the darcs-users mailing list