(speedbar-frame-mode): Only set the position on window systems.
[bpt/emacs.git] / PROBLEMS
1 This file describes various problems that have been encountered
2 in compiling, installing and running GNU Emacs.
3
4 * `Pid xxx killed due to text modification or page I/O error'
5
6 On HP/UX, you can get that error when the Emacs executable is on an NFS
7 file system. HP/UX responds this way if it tries to swap in a page and
8 does not get a response from the server within a timeout whose default
9 value is just ten seconds.
10
11 If this happens to you, extend the timeout period.
12
13 * `expand-file-name' fails to work on any but the machine you dumped Emacs on.
14
15 On Ultrix, if you use any of the functions which look up information
16 in the passwd database before dumping Emacs (say, by using
17 expand-file-name in site-init.el), then those functions will not work
18 in the dumped Emacs on any host but the one Emacs was dumped on.
19
20 The solution? Don't use expand-file-name in site-init.el, or in
21 anything it loads. Yuck - some solution.
22
23 I'm not sure why this happens; if you can find out exactly what is
24 going on, and perhaps find a fix or a workaround, please let us know.
25 Perhaps the YP functions cache some information, the cache is included
26 in the dumped Emacs, and is then inaccurate on any other host.
27
28 * On some variants of SVR4, Emacs does not work at all with X.
29
30 Try defining BROKEN_FIONREAD in your config.h file. If this solves
31 the problem, please send a bug report to tell us this is needed; be
32 sure to say exactly what type of machine and system you are using.
33
34 * Linking says that the functions insque and remque are undefined.
35
36 Change oldXMenu/Makefile by adding insque.o to the variable OBJS.
37
38 * Emacs fails to understand most Internet host names, even though
39 the names work properly with other programs on the same system.
40
41 This typically happens on Suns and other systems that use shared
42 libraries. The cause is that the site has installed a version of the
43 shared library which uses a name server--but has not installed a
44 similar version of the unshared library which Emacs uses.
45
46 The result is that most programs, using the shared library, work with
47 the nameserver, but Emacs does not.
48
49 The fix is to install an unshared library that corresponds to what you
50 installed in the shared library, and then relink Emacs.
51
52 * On a Sun running SunOS 4.1.1, you get this error message from GNU ld:
53
54 /lib/libc.a(_Q_sub.o): Undefined symbol __Q_get_rp_rd referenced from text segment
55
56 The problem is in the Sun shared C library, not in GNU ld.
57
58 The solution is to install Patch-ID# 100267-03 from Sun.
59
60 * Self documentation messages are garbled.
61
62 This means that the file `etc/DOC-...' doesn't properly correspond
63 with the Emacs executable. Redumping Emacs and then installing the
64 corresponding pair of files should fix the problem.
65
66 * Trouble using ptys on AIX.
67
68 People often install the pty devices on AIX incorrectly.
69 Use `smit pty' to reinstall them properly.
70
71 * Shell mode on HP/UX gives the message, "`tty`: Ambiguous".
72
73 christos@theory.tn.cornell.edu says:
74
75 The problem is that in your .cshrc you have something that tries to
76 execute `tty`. If you are not running the shell on a real tty then
77 tty will print "not a tty". Csh expects one word in some places,
78 but tty is giving it back 3.
79
80 The solution is to add a pair of quotes around `tty` to make it a single
81 word:
82
83 if (`tty` == "/dev/console")
84
85 should be changed to:
86
87 if ("`tty`" == "/dev/console")
88
89 Even better, move things that set up terminal sections out of .cshrc
90 and into .login.
91
92 * Using X Windows, control-shift-leftbutton makes Emacs hang.
93
94 Use the shell command `xset bc' to make the old X Menu package work.
95
96 * Emacs running under X Windows does not handle mouse clicks.
97 * `emacs -geometry 80x20' finds a file named `80x20'.
98
99 One cause of such problems is having (setq term-file-prefix nil) in
100 your .emacs file. Another cause is a bad value of EMACSLOADPATH in
101 the environment.
102
103 * Emacs starts in a directory other than the one that is current in the shell.
104
105 If the PWD environment variable exists, Emacs uses this variable as
106 the initial working directory.
107
108 Some shells automatically update this variable, while other shells fail
109 to do so. If you use two such shells in combination, the variable can
110 end up wrong. This confuses Emacs.
111
112 The solution is to put something in the start-up file for the shell
113 that does not update PWD, to get rid of that environment variable.
114 For example, in csh, use `unsetenv PWD'.
115
116 * Emacs gets error message from linker on Sun.
117
118 If the error message says that a symbol such as `f68881_used' or
119 `ffpa_used' or `start_float' is undefined, this probably indicates
120 that you have compiled some libraries, such as the X libraries,
121 with a floating point option other than the default.
122
123 It's not terribly hard to make this work with small changes in
124 crt0.c together with linking with Fcrt1.o, Wcrt1.o or Mcrt1.o.
125 However, the easiest approach is to build Xlib with the default
126 floating point option: -fsoft.
127
128 * Emacs fails to get default settings from X Windows server.
129
130 The X library in X11R4 has a bug; it interchanges the 2nd and 3rd
131 arguments to XGetDefaults. Define the macro XBACKWARDS in config.h to
132 tell Emacs to compensate for this.
133
134 I don't believe there is any way Emacs can determine for itself
135 whether this problem is present on a given system.
136
137 * Keyboard input gets confused after a beep when using a DECserver
138 as a concentrator.
139
140 This problem seems to be a matter of configuring the DECserver to use
141 7 bit characters rather than 8 bit characters.
142
143 * M-x shell persistently reports "Process shell exited abnormally with code 1".
144
145 This happened on Suns as a result of what is said to be a bug in Sunos
146 version 4.0.x. The only fix was to reboot the machine.
147
148 * Programs running under terminal emulator do not recognize `emacs'
149 terminal type.
150
151 The cause of this is a shell startup file that sets the TERMCAP
152 environment variable. The terminal emulator uses that variable to
153 provide the information on the special terminal type that Emacs
154 emulates.
155
156 Rewrite your shell startup file so that it does not change TERMCAP
157 in such a case. You could use the following conditional which sets
158 it only if it is undefined.
159
160 if ( ! ${?TERMCAP} ) setenv TERMCAP ~/my-termcap-file
161
162 Or you could set TERMCAP only when you set TERM--which should not
163 happen in a non-login shell.
164
165 * X Windows doesn't work if DISPLAY uses a hostname.
166
167 People have reported kernel bugs in certain systems that cause Emacs
168 not to work with X Windows if DISPLAY is set using a host name. But
169 the problem does not occur if DISPLAY is set to `unix:0.0'. I think
170 the bug has to do with SIGIO or FIONREAD.
171
172 You may be able to compensate for the bug by doing (set-input-mode nil nil).
173 However, that has the disadvantage of turning off interrupts, so that
174 you are unable to quit out of a Lisp program by typing C-g.
175
176 The easy way to do this is to put
177
178 (setq x-sigio-bug t)
179
180 in your site-init.el file.
181
182 * Problem with remote X server on Suns.
183
184 On a Sun, running Emacs on one machine with the X server on another
185 may not work if you have used the unshared system libraries. This
186 is because the unshared libraries fail to use YP for host name lookup.
187 As a result, the host name you specify may not be recognized.
188
189 * Watch out for .emacs files and EMACSLOADPATH environment vars
190
191 These control the actions of Emacs.
192 ~/.emacs is your Emacs init file.
193 EMACSLOADPATH overrides which directories the function
194 "load" will search.
195
196 If you observe strange problems, check for these and get rid
197 of them, then try again.
198
199 * Shell mode ignores interrupts on Apollo Domain
200
201 You may find that M-x shell prints the following message:
202
203 Warning: no access to tty; thus no job control in this shell...
204
205 This can happen if there are not enough ptys on your system.
206 Here is how to make more of them.
207
208 % cd /dev
209 % ls pty*
210 # shows how many pty's you have. I had 8, named pty0 to pty7)
211 % /etc/crpty 8
212 # creates eight new pty's
213
214 * Fatal signal in the command temacs -l loadup inc dump
215
216 This command is the final stage of building Emacs. It is run by the
217 Makefile in the src subdirectory, or by build.com on VMS.
218
219 It has been known to get fatal errors due to insufficient swapping
220 space available on the machine.
221
222 On 68000's, it has also happened because of bugs in the
223 subroutine `alloca'. Verify that `alloca' works right, even
224 for large blocks (many pages).
225
226 * test-distrib says that the distribution has been clobbered
227 * or, temacs prints "Command key out of range 0-127"
228 * or, temacs runs and dumps xemacs, but xemacs totally fails to work.
229 * or, temacs gets errors dumping xemacs
230
231 This can be because the .elc files have been garbled. Do not be
232 fooled by the fact that most of a .elc file is text: these are
233 binary files and can contain all 256 byte values.
234
235 In particular `shar' cannot be used for transmitting GNU Emacs.
236 It typically truncates "lines". What appear to be "lines" in
237 a binary file can of course be of any length. Even once `shar'
238 itself is made to work correctly, `sh' discards null characters
239 when unpacking the shell archive.
240
241 I have also seen character \177 changed into \377. I do not know
242 what transfer means caused this problem. Various network
243 file transfer programs are suspected of clobbering the high bit.
244
245 If you have a copy of Emacs that has been damaged in its
246 nonprinting characters, you can fix them:
247
248 1) Record the names of all the .elc files.
249 2) Delete all the .elc files.
250 3) Recompile alloc.c with a value of PURESIZE twice as large.
251 You might as well save the old alloc.o.
252 4) Remake xemacs. It should work now.
253 5) Running xemacs, do Meta-x byte-compile-file repeatedly
254 to recreate all the .elc files that used to exist.
255 You may need to increase the value of the variable
256 max-lisp-eval-depth to succeed in running the compiler interpreted
257 on certain .el files. 400 was sufficient as of last report.
258 6) Reinstall the old alloc.o (undoing changes to alloc.c if any)
259 and remake temacs.
260 7) Remake xemacs. It should work now, with valid .elc files.
261
262 * temacs prints "Pure Lisp storage exhausted"
263
264 This means that the Lisp code loaded from the .elc and .el
265 files during temacs -l loadup inc dump took up more
266 space than was allocated.
267
268 This could be caused by
269 1) adding code to the preloaded Lisp files
270 2) adding more preloaded files in loadup.el
271 3) having a site-init.el or site-load.el which loads files.
272 Note that ANY site-init.el or site-load.el is nonstandard;
273 if you have received Emacs from some other site
274 and it contains a site-init.el or site-load.el file, consider
275 deleting that file.
276 4) getting the wrong .el or .elc files
277 (not from the directory you expected).
278 5) deleting some .elc files that are supposed to exist.
279 This would cause the source files (.el files) to be
280 loaded instead. They take up more room, so you lose.
281 6) a bug in the Emacs distribution which underestimates
282 the space required.
283
284 If the need for more space is legitimate, change the definition
285 of PURESIZE in puresize.h.
286
287 But in some of the cases listed above, this problem is a consequence
288 of something else that is wrong. Be sure to check and fix the real
289 problem.
290
291 * Changes made to .el files do not take effect.
292
293 You may have forgotten to recompile them into .elc files.
294 Then the old .elc files will be loaded, and your changes
295 will not be seen. To fix this, do M-x byte-recompile-directory
296 and specify the directory that contains the Lisp files.
297
298 Emacs should print a warning when loading a .elc file which is older
299 than the corresponding .el file.
300
301 * The dumped Emacs (xemacs) crashes when run, trying to write pure data.
302
303 Two causes have been seen for such problems.
304
305 1) On a system where getpagesize is not a system call, it is defined
306 as a macro. If the definition (in both unexec.c and malloc.c) is wrong,
307 it can cause problems like this. You might be able to find the correct
308 value in the man page for a.out (5).
309
310 2) Some systems allocate variables declared static among the
311 initialized variables. Emacs makes all initialized variables in most
312 of its files pure after dumping, but the variables declared static and
313 not initialized are not supposed to be pure. On these systems you
314 may need to add "#define static" to the m- or the s- file.
315
316 * Compilation errors on VMS.
317
318 You will get warnings when compiling on VMS because there are
319 variable names longer than 32 (or whatever it is) characters.
320 This is not an error. Ignore it.
321
322 VAX C does not support #if defined(foo). Uses of this construct
323 were removed, but some may have crept back in. They must be rewritten.
324
325 There is a bug in the C compiler which fails to sign extend characters
326 in conditional expressions. The bug is:
327 char c = -1, d = 1;
328 int i;
329
330 i = d ? c : d;
331 The result is i == 255; the fix is to typecast the char in the
332 conditional expression as an (int). Known occurrences of such
333 constructs in Emacs have been fixed.
334
335 * rmail gets error getting new mail
336
337 rmail gets new mail from /usr/spool/mail/$USER using a program
338 called `movemail'. This program interlocks with /bin/mail using
339 the protocol defined by /bin/mail.
340
341 There are two different protocols in general use. One of them uses
342 the `flock' system call. The other involves creating a lock file;
343 `movemail' must be able to write in /usr/spool/mail in order to do
344 this. You control which one is used by defining, or not defining,
345 the macro MAIL_USE_FLOCK in config.h or the m- or s- file it includes.
346 IF YOU DON'T USE THE FORM OF INTERLOCKING THAT IS NORMAL ON YOUR
347 SYSTEM, YOU CAN LOSE MAIL!
348
349 If your system uses the lock file protocol, and fascist restrictions
350 prevent ordinary users from writing the lock files in /usr/spool/mail,
351 you may need to make `movemail' setgid to a suitable group such as
352 `mail'. You can use these commands (as root):
353
354 chgrp mail movemail
355 chmod 2755 movemail
356
357 * Emacs won't work with X-windows if the value of DISPLAY is HOSTNAME:0.
358 * GNUs can't make contact with the specified host for nntp.
359
360 Some people have found that Emacs was unable to connect to the local
361 host by name, as in DISPLAY=prep:0 if you are running on prep, but
362 could handle DISPLAY=unix:0. Here is what tale@rpi.edu said:
363
364 Seems as
365 though gethostbyname was bombing somewhere along the way. Well, we
366 had just upgrade from SunOS 3.5 (which X11 was built under) to SunOS
367 4.0.1. Any new X applications which tried to be built with the pre
368 OS-upgrade libraries had the same problems which Emacs was having.
369 Missing /etc/resolv.conf for a little while (when one of the libraries
370 was built?) also might have had a hand in it.
371
372 The result of all of this (with some speculation) was that we rebuilt
373 X and then rebuilt Emacs with the new libraries. Works as it should
374 now. Hoorah.
375
376 If you have already installed the name resolver in the file libresolv.a,
377 then you need to compile Emacs to use that library. The easiest way to
378 do this is to add to config.h a definition of LIBS_SYSTEM, LIBS_MACHINE
379 or LIB_STANDARD which uses -lresolv. Watch out! If you redefine a macro
380 that is already in use in your configuration to supply some other libraries,
381 be careful not to lose the others.
382
383 Thus, you could start by adding this to config.h:
384
385 #define LIBS_SYSTEM -lresolv
386
387 Then if this gives you an error for redefining a macro, and you see that
388 the s- file defines LIBS_SYSTEM as -lfoo -lbar, you could change config.h
389 again to say this:
390
391 #define LIBS_SYSTEM -lresolv -lfoo -lbar
392
393 * Emacs spontaneously displays "I-search: " at the bottom of the screen.
394
395 This means that Control-S/Control-Q "flow control" is being used.
396 C-s/C-q flow control is bad for Emacs editors because it takes away
397 C-s and C-q as user commands. Since editors do not output long streams
398 of text without user commands, there is no need for a user-issuable
399 "stop output" command in an editor; therefore, a properly designed
400 flow control mechanism would transmit all possible input characters
401 without interference. Designing such a mechanism is easy, for a person
402 with at least half a brain.
403
404 There are three possible reasons why flow control could be taking place:
405
406 1) Terminal has not been told to disable flow control
407 2) Insufficient padding for the terminal in use
408 3) Some sort of terminal concentrator or line switch is responsible
409
410 First of all, many terminals have a set-up mode which controls
411 whether they generate flow control characters. This must be
412 set to "no flow control" in order for Emacs to work. Sometimes
413 there is an escape sequence that the computer can send to turn
414 flow control off and on. If so, perhaps the termcap `ti' string
415 should turn flow control off, and the `te' string should turn it on.
416
417 Once the terminal has been told "no flow control", you may find it
418 needs more padding. The amount of padding Emacs sends is controlled
419 by the termcap entry for the terminal in use, and by the output baud
420 rate as known by the kernel. The shell command `stty' will print
421 your output baud rate; `stty' with suitable arguments will set it if
422 it is wrong. Setting to a higher speed causes increased padding. If
423 the results are wrong for the correct speed, there is probably a
424 problem in the termcap entry. You must speak to a local Unix wizard
425 to fix this. Perhaps you are just using the wrong terminal type.
426
427 For terminals that lack a "no flow control" mode, sometimes just
428 giving lots of padding will prevent actual generation of flow control
429 codes. You might as well try it.
430
431 If you are really unlucky, your terminal is connected to the computer
432 through a concentrator which sends flow control to the computer, or it
433 insists on sending flow control itself no matter how much padding you
434 give it. You are screwed! You should replace the terminal or
435 concentrator with a properly designed one. In the mean time,
436 some drastic measures can make Emacs semi-work.
437
438 One drastic measure to ignore C-s and C-q, while sending enough
439 padding that the terminal will not really lose any output. To make
440 such an adjustment, you need only invoke the function
441 enable-flow-control-on with a list of terminal types in your own
442 .emacs file. As arguments, give it the names of one or more terminal
443 types you use which require flow control adjustments.
444 Here's an example:
445
446 (enable-flow-control-on "vt200" "vt300" "vt101" "vt131")
447
448 An even more drastic measure is to make Emacs use flow control.
449 To do this, evaluate the Lisp expression (set-input-mode nil t).
450 Emacs will then interpret C-s and C-q as flow control commands. (More
451 precisely, it will allow the kernel to do so as it usually does.) You
452 will lose the ability to use them for Emacs commands. Also, as a
453 consequence of using CBREAK mode, the terminal's Meta-key, if any,
454 will not work, and C-g will be liable to cause a loss of output which
455 will produce garbage on the screen. (These problems apply to 4.2BSD;
456 they may not happen in 4.3 or VMS, and I don't know what would happen
457 in sysV.) You can use keyboard-translate-table, as shown above,
458 to map two other input characters (such as C-^ and C-\) into C-s and
459 C-q, so that you can still search and quote.
460
461 I have no intention of ever redesigning the Emacs command set for
462 the assumption that terminals use C-s/C-q flow control. This
463 flow control technique is a bad design, and terminals that need
464 it are bad merchandise and should not be purchased. If you can
465 get some use out of GNU Emacs on inferior terminals, I am glad,
466 but I will not make Emacs worse for properly designed systems
467 for the sake of inferior systems.
468
469 * Control-S and Control-Q commands are ignored completely.
470
471 For some reason, your system is using brain-damaged C-s/C-q flow
472 control despite Emacs's attempts to turn it off. Perhaps your
473 terminal is connected to the computer through a concentrator
474 that wants to use flow control.
475
476 You should first try to tell the concentrator not to use flow control.
477 If you succeed in this, try making the terminal work without
478 flow control, as described in the preceding section.
479
480 If that line of approach is not successful, map some other characters
481 into C-s and C-q using keyboard-translate-table. The example above
482 shows how to do this with C-^ and C-\.
483
484 * Control-S and Control-Q commands are ignored completely on a net connection.
485
486 Some versions of rlogin (and possibly telnet) do not pass flow
487 control characters to the remote system to which they connect.
488 On such systems, emacs on the remote system cannot disable flow
489 control on the local system.
490
491 One way to cure this is to disable flow control on the local host
492 (the one running rlogin, not the one running rlogind) using the
493 stty command, before starting the rlogin process. On many systems,
494 "stty start u stop u" will do this.
495
496 Some versions of tcsh will prevent even this from working. One way
497 around this is to start another shell before starting rlogin, and
498 issue the stty command to disable flow control from that shell.
499
500 * Screen is updated wrong, but only on one kind of terminal.
501
502 This could mean that the termcap entry you are using for that
503 terminal is wrong, or it could mean that Emacs has a bug handing
504 the combination of features specified for that terminal.
505
506 The first step in tracking this down is to record what characters
507 Emacs is sending to the terminal. Execute the Lisp expression
508 (open-termscript "./emacs-script") to make Emacs write all
509 terminal output into the file ~/emacs-script as well; then do
510 what makes the screen update wrong, and look at the file
511 and decode the characters using the manual for the terminal.
512 There are several possibilities:
513
514 1) The characters sent are correct, according to the terminal manual.
515
516 In this case, there is no obvious bug in Emacs, and most likely you
517 need more padding, or possibly the terminal manual is wrong.
518
519 2) The characters sent are incorrect, due to an obscure aspect
520 of the terminal behavior not described in an obvious way
521 by termcap.
522
523 This case is hard. It will be necessary to think of a way for
524 Emacs to distinguish between terminals with this kind of behavior
525 and other terminals that behave subtly differently but are
526 classified the same by termcap; or else find an algorithm for
527 Emacs to use that avoids the difference. Such changes must be
528 tested on many kinds of terminals.
529
530 3) The termcap entry is wrong.
531
532 See the file etc/TERMS for information on changes
533 that are known to be needed in commonly used termcap entries
534 for certain terminals.
535
536 4) The characters sent are incorrect, and clearly cannot be
537 right for any terminal with the termcap entry you were using.
538
539 This is unambiguously an Emacs bug, and can probably be fixed
540 in termcap.c, tparam.c, term.c, scroll.c, cm.c or dispnew.c.
541
542 * Output from Control-V is slow.
543
544 On many bit-map terminals, scrolling operations are fairly slow.
545 Often the termcap entry for the type of terminal in use fails
546 to inform Emacs of this. The two lines at the bottom of the screen
547 before a Control-V command are supposed to appear at the top after
548 the Control-V command. If Emacs thinks scrolling the lines is fast,
549 it will scroll them to the top of the screen.
550
551 If scrolling is slow but Emacs thinks it is fast, the usual reason is
552 that the termcap entry for the terminal you are using does not
553 specify any padding time for the `al' and `dl' strings. Emacs
554 concludes that these operations take only as much time as it takes to
555 send the commands at whatever line speed you are using. You must
556 fix the termcap entry to specify, for the `al' and `dl', as much
557 time as the operations really take.
558
559 Currently Emacs thinks in terms of serial lines which send characters
560 at a fixed rate, so that any operation which takes time for the
561 terminal to execute must also be padded. With bit-map terminals
562 operated across networks, often the network provides some sort of
563 flow control so that padding is never needed no matter how slow
564 an operation is. You must still specify a padding time if you want
565 Emacs to realize that the operation takes a long time. This will
566 cause padding characters to be sent unnecessarily, but they do
567 not really cost much. They will be transmitted while the scrolling
568 is happening and then discarded quickly by the terminal.
569
570 Most bit-map terminals provide commands for inserting or deleting
571 multiple lines at once. Define the `AL' and `DL' strings in the
572 termcap entry to say how to do these things, and you will have
573 fast output without wasted padding characters. These strings should
574 each contain a single %-spec saying how to send the number of lines
575 to be scrolled. These %-specs are like those in the termcap
576 `cm' string.
577
578 You should also define the `IC' and `DC' strings if your terminal
579 has a command to insert or delete multiple characters. These
580 take the number of positions to insert or delete as an argument.
581
582 A `cs' string to set the scrolling region will reduce the amount
583 of motion you see on the screen when part of the screen is scrolled.
584
585 * Your Delete key sends a Backspace to the terminal, using an AIXterm.
586
587 The solution is to include in your .Xdefaults the lines:
588
589 *aixterm.Translations: #override <Key>BackSpace: string(0x7f)
590 aixterm*ttyModes: erase ^?
591
592 This makes your Backspace key send DEL (ASCII 127).
593
594 * You type Control-H (Backspace) expecting to delete characters.
595
596 Put `stty dec' in your .login file and your problems will disappear
597 after a day or two.
598
599 The choice of Backspace for erasure was based on confusion, caused by
600 the fact that backspacing causes erasure (later, when you type another
601 character) on most display terminals. But it is a mistake. Deletion
602 of text is not the same thing as backspacing followed by failure to
603 overprint. I do not wish to propagate this confusion by conforming
604 to it.
605
606 For this reason, I believe `stty dec' is the right mode to use,
607 and I have designed Emacs to go with that. If there were a thousand
608 other control characters, I would define Control-h to delete as well;
609 but there are not very many other control characters, and I think
610 that providing the most mnemonic possible Help character is more
611 important than adapting to people who don't use `stty dec'.
612
613 If you are obstinate about confusing buggy overprinting with deletion,
614 you can redefine Backspace in your .emacs file:
615 (global-set-key "\b" 'delete-backward-char)
616 You may then wish to put the function help-command on some
617 other key. I leave to you the task of deciding which key.
618
619 * Editing files through RFS gives spurious "file has changed" warnings.
620 It is possible that a change in Emacs 18.37 gets around this problem,
621 but in case not, here is a description of how to fix the RFS bug that
622 causes it.
623
624 There was a serious pair of bugs in the handling of the fsync() system
625 call in the RFS server.
626
627 The first is that the fsync() call is handled as another name for the
628 close() system call (!!). It appears that fsync() is not used by very
629 many programs; Emacs version 18 does an fsync() before closing files
630 to make sure that the bits are on the disk.
631
632 This is fixed by the enclosed patch to the RFS server.
633
634 The second, more serious problem, is that fsync() is treated as a
635 non-blocking system call (i.e., it's implemented as a message that
636 gets sent to the remote system without waiting for a reply). Fsync is
637 a useful tool for building atomic file transactions. Implementing it
638 as a non-blocking RPC call (when the local call blocks until the sync
639 is done) is a bad idea; unfortunately, changing it will break the RFS
640 protocol. No fix was supplied for this problem.
641
642 (as always, your line numbers may vary)
643
644 % rcsdiff -c -r1.2 serversyscall.c
645 RCS file: RCS/serversyscall.c,v
646 retrieving revision 1.2
647 diff -c -r1.2 serversyscall.c
648 *** /tmp/,RCSt1003677 Wed Jan 28 15:15:02 1987
649 --- serversyscall.c Wed Jan 28 15:14:48 1987
650 ***************
651 *** 163,169 ****
652 /*
653 * No return sent for close or fsync!
654 */
655 ! if (syscall == RSYS_close || syscall == RSYS_fsync)
656 proc->p_returnval = deallocate_fd(proc, msg->m_args[0]);
657 else
658 {
659 --- 166,172 ----
660 /*
661 * No return sent for close or fsync!
662 */
663 ! if (syscall == RSYS_close)
664 proc->p_returnval = deallocate_fd(proc, msg->m_args[0]);
665 else
666 {
667
668 * Vax C compiler bugs affecting Emacs.
669
670 You may get one of these problems compiling Emacs:
671
672 foo.c line nnn: compiler error: no table entry for op STASG
673 foo.c: fatal error in /lib/ccom
674
675 These are due to bugs in the C compiler; the code is valid C.
676 Unfortunately, the bugs are unpredictable: the same construct
677 may compile properly or trigger one of these bugs, depending
678 on what else is in the source file being compiled. Even changes
679 in header files that should not affect the file being compiled
680 can affect whether the bug happens. In addition, sometimes files
681 that compile correctly on one machine get this bug on another machine.
682
683 As a result, it is hard for me to make sure this bug will not affect
684 you. I have attempted to find and alter these constructs, but more
685 can always appear. However, I can tell you how to deal with it if it
686 should happen. The bug comes from having an indexed reference to an
687 array of Lisp_Objects, as an argument in a function call:
688 Lisp_Object *args;
689 ...
690 ... foo (5, args[i], ...)...
691 putting the argument into a temporary variable first, as in
692 Lisp_Object *args;
693 Lisp_Object tem;
694 ...
695 tem = args[i];
696 ... foo (r, tem, ...)...
697 causes the problem to go away.
698 The `contents' field of a Lisp vector is an array of Lisp_Objects,
699 so you may see the problem happening with indexed references to that.
700
701 * 68000 C compiler problems
702
703 Various 68000 compilers have different problems.
704 These are some that have been observed.
705
706 ** Using value of assignment expression on union type loses.
707 This means that x = y = z; or foo (x = z); does not work
708 if x is of type Lisp_Object.
709
710 ** "cannot reclaim" error.
711
712 This means that an expression is too complicated. You get the correct
713 line number in the error message. The code must be rewritten with
714 simpler expressions.
715
716 ** XCONS, XSTRING, etc macros produce incorrect code.
717
718 If temacs fails to run at all, this may be the cause.
719 Compile this test program and look at the assembler code:
720
721 struct foo { char x; unsigned int y : 24; };
722
723 lose (arg)
724 struct foo arg;
725 {
726 test ((int *) arg.y);
727 }
728
729 If the code is incorrect, your compiler has this problem.
730 In the XCONS, etc., macros in lisp.h you must replace (a).u.val with
731 ((a).u.val + coercedummy) where coercedummy is declared as int.
732
733 This problem will not happen if the m-...h file for your type
734 of machine defines NO_UNION_TYPE. That is the recommended setting now.
735
736 * C compilers lose on returning unions
737
738 I hear that some C compilers cannot handle returning a union type.
739 Most of the functions in GNU Emacs return type Lisp_Object, which is
740 defined as a union on some rare architectures.
741
742 This problem will not happen if the m-...h file for your type
743 of machine defines NO_UNION_TYPE.
744