On Fri, Oct 01, 1999 at 11:39:01PM +0200, Jan Vroonhof wrote:
Isn't the problem simply that UNBOUNDP fails on the variable although
it is unbound? Does pdumping change the address of Qunbound without
changing the values of unbound variables?
Nicely spotted. Actually the problem was rather _not_ changing the
address of Qunbound :-)
It goes further now:
*** Error in XEmacs initialization
(error "No such face" bold)
*** Backtrace
really-early-error-handler((error "No such face" bold))
get-face(bold)
# bind (no-fallback default domain property face)
face-property-instance(bold foreground #<stream-device on "stdio"
0x120>)
# bind (domain prop face2 face1)
face-property-equal(bold default foreground #<stream-device on "stdio"
0x120>)
# bind (domain face2 face1 props)
face-equal-loop((foreground background font display-table underline dim) bold default
#<stream-device on "stdio" 0x120>)
# bind (device common-props win-props tty-props domain face2 face1)
face-equal(bold default #<stream-device on "stdio" 0x120>)
# bind (domain face)
face-differs-from-default-p(bold #<stream-device on "stdio" 0x120>)
# bind (device)
init-other-random-faces(#<stream-device on "stdio" 0x120>)
# bind (device)
init-device-faces(#<stream-device on "stdio" 0x120>)
*** Killing XEmacs
OG.
PS: I'm plonking that in the cvs now since it fixes the msw compile
problems.
1999-10-01 Olivier Galibert <galibert(a)pobox.com>
* lisp.h: Add pdump_wire declaration
* elhash.c (vars_of_elhash): Wire Vall_weak_hash_tables.
* symbols.c (init_symbols_once_early): Wire Qnil and Qunbound.
* alloc.c (pdump_wire): Added..
(pdump_load): Support it.
(pdump): Support it
* glyphs-msw.c
(reinit_image_instantiator_format_create_glyphs_mswindows): Fix
macro calls.