Mercurial > hg > xemacs-beta
annotate src/strcpy.c @ 844:047d37eb70d7
[xemacs-hg @ 2002-05-16 13:30:23 by ben]
ui fixes for things that were bothering me
bytecode.c, editfns.c, lisp.h, lread.c: Fix save-restriction to use markers rather than pseudo-markers
(integers representing the amount of text on either side of the
region). That way, all inserts are handled correctly, not just
those inside old restriction.
Add buffer argument to save_restriction_save().
process.c: Clean up very dirty and kludgy code that outputs into a buffer --
use proper unwind protects, etc.
font-lock.c: Do save-restriction/widen around the function -- otherwise, incorrect
results will ensue when a buffer has been narrowed before a call to
e.g. `buffer-syntactic-context' -- something that happens quite often.
fileio.c: Look for a handler for make-temp-name.
window.c, winslots.h: Try to solve this annoying problem: have two frames displaying the
buffer, in different places; in one, temporarily switch away to
another buffer and then back -- and you've lost your position;
it's reset to the other one in the other frame. My current
solution involves window-level caches of buffers and points (also
a cache for window-start); when set-window-buffer is called, it
looks to see if the buffer was previously visited in the window,
and if so, uses the most recent point at that time. (It's a
marker, so it handles changes.)
#### Note: It could be argued that doing it on the frame level
would be better -- e.g. if you visit a buffer temporarily through
a grep, and then go back to that buffer, you presumably want the
grep's position rather than some previous position provided
everything was in the same frame, even though the grep was in
another window in the frame. However, doing it on the frame level
fails when you have two windows on the same frame. Perhaps we
keep both a window and a frame cache, and use the frame cache if
there are no other windows on the frame showing the buffer, else
the window's cache? This is probably something to be configurable
using a specifier. Suggestions please please please?
window.c: Clean up a bit code that deals with the annoyance of window-point
vs. point.
dialog.el: Function to ask a
multiple-choice question, automatically choosing a dialog box or
minibuffer representation as necessary. Generalized version of
yes-or-no-p, y-or-n-p.
files.el: Use get-user-response to ask "yes/no/diff" question when recovering.
"diff" means that a diff is displayed between the current file and the
autosave. (Converts/deconverts escape-quoted as necessary. No more
complaints from you, Mr. Turnbull!) One known problem: when a dialog
is used, it's modal, so you can't scroll the diff. Will fix soon.
lisp-mode.el: If we're filling a string, don't treat semicolon as a comment,
which would give very unfriendly results.
Uses `buffer-syntactic-context'.
simple.el: all changes back to the beginning. (Useful if you've saved the file
in the middle of the changes.)
simple.el: Add option kill-word-into-kill-ring, which controls whether words
deleted with kill-word, backward-kill-word, etc. are "cut" into the
kill ring, or "cleared" into nothingness. (My preference is the
latter, by far. I'd almost go so far as suggesting we make it the
default, as you can always select a word and then cut it if you want
it cut.)
menubar-items.el: Add option corresponding to kill-word-into-kill-ring.
author | ben |
---|---|
date | Thu, 16 May 2002 13:30:58 +0000 |
parents | abe6d1db359e |
children |
rev | line source |
---|---|
428 | 1 /* This file is part of XEmacs. |
2 | |
3 XEmacs is free software; you can redistribute it and/or modify it | |
4 under the terms of the GNU General Public License as published by the | |
5 Free Software Foundation; either version 2, or (at your option) any | |
6 later version. | |
7 | |
8 XEmacs is distributed in the hope that it will be useful, but WITHOUT | |
9 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
10 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
11 for more details. | |
12 | |
13 You should have received a copy of the GNU General Public License | |
14 along with XEmacs; see the file COPYING. If not, write to | |
15 the Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
16 Boston, MA 02111-1307, USA. */ | |
17 | |
18 /* Synched up with: Not in FSF. */ | |
19 | |
20 /* In SunOS 4.1.1 the strcpy function references memory past the last byte of | |
21 the string! This will core dump if the memory following the last byte is | |
22 not mapped. | |
23 | |
24 Here are correct versions by hbs@lucid.com. | |
25 */ | |
26 | |
27 # include <config.h> | |
28 # ifndef REGISTER /* Strictly enforced in 20.3 */ | |
29 # define REGISTER | |
30 # endif | |
31 | |
32 #define ALIGNED(x) (!(((unsigned long) (x)) & (sizeof (unsigned long) - 1))) | |
33 | |
34 #define MAGIC 0x7efefeff | |
35 #define HIGH_BIT_P(c) ((c) & hi_bit) | |
36 #define HAS_ZERO(c) (((((c) + magic) ^ (c)) & not_magic) != not_magic) | |
37 | |
38 char * | |
39 strcpy (char *to, const char *from) | |
40 { | |
41 char *return_value = to; | |
42 if (to == from) | |
43 return to; | |
44 else if (ALIGNED (to) && ALIGNED (from)) | |
45 { | |
46 unsigned long *to1 = (unsigned long *) to; | |
47 const unsigned long *from1 = (const unsigned long *) from; | |
48 unsigned long c; | |
49 unsigned long magic = MAGIC; | |
50 unsigned long not_magic = ~magic; | |
51 /* unsigned long hi_bit = 0x80000000; */ | |
52 | |
53 while ((c = *from1) != 0) | |
54 { | |
55 if (HAS_ZERO(c)) | |
56 { | |
57 to = (char *) to1; | |
442 | 58 from = (const char *) from1; |
428 | 59 goto slow_loop; |
60 } | |
61 else | |
62 { | |
63 *to1 = c; | |
64 to1++; | |
65 from1++; | |
66 } | |
67 } | |
68 | |
69 to = (char *) to1; | |
70 *to = (char) 0; | |
71 return return_value; | |
72 } | |
73 else | |
74 { | |
75 char c; | |
76 | |
77 slow_loop: | |
78 | |
79 while ((c = *from) != 0) | |
80 { | |
81 *to = c; | |
82 to++; | |
83 from++; | |
84 } | |
85 *to = (char) 0; | |
86 } | |
87 return return_value; | |
88 } |