John Whitley <whitley(a)cse.Buffalo.EDU> writes in xemacs-beta(a)xemacs.org:
Someone buried in birdtracks wrote:
That would be me.
> >> Yes, but not with ChangeLogs in specific. Anything
font-locked takes
> >> longer to be font-locked after a revert because it appears to be doing
> >> it twice. I thought I fixed this at some point, but it does not
> >> appear to be working optimally now.
SL Baur writes:
> The code that sets major modes (and indirectly fires off font-lock) is
> twisted at best. Everything gets run at least twice, but there is
> very careful dance done to ensure some things don't get run twice.
Hah! No sooner was this posted than it bit me on the rear.
Tuareg mode, for editing Objective Caml programs, includes an
extension to font-lock called "sym-lock" which converts certain
keywords into atomic symbols under XEmacs.
With sym-lock enabled, vc-register causes the error message:
Symbol's function definition is void: t
Excellent! Can you please make Lisp backtraces for both of the times
the hook gets called?
In order to get the most information in the backtrace, evaluate in a
buffer each lisp file that appears (ie. make sure the backtrace
contains only stuff evaluated from .el's instead of .elc's). Thanks.