view src/devslots.h @ 1303:f99d3d25df86

[xemacs-hg @ 2003-02-15 10:15:54 by ben] autoload fixes, make-doc speed improvements Makefile.in.in: Run update-elc-2 with -no-autoloads to avoid multiple autoload-loading problem. configure.usage: Document quick-build better. make-docfile.el: Use `message' (defined in this file) in place of `princ'/`print', and put in a terpri, so that we get correct newline behavior. Rewrite if-progn -> when and a few similar stylistic niceties. And the big change: Allow MS Windows to specify the object files directly and frob them into C files here (formerly this was done in xemacs.mak, and very slooooooooooooooooooowly). Due to line-length limitations in CMD, we need to use a "response file" to hold the arguments, so when we see a response file argument (preceded by an @), read in the args (a bit of trickiness to do this), and process recursively. Also frob .obj -> .c as mentioned earlier and handle other junk dependencies that need to be removed (NEEDTODUMP, make-docfile.exe). update-elc-2.el: Use :test `equal' in call to set-difference. update-elc.el: Put back commented out kill-emacs, update header comment. xemacs.mak: Delete old unused code that checks SATISFIED. Move update-elc-2 up to be near update-elc. Run update-elc-2 with -no-autoloads to avoid multiple autoload-loading problem. Don't compute make-docfile args ourselves. Pass the raw objects to make-docfile.el, which does the computation (much faster than we could). Don't delete the DOC file, split the invocation into two calls to make-docfile.exe (one direct, one through make-docfile.el), etc. In general, all we do is call make-docfile. Add proper dependencies for DOC-file rebuilding so it doesn't get done when not necessary. Implement quick-building here: not building the DOC file unless it doesn't exist, as the quick-build docs say. Makefile.in.in: Don't delete the DOC file. Implement quick-building here: not building the DOC file unless it doesn't exist, as the quick-build docs say. config.h.in, emacs.c: Nothing but niggly spacing changes -- one space before a paren starting a function-call arglist, please.
author ben
date Sat, 15 Feb 2003 10:16:14 +0000
parents e22b0213b713
children 8b2f75cecb89
line wrap: on
line source

/* Definitions of marked slots in consoles
   Copyright (C) 1990, 1992, 1993 Free Software Foundation, Inc.
   Copyright (C) 2002 Ben Wing.

This file is part of XEmacs.

XEmacs is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.

XEmacs is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
for more details.

You should have received a copy of the GNU General Public License
along with XEmacs; see the file COPYING.  If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.  */

/* Synched up with: Not in FSF. */

/* We define the Lisp_Objects in the device structure in a separate file
   because there are numerous places we want to iterate over them, such
   as when defining them in the structure, initializing them, or marking
   them.

   To use, define MARKED_SLOT before including this file.  No need to
   undefine; that happens automatically. */

  /* Name of this device, for resourcing and printing purposes.
     If not explicitly given, it's initialized in a device-specific
     manner. */
  MARKED_SLOT (name)

  /* What this device is connected to */
  MARKED_SLOT (connection)

  /* A canonical name for the connection that is used to determine
     whether `make-device' is being called on an existing device. */
  MARKED_SLOT (canon_connection)

  /* List of frames on this device. */
  MARKED_SLOT (frame_list)

  /* The console this device is on. */
  MARKED_SLOT (console)

  /* Frame which is "currently selected".  This is what `selected-frame'
     returns and is the default frame for many operations.  This may
     not be the same as frame_with_focus `select-frame' changes the
     selected_frame but not the frame_with_focus.  However, eventually
     either the two values will be the same, or frame_with_focus will
     be nil: right before waiting for an event, the focus is changed
     to point to the selected_frame if XEmacs currently has the focus
     on this device.  Note that frame_with_focus may be nil (none of the
     frames on this device have the window-system focus), but
     selected_frame will never be nil if there are any frames on
     the device. */
  MARKED_SLOT (selected_frame)
  /* Frame that currently contains the window-manager focus, or none.
     Note that we've split frame_with_focus into two variables.
     frame_with_focus_real is the value we use most of the time,
     but frame_with_focus_for_hooks is used for running the select-frame-hook
     and deselect-frame-hook.  We do this because we split the focus handling
     into two parts: one part (deals with drawing the solid/box cursor)
     runs as soon as a focus event is received the other (running the
     hooks) runs after any pending sit-for/sleep-for/accept-process-output
     calls are done. */
  MARKED_SLOT (frame_with_focus_real)
  MARKED_SLOT (frame_with_focus_for_hooks)
  /* If we have recently issued a request to change the focus as a
     result of select-frame having been called, the following variable
     records the frame we are trying to focus on.  The reason for this
     is that the window manager may not grant our request to change
     the focus (so we can't just change frame_with_focus), and we don't
     want to keep sending requests again and again to the window manager.
     This variable is reset whenever a focus-change event is seen. */
  MARKED_SLOT (frame_that_ought_to_have_focus)

  /* Color class of this device. */
  MARKED_SLOT (device_class)

  /* Alist of values for user-defined tags in this device. */
  MARKED_SLOT (user_defined_tags)

  /* Hash tables for device-specific objects (fonts, colors, etc).
     These are key-weak hash tables (or hash tables containing key-weak
     hash tables) so that they disappear when the key goes away. */

  /* This is a simple key-weak hash table hashing color names to
     instances. */
  MARKED_SLOT (color_instance_cache)

  /* This is a simple key-weak hash table hashing font names to
     instances. */
  MARKED_SLOT (font_instance_cache)

#ifdef MULE
  /* This is a bi-level cache, where the hash table in this slot here
     indexes charset objects to key-weak hash tables, which in turn
     index font names to more specific font names that match the
     given charset's registry.  This speeds up the horrendously
     slow XListFonts() operation that needs to be done in order
     to determine an appropriate font. */
  MARKED_SLOT (charset_font_cache_stage_1)

  /* Similar cache for stage 2, if it exists.  See objects.c. */
  MARKED_SLOT (charset_font_cache_stage_2)
#endif

  /* This is a bi-level cache, where the hash table in this slot here
     indexes image-instance-type masks (there are currently 6
     image-instance types and thus 64 possible masks) to key-weak hash
     tables like the one for colors. */
  MARKED_SLOT (image_instance_cache)

#undef MARKED_SLOT