Debugging 64bit xemacs (beta, latest cvs) on sparc64 (Sun T1000)

[Andrew Walrond]

Hi Malcolm,

Bad news I'm afraid. After further testing, I can get the patched xemacs 
to crash very easily, just by scrolling around with some ruby code. It 
is never exactly repeatable, which makes things a lot more difficult to 
nail down, but I can always get it to crash within a few 10s of seconds.

:(

andrew at spawn1 xemacs-21.5-work $ cvs -q diff

Index: configure
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/configure,v
retrieving revision 1.291
diff -r1.291 configure
4393c4393
<   sparc-*-*        ) machine=sparc ;;
---
 >   sparc*-*-*       ) machine=sparc ;;
Index: configure.ac
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/configure.ac,v
retrieving revision 1.54
diff -r1.54 configure.ac
1287c1287
<   sparc-*-*        ) machine=sparc ;;
---
 >   sparc*-*-*       ) machine=sparc ;;

andrew at spawn1 xemacs-21.5-work $ xemacs --version
XEmacs 21.5  (beta27) "fiddleheads" (+CVS-20070101) [Lucid] 
(sparc64-unknown-linux) of Tue Jan  2 2007 on spawn1.secure
andrew at spawn1 xemacs-21.5-work $

Typical crash:

Lisp backtrace follows:

   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # bind (print-message-label)
   display-error((end-of-buffer) t)
   # bind (etype debug-on-error inhibit-quit error-object)
   command-error((end-of-buffer))
   # (catch top-level ...)

Segmentation fault

Everything else works peachy on this pure 64bit distro; only xemacs 
exhibits any problems. Not a high priority I guess since I am probably 
the only one building a 64bit version on sparc64, but I can give ssh 
access to a machine setup ready for debugging if it would help.

Best,

Andrew Walrond

Malcolm Purvis wrote:
> Dear Andrew,
> 
> Sorry for talking so long to look at your problems with the sparc64.  Are you
> still experiencing crashes?  If so, I'm wondering if the patch below, when
> applied to the latest CVS, improves things.
> 
> BTW, I also notice from the Installation file you posted at the beginning of
> December that you do not build XEmacs with X support.  Is that intentional?
> 
> Malcolm
> 
> Index: configure
> ===================================================================
> RCS file: /pack/xemacscvs/XEmacs/xemacs/configure,v
> retrieving revision 1.291
> diff -u -r1.291 configure
> --- configure	2007/01/01 10:03:53	1.291
> +++ configure	2007/01/01 11:16:38
> @@ -4390,7 +4390,7 @@
>  machine='' opsys=''
>  
>  case "$ac_cv_build" in
> -  sparc-*-*        ) machine=sparc ;;
> +  sparc*-*-*       ) machine=sparc ;;
>    alpha*-*-*       ) machine=alpha ;;
>    vax-*-*          ) machine=vax ;;
>    mips-dec-*       ) machine=pmax ;;
> Index: configure.ac
> ===================================================================
> RCS file: /pack/xemacscvs/XEmacs/xemacs/configure.ac,v
> retrieving revision 1.54
> diff -u -r1.54 configure.ac
> --- configure.ac	2006/12/28 12:56:04	1.54
> +++ configure.ac	2007/01/01 11:16:46
> @@ -1284,7 +1284,7 @@
>  
>  dnl Straightforward machine determination
>  case "$ac_cv_build" in
> -  sparc-*-*        ) machine=sparc ;;
> +  sparc*-*-*       ) machine=sparc ;;
>    alpha*-*-*       ) machine=alpha ;;
>    vax-*-*          ) machine=vax ;;
>    mips-dec-*       ) machine=pmax ;;
>