On Nov 28, 2007 2:09 AM, Michael Sperber <sperber(a)deinprogramm.de> wrote:
Thanks for hanging in there!
"Vin Shelton" <acs(a)alumni.princeton.edu> writes:
> Here's the latest error for a 21.4 non-mule build:
> [...]
> Symbol's function definition is void: defclass
> xemacs exiting
I believe this is because 21.4 doesn't have this patch yet:
http://calypso.tux.org/pipermail/xemacs-patches/2007-October/000924.html
Will a compiled .elc run in a 21.4 without that patch? Otherwise,
you've introduced a backward incompatibility that's very serious, and
we'd have to discuss how to fix this.
> Loading
/usr/local/build/xemacs-packages-2007-11-27/xemacs-packages/semantic/auto-autoloads.el...
You need to have compiled semantic (a dependency) before doing
this---byte-compiling its auto-autoloads.el will remove the missing
`function-overload-p'.
Actually, IMO *you* need to fix this in the build procedure. 'make'
at the top level must take care of this, but I think it already does,
no?
> Perhaps you should try building all the packages from clean CVS
> sources yourself?
I did, about 200 times, but it's so difficult to get consistent results
with Lisp's crazy compilation infrastructure. Just now, I found and
fixed a dependency problem with the ede package, which cogre
transitively depends on---you shouldn't even have gotten to the point
you did.
Granted that things could be more robust in this area, but the nightly
package build can help in this area. (Thank you, Mats.)
I'm running a clean build with the patched ede Makefile using 21.5
right now. I may or may not have results before I fly home in a few
hours.
Regards,
Vin
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-patches