Re: crash on 21.5-b16 i386-pc-solaris2.9
Hi Steve,
thanks for the quick response. Responses/comments etc inline below:
Steve Youngs wrote:
|--==> "JCM" == James C McPherson
<james.mcpherson(a)optusnet.com.au> writes:
JCM> Operation: start xemacs. Select
Tools-> Packages->Set Download Site->Official Releases->(Australian sites)
JCM> ... Update Package Index
JCM> ... /ftp.xemacs.org:pub/xemacs/packages/package-index.LATEST.gpg (click OK)
JCM> Reproducible: always. dies with US locations and Australian.
JCM> ... dies as ordinary user and root as well.
I can't reproduce it here on Linux.
JCM> espresso:jmcp $ xemacs -geometry +0+0 &
Do you get the same thing if you invoke XEmacs like...
xemacs -vanilla -geometry +0+0 &
If using `-vanilla' fixes the problem, then something in _your_ setup is
messing it up.
I tried it again on the as-distributed 21.5.b16 with -vanilla and the
crash still occurred.
JCM> ./configure '--prefix=/d00/xemacs'
'--debug=no' '--pdump=no'
JCM> '--use-kkcc=yes' '--with-gpm=no' '--with-workshop'
'--with-gcc=no'
JCM> '--with-ldap=no'
Compiling with debugging symbols would give the bug-catching boys and
girls much more of a fighting chance. Drop that `=no' off of the end
of `--debug'. :-)
well the configure script turned on debugging symbols by default - or
at least, it said it did due to 21.5.b16 being a beta...
JCM> XEmacs 21.5-b16 "celeriac" configured for
`i386-pc-solaris2.9'.
Can you update from CVS? [1] (If not, let me know and I'll get a patch
to you that will bring your XEmacs up to date.) It's been a fairly
long time since -b16 was released and there's been quite a number of
changes made since then.
JCM> Using Motif dialog boxes.
JCM> Using Motif native widgets.
This always frightens me. Is it just a coincidence that "Motif"
contains two thirds of "Not"? :-)
Yeah, and that it's a bastard child of HP, IBM, Sun and few other
now non-existent companies trying to emulate Windows3.11 ;| Gtk is
a whole heap nicer but for some reason which I couldn't be bothered
getting figured out yesterday configure wasn't finding my gtk libs.
Anyway, pulled current from cvs, built it -- still got the failure
with Snarf-documentation -- but the crash described earlier did not
occur. Which is definitely nice ;)
If you're tracking this issue in a db somewhere then I think you
can close it off with all due speed.
thanks again,
James C. McPherson