-----Original Message----- From: xemacs-beta-admin(a)xemacs.org [mailto:xemacs-beta-admin＠xemacs.org]On Behalf Of William M. Perry Sent: Wednesday, December 12, 2001 9:00 PM To: xemacs-beta(a)xemacs.org Cc: xemacs-patches(a)xemacs.org Subject: [PATCH] generalizing redisplay_output_display_block() One of the things that confused me when doing the GTK work initially was that xxx_output_display_block was 99% the same on all the various backends. The only thing different were calls to output the horizontal lines, eol cursor, and blank areas. This patch fixes the problem for X11 & GTK by introducing a few new methods to the device/console structure: output_eol_cursor(struct window *w, struct display_line *dl, int xpos, face_index findex) output_blank(struct window *w, struct display_line *dl, struct rune *rb, int start_pixpos, int cursor_start, int cursor_width) output_horizontal_line(struct window *w, struct display_line *dl, struct rune *rb) The X and GTK builds both run and seem to display things quite nicely with this. Layouts, text, toolbars, gutters, etc. The mswindows stuff is partially done, but not in this patch. The cursor handling and output_string would need to be modified more than I am comfortable with at the moment. The TTY redisplay would need to change more than I am comfortable with as well. This falls back and uses the device-specific output_display_block() method if it is available. So this code uses the generic code for the X display, but tty specific routines on the TTY. Comments? Vetoes? Anybody willing to fix up the MSW and TTY backends to deal with this? This would allow the TTY code to spit out at least the text strings from layouts, etc, if it exposed output_string(). -bp -- Ceterum censeo vi esse delendam

I've just completed some more merging... I'll send out a patch later today, but basically toolbar-(x|gtk).c were using some functions directory out of redisplay-(x|gtk).c instead of going thru the common redisplay routines and device methods. Changing them to do this was pretty simple, and I've verified that X & GTK toolbars work as advertised. The big plus is that if someone ports to a new OS and doesn't use native toolbars (which is more difficult), they can just link with this file. I'm attaching the new file if people want to look at it. I'll probably commit the GTK and common code but not futz with the X stuff until I hear from others. I am trying to get as much code between GTK & X shared as possible so that the odds of bitrot are slimmer. I think the event loop is next, then maybe redisplay.