core dump in cvsup caused by _once()?

[Gary Jennejohn]

On Mon, 30 Nov 2009 15:45:39 +0200
Andriy Gapon <avg at icyb.net.ua> wrote:

> on 28/11/2009 12:15 Gary Jennejohn said the following:
> > Since I installed a new world and kernel on November 26 I'm seeing
> > core dumps with cvsup, even though I reinstalled cvsup yesterday.
> > 
> > Here the output from a gdb session without any debugging symbols:
> > 
> > Core was generated by `cvsup'.
> > Program terminated with signal 4, Illegal instruction.
> 
> 
> Hmm, 'Illegal instruction' is interesting.
> It would be interesting to see diagnostics with debug enabled in libc and, perhps,
> libthr.
>

I'll look into generating libc and cvsup with debugging.

As I wrote in a followup mail, cvsup works OK to update the CVS tree;
it dumps core when it connects to the local cvsupd to update ports/src.

> BTW, is cvsup threaded (linked with a threads lib)?
> 

Doesn't look like it:
garyj:ernst:~:-bash:12> ldd /usr/local/bin/cvsup
/usr/local/bin/cvsup:
        libz.so.5 => /lib/libz.so.5 (0x8006ea000)
        libm.so.5 => /lib/libm.so.5 (0x8007fe000)
        libc.so.7 => /lib/libc.so.7 (0x80091d000)

> P.S. what is your system, both OS and hardware?
> 

garyj:ernst:~:-bash:14> uname -a
FreeBSD ernst.jennejohn.org 9.0-CURRENT FreeBSD 9.0-CURRENT #28:
Thu Nov 26 14:55:58 CET 2009 [...] amd64

It's a AMD64 X2 CPU.

---
Gary Jennejohn