annotate src/s/dgux4.h @ 69579:f53a6a5430a7

Remove @headitem as it is not yet supported on SourceForge.
author Bill Wohler <wohler@newt.com>
date Sun, 19 Mar 2006 20:46:20 +0000
parents 3bd95f4f2941
children e90d04cd455a c5406394f567
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
1 /* Definitions file for GNU Emacs running on Data General's DG/UX
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
2 Release 4.10 and above.
68651
3bd95f4f2941 Update years in copyright notice; nfc.
Thien-Thi Nguyen <ttn@gnuvola.org>
parents: 64769
diff changeset
3 Copyright (C) 1996, 2002, 2003, 2004, 2005,
3bd95f4f2941 Update years in copyright notice; nfc.
Thien-Thi Nguyen <ttn@gnuvola.org>
parents: 64769
diff changeset
4 2006 Free Software Foundation, Inc.
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
5
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
6 This file is part of GNU Emacs.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
7
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
8 GNU Emacs is free software; you can redistribute it and/or modify
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
9 it under the terms of the GNU General Public License as published by
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
10 the Free Software Foundation; either version 2, or (at your option)
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
11 any later version.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
12
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
13 GNU Emacs is distributed in the hope that it will be useful,
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
16 GNU General Public License for more details.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
17
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
18 You should have received a copy of the GNU General Public License
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
19 along with GNU Emacs; see the file COPYING. If not, write to
64092
e055d02d8b97 Update FSF's address.
Lute Kamstra <lute@gnu.org>
parents: 52401
diff changeset
20 the Free Software Foundation, 51 Franklin Street, Fifth Floor,
e055d02d8b97 Update FSF's address.
Lute Kamstra <lute@gnu.org>
parents: 52401
diff changeset
21 Boston, MA 02110-1301, USA. */
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
22
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
23 /* This file was written by Roderick Schertler <roderick@ibcinc.com>,
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
24 contact me if you have problems with or comments about running Emacs
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
25 on dgux.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
26
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
27 A number of things in the older dgux*.h files don't make sense to me,
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
28 but since I'm relying on memory and I don't have any older dgux
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
29 systems installed on which to test changes I'm undoing or fixing them
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
30 here rather than fixing them at the source. */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
31
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
32 /* In dgux.h it says "Can't use sys_signal because then etc/server.c
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
33 would need sysdep.o." and then it #defines signal() to be
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
34 berk_signal(), but emacsserver.c does `#undef signal' anyway, so that
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
35 doesn't make sense.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
36
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
37 Further, sys_signal() in sysdep.c already had a special case for
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
38 #ifdef DGUX, it called berk_signal() explicitly. I've removed that
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
39 special case because it also didn't make sense: All versions of dgux
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
40 which the dgux*.h headers take into account have POSIX signals
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
41 (POSIX_SIGNALS is #defined in dgux.h). The comments in sys_signal()
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
42 even acknowledged this (saying that the special berk_signal() case
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
43 wasn't really necessary), they said that sys_signal() was using
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
44 berk_signal() instead of sigaction() for efficiency. Since both give
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
45 reliable signals neither has to be invoked within the handler. If
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
46 the efficiency that the comments were talking about is the overhead
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
47 of setting up the sigaction struct rather than just passing the
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
48 function pointer in (which is the only efficiency I can think of)
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
49 then that's a needless optimization, the Emacs sources do better
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
50 without the special case.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
51
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
52 The following definition will prevent dgux.h from re-defining
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
53 signal(). I can't just say `#undef signal' after including dgux.h
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
54 because signal() is already a macro, defined in <sys/signal.h>, and
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
55 the original definition would be lost. */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
56 #define NO_DGUX_SIGNAL_REDEF
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
57
20002
e89847e2df84 Fix name of include file to match 1996-08-24 renaming.
Karl Heuer <kwzh@gnu.org>
parents: 18097
diff changeset
58 #include "dgux5-4-3.h"
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
59
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
60 #define LIBS_DEBUG /* nothing, -lg doesn't exist */
20289
d9e4922a5ace (LIBS_SYSTEM): Define this.
Karl Heuer <kwzh@gnu.org>
parents: 20002
diff changeset
61 #define LIBS_SYSTEM -lsocket -lnsl
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
62
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
63 #ifndef NOT_C_CODE
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
64
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
65 /* dgux.h defines _setjmp() to be sigsetjmp(), but it defines _longjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
66 to be longjmp() rather than siglongjmp(). Further, it doesn't define
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
67 jmp_buf, so sigsetjmp() is being called with a jmp_buf rather than a
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
68 sigjmp_buf, and the buffer is then passed to vanilla longjmp(). This
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
69 provides a more complete emulation of the Berkeley semantics. */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
70
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
71 #include <setjmp.h>
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
72 #undef jmp_buf
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
73 #undef _setjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
74 #undef setjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
75 #undef _longjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
76 #undef longjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
77 #define jmp_buf sigjmp_buf
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
78 #define _setjmp(env) sigsetjmp(env, 0)
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
79 #define setjmp(env) sigsetjmp(env, 1)
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
80 #define _longjmp siglongjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
81 #define longjmp siglongjmp
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
82
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
83 /* The BAUD_CONVERT definition in dgux.h is wrong with this version
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
84 of dgux, but I'm not sure when it changed.
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
85
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
86 With the current system Emacs' standard handling of ospeed and
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
87 baud_rate don't work. The baud values (B9600 and so on) returned by
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
88 cfgetospeed() aren't compatible with those used by ospeed. speed_t,
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
89 the type returned by cfgetospeed(), is unsigned long and speed_t
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
90 values are large. Further, it isn't possible to get at both the
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
91 SysV3 (ospeed) and POSIX (cfgetospeed()) values through symbolic
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
92 constants simultaneously because they both use the same names
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
93 (B9600). To get both baud_rate and ospeed right at the same time
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
94 it's necessary to hardcode the values for one set of values, here I'm
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
95 hardcoding ospeed. */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
96 #undef BAUD_CONVERT
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
97 #define INIT_BAUD_RATE() \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
98 struct termios sg; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
99 \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
100 tcgetattr (input_fd, &sg); \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
101 switch (cfgetospeed (&sg)) { \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
102 case B50: baud_rate = 50; ospeed = 0x1; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
103 case B75: baud_rate = 75; ospeed = 0x2; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
104 case B110: baud_rate = 110; ospeed = 0x3; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
105 case B134: baud_rate = 134; ospeed = 0x4; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
106 case B150: baud_rate = 150; ospeed = 0x5; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
107 case B200: baud_rate = 200; ospeed = 0x6; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
108 case B300: baud_rate = 300; ospeed = 0x7; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
109 case B600: baud_rate = 600; ospeed = 0x8; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
110 default: \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
111 case B1200: baud_rate = 1200; ospeed = 0x9; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
112 case B1800: baud_rate = 1800; ospeed = 0xa; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
113 case B2400: baud_rate = 2400; ospeed = 0xb; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
114 case B4800: baud_rate = 4800; ospeed = 0xc; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
115 case B9600: baud_rate = 9600; ospeed = 0xd; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
116 case B19200: baud_rate = 19200; ospeed = 0xe; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
117 case B38400: baud_rate = 38400; ospeed = 0xf; break; \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
118 } \
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
119 return;
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
120
23695
a9640ecf5ab4 (BROKEN_FIONREAD): #undef commented out.
Karl Heuer <kwzh@gnu.org>
parents: 20289
diff changeset
121
a9640ecf5ab4 (BROKEN_FIONREAD): #undef commented out.
Karl Heuer <kwzh@gnu.org>
parents: 20289
diff changeset
122 #if 0 /* Ehud Karni <ehud@unix.simonwiesel.co.il> says that the problem
a9640ecf5ab4 (BROKEN_FIONREAD): #undef commented out.
Karl Heuer <kwzh@gnu.org>
parents: 20289
diff changeset
123 still exists on m88k-dg-dguxR4.11MU04 and i586-dg-dguxR4.11MU04. */
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
124 /* The `stop on tty output' problem which occurs when using
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
125 INTERRUPT_INPUT and when Emacs is invoked under X11 using a job
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
126 control shell (csh, ksh, etc.) in the background doesn't look to be
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
127 present in R4.11. (At least, I can't reproduce it using jsh, csh,
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
128 ksh or zsh.) */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
129 #undef BROKEN_FIONREAD
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
130 #define INTERRUPT_INPUT
23695
a9640ecf5ab4 (BROKEN_FIONREAD): #undef commented out.
Karl Heuer <kwzh@gnu.org>
parents: 20289
diff changeset
131 #endif /* 0 - never */
18097
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
132
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
133 /* In R4.11 (or maybe R4.10, I don't have a system with that version
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
134 loaded) some of the internal stdio semantics were changed. One I
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
135 found while working on MH is that _cnt has to be 0 before _filbuf()
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
136 is called. Another is that (_ptr - _base) doesn't indicate how many
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
137 characters are waiting to be sent. I can't spot a good way to get
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
138 that info from the FILE internals. */
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
139 #define PENDING_OUTPUT_COUNT(FILE) (1)
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
140
ce2004da4ccd Initial revision
Richard M. Stallman <rms@gnu.org>
parents:
diff changeset
141 #endif /* NOT_C_CODE */
52401
695cf19ef79e Add arch taglines
Miles Bader <miles@gnu.org>
parents: 23695
diff changeset
142
695cf19ef79e Add arch taglines
Miles Bader <miles@gnu.org>
parents: 23695
diff changeset
143 /* arch-tag: c7013e7b-6e2e-44f2-ba61-90b6d5e2ea45
695cf19ef79e Add arch taglines
Miles Bader <miles@gnu.org>
parents: 23695
diff changeset
144 (do not change this comment) */