rewrite dynarr code -------------------- ChangeLog entries follow: -------------------- src/ChangeLog addition: 2010-02-03 Ben Wing <ben@xemacs.org> * device-x.c (x_get_resource_prefix): * device-x.c (Fx_get_resource): * device-x.c (Fx_get_resource_prefix): * device-x.c (Fx_put_resource): * dialog-msw.c: * dialog-msw.c (handle_question_dialog_box): * dired-msw.c (mswindows_sort_files): * dired-msw.c (mswindows_get_files): * extents.c (extent_fragment_sort_by_priority): * extents.c (Fset_extent_parent): * file-coding.c (coding_reader): * file-coding.c (coding_writer): * file-coding.c (gzip_convert): * frame.c (generate_title_string): * gutter.c (calculate_gutter_size_from_display_lines): * indent.c (vmotion_1): * lread.c (read_bit_vector): * mule-coding.c (iso2022_decode): * rangetab.c: * rangetab.c (Fcopy_range_table): * rangetab.c (Fget_range_table): * rangetab.c (unified_range_table_copy_data): * redisplay-msw.c (mswindows_output_string): * redisplay-output.c (output_display_line): * redisplay-output.c (redisplay_move_cursor): * redisplay-output.c (redisplay_clear_bottom_of_window): * redisplay-tty.c (tty_output_ichar_dynarr): * redisplay-tty.c (set_foreground_to): * redisplay-tty.c (set_background_to): * redisplay-xlike-inc.c (XLIKE_output_string): * redisplay.c (redisplay_window_text_width_string): * redisplay.c (redisplay_text_width_string): * redisplay.c (create_text_block): * redisplay.c (SET_CURRENT_MODE_CHARS_PIXSIZE): * redisplay.c (generate_fstring_runes): * redisplay.c (regenerate_modeline): * redisplay.c (ensure_modeline_generated): * redisplay.c (real_current_modeline_height): * redisplay.c (create_string_text_block): * redisplay.c (regenerate_window): * redisplay.c (REGEN_INC_FIND_START_END): * redisplay.c (point_visible): * redisplay.c (redisplay_window): * redisplay.c (mark_glyph_block_dynarr): * redisplay.c (line_start_cache_start): * redisplay.c (start_with_line_at_pixpos): * redisplay.c (update_line_start_cache): * redisplay.c (glyph_to_pixel_translation): * redisplay.c (pixel_to_glyph_translation): * sysdep.c (qxe_readdir): * text.c (dfc_convert_to_external_format): * text.c (dfc_convert_to_internal_format): * toolbar-common.c (common_output_toolbar_button): * window.c (window_modeline_height): * window.c (Fwindow_last_line_visible_height): * window.c (window_displayed_height): * window.c (window_scroll): * window.c (get_current_pixel_pos): Use Dynarr_begin() in place of Dynarr_atp (foo, 0). * dynarr.c (Dynarr_realloc): * dynarr.c (Dynarr_lisp_realloc): * dynarr.c (Dynarr_resize): * dynarr.c (Dynarr_insert_many): * dynarr.c (Dynarr_delete_many): * dynarr.c (Dynarr_memory_usage): * dynarr.c (stack_like_malloc): * dynarr.c (stack_like_free): * lisp.h: * lisp.h (DECLARE_DYNARR_LISP_IMP): * lisp.h (XD_DYNARR_DESC): * lisp.h (Dynarr_pop): * gutter.c (output_gutter): * redisplay-output.c (sync_rune_structs): * redisplay-output.c (redisplay_output_window): Redo the dynarr code, add greater checks. Rename the `len', `largest' and `max' members to `len_', `largest_' and `max_' to try and catch existing places that might directly modify these values. Make new accessors Dynarr_largest() and Dynarr_max() and make them and existing Dynarr_length() be non-lvalues by adding '+ 0' to them; fix a couple of places in the redisplay code that tried to modify the length directly by setting Dynarr_length(). Use the accessors whenever possible even in the dynarr code itself. The accessors also verify that 0 <= len <= largest <= max. Rename settor function Dynarr_set_size() to Dynarr_set_length() and use it more consistently; also create lower-level Dynarr_set_length_1(). This latter function should be the only function that directly modifies the `len_' member of a Dynarr, and in the process makes sure that the `largest' value is kept correct. Consistently use ERROR_CHECK_STRUCTURES instead of ERROR_CHECK_TYPES for error-checking code. Reintroduce the temporarily disabled verification code on the positions of Dynarr_at(), Dynarr_atp() and Dynarr_atp_past_end(). Also create Dynarr_resize_if() in place of a repeated code fragment. Clean up all the functions that modify Dynarrs to use the new macros and functions and verify the correctness of the Dynarr both before and after the change. Note that there are two kinds of verification -- one for accessing and one for modifying. The difference is that the modify verification additionally checks to make sure that the Dynarr isn't locked. (This is used in redisplay to check for problems with reentrancy.) * lrecord.h: Move XD_DYNARR_DESC to lisp.h, grouping with the dynarr code.

/* code is all from loser.c and loser.el by Mly

Copyright (C) 2002 Richard Mlynarik <mly@pobox.com>

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., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301, USA.

Commentary:

Compile this file.  Run it in the background giving it a command line
argument PORT which is a positive integer 1024 < PORT < 32768 (avoid the
numbers assigned in /etc/services).

Then start up a fresh (you're going to crash) XEmacs.  Execute the following

(defun lose (port)
  (interactive "nUrk: ")
  (require 'comint)
  (while t
    (condition-case e
        (let* ((name "*lose*")
	       (b (get-buffer-create name)))
          (switch-to-buffer b)
          (comint-mode)
          (comint-exec b name (cons "127.0.0.1" port) nil '())
          (process-send-string (get-buffer-process b) "\377\373\001")
          (process-send-string (get-buffer-process b) "\377\373\001"))
      (error (message "URK: %s" e)) (sit-for 1))))

Then M-x lose RET PORT RET and you lose big (in XEmacs 21.1, anyway).
Note: the error messages are proper functioning.  What should eventually
happen after a number of SIGPIPEs is that you get a SIGSEGV and life is
bad and XEmacs is dead.
*/

#include <arpa/inet.h>

int
main (int argc, char **argv)
{
  struct sockaddr_in junk;
  int s;

  memset (&junk, 0, sizeof (junk));

  junk.sin_family = AF_INET;
  junk.sin_addr.s_addr = htonl (INADDR_ANY); /* un*x sucks */
  junk.sin_port = htons (atoi (argv[1])); /* un*x blows */

  s = socket (PF_INET, SOCK_STREAM, 0);

  bind (s, (struct sockaddr *)&junk, sizeof (junk));

  listen (s, 1);
  
  for (;;)
  {
    int loser = accept (s, NULL, 0);
    close (loser);
  }
}