Mercurial > emacs
annotate src/blockinput.h @ 18291:130a48e6cc13
(ffap-soft-value): Make this a function again; the macro
version does intern-soft too early. Deleted XEmacs-specific code.
(ffap-string-at-point-mode-alist): added "=" and
"&" to the url syntax, as suggested by SJE.
(ffap-read-file-or-url): fixed the HIST argument to
completing-read (only visible in XEmacs?), as reported by
Christoph Wedler <wedler@fmi.uni-passau.de>.
(ffap-kpathsea-expand-path) New func, replaces ffap-add-subdirs,
a first attempt at kpathsea emulation. Also convert "" to "." in
path lists, for XEmacs. Suggestions from SJE.
Added mouse-track support (but no binding), as
suggested by MDB. Moved Emacs mouse bindings from
"down-mouse" events to ordinary mouse events.
(ffap-alist): added ffap-fortran-mode, as requested by MDB.
Rewrote and merged XEmacs support, eliminating file
ffap-xe.el. Modified ffap-other-frame to work in dedicated
frames, fixing a bug reported by JENS.
(ffap-menu-rescan): avoid modifying the buffer.
Two bugs reported by Christoph Wedler <wedler@fmi.uni-passau.de>:
(ffap-fixup-url): avoid autoloading through url-normalize-url.
(ffap-read-file-or-url): for XEmacs, give extra HACK-HOMEDIR arg
to `abbreviate-file-name'.
(ffap-file-at-point): suppress errors from `ffap-alist'.
(ffap-url-at-point): modified regexp to accept
mail hostnames ending with a digit. Fixes bug report of SJE.
(ffap-url-at-point): use higher level function
(w3-view-this-url t) suggested by wmperry, instead of
w3-zone-at/w3-zone-data or widget-at/widget-get.
(ffap-url-at-point): modified to work with
w3-version "WWW 2.3.64 1996/06/02 06:20:23" alpha, which
uses the 'widget package rather than the old w3-zone-at.
Bug was reported by JENS.
Adopted comments and doc strings to Emacs coding
conventions. Reorganized. Retired v18 support.
(ffap-bindings): Offers a default installation.
(ffap-string-at-point): Modified arguments.
(ffap-gnus-hook): Updated for Gnus 5.
(ffap-tex-init): Delayed initialization of `ffap-tex-path'.
(ffap-dired): New entry in `ffap-alist'.
(ffap-menu-rescan): May fontify the choices in buffer.
(ffap-read-file-or-url): `PC-completion-as-file-name-predicate'
used if available, to work with complete.el.
author | Karl Heuer <kwzh@gnu.org> |
---|---|
date | Wed, 18 Jun 1997 04:24:37 +0000 |
parents | ee40177f6c68 |
children | fb4c986db0e2 |
rev | line source |
---|---|
2439
b6c62e4abf59
Put interrupt input blocking in a separate file from xterm.h.
Jim Blandy <jimb@redhat.com>
parents:
2437
diff
changeset
|
1 /* blockinput.h - interface to blocking complicated interrupt-driven input. |
2437 | 2 Copyright (C) 1989, 1993 Free Software Foundation, Inc. |
3 | |
4 This file is part of GNU Emacs. | |
5 | |
6 GNU Emacs is free software; you can redistribute it and/or modify | |
7 it under the terms of the GNU General Public License as published by | |
8 the Free Software Foundation; either version 2, or (at your option) | |
9 any later version. | |
10 | |
11 GNU Emacs is distributed in the hope that it will be useful, | |
12 but WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 GNU General Public License for more details. | |
15 | |
16 You should have received a copy of the GNU General Public License | |
17 along with GNU Emacs; see the file COPYING. If not, write to | |
14186
ee40177f6c68
Update FSF's address in the preamble.
Erik Naggum <erik@naggum.no>
parents:
2721
diff
changeset
|
18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
ee40177f6c68
Update FSF's address in the preamble.
Erik Naggum <erik@naggum.no>
parents:
2721
diff
changeset
|
19 Boston, MA 02111-1307, USA. */ |
2437 | 20 |
21 | |
22 /* When Emacs is using signal-driven input, the processing of those | |
23 input signals can get pretty hairy. For example, when Emacs is | |
24 running under X windows, handling an input signal can entail | |
25 retrieving events from the X event queue, or making other X calls. | |
26 | |
27 If an input signal occurs while Emacs is in the midst of some | |
28 non-reentrant code, and the signal processing invokes that same | |
29 code, we lose. For example, malloc and the Xlib functions aren't | |
30 usually re-entrant, and both are used by the X input signal handler | |
31 - if we try to process an input signal in the midst of executing | |
32 any of these functions, we'll lose. | |
33 | |
34 To avoid this, we make the following requirements: | |
35 | |
36 * Everyone must evaluate BLOCK_INPUT before entering these functions, | |
37 and then call UNBLOCK_INPUT after performing them. Calls | |
38 BLOCK_INPUT and UNBLOCK_INPUT may be nested. | |
39 | |
40 * Any complicated interrupt handling code should test | |
41 interrupt_input_blocked, and put off its work until later. | |
42 | |
43 * If the interrupt handling code wishes, it may set | |
44 interrupt_input_pending to a non-zero value. If that flag is set | |
45 when input becomes unblocked, UNBLOCK_INPUT will send a new SIGIO. */ | |
46 | |
2479
6d011b24f6fe
(interrupt_input_blocked): Make this signed int.
Richard M. Stallman <rms@gnu.org>
parents:
2477
diff
changeset
|
47 extern int interrupt_input_blocked; |
2437 | 48 |
49 /* Nonzero means an input interrupt has arrived | |
50 during the current critical section. */ | |
51 extern int interrupt_input_pending; | |
52 | |
53 /* Begin critical section. */ | |
54 #define BLOCK_INPUT (interrupt_input_blocked++) | |
55 | |
2721
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
56 /* End critical section. |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
57 |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
58 If doing signal-driven input, and a signal came in when input was |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
59 blocked, reinvoke the signal handler now to deal with it. |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
60 |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
61 We used to have two possible definitions of this macro - one for |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
62 when SIGIO was #defined, and one for when it wasn't; when SIGIO |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
63 wasn't #defined, we wouldn't bother to check if we should re-invoke |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
64 the signal handler. But that doesn't work very well; some of the |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
65 files which use this macro don't #include the right files to get |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
66 SIGIO. |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
67 |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
68 So, we always test interrupt_input_pending now; that's not too |
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
69 expensive, and it'll never get set if we don't need to resignal. */ |
2437 | 70 #define UNBLOCK_INPUT \ |
71 (interrupt_input_blocked--, \ | |
72 (interrupt_input_blocked < 0 ? (abort (), 0) : 0), \ | |
2477
4d0d858bc33f
(UNBLOCK_INPUT): Fix typo.
Richard M. Stallman <rms@gnu.org>
parents:
2439
diff
changeset
|
73 ((interrupt_input_blocked == 0 && interrupt_input_pending != 0) \ |
2721
ca5cb6cafc40
* blockinput.h (UNBLOCK_INPUT): We cannot assume that SIGIO is
Jim Blandy <jimb@redhat.com>
parents:
2479
diff
changeset
|
74 ? (reinvoke_input_signal (), 0) \ |
2437 | 75 : 0)) |
76 | |
77 #define TOTALLY_UNBLOCK_INPUT (interrupt_input_blocked = 0) | |
78 #define UNBLOCK_INPUT_RESIGNAL UNBLOCK_INPUT |