Re: pkg/45589: pkgsrc/editors/emacs build failure

[bch%methodlogic.net@localhost]

On Tue, Nov 08, 2011 at 03:19:28PM +0000, bch%methodlogic.net@localhost wrote:
> On Tue, Nov 08, 2011 at 01:10:06PM +0000, David Holland wrote:
> > The following reply was made to PR pkg/45589; it has been noted by GNATS.
> > 
> > From: David Holland <dholland-pbugs%netbsd.org@localhost>
> > To: gnats-bugs%NetBSD.org@localhost
> > Cc: 
> > Subject: Re: pkg/45589: pkgsrc/editors/emacs build failure
> > Date: Tue, 8 Nov 2011 13:05:14 +0000
> > 
> >  On Tue, Nov 08, 2011 at 06:20:01AM +0000, bch%methodlogic.net@localhost 
> > wrote:
> >   >         Fatal error (6)[1]   Abort trap (core dumped) 
> > EMACSLOADPATH=/u...
> >   >         gmake[3]: *** 
> > [/usr/pkgsrc/editors/emacs/work/emacs-23.3/lisp/emacs-lisp/bytecomp.elc] 
> > Error 134
> >  
> >  This is almost certainly the same problem that's been reported several
> >  times elsewhere, e.g. in PR 45405; PR 40166 may be the same too.
> >  
> >  It doesn't affect everyone and there's no pattern anyone's been able
> >  to figure out.
> 
> "cjbaird" in #netbsd has same problem (-current on amd64 as well),
> but apparently is so incapacitated that can't post to ML.
> 
> Datapoint (which may be closer to core issue): I restored a runnable emacs 
> from a package I'd built, but noticed that it dumps core on exit (C-x C-c).
> eg:
> ======================
> kamloops$ emacs . 
> Fatal error (6)[1]   Abort trap (core dumped) emacs .
> ======================
> 
> I can't see anything obvious in the core file, but I haven't _really_
> dug in; gdb problem solving for me is usually pretty obvious and simple,
> so there may be features to get better info I'm just not familiar
> with. Emacs itself is doing the compiling, a bad exit like the above

Too early, no coffee: Let me rewrite.

"In the build process failure, Emacs is doing it's own compiling,
but the failure on exit alone (observed from package I installed too)
is what's causing the build to fail. In that case we can start/stop an
installed Emacs at will and observe failure, run traces, etc."

I've got such a beast now, and can provide core files or analysis with
guidance.

cd ~coffee

-bch

> is probably what's triggering the failure. In that case, we can trigger
> failures at will (start/stop emacs) and observe.
> 
> -bch
> 
> >  -- 
> >  David A. Holland
> >  dholland%netbsd.org@localhost
> >  
> 
> -- 
> Brad Harder
> Method Logic Digital Consulting
> http://methodlogic.net/
> http://twitter.com/bcharder
> 

-- 
Brad Harder
Method Logic Digital Consulting
http://methodlogic.net/
http://twitter.com/bcharder