kkm> Some time ago, Stephen J. Turnbull wrote... kkm> Stephen, what comment are you referring to? There's a comment at line 6281 if (preemption_check) { /* The preemption check itself takes a lot of time, so normally don't do it here. We do it if called from Lisp, though (`redisplay-frame'). */ int preempted; REDISPLAY_PREEMPTION_CHECK; if (preempted) return 1; } An identical check is done in redisplay_device() at l 6457, basically the first thing that redisplay_device() does: REDISPLAY_PREEMPTION_CHECK; if (preempted) return 1; kkm> I may have modified the file, so line numbers are off. The lines in your patch were exact, so I assumed redisplay.c is the same in 21.4.3 and 21.5 (the above excerpts are from 21.4.3). kkm> because there was obvious nonsense in the above lines: As I say, it appears to be intentional. A misjudgement, I think after light testing, but intentional. kkm> r_f() never gets preempted, and thus never returns non-zero when kkm> its second parameter equal 0. kkm> That is, preemption never happened! CVS revision 1.1.1.1, kkm> the oldest one, also has the preemption check turned off. AFAICT it happens once per call of redisplay_device(). kkm> What I meant by testing was that I wanted everybody to keep kkm> an eye open on glitches in display appearance. Oh I understood that. What I meant is that "count(everybody)" is a lot bigger on xemacs-beta. Everybody on xemacs-beta (pretty much) can read patches; with that and the description they can probably give it a lot more thorough workout than the few on xemacs-review. Especially with the commit-for-review policy, it's important to keep the testers aware of changes where the developers don't fully understand the consequences. kkm> Thank you for checking it, in any case! I'm in swap now, and VM still seems faster. :-) I like this patch, a lot! -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."