rsync fails with 2 sources

[Andreas Ley]

>I haven't been able to duplicate the crash you two are seeing, so if
>either of you can send me a backtrace of the crash (either from a core
>dump or by doing some creative gdb attaching), I'd appreciate it.

I didn't get a core, but suceeded tu run gdbserver:

#0  0x400a46d9 in strncmp () from /lib/libc.so.6
#1  0x08059878 in make_file (fname=0xbfffb720 "usr/local/lib/.", 
    flist=0x8096dc8, exclude_level=2) at flist.c:818
#2  0x08059cd9 in send_file_name (f=-1, flist=0x8096dc8, 
    fname=0xbfffb720 "usr/local/lib/.", recursive=1, base_flags=1)
    at flist.c:929
#3  0x0805a7fc in send_file_list (f=-1, argc=0, argv=0xbffff7d0)
    at flist.c:1178
#4  0x0804c1e0 in delete_files (flist=0x8094b00) at receiver.c:102
#5  0x080519b6 in do_recv (f_in=0, f_out=1, flist=0x8094b00, local_name=0x0)
    at main.c:426
#6  0x08051dee in do_server_recv (f_in=0, f_out=1, argc=1, argv=0x80919fc)
    at main.c:547
#7  0x08051ee5 in start_server (f_in=0, f_out=1, argc=2, argv=0x80919f8)
    at main.c:577
#8  0x08052da9 in main (argc=2, argv=0x80919f8) at main.c:1041

Looks like a bug in libc6 - ouch. This is debian woody with libc6 2.2.5-11.5.
Since the hosts are productive mail servers, I can't easily test different
libc6 versions, so I hope the stack trace is sufficient to locate the error.

Thnx, Andy

-- 
Andreas Ley, Rechenzentrum, Universitaet Karlsruhe, D-76128 Karlsruhe, Germany
Andreas.Ley at rz.uni-karlsruhe.de, Phone: +49 721 608 6341, Fax: +49 721 32550
"Those who do not understand Unix are condemned to reinvent it, poorly"
								-- Henry Spencer