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