| 1 | Known Problems with GNU Emacs |
| 2 | |
| 3 | Copyright (C) 1987, 1988, 1989, 1993, 1994, 1995, 1996, 1997, 1998, 1999, |
| 4 | 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 |
| 5 | Free Software Foundation, Inc. |
| 6 | See the end of the file for license conditions. |
| 7 | |
| 8 | |
| 9 | This file describes various problems that have been encountered |
| 10 | in compiling, installing and running GNU Emacs. Try doing C-c C-t |
| 11 | and browsing through the outline headers. (See C-h m for help on |
| 12 | Outline mode.) |
| 13 | |
| 14 | * Mule-UCS doesn't work in Emacs 23. |
| 15 | |
| 16 | It's completely redundant now, as far as we know. |
| 17 | |
| 18 | * Emacs startup failures |
| 19 | |
| 20 | ** Emacs fails to start, complaining about missing fonts. |
| 21 | |
| 22 | A typical error message might be something like |
| 23 | |
| 24 | No fonts match `-*-fixed-medium-r-*--6-*-*-*-*-*-iso8859-1' |
| 25 | |
| 26 | This happens because some X resource specifies a bad font family for |
| 27 | Emacs to use. The possible places where this specification might be |
| 28 | are: |
| 29 | |
| 30 | - in your ~/.Xdefaults file |
| 31 | |
| 32 | - client-side X resource file, such as ~/Emacs or |
| 33 | /usr/X11R6/lib/app-defaults/Emacs or |
| 34 | /usr/X11R6/lib/X11/app-defaults/Emacs |
| 35 | |
| 36 | One of these files might have bad or malformed specification of a |
| 37 | fontset that Emacs should use. To fix the problem, you need to find |
| 38 | the problematic line(s) and correct them. |
| 39 | |
| 40 | ** Emacs aborts while starting up, only when run without X. |
| 41 | |
| 42 | This problem often results from compiling Emacs with GCC when GCC was |
| 43 | installed incorrectly. The usual error in installing GCC is to |
| 44 | specify --includedir=/usr/include. Installation of GCC makes |
| 45 | corrected copies of the system header files. GCC is supposed to use |
| 46 | the corrected copies in preference to the original system headers. |
| 47 | Specifying --includedir=/usr/include causes the original system header |
| 48 | files to be used. On some systems, the definition of ioctl in the |
| 49 | original system header files is invalid for ANSI C and causes Emacs |
| 50 | not to work. |
| 51 | |
| 52 | The fix is to reinstall GCC, and this time do not specify --includedir |
| 53 | when you configure it. Then recompile Emacs. Specifying --includedir |
| 54 | is appropriate only in very special cases and it should *never* be the |
| 55 | same directory where system header files are kept. |
| 56 | |
| 57 | ** Emacs does not start, complaining that it cannot open termcap database file. |
| 58 | |
| 59 | If your system uses Terminfo rather than termcap (most modern |
| 60 | systems do), this could happen if the proper version of |
| 61 | ncurses is not visible to the Emacs configure script (i.e. it |
| 62 | cannot be found along the usual path the linker looks for |
| 63 | libraries). It can happen because your version of ncurses is |
| 64 | obsolete, or is available only in form of binaries. |
| 65 | |
| 66 | The solution is to install an up-to-date version of ncurses in |
| 67 | the developer's form (header files, static libraries and |
| 68 | symbolic links); in some GNU/Linux distributions (e.g. Debian) |
| 69 | it constitutes a separate package. |
| 70 | |
| 71 | ** Emacs 20 and later fails to load Lisp files at startup. |
| 72 | |
| 73 | The typical error message might be like this: |
| 74 | |
| 75 | "Cannot open load file: fontset" |
| 76 | |
| 77 | This could happen if you compress the file lisp/subdirs.el. That file |
| 78 | tells Emacs what are the directories where it should look for Lisp |
| 79 | files. Emacs cannot work with subdirs.el compressed, since the |
| 80 | Auto-compress mode it needs for this will not be loaded until later, |
| 81 | when your .emacs file is processed. (The package `fontset.el' is |
| 82 | required to set up fonts used to display text on window systems, and |
| 83 | it's loaded very early in the startup procedure.) |
| 84 | |
| 85 | Similarly, any other .el file for which there's no corresponding .elc |
| 86 | file could fail to load if it is compressed. |
| 87 | |
| 88 | The solution is to uncompress all .el files which don't have a .elc |
| 89 | file. |
| 90 | |
| 91 | Another possible reason for such failures is stale *.elc files |
| 92 | lurking somewhere on your load-path. The following command will |
| 93 | print any duplicate Lisp files that are present in load-path: |
| 94 | |
| 95 | emacs -q -batch -f list-load-path-shadows |
| 96 | |
| 97 | If this command prints any file names, some of these files are stale, |
| 98 | and should be deleted or their directories removed from your |
| 99 | load-path. |
| 100 | |
| 101 | ** Emacs prints an error at startup after upgrading from an earlier version. |
| 102 | |
| 103 | An example of such an error is: |
| 104 | |
| 105 | x-complement-fontset-spec: "Wrong type argument: stringp, nil" |
| 106 | |
| 107 | This can be another symptom of stale *.elc files in your load-path. |
| 108 | The following command will print any duplicate Lisp files that are |
| 109 | present in load-path: |
| 110 | |
| 111 | emacs -q -batch -f list-load-path-shadows |
| 112 | |
| 113 | If this command prints any file names, some of these files are stale, |
| 114 | and should be deleted or their directories removed from your |
| 115 | load-path. |
| 116 | |
| 117 | ** With X11R6.4, public-patch-3, Emacs crashes at startup. |
| 118 | |
| 119 | Reportedly this patch in X fixes the problem. |
| 120 | |
| 121 | --- xc/lib/X11/imInt.c~ Wed Jun 30 13:31:56 1999 |
| 122 | +++ xc/lib/X11/imInt.c Thu Jul 1 15:10:27 1999 |
| 123 | @@ -1,4 +1,4 @@ |
| 124 | -/* $TOG: imInt.c /main/5 1998/05/30 21:11:16 kaleb $ */ |
| 125 | +/* $TOG: imInt.c /main/5 1998/05/30 21:11:16 kaleb $ */ |
| 126 | /****************************************************************** |
| 127 | |
| 128 | Copyright 1992, 1993, 1994 by FUJITSU LIMITED |
| 129 | @@ -166,8 +166,8 @@ |
| 130 | _XimMakeImName(lcd) |
| 131 | XLCd lcd; |
| 132 | { |
| 133 | - char* begin; |
| 134 | - char* end; |
| 135 | + char* begin = NULL; |
| 136 | + char* end = NULL; |
| 137 | char* ret; |
| 138 | int i = 0; |
| 139 | char* ximmodifier = XIMMODIFIER; |
| 140 | @@ -182,7 +182,11 @@ |
| 141 | } |
| 142 | ret = Xmalloc(end - begin + 2); |
| 143 | if (ret != NULL) { |
| 144 | - (void)strncpy(ret, begin, end - begin + 1); |
| 145 | + if (begin != NULL) { |
| 146 | + (void)strncpy(ret, begin, end - begin + 1); |
| 147 | + } else { |
| 148 | + ret[0] = '\0'; |
| 149 | + } |
| 150 | ret[end - begin + 1] = '\0'; |
| 151 | } |
| 152 | return ret; |
| 153 | |
| 154 | ** Emacs crashes on startup after a glibc upgrade. |
| 155 | |
| 156 | This is caused by a binary incompatible change to the malloc |
| 157 | implementation in glibc 2.5.90-22. As a result, Emacs binaries built |
| 158 | using prior versions of glibc crash when run under 2.5.90-22. |
| 159 | |
| 160 | This problem was first seen in pre-release versions of Fedora 7, and |
| 161 | may be fixed in the final Fedora 7 release. To stop the crash from |
| 162 | happening, first try upgrading to the newest version of glibc; if this |
| 163 | does not work, rebuild Emacs with the same version of glibc that you |
| 164 | will run it under. For details, see |
| 165 | |
| 166 | https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=239344 |
| 167 | |
| 168 | * Crash bugs |
| 169 | |
| 170 | ** Emacs crashes when running in a terminal, if compiled with GCC 4.5.0 |
| 171 | This version of GCC is buggy: see |
| 172 | |
| 173 | http://debbugs.gnu.org/cgi/bugreport.cgi?bug=6031 |
| 174 | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43904 |
| 175 | |
| 176 | You can work around this error in gcc-4.5 by omitting sibling call |
| 177 | optimization. To do this, configure Emacs with |
| 178 | |
| 179 | CFLAGS="-g -O2 -fno-optimize-sibling-calls" ./configure |
| 180 | |
| 181 | ** Emacs crashes in x-popup-dialog. |
| 182 | |
| 183 | This can happen if the dialog widget cannot find the font it wants to |
| 184 | use. You can work around the problem by specifying another font with |
| 185 | an X resource--for example, `Emacs.dialog*.font: 9x15' (or any font that |
| 186 | happens to exist on your X server). |
| 187 | |
| 188 | ** Emacs crashes when you use Bibtex mode. |
| 189 | |
| 190 | This happens if your system puts a small limit on stack size. You can |
| 191 | prevent the problem by using a suitable shell command (often `ulimit') |
| 192 | to raise the stack size limit before you run Emacs. |
| 193 | |
| 194 | Patches to raise the stack size limit automatically in `main' |
| 195 | (src/emacs.c) on various systems would be greatly appreciated. |
| 196 | |
| 197 | ** Error message `Symbol's value as variable is void: x', followed by |
| 198 | a segmentation fault and core dump. |
| 199 | |
| 200 | This has been tracked to a bug in tar! People report that tar erroneously |
| 201 | added a line like this at the beginning of files of Lisp code: |
| 202 | |
| 203 | x FILENAME, N bytes, B tape blocks |
| 204 | |
| 205 | If your tar has this problem, install GNU tar--if you can manage to |
| 206 | untar it :-). |
| 207 | |
| 208 | ** Crashes when displaying GIF images in Emacs built with version |
| 209 | libungif-4.1.0 are resolved by using version libungif-4.1.0b1. |
| 210 | Configure checks for the correct version, but this problem could occur |
| 211 | if a binary built against a shared libungif is run on a system with an |
| 212 | older version. |
| 213 | |
| 214 | ** Emacs aborts inside the function `tparam1'. |
| 215 | |
| 216 | This can happen if Emacs was built without terminfo support, but the |
| 217 | terminal's capabilities use format that is only supported by terminfo. |
| 218 | If your system has ncurses installed, this might happen if your |
| 219 | version of ncurses is broken; upgrading to a newer version of ncurses |
| 220 | and reconfiguring and rebuilding Emacs should solve this. |
| 221 | |
| 222 | All modern systems support terminfo, so even if ncurses is not the |
| 223 | problem, you should look for a way to configure Emacs so that it uses |
| 224 | terminfo when built. |
| 225 | |
| 226 | ** Emacs crashes when using some version of the Exceed X server. |
| 227 | |
| 228 | Upgrading to a newer version of Exceed has been reported to prevent |
| 229 | these crashes. You should consider switching to a free X server, such |
| 230 | as Xming or Cygwin/X. |
| 231 | |
| 232 | ** Emacs crashes with SIGSEGV in XtInitializeWidgetClass. |
| 233 | |
| 234 | It crashes on X, but runs fine when called with option "-nw". |
| 235 | |
| 236 | This has been observed when Emacs is linked with GNU ld but without passing |
| 237 | the -z nocombreloc flag. Emacs normally knows to pass the -z nocombreloc |
| 238 | flag when needed, so if you come across a situation where the flag is |
| 239 | necessary but missing, please report it via M-x report-emacs-bug. |
| 240 | |
| 241 | On platforms such as Solaris, you can also work around this problem by |
| 242 | configuring your compiler to use the native linker instead of GNU ld. |
| 243 | |
| 244 | ** Emacs compiled with Gtk+ crashes when closing a display (x-close-connection). |
| 245 | |
| 246 | This happens because of bugs in Gtk+. Gtk+ 2.10 seems to be OK. See bug |
| 247 | http://bugzilla.gnome.org/show_bug.cgi?id=85715. |
| 248 | |
| 249 | ** Emacs compiled with Gtk+ may loop forever if a display crashes. |
| 250 | |
| 251 | This is related to the bug above. A scenario for this is when emacs is run |
| 252 | as a server, and an X frame is created. If the X server for the frame |
| 253 | crashes or exits unexpectedly and an attempt is made to create a new |
| 254 | frame on another X display, then a Gtk+ error happens in the emacs |
| 255 | server that results in an endless loop. This is not fixed in any known |
| 256 | Gtk+ version (2.14.4 being current). |
| 257 | |
| 258 | * General runtime problems |
| 259 | |
| 260 | ** Lisp problems |
| 261 | |
| 262 | *** Changes made to .el files do not take effect. |
| 263 | |
| 264 | You may have forgotten to recompile them into .elc files. |
| 265 | Then the old .elc files will be loaded, and your changes |
| 266 | will not be seen. To fix this, do M-x byte-recompile-directory |
| 267 | and specify the directory that contains the Lisp files. |
| 268 | |
| 269 | Emacs should print a warning when loading a .elc file which is older |
| 270 | than the corresponding .el file. |
| 271 | |
| 272 | *** Watch out for .emacs files and EMACSLOADPATH environment vars. |
| 273 | |
| 274 | These control the actions of Emacs. |
| 275 | ~/.emacs is your Emacs init file. |
| 276 | EMACSLOADPATH overrides which directories the function |
| 277 | "load" will search. |
| 278 | |
| 279 | If you observe strange problems, check for these and get rid |
| 280 | of them, then try again. |
| 281 | |
| 282 | *** Using epop3.el package causes Emacs to signal an error. |
| 283 | |
| 284 | The error message might be something like this: |
| 285 | |
| 286 | "Lisp nesting exceeds max-lisp-eval-depth" |
| 287 | |
| 288 | This happens because epop3 redefines the function gethash, which is a |
| 289 | built-in primitive beginning with Emacs 21.1. We don't have a patch |
| 290 | for epop3 that fixes this, but perhaps a newer version of epop3 |
| 291 | corrects that. |
| 292 | |
| 293 | *** Buffers from `with-output-to-temp-buffer' get set up in Help mode. |
| 294 | |
| 295 | Changes in Emacs 20.4 to the hooks used by that function cause |
| 296 | problems for some packages, specifically BBDB. See the function's |
| 297 | documentation for the hooks involved. BBDB 2.00.06 fixes the problem. |
| 298 | |
| 299 | *** The Hyperbole package causes *Help* buffers not to be displayed in |
| 300 | Help mode due to setting `temp-buffer-show-hook' rather than using |
| 301 | `add-hook'. Using `(add-hook 'temp-buffer-show-hook |
| 302 | 'help-mode-maybe)' after loading Hyperbole should fix this. |
| 303 | |
| 304 | ** Keyboard problems |
| 305 | |
| 306 | *** "Compose Character" key does strange things when used as a Meta key. |
| 307 | |
| 308 | If you define one key to serve as both Meta and Compose Character, you |
| 309 | will get strange results. In previous Emacs versions, this "worked" |
| 310 | in that the key acted as Meta--that's because the older Emacs versions |
| 311 | did not try to support Compose Character. Now Emacs tries to do |
| 312 | character composition in the standard X way. This means that you |
| 313 | must pick one meaning or the other for any given key. |
| 314 | |
| 315 | You can use both functions (Meta, and Compose Character) if you assign |
| 316 | them to two different keys. |
| 317 | |
| 318 | *** C-z just refreshes the screen instead of suspending Emacs. |
| 319 | |
| 320 | You are probably using a shell that doesn't support job control, even |
| 321 | though the system itself is capable of it. Either use a different shell, |
| 322 | or set the variable `cannot-suspend' to a non-nil value. |
| 323 | |
| 324 | *** With M-x enable-flow-control, you need to type C-\ twice |
| 325 | to do incremental search--a single C-\ gets no response. |
| 326 | |
| 327 | This has been traced to communicating with your machine via kermit, |
| 328 | with C-\ as the kermit escape character. One solution is to use |
| 329 | another escape character in kermit. One user did |
| 330 | |
| 331 | set escape-character 17 |
| 332 | |
| 333 | in his .kermrc file, to make C-q the kermit escape character. |
| 334 | |
| 335 | ** Mailers and other helper programs |
| 336 | |
| 337 | *** movemail compiled with POP support can't connect to the POP server. |
| 338 | |
| 339 | Make sure that the `pop' entry in /etc/services, or in the services |
| 340 | NIS map if your machine uses NIS, has the same port number as the |
| 341 | entry on the POP server. A common error is for the POP server to be |
| 342 | listening on port 110, the assigned port for the POP3 protocol, while |
| 343 | the client is trying to connect on port 109, the assigned port for the |
| 344 | old POP protocol. |
| 345 | |
| 346 | *** RMAIL gets error getting new mail. |
| 347 | |
| 348 | RMAIL gets new mail from /usr/spool/mail/$USER using a program |
| 349 | called `movemail'. This program interlocks with /bin/mail using |
| 350 | the protocol defined by /bin/mail. |
| 351 | |
| 352 | There are two different protocols in general use. One of them uses |
| 353 | the `flock' system call. The other involves creating a lock file; |
| 354 | `movemail' must be able to write in /usr/spool/mail in order to do |
| 355 | this. You control which one is used by defining, or not defining, |
| 356 | the macro MAIL_USE_FLOCK in config.h or the m- or s- file it includes. |
| 357 | IF YOU DON'T USE THE FORM OF INTERLOCKING THAT IS NORMAL ON YOUR |
| 358 | SYSTEM, YOU CAN LOSE MAIL! |
| 359 | |
| 360 | If your system uses the lock file protocol, and fascist restrictions |
| 361 | prevent ordinary users from writing the lock files in /usr/spool/mail, |
| 362 | you may need to make `movemail' setgid to a suitable group such as |
| 363 | `mail'. To do this, use the following commands (as root) after doing the |
| 364 | make install. |
| 365 | |
| 366 | chgrp mail movemail |
| 367 | chmod 2755 movemail |
| 368 | |
| 369 | Installation normally copies movemail from the build directory to an |
| 370 | installation directory which is usually under /usr/local/lib. The |
| 371 | installed copy of movemail is usually in the directory |
| 372 | /usr/local/lib/emacs/VERSION/TARGET. You must change the group and |
| 373 | mode of the installed copy; changing the group and mode of the build |
| 374 | directory copy is ineffective. |
| 375 | |
| 376 | *** rcs2log gives you the awk error message "too many fields". |
| 377 | |
| 378 | This is due to an arbitrary limit in certain versions of awk. |
| 379 | The solution is to use gawk (GNU awk). |
| 380 | |
| 381 | ** Problems with hostname resolution |
| 382 | |
| 383 | *** Emacs fails to understand most Internet host names, even though |
| 384 | the names work properly with other programs on the same system. |
| 385 | *** Emacs won't work with X-windows if the value of DISPLAY is HOSTNAME:0. |
| 386 | *** Gnus can't make contact with the specified host for nntp. |
| 387 | |
| 388 | This typically happens on Suns and other systems that use shared |
| 389 | libraries. The cause is that the site has installed a version of the |
| 390 | shared library which uses a name server--but has not installed a |
| 391 | similar version of the unshared library which Emacs uses. |
| 392 | |
| 393 | The result is that most programs, using the shared library, work with |
| 394 | the nameserver, but Emacs does not. |
| 395 | |
| 396 | The fix is to install an unshared library that corresponds to what you |
| 397 | installed in the shared library, and then relink Emacs. |
| 398 | |
| 399 | If you have already installed the name resolver in the file libresolv.a, |
| 400 | then you need to compile Emacs to use that library. The easiest way to |
| 401 | do this is to add to config.h a definition of LIBS_SYSTEM, LIBS_MACHINE |
| 402 | or LIB_STANDARD which uses -lresolv. Watch out! If you redefine a macro |
| 403 | that is already in use in your configuration to supply some other libraries, |
| 404 | be careful not to lose the others. |
| 405 | |
| 406 | Thus, you could start by adding this to config.h: |
| 407 | |
| 408 | #define LIBS_SYSTEM -lresolv |
| 409 | |
| 410 | Then if this gives you an error for redefining a macro, and you see that |
| 411 | the s- file defines LIBS_SYSTEM as -lfoo -lbar, you could change config.h |
| 412 | again to say this: |
| 413 | |
| 414 | #define LIBS_SYSTEM -lresolv -lfoo -lbar |
| 415 | |
| 416 | *** Emacs does not know your host's fully-qualified domain name. |
| 417 | |
| 418 | For example, (system-name) returns some variation on |
| 419 | "localhost.localdomain", rather the name you were expecting. |
| 420 | |
| 421 | You need to configure your machine with a fully qualified domain name, |
| 422 | (i.e. a name with at least one ".") either in /etc/hosts, |
| 423 | /etc/hostname, the NIS, or wherever your system calls for specifying |
| 424 | this. |
| 425 | |
| 426 | If you cannot fix the configuration, you can set the Lisp variable |
| 427 | mail-host-address to the value you want. |
| 428 | |
| 429 | ** NFS and RFS |
| 430 | |
| 431 | *** Emacs says it has saved a file, but the file does not actually |
| 432 | appear on disk. |
| 433 | |
| 434 | This can happen on certain systems when you are using NFS, if the |
| 435 | remote disk is full. It is due to a bug in NFS (or certain NFS |
| 436 | implementations), and there is apparently nothing Emacs can do to |
| 437 | detect the problem. Emacs checks the failure codes of all the system |
| 438 | calls involved in writing a file, including `close'; but in the case |
| 439 | where the problem occurs, none of those system calls fails. |
| 440 | |
| 441 | *** Editing files through RFS gives spurious "file has changed" warnings. |
| 442 | It is possible that a change in Emacs 18.37 gets around this problem, |
| 443 | but in case not, here is a description of how to fix the RFS bug that |
| 444 | causes it. |
| 445 | |
| 446 | There was a serious pair of bugs in the handling of the fsync() system |
| 447 | call in the RFS server. |
| 448 | |
| 449 | The first is that the fsync() call is handled as another name for the |
| 450 | close() system call (!!). It appears that fsync() is not used by very |
| 451 | many programs; Emacs version 18 does an fsync() before closing files |
| 452 | to make sure that the bits are on the disk. |
| 453 | |
| 454 | This is fixed by the enclosed patch to the RFS server. |
| 455 | |
| 456 | The second, more serious problem, is that fsync() is treated as a |
| 457 | non-blocking system call (i.e., it's implemented as a message that |
| 458 | gets sent to the remote system without waiting for a reply). Fsync is |
| 459 | a useful tool for building atomic file transactions. Implementing it |
| 460 | as a non-blocking RPC call (when the local call blocks until the sync |
| 461 | is done) is a bad idea; unfortunately, changing it will break the RFS |
| 462 | protocol. No fix was supplied for this problem. |
| 463 | |
| 464 | (as always, your line numbers may vary) |
| 465 | |
| 466 | % rcsdiff -c -r1.2 serversyscall.c |
| 467 | RCS file: RCS/serversyscall.c,v |
| 468 | retrieving revision 1.2 |
| 469 | diff -c -r1.2 serversyscall.c |
| 470 | *** /tmp/,RCSt1003677 Wed Jan 28 15:15:02 1987 |
| 471 | --- serversyscall.c Wed Jan 28 15:14:48 1987 |
| 472 | *************** |
| 473 | *** 163,169 **** |
| 474 | /* |
| 475 | * No return sent for close or fsync! |
| 476 | */ |
| 477 | ! if (syscall == RSYS_close || syscall == RSYS_fsync) |
| 478 | proc->p_returnval = deallocate_fd(proc, msg->m_args[0]); |
| 479 | else |
| 480 | { |
| 481 | --- 166,172 ---- |
| 482 | /* |
| 483 | * No return sent for close or fsync! |
| 484 | */ |
| 485 | ! if (syscall == RSYS_close) |
| 486 | proc->p_returnval = deallocate_fd(proc, msg->m_args[0]); |
| 487 | else |
| 488 | { |
| 489 | |
| 490 | ** PSGML |
| 491 | |
| 492 | *** Old versions of the PSGML package use the obsolete variables |
| 493 | `before-change-function' and `after-change-function', which are no |
| 494 | longer used by Emacs. Please use PSGML 1.2.3 or later. |
| 495 | |
| 496 | *** PSGML conflicts with sgml-mode. |
| 497 | |
| 498 | PSGML package uses the same names of some variables (like keymap) |
| 499 | as built-in sgml-mode.el because it was created as a replacement |
| 500 | of that package. The conflict will be shown if you load |
| 501 | sgml-mode.el before psgml.el. E.g. this could happen if you edit |
| 502 | HTML page and then start to work with SGML or XML file. html-mode |
| 503 | (from sgml-mode.el) is used for HTML file and loading of psgml.el |
| 504 | (for sgml-mode or xml-mode) will cause an error. |
| 505 | |
| 506 | *** Versions of the PSGML package earlier than 1.0.3 (stable) or 1.1.2 |
| 507 | (alpha) fail to parse DTD files correctly in Emacs 20.3 and later. |
| 508 | Here is a patch for psgml-parse.el from PSGML 1.0.1 and, probably, |
| 509 | earlier versions. |
| 510 | |
| 511 | --- psgml-parse.el 1998/08/21 19:18:18 1.1 |
| 512 | +++ psgml-parse.el 1998/08/21 19:20:00 |
| 513 | @@ -2383,7 +2383,7 @@ (defun sgml-push-to-entity (entity &opti |
| 514 | (setq sgml-buffer-parse-state nil)) |
| 515 | (cond |
| 516 | ((stringp entity) ; a file name |
| 517 | - (save-excursion (insert-file-contents entity)) |
| 518 | + (insert-file-contents entity) |
| 519 | (setq default-directory (file-name-directory entity))) |
| 520 | ((consp (sgml-entity-text entity)) ; external id? |
| 521 | (let* ((extid (sgml-entity-text entity)) |
| 522 | |
| 523 | ** AUCTeX |
| 524 | |
| 525 | You should not be using a version older than 11.52 if you can avoid |
| 526 | it. |
| 527 | |
| 528 | *** Emacs 21 freezes when visiting a TeX file with AUCTeX installed. |
| 529 | |
| 530 | Emacs 21 needs version 10 or later of AUCTeX; upgrading should solve |
| 531 | these problems. |
| 532 | |
| 533 | *** No colors in AUCTeX with Emacs 21. |
| 534 | |
| 535 | Upgrade to AUC TeX version 10 or later, and make sure it is |
| 536 | byte-compiled with Emacs 21. |
| 537 | |
| 538 | ** PCL-CVS |
| 539 | |
| 540 | *** Lines are not updated or new lines are added in the buffer upon commit. |
| 541 | |
| 542 | When committing files located higher in the hierarchy than the examined |
| 543 | directory, some versions of the CVS program return an ambiguous message |
| 544 | from which PCL-CVS cannot extract the full location of the committed |
| 545 | files. As a result, the corresponding lines in the PCL-CVS buffer are |
| 546 | not updated with the new revision of these files, and new lines are |
| 547 | added to the top-level directory. |
| 548 | |
| 549 | This can happen with CVS versions 1.12.8 and 1.12.9. Upgrade to CVS |
| 550 | 1.12.10 or newer to fix this problem. |
| 551 | |
| 552 | ** Miscellaneous problems |
| 553 | |
| 554 | *** Emacs uses 100% of CPU time |
| 555 | |
| 556 | This is a known problem with some versions of the Semantic package. |
| 557 | The solution is to upgrade Semantic to version 2.0pre4 (distributed |
| 558 | with CEDET 1.0pre4) or later. |
| 559 | |
| 560 | *** Self-documentation messages are garbled. |
| 561 | |
| 562 | This means that the file `etc/DOC-...' doesn't properly correspond |
| 563 | with the Emacs executable. Redumping Emacs and then installing the |
| 564 | corresponding pair of files should fix the problem. |
| 565 | |
| 566 | *** Programs running under terminal emulator do not recognize `emacs' |
| 567 | terminal type. |
| 568 | |
| 569 | The cause of this is a shell startup file that sets the TERMCAP |
| 570 | environment variable. The terminal emulator uses that variable to |
| 571 | provide the information on the special terminal type that Emacs |
| 572 | emulates. |
| 573 | |
| 574 | Rewrite your shell startup file so that it does not change TERMCAP |
| 575 | in such a case. You could use the following conditional which sets |
| 576 | it only if it is undefined. |
| 577 | |
| 578 | if ( ! ${?TERMCAP} ) setenv TERMCAP ~/my-termcap-file |
| 579 | |
| 580 | Or you could set TERMCAP only when you set TERM--which should not |
| 581 | happen in a non-login shell. |
| 582 | |
| 583 | *** In Shell mode, you get a ^M at the end of every line. |
| 584 | |
| 585 | This happens to people who use tcsh, because it is trying to be too |
| 586 | smart. It sees that the Shell uses terminal type `unknown' and turns |
| 587 | on the flag to output ^M at the end of each line. You can fix the |
| 588 | problem by adding this to your .cshrc file: |
| 589 | |
| 590 | if ($?EMACS) then |
| 591 | if ("$EMACS" =~ /*) then |
| 592 | unset edit |
| 593 | stty -icrnl -onlcr -echo susp ^Z |
| 594 | endif |
| 595 | endif |
| 596 | |
| 597 | *** Emacs startup on GNU/Linux systems (and possibly other systems) is slow. |
| 598 | |
| 599 | This can happen if the system is misconfigured and Emacs can't get the |
| 600 | full qualified domain name, FQDN. You should have your FQDN in the |
| 601 | /etc/hosts file, something like this: |
| 602 | |
| 603 | 127.0.0.1 localhost |
| 604 | 129.187.137.82 nuc04.t30.physik.tu-muenchen.de nuc04 |
| 605 | |
| 606 | The way to set this up may vary on non-GNU systems. |
| 607 | |
| 608 | *** Attempting to visit remote files via ange-ftp fails. |
| 609 | |
| 610 | If the error message is "ange-ftp-file-modtime: Specified time is not |
| 611 | representable", then this could happen when `lukemftp' is used as the |
| 612 | ftp client. This was reported to happen on Debian GNU/Linux, kernel |
| 613 | version 2.4.3, with `lukemftp' 1.5-5, but might happen on other |
| 614 | systems as well. To avoid this problem, switch to using the standard |
| 615 | ftp client. On a Debian system, type |
| 616 | |
| 617 | update-alternatives --config ftp |
| 618 | |
| 619 | and then choose /usr/bin/netkit-ftp. |
| 620 | |
| 621 | *** JPEG images aren't displayed. |
| 622 | |
| 623 | This has been reported when Emacs is built with jpeg-6a library. |
| 624 | Upgrading to jpeg-6b solves the problem. Configure checks for the |
| 625 | correct version, but this problem could occur if a binary built |
| 626 | against a shared libjpeg is run on a system with an older version. |
| 627 | |
| 628 | *** Dired is very slow. |
| 629 | |
| 630 | This could happen if invocation of the `df' program takes a long |
| 631 | time. Possible reasons for this include: |
| 632 | |
| 633 | - ClearCase mounted filesystems (VOBs) that sometimes make `df' |
| 634 | response time extremely slow (dozens of seconds); |
| 635 | |
| 636 | - slow automounters on some old versions of Unix; |
| 637 | |
| 638 | - slow operation of some versions of `df'. |
| 639 | |
| 640 | To work around the problem, you could either (a) set the variable |
| 641 | `directory-free-space-program' to nil, and thus prevent Emacs from |
| 642 | invoking `df'; (b) use `df' from the GNU Fileutils package; or |
| 643 | (c) use CVS, which is Free Software, instead of ClearCase. |
| 644 | |
| 645 | *** Versions of the W3 package released before Emacs 21.1 don't run |
| 646 | under Emacs 21. This fixed in W3 version 4.0pre.47. |
| 647 | |
| 648 | *** The LDAP support rely on ldapsearch program from OpenLDAP version 2. |
| 649 | |
| 650 | It can fail to work with ldapsearch program from OpenLDAP version 1. |
| 651 | Version 1 of OpenLDAP is now deprecated. If you are still using it, |
| 652 | please upgrade to version 2. As a temporary workaround, remove |
| 653 | argument "-x" from the variable `ldap-ldapsearch-args'. |
| 654 | |
| 655 | *** ps-print commands fail to find prologue files ps-prin*.ps. |
| 656 | |
| 657 | This can happen if you use an old version of X-Symbol package: it |
| 658 | defines compatibility functions which trick ps-print into thinking it |
| 659 | runs in XEmacs, and look for the prologue files in a wrong directory. |
| 660 | |
| 661 | The solution is to upgrade X-Symbol to a later version. |
| 662 | |
| 663 | *** On systems with shared libraries you might encounter run-time errors |
| 664 | from the dynamic linker telling you that it is unable to find some |
| 665 | shared libraries, for instance those for Xaw3d or image support. |
| 666 | These errors mean Emacs has been linked with a library whose shared |
| 667 | library is not in the default search path of the dynamic linker. |
| 668 | |
| 669 | Similar problems could prevent Emacs from building, since the build |
| 670 | process invokes Emacs several times. |
| 671 | |
| 672 | On many systems, it is possible to set LD_LIBRARY_PATH in your |
| 673 | environment to specify additional directories where shared libraries |
| 674 | can be found. |
| 675 | |
| 676 | Other systems allow to set LD_RUN_PATH in a similar way, but before |
| 677 | Emacs is linked. With LD_RUN_PATH set, the linker will include a |
| 678 | specified run-time search path in the executable. |
| 679 | |
| 680 | On some systems, Emacs can crash due to problems with dynamic |
| 681 | linking. Specifically, on SGI Irix 6.5, crashes were reported with |
| 682 | backtraces like this: |
| 683 | |
| 684 | (dbx) where |
| 685 | 0 strcmp(0xf49239d, 0x4031184, 0x40302b4, 0x12, 0xf0000000, 0xf4923aa, 0x0, 0x492ddb2) ["/xlv22/ficus-jan23/work/irix/lib/libc/libc_n32_M3_ns/strings/strcmp.s":35, 0xfb7e480] |
| 686 | 1 general_find_symbol(0xf49239d, 0x0, 0x0, 0x0, 0xf0000000, 0xf4923aa, 0x0, 0x492ddb2) |
| 687 | ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":2140, 0xfb65a98] |
| 688 | 2 resolve_symbol(0xf49239d, 0x4031184, 0x0, 0xfbdd438, 0x0, 0xf4923aa, 0x0, 0x492ddb2) |
| 689 | ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":1947, 0xfb657e4] |
| 690 | 3 lazy_text_resolve(0xd18, 0x1a3, 0x40302b4, 0x12, 0xf0000000, 0xf4923aa, 0x0, 0x492ddb2) |
| 691 | ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":997, 0xfb64d44] |
| 692 | 4 _rld_text_resolve(0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0) |
| 693 | ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld_bridge.s":175, 0xfb6032c] |
| 694 | |
| 695 | (`rld' is the dynamic linker.) We don't know yet why this |
| 696 | happens, but setting the environment variable LD_BIND_NOW to 1 (which |
| 697 | forces the dynamic linker to bind all shared objects early on) seems |
| 698 | to work around the problem. |
| 699 | |
| 700 | Please refer to the documentation of your dynamic linker for details. |
| 701 | |
| 702 | *** You request inverse video, and the first Emacs frame is in inverse |
| 703 | video, but later frames are not in inverse video. |
| 704 | |
| 705 | This can happen if you have an old version of the custom library in |
| 706 | your search path for Lisp packages. Use M-x list-load-path-shadows to |
| 707 | check whether this is true. If it is, delete the old custom library. |
| 708 | |
| 709 | *** When you run Ispell from Emacs, it reports a "misalignment" error. |
| 710 | |
| 711 | This can happen if you compiled the Ispell program to use ASCII |
| 712 | characters only and then try to use it from Emacs with non-ASCII |
| 713 | characters, like Latin-1. The solution is to recompile Ispell with |
| 714 | support for 8-bit characters. |
| 715 | |
| 716 | To see whether your Ispell program supports 8-bit characters, type |
| 717 | this at your shell's prompt: |
| 718 | |
| 719 | ispell -vv |
| 720 | |
| 721 | and look in the output for the string "NO8BIT". If Ispell says |
| 722 | "!NO8BIT (8BIT)", your speller supports 8-bit characters; otherwise it |
| 723 | does not. |
| 724 | |
| 725 | To rebuild Ispell with 8-bit character support, edit the local.h file |
| 726 | in the Ispell distribution and make sure it does _not_ define NO8BIT. |
| 727 | Then rebuild the speller. |
| 728 | |
| 729 | Another possible cause for "misalignment" error messages is that the |
| 730 | version of Ispell installed on your machine is old. Upgrade. |
| 731 | |
| 732 | Yet another possibility is that you are trying to spell-check a word |
| 733 | in a language that doesn't fit the dictionary you choose for use by |
| 734 | Ispell. (Ispell can only spell-check one language at a time, because |
| 735 | it uses a single dictionary.) Make sure that the text you are |
| 736 | spelling and the dictionary used by Ispell conform to each other. |
| 737 | |
| 738 | If your spell-checking program is Aspell, it has been reported that if |
| 739 | you have a personal configuration file (normally ~/.aspell.conf), it |
| 740 | can cause this error. Remove that file, execute `ispell-kill-ispell' |
| 741 | in Emacs, and then try spell-checking again. |
| 742 | |
| 743 | * Runtime problems related to font handling |
| 744 | |
| 745 | ** Characters are displayed as empty boxes or with wrong font under X. |
| 746 | |
| 747 | *** This can occur when two different versions of FontConfig are used. |
| 748 | For example, XFree86 4.3.0 has one version and Gnome usually comes |
| 749 | with a newer version. Emacs compiled with Gtk+ will then use the |
| 750 | newer version. In most cases the problem can be temporarily fixed by |
| 751 | stopping the application that has the error (it can be Emacs or any |
| 752 | other application), removing ~/.fonts.cache-1, and then start the |
| 753 | application again. If removing ~/.fonts.cache-1 and restarting |
| 754 | doesn't help, the application with problem must be recompiled with the |
| 755 | same version of FontConfig as the rest of the system uses. For KDE, |
| 756 | it is sufficient to recompile Qt. |
| 757 | |
| 758 | *** Some fonts have a missing glyph and no default character. This is |
| 759 | known to occur for character number 160 (no-break space) in some |
| 760 | fonts, such as Lucida but Emacs sets the display table for the unibyte |
| 761 | and Latin-1 version of this character to display a space. |
| 762 | |
| 763 | *** Some of the fonts called for in your fontset may not exist on your |
| 764 | X server. |
| 765 | |
| 766 | Each X11 font covers just a fraction of the characters that Emacs |
| 767 | supports. To display the whole range of Emacs characters requires |
| 768 | many different fonts, collected into a fontset. You can remedy the |
| 769 | problem by installing additional fonts. |
| 770 | |
| 771 | The intlfonts distribution includes a full spectrum of fonts that can |
| 772 | display all the characters Emacs supports. The etl-unicode collection |
| 773 | of fonts (available from <URL:ftp://ftp.x.org/contrib/fonts/> and |
| 774 | <URL:ftp://ftp.xfree86.org/pub/mirror/X.Org/contrib/fonts/>) includes |
| 775 | fonts that can display many Unicode characters; they can also be used |
| 776 | by ps-print and ps-mule to print Unicode characters. |
| 777 | |
| 778 | ** Under X11, some characters appear improperly aligned in their lines. |
| 779 | |
| 780 | You may have bad X11 fonts; try installing the intlfonts distribution |
| 781 | or the etl-unicode collection (see above). |
| 782 | |
| 783 | ** Under X, an unexpected monospace font is used as the default font. |
| 784 | |
| 785 | When compiled with XFT, Emacs tries to use a default font named |
| 786 | "monospace". This is a "virtual font", which the operating system |
| 787 | (Fontconfig) redirects to a suitable font such as DejaVu Sans Mono. |
| 788 | On some systems, there exists a font that is actually named Monospace, |
| 789 | which takes over the virtual font. This is considered an operating |
| 790 | system bug; see |
| 791 | |
| 792 | http://lists.gnu.org/archive/html/emacs-devel/2008-10/msg00696.html |
| 793 | |
| 794 | If you encounter this problem, set the default font to a specific font |
| 795 | in your .Xresources or initialization file. For instance, you can put |
| 796 | the following in your .Xresources: |
| 797 | |
| 798 | Emacs.font: DejaVu Sans Mono 12 |
| 799 | |
| 800 | ** Certain fonts make each line take one pixel more than it should. |
| 801 | |
| 802 | This is because these fonts contain characters a little taller than |
| 803 | the font's nominal height. Emacs needs to make sure that lines do not |
| 804 | overlap. |
| 805 | |
| 806 | ** Loading fonts is very slow. |
| 807 | |
| 808 | You might be getting scalable fonts instead of precomputed bitmaps. |
| 809 | Known scalable font directories are "Type1" and "Speedo". A font |
| 810 | directory contains scalable fonts if it contains the file |
| 811 | "fonts.scale". |
| 812 | |
| 813 | If this is so, re-order your X windows font path to put the scalable |
| 814 | font directories last. See the documentation of `xset' for details. |
| 815 | |
| 816 | With some X servers, it may be necessary to take the scalable font |
| 817 | directories out of your path entirely, at least for Emacs 19.26. |
| 818 | Changes in the future may make this unnecessary. |
| 819 | |
| 820 | ** Font Lock displays portions of the buffer in incorrect faces. |
| 821 | |
| 822 | By far the most frequent cause of this is a parenthesis `(' or a brace |
| 823 | `{' in column zero. Font Lock assumes that such a paren is outside of |
| 824 | any comment or string. This is of course not true in general, but the |
| 825 | vast majority of well-formatted program source files don't have such |
| 826 | parens, and therefore this assumption is used to allow optimizations |
| 827 | in Font Lock's syntactical analysis. These optimizations avoid some |
| 828 | pathological cases where jit-lock, the Just-in-Time fontification |
| 829 | introduced with Emacs 21.1, could significantly slow down scrolling |
| 830 | through the buffer, especially scrolling backwards, and also jumping |
| 831 | to the end of a very large buffer. |
| 832 | |
| 833 | Beginning with version 22.1, a parenthesis or a brace in column zero |
| 834 | is highlighted in bold-red face if it is inside a string or a comment, |
| 835 | to indicate that it could interfere with Font Lock (and also with |
| 836 | indentation) and should be moved or escaped with a backslash. |
| 837 | |
| 838 | If you don't use large buffers, or have a very fast machine which |
| 839 | makes the delays insignificant, you can avoid the incorrect |
| 840 | fontification by setting the variable |
| 841 | `font-lock-beginning-of-syntax-function' to a nil value. (This must |
| 842 | be done _after_ turning on Font Lock.) |
| 843 | |
| 844 | Another alternative is to avoid a paren in column zero. For example, |
| 845 | in a Lisp string you could precede the paren with a backslash. |
| 846 | |
| 847 | ** With certain fonts, when the cursor appears on a character, the |
| 848 | character doesn't appear--you get a solid box instead. |
| 849 | |
| 850 | One user on a Linux-based GNU system reported that this problem went |
| 851 | away with installation of a new X server. The failing server was |
| 852 | XFree86 3.1.1. XFree86 3.1.2 works. |
| 853 | |
| 854 | ** Emacs pauses for several seconds when changing the default font. |
| 855 | |
| 856 | This has been reported for fvwm 2.2.5 and the window manager of KDE |
| 857 | 2.1. The reason for the pause is Xt waiting for a ConfigureNotify |
| 858 | event from the window manager, which the window manager doesn't send. |
| 859 | Xt stops waiting after a default timeout of usually 5 seconds. |
| 860 | |
| 861 | A workaround for this is to add something like |
| 862 | |
| 863 | emacs.waitForWM: false |
| 864 | |
| 865 | to your X resources. Alternatively, add `(wait-for-wm . nil)' to a |
| 866 | frame's parameter list, like this: |
| 867 | |
| 868 | (modify-frame-parameters nil '((wait-for-wm . nil))) |
| 869 | |
| 870 | (this should go into your `.emacs' file). |
| 871 | |
| 872 | ** Underlines appear at the wrong position. |
| 873 | |
| 874 | This is caused by fonts having a wrong UNDERLINE_POSITION property. |
| 875 | Examples are the font 7x13 on XFree prior to version 4.1, or the jmk |
| 876 | neep font from the Debian xfonts-jmk package prior to version 3.0.17. |
| 877 | To circumvent this problem, set x-use-underline-position-properties |
| 878 | to nil in your `.emacs'. |
| 879 | |
| 880 | To see what is the value of UNDERLINE_POSITION defined by the font, |
| 881 | type `xlsfonts -lll FONT' and look at the font's UNDERLINE_POSITION |
| 882 | property. |
| 883 | |
| 884 | ** When using Exceed, fonts sometimes appear too tall. |
| 885 | |
| 886 | When the display is set to an Exceed X-server and fonts are specified |
| 887 | (either explicitly with the -fn option or implicitly with X resources) |
| 888 | then the fonts may appear "too tall". The actual character sizes are |
| 889 | correct but there is too much vertical spacing between rows, which |
| 890 | gives the appearance of "double spacing". |
| 891 | |
| 892 | To prevent this, turn off the Exceed's "automatic font substitution" |
| 893 | feature (in the font part of the configuration window). |
| 894 | |
| 895 | ** Subscript/superscript text in TeX is hard to read. |
| 896 | |
| 897 | If `tex-fontify-script' is non-nil, tex-mode displays |
| 898 | subscript/superscript text in the faces subscript/superscript, which |
| 899 | are smaller than the normal font and lowered/raised. With some fonts, |
| 900 | nested superscripts (say) can be hard to read. Switching to a |
| 901 | different font, or changing your antialiasing setting (on an LCD |
| 902 | screen), can both make the problem disappear. Alternatively, customize |
| 903 | the following variables: tex-font-script-display (how much to |
| 904 | lower/raise); tex-suscript-height-ratio (how much smaller than |
| 905 | normal); tex-suscript-height-minimum (minimum height). |
| 906 | |
| 907 | * Internationalization problems |
| 908 | |
| 909 | ** M-{ does not work on a Spanish PC keyboard. |
| 910 | |
| 911 | Many Spanish keyboards seem to ignore that combination. Emacs can't |
| 912 | do anything about it. |
| 913 | |
| 914 | ** International characters aren't displayed under X. |
| 915 | |
| 916 | *** Missing X fonts |
| 917 | |
| 918 | XFree86 4 contains many fonts in iso10646-1 encoding which have |
| 919 | minimal character repertoires (whereas the encoding part of the font |
| 920 | name is meant to be a reasonable indication of the repertoire |
| 921 | according to the XLFD spec). Emacs may choose one of these to display |
| 922 | characters from the mule-unicode charsets and then typically won't be |
| 923 | able to find the glyphs to display many characters. (Check with C-u |
| 924 | C-x = .) To avoid this, you may need to use a fontset which sets the |
| 925 | font for the mule-unicode sets explicitly. E.g. to use GNU unifont, |
| 926 | include in the fontset spec: |
| 927 | |
| 928 | mule-unicode-2500-33ff:-gnu-unifont-*-iso10646-1,\ |
| 929 | mule-unicode-e000-ffff:-gnu-unifont-*-iso10646-1,\ |
| 930 | mule-unicode-0100-24ff:-gnu-unifont-*-iso10646-1 |
| 931 | |
| 932 | *** Athena/Lucid toolkit limitations |
| 933 | |
| 934 | The Athena/Lucid toolkit cannot display UTF-8 strings in the menu, so |
| 935 | if you have UTF-8 buffer names, the buffer menu won't display the |
| 936 | names properly. The GTK+ toolkit works properly. |
| 937 | |
| 938 | ** The UTF-8/16/7 coding systems don't encode CJK (Far Eastern) characters. |
| 939 | |
| 940 | Emacs directly supports the Unicode BMP whose code points are in the |
| 941 | ranges 0000-33ff and e000-ffff, and indirectly supports the parts of |
| 942 | CJK characters belonging to these legacy charsets: |
| 943 | |
| 944 | GB2312, Big5, JISX0208, JISX0212, JISX0213-1, JISX0213-2, KSC5601 |
| 945 | |
| 946 | The latter support is done in Utf-Translate-Cjk mode (turned on by |
| 947 | default). Which Unicode CJK characters are decoded into which Emacs |
| 948 | charset is decided by the current language environment. For instance, |
| 949 | in Chinese-GB, most of them are decoded into chinese-gb2312. |
| 950 | |
| 951 | If you read UTF-8 data with code points outside these ranges, the |
| 952 | characters appear in the buffer as raw bytes of the original UTF-8 |
| 953 | (composed into a single quasi-character) and they will be written back |
| 954 | correctly as UTF-8, assuming you don't break the composed sequences. |
| 955 | If you read such characters from UTF-16 or UTF-7 data, they are |
| 956 | substituted with the Unicode `replacement character', and you lose |
| 957 | information. |
| 958 | |
| 959 | ** Accented ISO-8859-1 characters are displayed as | or _. |
| 960 | |
| 961 | Try other font set sizes (S-mouse-1). If the problem persists with |
| 962 | other sizes as well, your text is corrupted, probably through software |
| 963 | that is not 8-bit clean. If the problem goes away with another font |
| 964 | size, it's probably because some fonts pretend to be ISO-8859-1 fonts |
| 965 | when they are really ASCII fonts. In particular the schumacher-clean |
| 966 | fonts have this bug in some versions of X. |
| 967 | |
| 968 | To see what glyphs are included in a font, use `xfd', like this: |
| 969 | |
| 970 | xfd -fn -schumacher-clean-medium-r-normal--12-120-75-75-c-60-iso8859-1 |
| 971 | |
| 972 | If this shows only ASCII glyphs, the font is indeed the source of the |
| 973 | problem. |
| 974 | |
| 975 | The solution is to remove the corresponding lines from the appropriate |
| 976 | `fonts.alias' file, then run `mkfontdir' in that directory, and then run |
| 977 | `xset fp rehash'. |
| 978 | |
| 979 | ** The `oc-unicode' package doesn't work with Emacs 21. |
| 980 | |
| 981 | This package tries to define more private charsets than there are free |
| 982 | slots now. The current built-in Unicode support is actually more |
| 983 | flexible. (Use option `utf-translate-cjk-mode' if you need CJK |
| 984 | support.) Files encoded as emacs-mule using oc-unicode aren't |
| 985 | generally read correctly by Emacs 21. |
| 986 | |
| 987 | ** After a while, Emacs slips into unibyte mode. |
| 988 | |
| 989 | The VM mail package, which is not part of Emacs, sometimes does |
| 990 | (standard-display-european t) |
| 991 | That should be changed to |
| 992 | (standard-display-european 1 t) |
| 993 | |
| 994 | * X runtime problems |
| 995 | |
| 996 | ** X keyboard problems |
| 997 | |
| 998 | *** You "lose characters" after typing Compose Character key. |
| 999 | |
| 1000 | This is because the Compose Character key is defined as the keysym |
| 1001 | Multi_key, and Emacs (seeing that) does the proper X11 |
| 1002 | character-composition processing. If you don't want your Compose key |
| 1003 | to do that, you can redefine it with xmodmap. |
| 1004 | |
| 1005 | For example, here's one way to turn it into a Meta key: |
| 1006 | |
| 1007 | xmodmap -e "keysym Multi_key = Meta_L" |
| 1008 | |
| 1009 | If all users at your site of a particular keyboard prefer Meta to |
| 1010 | Compose, you can make the remapping happen automatically by adding the |
| 1011 | xmodmap command to the xdm setup script for that display. |
| 1012 | |
| 1013 | *** Using X Windows, control-shift-leftbutton makes Emacs hang. |
| 1014 | |
| 1015 | Use the shell command `xset bc' to make the old X Menu package work. |
| 1016 | |
| 1017 | *** C-SPC fails to work on Fedora GNU/Linux (or with fcitx input method). |
| 1018 | |
| 1019 | Fedora Core 4 steals the C-SPC key by default for the `iiimx' program |
| 1020 | which is the input method for some languages. It blocks Emacs users |
| 1021 | from using the C-SPC key for `set-mark-command'. |
| 1022 | |
| 1023 | One solutions is to remove the `<Ctrl>space' from the `Iiimx' file |
| 1024 | which can be found in the `/usr/lib/X11/app-defaults' directory. |
| 1025 | However, that requires root access. |
| 1026 | |
| 1027 | Another is to specify `Emacs*useXIM: false' in your X resources. |
| 1028 | |
| 1029 | Another is to build Emacs with the `--without-xim' configure option. |
| 1030 | |
| 1031 | The same problem happens on any other system if you are using fcitx |
| 1032 | (Chinese input method) which by default use C-SPC for toggling. If |
| 1033 | you want to use fcitx with Emacs, you have two choices. Toggle fcitx |
| 1034 | by another key (e.g. C-\) by modifying ~/.fcitx/config, or be |
| 1035 | accustomed to use C-@ for `set-mark-command'. |
| 1036 | |
| 1037 | *** M-SPC seems to be ignored as input. |
| 1038 | |
| 1039 | See if your X server is set up to use this as a command |
| 1040 | for character composition. |
| 1041 | |
| 1042 | *** The S-C-t key combination doesn't get passed to Emacs on X. |
| 1043 | |
| 1044 | This happens because some X configurations assign the Ctrl-Shift-t |
| 1045 | combination the same meaning as the Multi_key. The offending |
| 1046 | definition is in the file `...lib/X11/locale/iso8859-1/Compose'; there |
| 1047 | might be other similar combinations which are grabbed by X for similar |
| 1048 | purposes. |
| 1049 | |
| 1050 | We think that this can be countermanded with the `xmodmap' utility, if |
| 1051 | you want to be able to bind one of these key sequences within Emacs. |
| 1052 | |
| 1053 | *** Under X, C-v and/or other keys don't work. |
| 1054 | |
| 1055 | These may have been intercepted by your window manager. In |
| 1056 | particular, AfterStep 1.6 is reported to steal C-v in its default |
| 1057 | configuration. Various Meta keys are also likely to be taken by the |
| 1058 | configuration of the `feel'. See the WM's documentation for how to |
| 1059 | change this. |
| 1060 | |
| 1061 | *** Clicking C-mouse-2 in the scroll bar doesn't split the window. |
| 1062 | |
| 1063 | This currently doesn't work with scroll-bar widgets (and we don't know |
| 1064 | a good way of implementing it with widgets). If Emacs is configured |
| 1065 | --without-toolkit-scroll-bars, C-mouse-2 on the scroll bar does work. |
| 1066 | |
| 1067 | *** Inability to send an Alt-modified key, when Emacs is communicating |
| 1068 | directly with an X server. |
| 1069 | |
| 1070 | If you have tried to bind an Alt-modified key as a command, and it |
| 1071 | does not work to type the command, the first thing you should check is |
| 1072 | whether the key is getting through to Emacs. To do this, type C-h c |
| 1073 | followed by the Alt-modified key. C-h c should say what kind of event |
| 1074 | it read. If it says it read an Alt-modified key, then make sure you |
| 1075 | have made the key binding correctly. |
| 1076 | |
| 1077 | If C-h c reports an event that doesn't have the Alt modifier, it may |
| 1078 | be because your X server has no key for the Alt modifier. The X |
| 1079 | server that comes from MIT does not set up the Alt modifier by |
| 1080 | default. |
| 1081 | |
| 1082 | If your keyboard has keys named Alt, you can enable them as follows: |
| 1083 | |
| 1084 | xmodmap -e 'add mod2 = Alt_L' |
| 1085 | xmodmap -e 'add mod2 = Alt_R' |
| 1086 | |
| 1087 | If the keyboard has just one key named Alt, then only one of those |
| 1088 | commands is needed. The modifier `mod2' is a reasonable choice if you |
| 1089 | are using an unmodified MIT version of X. Otherwise, choose any |
| 1090 | modifier bit not otherwise used. |
| 1091 | |
| 1092 | If your keyboard does not have keys named Alt, you can use some other |
| 1093 | keys. Use the keysym command in xmodmap to turn a function key (or |
| 1094 | some other 'spare' key) into Alt_L or into Alt_R, and then use the |
| 1095 | commands show above to make them modifier keys. |
| 1096 | |
| 1097 | Note that if you have Alt keys but no Meta keys, Emacs translates Alt |
| 1098 | into Meta. This is because of the great importance of Meta in Emacs. |
| 1099 | |
| 1100 | ** Window-manager and toolkit-related problems |
| 1101 | |
| 1102 | *** Metacity: Resizing Emacs or ALT-Tab causes X to be unresponsive. |
| 1103 | |
| 1104 | This happens sometimes when using Metacity. Resizing Emacs or ALT-Tab:bing |
| 1105 | makes the system unresponsive to the mouse or the keyboard. Killing Emacs |
| 1106 | or shifting out from X11 and back again usually cures it (i.e. Ctrl-Alt-F1 |
| 1107 | and then Alt-F7). A bug for it is here: |
| 1108 | https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/231034. |
| 1109 | Note that a permanent fix seems to be to disable "assistive technologies". |
| 1110 | |
| 1111 | *** Gnome: Emacs receives input directly from the keyboard, bypassing XIM. |
| 1112 | |
| 1113 | This seems to happen when gnome-settings-daemon version 2.12 or later |
| 1114 | is running. If gnome-settings-daemon is not running, Emacs receives |
| 1115 | input through XIM without any problem. Furthermore, this seems only |
| 1116 | to happen in *.UTF-8 locales; zh_CN.GB2312 and zh_CN.GBK locales, for |
| 1117 | example, work fine. A bug report has been filed in the Gnome |
| 1118 | bugzilla: http://bugzilla.gnome.org/show_bug.cgi?id=357032 |
| 1119 | |
| 1120 | *** Gnome: Emacs' xterm-mouse-mode doesn't work on the Gnome terminal. |
| 1121 | |
| 1122 | A symptom of this bug is that double-clicks insert a control sequence |
| 1123 | into the buffer. The reason this happens is an apparent |
| 1124 | incompatibility of the Gnome terminal with Xterm, which also affects |
| 1125 | other programs using the Xterm mouse interface. A problem report has |
| 1126 | been filed. |
| 1127 | |
| 1128 | *** KDE: When running on KDE, colors or fonts are not as specified for Emacs, |
| 1129 | or messed up. |
| 1130 | |
| 1131 | For example, you could see background you set for Emacs only in the |
| 1132 | empty portions of the Emacs display, while characters have some other |
| 1133 | background. |
| 1134 | |
| 1135 | This happens because KDE's defaults apply its color and font |
| 1136 | definitions even to applications that weren't compiled for KDE. The |
| 1137 | solution is to uncheck the "Apply fonts and colors to non-KDE apps" |
| 1138 | option in Preferences->Look&Feel->Style (KDE 2). In KDE 3, this option |
| 1139 | is in the "Colors" section, rather than "Style". |
| 1140 | |
| 1141 | Alternatively, if you do want the KDE defaults to apply to other |
| 1142 | applications, but not to Emacs, you could modify the file `Emacs.ad' |
| 1143 | (should be in the `/usr/share/apps/kdisplay/app-defaults/' directory) |
| 1144 | so that it doesn't set the default background and foreground only for |
| 1145 | Emacs. For example, make sure the following resources are either not |
| 1146 | present or commented out: |
| 1147 | |
| 1148 | Emacs.default.attributeForeground |
| 1149 | Emacs.default.attributeBackground |
| 1150 | Emacs*Foreground |
| 1151 | Emacs*Background |
| 1152 | |
| 1153 | It is also reported that a bug in the gtk-engines-qt engine can cause this if |
| 1154 | Emacs is compiled with Gtk+. |
| 1155 | The bug is fixed in version 0.7 or newer of gtk-engines-qt. |
| 1156 | |
| 1157 | *** KDE: Emacs hangs on KDE when a large portion of text is killed. |
| 1158 | |
| 1159 | This is caused by a bug in the KDE applet `klipper' which periodically |
| 1160 | requests the X clipboard contents from applications. Early versions |
| 1161 | of klipper don't implement the ICCCM protocol for large selections, |
| 1162 | which leads to Emacs being flooded with selection requests. After a |
| 1163 | while, Emacs may print a message: |
| 1164 | |
| 1165 | Timed out waiting for property-notify event |
| 1166 | |
| 1167 | A workaround is to not use `klipper'. An upgrade to the `klipper' that |
| 1168 | comes with KDE 3.3 or later also solves the problem. |
| 1169 | |
| 1170 | *** CDE: Frames may cover dialogs they created when using CDE. |
| 1171 | |
| 1172 | This can happen if you have "Allow Primary Windows On Top" enabled which |
| 1173 | seems to be the default in the Common Desktop Environment. |
| 1174 | To change, go in to "Desktop Controls" -> "Window Style Manager" |
| 1175 | and uncheck "Allow Primary Windows On Top". |
| 1176 | |
| 1177 | *** Xaw3d : When using Xaw3d scroll bars without arrows, the very first mouse |
| 1178 | click in a scroll bar might be ignored by the scroll bar widget. This |
| 1179 | is probably a bug in Xaw3d; when Xaw3d is compiled with arrows, the |
| 1180 | problem disappears. |
| 1181 | |
| 1182 | *** Xaw: There are known binary incompatibilities between Xaw, Xaw3d, neXtaw, |
| 1183 | XawM and the few other derivatives of Xaw. So when you compile with |
| 1184 | one of these, it may not work to dynamically link with another one. |
| 1185 | For example, strange problems, such as Emacs exiting when you type |
| 1186 | "C-x 1", were reported when Emacs compiled with Xaw3d and libXaw was |
| 1187 | used with neXtaw at run time. |
| 1188 | |
| 1189 | The solution is to rebuild Emacs with the toolkit version you actually |
| 1190 | want to use, or set LD_PRELOAD to preload the same toolkit version you |
| 1191 | built Emacs with. |
| 1192 | |
| 1193 | *** Open Motif: Problems with file dialogs in Emacs built with Open Motif. |
| 1194 | |
| 1195 | When Emacs 21 is built with Open Motif 2.1, it can happen that the |
| 1196 | graphical file dialog boxes do not work properly. The "OK", "Filter" |
| 1197 | and "Cancel" buttons do not respond to mouse clicks. Dragging the |
| 1198 | file dialog window usually causes the buttons to work again. |
| 1199 | |
| 1200 | The solution is to use LessTif instead. LessTif is a free replacement |
| 1201 | for Motif. See the file INSTALL for information on how to do this. |
| 1202 | |
| 1203 | Another workaround is not to use the mouse to trigger file prompts, |
| 1204 | but to use the keyboard. This way, you will be prompted for a file in |
| 1205 | the minibuffer instead of a graphical file dialog. |
| 1206 | |
| 1207 | *** LessTif: Problems in Emacs built with LessTif. |
| 1208 | |
| 1209 | The problems seem to depend on the version of LessTif and the Motif |
| 1210 | emulation for which it is set up. |
| 1211 | |
| 1212 | Only the Motif 1.2 emulation seems to be stable enough in LessTif. |
| 1213 | LessTif 0.92-17's Motif 1.2 emulation seems to work okay on FreeBSD. |
| 1214 | On GNU/Linux systems, lesstif-0.92.6 configured with "./configure |
| 1215 | --enable-build-12 --enable-default-12" is reported to be the most |
| 1216 | successful. The binary GNU/Linux package |
| 1217 | lesstif-devel-0.92.0-1.i386.rpm was reported to have problems with |
| 1218 | menu placement. |
| 1219 | |
| 1220 | On some systems, even with Motif 1.2 emulation, Emacs occasionally |
| 1221 | locks up, grabbing all mouse and keyboard events. We still don't know |
| 1222 | what causes these problems; they are not reproducible by Emacs |
| 1223 | developers. |
| 1224 | |
| 1225 | *** Motif: The Motif version of Emacs paints the screen a solid color. |
| 1226 | |
| 1227 | This has been observed to result from the following X resource: |
| 1228 | |
| 1229 | Emacs*default.attributeFont: -*-courier-medium-r-*-*-*-140-*-*-*-*-iso8859-* |
| 1230 | |
| 1231 | That the resource has this effect indicates a bug in something, but we |
| 1232 | do not yet know what. If it is an Emacs bug, we hope someone can |
| 1233 | explain what the bug is so we can fix it. In the mean time, removing |
| 1234 | the resource prevents the problem. |
| 1235 | |
| 1236 | ** General X problems |
| 1237 | |
| 1238 | *** Redisplay using X11 is much slower than previous Emacs versions. |
| 1239 | |
| 1240 | We've noticed that certain X servers draw the text much slower when |
| 1241 | scroll bars are on the left. We don't know why this happens. If this |
| 1242 | happens to you, you can work around it by putting the scroll bars |
| 1243 | on the right (as they were in Emacs 19). |
| 1244 | |
| 1245 | Here's how to do this: |
| 1246 | |
| 1247 | (set-scroll-bar-mode 'right) |
| 1248 | |
| 1249 | If you're not sure whether (or how much) this problem affects you, |
| 1250 | try that and see how much difference it makes. To set things back |
| 1251 | to normal, do |
| 1252 | |
| 1253 | (set-scroll-bar-mode 'left) |
| 1254 | |
| 1255 | *** Error messages about undefined colors on X. |
| 1256 | |
| 1257 | The messages might say something like this: |
| 1258 | |
| 1259 | Unable to load color "grey95" |
| 1260 | |
| 1261 | (typically, in the `*Messages*' buffer), or something like this: |
| 1262 | |
| 1263 | Error while displaying tooltip: (error Undefined color lightyellow) |
| 1264 | |
| 1265 | These problems could happen if some other X program has used up too |
| 1266 | many colors of the X palette, leaving Emacs with insufficient system |
| 1267 | resources to load all the colors it needs. |
| 1268 | |
| 1269 | A solution is to exit the offending X programs before starting Emacs. |
| 1270 | |
| 1271 | "undefined color" messages can also occur if the RgbPath entry in the |
| 1272 | X configuration file is incorrect, or the rgb.txt file is not where |
| 1273 | X expects to find it. |
| 1274 | |
| 1275 | *** Improving performance with slow X connections. |
| 1276 | |
| 1277 | There are several ways to improve this performance, any subset of which can |
| 1278 | be carried out at the same time: |
| 1279 | |
| 1280 | 1) If you don't need X Input Methods (XIM) for entering text in some |
| 1281 | language you use, you can improve performance on WAN links by using |
| 1282 | the X resource useXIM to turn off use of XIM. This does not affect |
| 1283 | the use of Emacs' own input methods, which are part of the Leim |
| 1284 | package. |
| 1285 | |
| 1286 | 2) If the connection is very slow, you might also want to consider |
| 1287 | switching off scroll bars, menu bar, and tool bar. Adding the |
| 1288 | following forms to your .emacs file will accomplish that, but only |
| 1289 | after the the initial frame is displayed: |
| 1290 | |
| 1291 | (scroll-bar-mode -1) |
| 1292 | (menu-bar-mode -1) |
| 1293 | (tool-bar-mode -1) |
| 1294 | |
| 1295 | For still quicker startup, put these X resources in your .Xdefaults |
| 1296 | file: |
| 1297 | |
| 1298 | Emacs.verticalScrollBars: off |
| 1299 | Emacs.menuBar: off |
| 1300 | Emacs.toolBar: off |
| 1301 | |
| 1302 | 3) Use ssh to forward the X connection, and enable compression on this |
| 1303 | forwarded X connection (ssh -XC remotehostname emacs ...). |
| 1304 | |
| 1305 | 4) Use lbxproxy on the remote end of the connection. This is an interface |
| 1306 | to the low bandwidth X extension in most modern X servers, which |
| 1307 | improves performance dramatically, at the slight expense of correctness |
| 1308 | of the X protocol. lbxproxy acheives the performance gain by grouping |
| 1309 | several X requests in one TCP packet and sending them off together, |
| 1310 | instead of requiring a round-trip for each X request in a separate |
| 1311 | packet. The switches that seem to work best for emacs are: |
| 1312 | -noatomsfile -nowinattr -cheaterrors -cheatevents |
| 1313 | Note that the -nograbcmap option is known to cause problems. |
| 1314 | For more about lbxproxy, see: |
| 1315 | http://www.xfree86.org/4.3.0/lbxproxy.1.html |
| 1316 | |
| 1317 | 5) If copying and killing is slow, try to disable the interaction with the |
| 1318 | native system's clipboard by adding these lines to your .emacs file: |
| 1319 | (setq interprogram-cut-function nil) |
| 1320 | (setq interprogram-paste-function nil) |
| 1321 | |
| 1322 | *** Emacs gives the error, Couldn't find per display information. |
| 1323 | |
| 1324 | This can result if the X server runs out of memory because Emacs uses |
| 1325 | a large number of fonts. On systems where this happens, C-h h is |
| 1326 | likely to cause it. |
| 1327 | |
| 1328 | We do not know of a way to prevent the problem. |
| 1329 | |
| 1330 | *** Emacs does not notice when you release the mouse. |
| 1331 | |
| 1332 | There are reports that this happened with (some) Microsoft mice and |
| 1333 | that replacing the mouse made it stop. |
| 1334 | |
| 1335 | *** You can't select from submenus (in the X toolkit version). |
| 1336 | |
| 1337 | On certain systems, mouse-tracking and selection in top-level menus |
| 1338 | works properly with the X toolkit, but neither of them works when you |
| 1339 | bring up a submenu (such as Bookmarks or Compare or Apply Patch, in |
| 1340 | the Files menu). |
| 1341 | |
| 1342 | This works on most systems. There is speculation that the failure is |
| 1343 | due to bugs in old versions of X toolkit libraries, but no one really |
| 1344 | knows. If someone debugs this and finds the precise cause, perhaps a |
| 1345 | workaround can be found. |
| 1346 | |
| 1347 | *** An error message such as `X protocol error: BadMatch (invalid |
| 1348 | parameter attributes) on protocol request 93'. |
| 1349 | |
| 1350 | This comes from having an invalid X resource, such as |
| 1351 | emacs*Cursor: black |
| 1352 | (which is invalid because it specifies a color name for something |
| 1353 | that isn't a color.) |
| 1354 | |
| 1355 | The fix is to correct your X resources. |
| 1356 | |
| 1357 | *** Slow startup on X11R6 with X windows. |
| 1358 | |
| 1359 | If Emacs takes two minutes to start up on X11R6, see if your X |
| 1360 | resources specify any Adobe fonts. That causes the type-1 font |
| 1361 | renderer to start up, even if the font you asked for is not a type-1 |
| 1362 | font. |
| 1363 | |
| 1364 | One way to avoid this problem is to eliminate the type-1 fonts from |
| 1365 | your font path, like this: |
| 1366 | |
| 1367 | xset -fp /usr/X11R6/lib/X11/fonts/Type1/ |
| 1368 | |
| 1369 | *** Pull-down menus appear in the wrong place, in the toolkit version of Emacs. |
| 1370 | |
| 1371 | An X resource of this form can cause the problem: |
| 1372 | |
| 1373 | Emacs*geometry: 80x55+0+0 |
| 1374 | |
| 1375 | This resource is supposed to apply, and does apply, to the menus |
| 1376 | individually as well as to Emacs frames. If that is not what you |
| 1377 | want, rewrite the resource. |
| 1378 | |
| 1379 | To check thoroughly for such resource specifications, use `xrdb |
| 1380 | -query' to see what resources the X server records, and also look at |
| 1381 | the user's ~/.Xdefaults and ~/.Xdefaults-* files. |
| 1382 | |
| 1383 | *** Emacs running under X Windows does not handle mouse clicks. |
| 1384 | *** `emacs -geometry 80x20' finds a file named `80x20'. |
| 1385 | |
| 1386 | One cause of such problems is having (setq term-file-prefix nil) in |
| 1387 | your .emacs file. Another cause is a bad value of EMACSLOADPATH in |
| 1388 | the environment. |
| 1389 | |
| 1390 | *** Emacs fails to get default settings from X Windows server. |
| 1391 | |
| 1392 | The X library in X11R4 has a bug; it interchanges the 2nd and 3rd |
| 1393 | arguments to XGetDefaults. Define the macro XBACKWARDS in config.h to |
| 1394 | tell Emacs to compensate for this. |
| 1395 | |
| 1396 | I don't believe there is any way Emacs can determine for itself |
| 1397 | whether this problem is present on a given system. |
| 1398 | |
| 1399 | *** X Windows doesn't work if DISPLAY uses a hostname. |
| 1400 | |
| 1401 | People have reported kernel bugs in certain systems that cause Emacs |
| 1402 | not to work with X Windows if DISPLAY is set using a host name. But |
| 1403 | the problem does not occur if DISPLAY is set to `unix:0.0'. I think |
| 1404 | the bug has to do with SIGIO or FIONREAD. |
| 1405 | |
| 1406 | You may be able to compensate for the bug by doing (set-input-mode nil nil). |
| 1407 | However, that has the disadvantage of turning off interrupts, so that |
| 1408 | you are unable to quit out of a Lisp program by typing C-g. |
| 1409 | |
| 1410 | The easy way to do this is to put |
| 1411 | |
| 1412 | (setq x-sigio-bug t) |
| 1413 | |
| 1414 | in your site-init.el file. |
| 1415 | |
| 1416 | *** Prevent double pastes in X |
| 1417 | |
| 1418 | The problem: a region, such as a command, is pasted twice when you copy |
| 1419 | it with your mouse from GNU Emacs to an xterm or an RXVT shell in X. |
| 1420 | The solution: try the following in your X configuration file, |
| 1421 | /etc/X11/xorg.conf This should enable both PS/2 and USB mice for |
| 1422 | single copies. You do not need any other drivers or options. |
| 1423 | |
| 1424 | Section "InputDevice" |
| 1425 | Identifier "Generic Mouse" |
| 1426 | Driver "mousedev" |
| 1427 | Option "Device" "/dev/input/mice" |
| 1428 | EndSection |
| 1429 | |
| 1430 | * Runtime problems on character terminals |
| 1431 | |
| 1432 | ** The meta key does not work on xterm. |
| 1433 | Typing M-x rings the terminal bell, and inserts a string like ";120~". |
| 1434 | For recent xterm versions (>= 216), Emacs uses xterm's modifyOtherKeys |
| 1435 | feature to generate strings for key combinations that are not |
| 1436 | otherwise usable. One circumstance in which this can cause problems |
| 1437 | is if you have specified the X resource |
| 1438 | |
| 1439 | xterm*VT100.Translations |
| 1440 | |
| 1441 | to contain translations that use the meta key. Then xterm will not |
| 1442 | use meta in modified function-keys, which confuses Emacs. To fix |
| 1443 | this, you can remove the X resource or put this in your init file: |
| 1444 | |
| 1445 | (xterm-remove-modify-other-keys) |
| 1446 | |
| 1447 | ** Emacs spontaneously displays "I-search: " at the bottom of the screen. |
| 1448 | |
| 1449 | This means that Control-S/Control-Q (XON/XOFF) "flow control" is being |
| 1450 | used. C-s/C-q flow control is bad for Emacs editors because it takes |
| 1451 | away C-s and C-q as user commands. Since editors do not output long |
| 1452 | streams of text without user commands, there is no need for a |
| 1453 | user-issuable "stop output" command in an editor; therefore, a |
| 1454 | properly designed flow control mechanism would transmit all possible |
| 1455 | input characters without interference. Designing such a mechanism is |
| 1456 | easy, for a person with at least half a brain. |
| 1457 | |
| 1458 | There are three possible reasons why flow control could be taking place: |
| 1459 | |
| 1460 | 1) Terminal has not been told to disable flow control |
| 1461 | 2) Insufficient padding for the terminal in use |
| 1462 | 3) Some sort of terminal concentrator or line switch is responsible |
| 1463 | |
| 1464 | First of all, many terminals have a set-up mode which controls whether |
| 1465 | they generate XON/XOFF flow control characters. This must be set to |
| 1466 | "no XON/XOFF" in order for Emacs to work. (For example, on a VT220 |
| 1467 | you may select "No XOFF" in the setup menu.) Sometimes there is an |
| 1468 | escape sequence that the computer can send to turn flow control off |
| 1469 | and on. If so, perhaps the termcap `ti' string should turn flow |
| 1470 | control off, and the `te' string should turn it on. |
| 1471 | |
| 1472 | Once the terminal has been told "no flow control", you may find it |
| 1473 | needs more padding. The amount of padding Emacs sends is controlled |
| 1474 | by the termcap entry for the terminal in use, and by the output baud |
| 1475 | rate as known by the kernel. The shell command `stty' will print |
| 1476 | your output baud rate; `stty' with suitable arguments will set it if |
| 1477 | it is wrong. Setting to a higher speed causes increased padding. If |
| 1478 | the results are wrong for the correct speed, there is probably a |
| 1479 | problem in the termcap entry. You must speak to a local Unix wizard |
| 1480 | to fix this. Perhaps you are just using the wrong terminal type. |
| 1481 | |
| 1482 | For terminals that lack a "no flow control" mode, sometimes just |
| 1483 | giving lots of padding will prevent actual generation of flow control |
| 1484 | codes. You might as well try it. |
| 1485 | |
| 1486 | If you are really unlucky, your terminal is connected to the computer |
| 1487 | through a concentrator which sends XON/XOFF flow control to the |
| 1488 | computer, or it insists on sending flow control itself no matter how |
| 1489 | much padding you give it. Unless you can figure out how to turn flow |
| 1490 | control off on this concentrator (again, refer to your local wizard), |
| 1491 | you are screwed! You should have the terminal or concentrator |
| 1492 | replaced with a properly designed one. In the mean time, some drastic |
| 1493 | measures can make Emacs semi-work. |
| 1494 | |
| 1495 | You can make Emacs ignore C-s and C-q and let the operating system |
| 1496 | handle them. To do this on a per-session basis, just type M-x |
| 1497 | enable-flow-control RET. You will see a message that C-\ and C-^ are |
| 1498 | now translated to C-s and C-q. (Use the same command M-x |
| 1499 | enable-flow-control to turn *off* this special mode. It toggles flow |
| 1500 | control handling.) |
| 1501 | |
| 1502 | If C-\ and C-^ are inconvenient for you (for example, if one of them |
| 1503 | is the escape character of your terminal concentrator), you can choose |
| 1504 | other characters by setting the variables flow-control-c-s-replacement |
| 1505 | and flow-control-c-q-replacement. But choose carefully, since all |
| 1506 | other control characters are already used by emacs. |
| 1507 | |
| 1508 | IMPORTANT: if you type C-s by accident while flow control is enabled, |
| 1509 | Emacs output will freeze, and you will have to remember to type C-q in |
| 1510 | order to continue. |
| 1511 | |
| 1512 | If you work in an environment where a majority of terminals of a |
| 1513 | certain type are flow control hobbled, you can use the function |
| 1514 | `enable-flow-control-on' to turn on this flow control avoidance scheme |
| 1515 | automatically. Here is an example: |
| 1516 | |
| 1517 | (enable-flow-control-on "vt200" "vt300" "vt101" "vt131") |
| 1518 | |
| 1519 | If this isn't quite correct (e.g. you have a mixture of flow-control hobbled |
| 1520 | and good vt200 terminals), you can still run enable-flow-control |
| 1521 | manually. |
| 1522 | |
| 1523 | I have no intention of ever redesigning the Emacs command set for the |
| 1524 | assumption that terminals use C-s/C-q flow control. XON/XOFF flow |
| 1525 | control technique is a bad design, and terminals that need it are bad |
| 1526 | merchandise and should not be purchased. Now that X is becoming |
| 1527 | widespread, XON/XOFF seems to be on the way out. If you can get some |
| 1528 | use out of GNU Emacs on inferior terminals, more power to you, but I |
| 1529 | will not make Emacs worse for properly designed systems for the sake |
| 1530 | of inferior systems. |
| 1531 | |
| 1532 | ** Control-S and Control-Q commands are ignored completely. |
| 1533 | |
| 1534 | For some reason, your system is using brain-damaged C-s/C-q flow |
| 1535 | control despite Emacs's attempts to turn it off. Perhaps your |
| 1536 | terminal is connected to the computer through a concentrator |
| 1537 | that wants to use flow control. |
| 1538 | |
| 1539 | You should first try to tell the concentrator not to use flow control. |
| 1540 | If you succeed in this, try making the terminal work without |
| 1541 | flow control, as described in the preceding section. |
| 1542 | |
| 1543 | If that line of approach is not successful, map some other characters |
| 1544 | into C-s and C-q using keyboard-translate-table. The example above |
| 1545 | shows how to do this with C-^ and C-\. |
| 1546 | |
| 1547 | ** Screen is updated wrong, but only on one kind of terminal. |
| 1548 | |
| 1549 | This could mean that the termcap entry you are using for that |
| 1550 | terminal is wrong, or it could mean that Emacs has a bug handing |
| 1551 | the combination of features specified for that terminal. |
| 1552 | |
| 1553 | The first step in tracking this down is to record what characters |
| 1554 | Emacs is sending to the terminal. Execute the Lisp expression |
| 1555 | (open-termscript "./emacs-script") to make Emacs write all |
| 1556 | terminal output into the file ~/emacs-script as well; then do |
| 1557 | what makes the screen update wrong, and look at the file |
| 1558 | and decode the characters using the manual for the terminal. |
| 1559 | There are several possibilities: |
| 1560 | |
| 1561 | 1) The characters sent are correct, according to the terminal manual. |
| 1562 | |
| 1563 | In this case, there is no obvious bug in Emacs, and most likely you |
| 1564 | need more padding, or possibly the terminal manual is wrong. |
| 1565 | |
| 1566 | 2) The characters sent are incorrect, due to an obscure aspect |
| 1567 | of the terminal behavior not described in an obvious way |
| 1568 | by termcap. |
| 1569 | |
| 1570 | This case is hard. It will be necessary to think of a way for |
| 1571 | Emacs to distinguish between terminals with this kind of behavior |
| 1572 | and other terminals that behave subtly differently but are |
| 1573 | classified the same by termcap; or else find an algorithm for |
| 1574 | Emacs to use that avoids the difference. Such changes must be |
| 1575 | tested on many kinds of terminals. |
| 1576 | |
| 1577 | 3) The termcap entry is wrong. |
| 1578 | |
| 1579 | See the file etc/TERMS for information on changes |
| 1580 | that are known to be needed in commonly used termcap entries |
| 1581 | for certain terminals. |
| 1582 | |
| 1583 | 4) The characters sent are incorrect, and clearly cannot be |
| 1584 | right for any terminal with the termcap entry you were using. |
| 1585 | |
| 1586 | This is unambiguously an Emacs bug, and can probably be fixed |
| 1587 | in termcap.c, tparam.c, term.c, scroll.c, cm.c or dispnew.c. |
| 1588 | |
| 1589 | ** Control-S and Control-Q commands are ignored completely on a net connection. |
| 1590 | |
| 1591 | Some versions of rlogin (and possibly telnet) do not pass flow |
| 1592 | control characters to the remote system to which they connect. |
| 1593 | On such systems, emacs on the remote system cannot disable flow |
| 1594 | control on the local system. Sometimes `rlogin -8' will avoid this |
| 1595 | problem. |
| 1596 | |
| 1597 | One way to cure this is to disable flow control on the local host |
| 1598 | (the one running rlogin, not the one running rlogind) using the |
| 1599 | stty command, before starting the rlogin process. On many systems, |
| 1600 | "stty start u stop u" will do this. On some systems, use |
| 1601 | "stty -ixon" instead. |
| 1602 | |
| 1603 | Some versions of tcsh will prevent even this from working. One way |
| 1604 | around this is to start another shell before starting rlogin, and |
| 1605 | issue the stty command to disable flow control from that shell. |
| 1606 | |
| 1607 | If none of these methods work, the best solution is to type |
| 1608 | M-x enable-flow-control at the beginning of your emacs session, or |
| 1609 | if you expect the problem to continue, add a line such as the |
| 1610 | following to your .emacs (on the host running rlogind): |
| 1611 | |
| 1612 | (enable-flow-control-on "vt200" "vt300" "vt101" "vt131") |
| 1613 | |
| 1614 | See the entry about spontaneous display of I-search (above) for more |
| 1615 | info. |
| 1616 | |
| 1617 | ** Output from Control-V is slow. |
| 1618 | |
| 1619 | On many bit-map terminals, scrolling operations are fairly slow. |
| 1620 | Often the termcap entry for the type of terminal in use fails |
| 1621 | to inform Emacs of this. The two lines at the bottom of the screen |
| 1622 | before a Control-V command are supposed to appear at the top after |
| 1623 | the Control-V command. If Emacs thinks scrolling the lines is fast, |
| 1624 | it will scroll them to the top of the screen. |
| 1625 | |
| 1626 | If scrolling is slow but Emacs thinks it is fast, the usual reason is |
| 1627 | that the termcap entry for the terminal you are using does not |
| 1628 | specify any padding time for the `al' and `dl' strings. Emacs |
| 1629 | concludes that these operations take only as much time as it takes to |
| 1630 | send the commands at whatever line speed you are using. You must |
| 1631 | fix the termcap entry to specify, for the `al' and `dl', as much |
| 1632 | time as the operations really take. |
| 1633 | |
| 1634 | Currently Emacs thinks in terms of serial lines which send characters |
| 1635 | at a fixed rate, so that any operation which takes time for the |
| 1636 | terminal to execute must also be padded. With bit-map terminals |
| 1637 | operated across networks, often the network provides some sort of |
| 1638 | flow control so that padding is never needed no matter how slow |
| 1639 | an operation is. You must still specify a padding time if you want |
| 1640 | Emacs to realize that the operation takes a long time. This will |
| 1641 | cause padding characters to be sent unnecessarily, but they do |
| 1642 | not really cost much. They will be transmitted while the scrolling |
| 1643 | is happening and then discarded quickly by the terminal. |
| 1644 | |
| 1645 | Most bit-map terminals provide commands for inserting or deleting |
| 1646 | multiple lines at once. Define the `AL' and `DL' strings in the |
| 1647 | termcap entry to say how to do these things, and you will have |
| 1648 | fast output without wasted padding characters. These strings should |
| 1649 | each contain a single %-spec saying how to send the number of lines |
| 1650 | to be scrolled. These %-specs are like those in the termcap |
| 1651 | `cm' string. |
| 1652 | |
| 1653 | You should also define the `IC' and `DC' strings if your terminal |
| 1654 | has a command to insert or delete multiple characters. These |
| 1655 | take the number of positions to insert or delete as an argument. |
| 1656 | |
| 1657 | A `cs' string to set the scrolling region will reduce the amount |
| 1658 | of motion you see on the screen when part of the screen is scrolled. |
| 1659 | |
| 1660 | ** You type Control-H (Backspace) expecting to delete characters. |
| 1661 | |
| 1662 | Put `stty dec' in your .login file and your problems will disappear |
| 1663 | after a day or two. |
| 1664 | |
| 1665 | The choice of Backspace for erasure was based on confusion, caused by |
| 1666 | the fact that backspacing causes erasure (later, when you type another |
| 1667 | character) on most display terminals. But it is a mistake. Deletion |
| 1668 | of text is not the same thing as backspacing followed by failure to |
| 1669 | overprint. I do not wish to propagate this confusion by conforming |
| 1670 | to it. |
| 1671 | |
| 1672 | For this reason, I believe `stty dec' is the right mode to use, |
| 1673 | and I have designed Emacs to go with that. If there were a thousand |
| 1674 | other control characters, I would define Control-h to delete as well; |
| 1675 | but there are not very many other control characters, and I think |
| 1676 | that providing the most mnemonic possible Help character is more |
| 1677 | important than adapting to people who don't use `stty dec'. |
| 1678 | |
| 1679 | If you are obstinate about confusing buggy overprinting with deletion, |
| 1680 | you can redefine Backspace in your .emacs file: |
| 1681 | (global-set-key "\b" 'delete-backward-char) |
| 1682 | You can probably access help-command via f1. |
| 1683 | |
| 1684 | ** Colors are not available on a tty or in xterm. |
| 1685 | |
| 1686 | Emacs 21 supports colors on character terminals and terminal |
| 1687 | emulators, but this support relies on the terminfo or termcap database |
| 1688 | entry to specify that the display supports color. Emacs looks at the |
| 1689 | "Co" capability for the terminal to find out how many colors are |
| 1690 | supported; it should be non-zero to activate the color support within |
| 1691 | Emacs. (Most color terminals support 8 or 16 colors.) If your system |
| 1692 | uses terminfo, the name of the capability equivalent to "Co" is |
| 1693 | "colors". |
| 1694 | |
| 1695 | In addition to the "Co" capability, Emacs needs the "op" (for |
| 1696 | ``original pair'') capability, which tells how to switch the terminal |
| 1697 | back to the default foreground and background colors. Emacs will not |
| 1698 | use colors if this capability is not defined. If your terminal entry |
| 1699 | doesn't provide such a capability, try using the ANSI standard escape |
| 1700 | sequence \E[00m (that is, define a new termcap/terminfo entry and make |
| 1701 | it use your current terminal's entry plus \E[00m for the "op" |
| 1702 | capability). |
| 1703 | |
| 1704 | Finally, the "NC" capability (terminfo name: "ncv") tells Emacs which |
| 1705 | attributes cannot be used with colors. Setting this capability |
| 1706 | incorrectly might have the effect of disabling colors; try setting |
| 1707 | this capability to `0' (zero) and see if that helps. |
| 1708 | |
| 1709 | Emacs uses the database entry for the terminal whose name is the value |
| 1710 | of the environment variable TERM. With `xterm', a common terminal |
| 1711 | entry that supports color is `xterm-color', so setting TERM's value to |
| 1712 | `xterm-color' might activate the color support on an xterm-compatible |
| 1713 | emulator. |
| 1714 | |
| 1715 | Beginning with version 22.1, Emacs supports the --color command-line |
| 1716 | option which may be used to force Emacs to use one of a few popular |
| 1717 | modes for getting colors on a tty. For example, --color=ansi8 sets up |
| 1718 | for using the ANSI-standard escape sequences that support 8 colors. |
| 1719 | |
| 1720 | Some modes do not use colors unless you turn on the Font-lock mode. |
| 1721 | Some people have long ago set their `~/.emacs' files to turn on |
| 1722 | Font-lock on X only, so they won't see colors on a tty. The |
| 1723 | recommended way of turning on Font-lock is by typing "M-x |
| 1724 | global-font-lock-mode RET" or by customizing the variable |
| 1725 | `global-font-lock-mode'. |
| 1726 | |
| 1727 | * Runtime problems specific to individual Unix variants |
| 1728 | |
| 1729 | ** GNU/Linux |
| 1730 | |
| 1731 | *** GNU/Linux: Process output is corrupted. |
| 1732 | |
| 1733 | There is a bug in Linux kernel 2.6.10 PTYs that can cause emacs to |
| 1734 | read corrupted process output. |
| 1735 | |
| 1736 | *** GNU/Linux: Remote access to CVS with SSH causes file corruption. |
| 1737 | |
| 1738 | If you access a remote CVS repository via SSH, files may be corrupted |
| 1739 | due to bad interaction between CVS, SSH, and libc. |
| 1740 | |
| 1741 | To fix the problem, save the following script into a file, make it |
| 1742 | executable, and set CVS_RSH environment variable to the file name of |
| 1743 | the script: |
| 1744 | |
| 1745 | #!/bin/bash |
| 1746 | exec 2> >(exec cat >&2 2>/dev/null) |
| 1747 | exec ssh "$@" |
| 1748 | |
| 1749 | *** GNU/Linux: On Linux-based GNU systems using libc versions 5.4.19 through |
| 1750 | 5.4.22, Emacs crashes at startup with a segmentation fault. |
| 1751 | |
| 1752 | This problem happens if libc defines the symbol __malloc_initialized. |
| 1753 | One known solution is to upgrade to a newer libc version. 5.4.33 is |
| 1754 | known to work. |
| 1755 | |
| 1756 | *** GNU/Linux: After upgrading to a newer version of Emacs, |
| 1757 | the Meta key stops working. |
| 1758 | |
| 1759 | This was reported to happen on a GNU/Linux system distributed by |
| 1760 | Mandrake. The reason is that the previous version of Emacs was |
| 1761 | modified by Mandrake to make the Alt key act as the Meta key, on a |
| 1762 | keyboard where the Windows key is the one which produces the Meta |
| 1763 | modifier. A user who started using a newer version of Emacs, which |
| 1764 | was not hacked by Mandrake, expected the Alt key to continue to act as |
| 1765 | Meta, and was astonished when that didn't happen. |
| 1766 | |
| 1767 | The solution is to find out what key on your keyboard produces the Meta |
| 1768 | modifier, and use that key instead. Try all of the keys to the left |
| 1769 | and to the right of the space bar, together with the `x' key, and see |
| 1770 | which combination produces "M-x" in the echo area. You can also use |
| 1771 | the `xmodmap' utility to show all the keys which produce a Meta |
| 1772 | modifier: |
| 1773 | |
| 1774 | xmodmap -pk | egrep -i "meta|alt" |
| 1775 | |
| 1776 | A more convenient way of finding out which keys produce a Meta modifier |
| 1777 | is to use the `xkbprint' utility, if it's available on your system: |
| 1778 | |
| 1779 | xkbprint 0:0 /tmp/k.ps |
| 1780 | |
| 1781 | This produces a PostScript file `/tmp/k.ps' with a picture of your |
| 1782 | keyboard; printing that file on a PostScript printer will show what |
| 1783 | keys can serve as Meta. |
| 1784 | |
| 1785 | The `xkeycaps' also shows a visual representation of the current |
| 1786 | keyboard settings. It also allows to modify them. |
| 1787 | |
| 1788 | *** GNU/Linux: slow startup on Linux-based GNU systems. |
| 1789 | |
| 1790 | People using systems based on the Linux kernel sometimes report that |
| 1791 | startup takes 10 to 15 seconds longer than `usual'. |
| 1792 | |
| 1793 | This is because Emacs looks up the host name when it starts. |
| 1794 | Normally, this takes negligible time; the extra delay is due to |
| 1795 | improper system configuration. This problem can occur for both |
| 1796 | networked and non-networked machines. |
| 1797 | |
| 1798 | Here is how to fix the configuration. It requires being root. |
| 1799 | |
| 1800 | **** Networked Case. |
| 1801 | |
| 1802 | First, make sure the files `/etc/hosts' and `/etc/host.conf' both |
| 1803 | exist. The first line in the `/etc/hosts' file should look like this |
| 1804 | (replace HOSTNAME with your host name): |
| 1805 | |
| 1806 | 127.0.0.1 HOSTNAME |
| 1807 | |
| 1808 | Also make sure that the `/etc/host.conf' files contains the following |
| 1809 | lines: |
| 1810 | |
| 1811 | order hosts, bind |
| 1812 | multi on |
| 1813 | |
| 1814 | Any changes, permanent and temporary, to the host name should be |
| 1815 | indicated in the `/etc/hosts' file, since it acts a limited local |
| 1816 | database of addresses and names (e.g., some SLIP connections |
| 1817 | dynamically allocate ip addresses). |
| 1818 | |
| 1819 | **** Non-Networked Case. |
| 1820 | |
| 1821 | The solution described in the networked case applies here as well. |
| 1822 | However, if you never intend to network your machine, you can use a |
| 1823 | simpler solution: create an empty `/etc/host.conf' file. The command |
| 1824 | `touch /etc/host.conf' suffices to create the file. The `/etc/hosts' |
| 1825 | file is not necessary with this approach. |
| 1826 | |
| 1827 | *** GNU/Linux: Emacs on a tty switches the cursor to large blinking block. |
| 1828 | |
| 1829 | This was reported to happen on some GNU/Linux systems which use |
| 1830 | ncurses version 5.0, but could be relevant for other versions as well. |
| 1831 | These versions of ncurses come with a `linux' terminfo entry, where |
| 1832 | the "cvvis" capability (termcap "vs") is defined as "\E[?25h\E[?8c" |
| 1833 | (show cursor, change size). This escape sequence switches on a |
| 1834 | blinking hardware text-mode cursor whose size is a full character |
| 1835 | cell. This blinking cannot be stopped, since a hardware cursor |
| 1836 | always blinks. |
| 1837 | |
| 1838 | A work-around is to redefine the "cvvis" capability so that it |
| 1839 | enables a *software* cursor. The software cursor works by inverting |
| 1840 | the colors of the character at point, so what you see is a block |
| 1841 | cursor that doesn't blink. For this to work, you need to redefine |
| 1842 | the "cnorm" capability as well, so that it operates on the software |
| 1843 | cursor instead of the hardware cursor. |
| 1844 | |
| 1845 | To this end, run "infocmp linux > linux-term", edit the file |
| 1846 | `linux-term' to make both the "cnorm" and "cvvis" capabilities send |
| 1847 | the sequence "\E[?25h\E[?17;0;64c", and then run "tic linux-term" to |
| 1848 | produce a modified terminfo entry. |
| 1849 | |
| 1850 | Alternatively, if you want a blinking underscore as your Emacs cursor, |
| 1851 | change the "cvvis" capability to send the "\E[?25h\E[?0c" command. |
| 1852 | |
| 1853 | *** GNU/Linux: Error messages `internal facep []' happen on GNU/Linux systems. |
| 1854 | |
| 1855 | There is a report that replacing libc.so.5.0.9 with libc.so.5.2.16 |
| 1856 | caused this to start happening. People are not sure why, but the |
| 1857 | problem seems unlikely to be in Emacs itself. Some suspect that it |
| 1858 | is actually Xlib which won't work with libc.so.5.2.16. |
| 1859 | |
| 1860 | Using the old library version is a workaround. |
| 1861 | |
| 1862 | ** FreeBSD |
| 1863 | |
| 1864 | *** FreeBSD 2.1.5: useless symbolic links remain in /tmp or other |
| 1865 | directories that have the +t bit. |
| 1866 | |
| 1867 | This is because of a kernel bug in FreeBSD 2.1.5 (fixed in 2.2). |
| 1868 | Emacs uses symbolic links to implement file locks. In a directory |
| 1869 | with +t bit, the directory owner becomes the owner of the symbolic |
| 1870 | link, so that it cannot be removed by anyone else. |
| 1871 | |
| 1872 | If you don't like those useless links, you can let Emacs not to using |
| 1873 | file lock by adding #undef CLASH_DETECTION to config.h. |
| 1874 | |
| 1875 | *** FreeBSD: Getting a Meta key on the console. |
| 1876 | |
| 1877 | By default, neither Alt nor any other key acts as a Meta key on |
| 1878 | FreeBSD, but this can be changed using kbdcontrol(1). Dump the |
| 1879 | current keymap to a file with the command |
| 1880 | |
| 1881 | $ kbdcontrol -d >emacs.kbd |
| 1882 | |
| 1883 | Edit emacs.kbd, and give the key you want to be the Meta key the |
| 1884 | definition `meta'. For instance, if your keyboard has a ``Windows'' |
| 1885 | key with scan code 105, change the line for scan code 105 in emacs.kbd |
| 1886 | to look like this |
| 1887 | |
| 1888 | 105 meta meta meta meta meta meta meta meta O |
| 1889 | |
| 1890 | to make the Windows key the Meta key. Load the new keymap with |
| 1891 | |
| 1892 | $ kbdcontrol -l emacs.kbd |
| 1893 | |
| 1894 | ** HP-UX |
| 1895 | |
| 1896 | *** HP/UX : Shell mode gives the message, "`tty`: Ambiguous". |
| 1897 | |
| 1898 | christos@theory.tn.cornell.edu says: |
| 1899 | |
| 1900 | The problem is that in your .cshrc you have something that tries to |
| 1901 | execute `tty`. If you are not running the shell on a real tty then |
| 1902 | tty will print "not a tty". Csh expects one word in some places, |
| 1903 | but tty is giving it back 3. |
| 1904 | |
| 1905 | The solution is to add a pair of quotes around `tty` to make it a single |
| 1906 | word: |
| 1907 | |
| 1908 | if (`tty` == "/dev/console") |
| 1909 | |
| 1910 | should be changed to: |
| 1911 | |
| 1912 | if ("`tty`" == "/dev/console") |
| 1913 | |
| 1914 | Even better, move things that set up terminal sections out of .cshrc |
| 1915 | and into .login. |
| 1916 | |
| 1917 | *** HP/UX: `Pid xxx killed due to text modification or page I/O error'. |
| 1918 | |
| 1919 | On HP/UX, you can get that error when the Emacs executable is on an NFS |
| 1920 | file system. HP/UX responds this way if it tries to swap in a page and |
| 1921 | does not get a response from the server within a timeout whose default |
| 1922 | value is just ten seconds. |
| 1923 | |
| 1924 | If this happens to you, extend the timeout period. |
| 1925 | |
| 1926 | *** HP/UX: The right Alt key works wrong on German HP keyboards (and perhaps |
| 1927 | other non-English HP keyboards too). |
| 1928 | |
| 1929 | This is because HP-UX defines the modifiers wrong in X. Here is a |
| 1930 | shell script to fix the problem; be sure that it is run after VUE |
| 1931 | configures the X server. |
| 1932 | |
| 1933 | xmodmap 2> /dev/null - << EOF |
| 1934 | keysym Alt_L = Meta_L |
| 1935 | keysym Alt_R = Meta_R |
| 1936 | EOF |
| 1937 | |
| 1938 | xmodmap - << EOF |
| 1939 | clear mod1 |
| 1940 | keysym Mode_switch = NoSymbol |
| 1941 | add mod1 = Meta_L |
| 1942 | keysym Meta_R = Mode_switch |
| 1943 | add mod2 = Mode_switch |
| 1944 | EOF |
| 1945 | |
| 1946 | *** HP/UX: "Cannot find callback list" messages from dialog boxes in |
| 1947 | Emacs built with Motif. |
| 1948 | |
| 1949 | This problem resulted from a bug in GCC 2.4.5. Newer GCC versions |
| 1950 | such as 2.7.0 fix the problem. |
| 1951 | |
| 1952 | *** HP/UX: Emacs does not recognize the AltGr key. |
| 1953 | |
| 1954 | To fix this, set up a file ~/.dt/sessions/sessionetc with executable |
| 1955 | rights, containing this text: |
| 1956 | |
| 1957 | -------------------------------- |
| 1958 | xmodmap 2> /dev/null - << EOF |
| 1959 | keysym Alt_L = Meta_L |
| 1960 | keysym Alt_R = Meta_R |
| 1961 | EOF |
| 1962 | |
| 1963 | xmodmap - << EOF |
| 1964 | clear mod1 |
| 1965 | keysym Mode_switch = NoSymbol |
| 1966 | add mod1 = Meta_L |
| 1967 | keysym Meta_R = Mode_switch |
| 1968 | add mod2 = Mode_switch |
| 1969 | EOF |
| 1970 | -------------------------------- |
| 1971 | |
| 1972 | *** HP/UX 11.0: Emacs makes HP/UX 11.0 crash. |
| 1973 | |
| 1974 | This is a bug in HPUX; HPUX patch PHKL_16260 is said to fix it. |
| 1975 | |
| 1976 | ** AIX |
| 1977 | |
| 1978 | *** AIX: Trouble using ptys. |
| 1979 | |
| 1980 | People often install the pty devices on AIX incorrectly. |
| 1981 | Use `smit pty' to reinstall them properly. |
| 1982 | |
| 1983 | *** AIXterm: Your Delete key sends a Backspace to the terminal. |
| 1984 | |
| 1985 | The solution is to include in your .Xdefaults the lines: |
| 1986 | |
| 1987 | *aixterm.Translations: #override <Key>BackSpace: string(0x7f) |
| 1988 | aixterm*ttyModes: erase ^? |
| 1989 | |
| 1990 | This makes your Backspace key send DEL (ASCII 127). |
| 1991 | |
| 1992 | *** AIX: If linking fails because libXbsd isn't found, check if you |
| 1993 | are compiling with the system's `cc' and CFLAGS containing `-O5'. If |
| 1994 | so, you have hit a compiler bug. Please make sure to re-configure |
| 1995 | Emacs so that it isn't compiled with `-O5'. |
| 1996 | |
| 1997 | *** AIX 4.3.x or 4.4: Compiling fails. |
| 1998 | |
| 1999 | This could happen if you use /bin/c89 as your compiler, instead of |
| 2000 | the default `cc'. /bin/c89 treats certain warnings, such as benign |
| 2001 | redefinitions of macros, as errors, and fails the build. A solution |
| 2002 | is to use the default compiler `cc'. |
| 2003 | |
| 2004 | *** AIX 4: Some programs fail when run in a Shell buffer |
| 2005 | with an error message like No terminfo entry for "unknown". |
| 2006 | |
| 2007 | On AIX, many terminal type definitions are not installed by default. |
| 2008 | `unknown' is one of them. Install the "Special Generic Terminal |
| 2009 | Definitions" to make them defined. |
| 2010 | |
| 2011 | ** Solaris |
| 2012 | |
| 2013 | We list bugs in current versions here. Solaris 2.x and 4.x are covered in the |
| 2014 | section on legacy systems. |
| 2015 | |
| 2016 | *** On Solaris, C-x doesn't get through to Emacs when you use the console. |
| 2017 | |
| 2018 | This is a Solaris feature (at least on Intel x86 cpus). Type C-r |
| 2019 | C-r C-t, to toggle whether C-x gets through to Emacs. |
| 2020 | |
| 2021 | *** Problem with remote X server on Suns. |
| 2022 | |
| 2023 | On a Sun, running Emacs on one machine with the X server on another |
| 2024 | may not work if you have used the unshared system libraries. This |
| 2025 | is because the unshared libraries fail to use YP for host name lookup. |
| 2026 | As a result, the host name you specify may not be recognized. |
| 2027 | |
| 2028 | *** Solaris 2,6: Emacs crashes with SIGBUS or SIGSEGV on Solaris after you delete a frame. |
| 2029 | |
| 2030 | We suspect that this is a bug in the X libraries provided by |
| 2031 | Sun. There is a report that one of these patches fixes the bug and |
| 2032 | makes the problem stop: |
| 2033 | |
| 2034 | 105216-01 105393-01 105518-01 105621-01 105665-01 105615-02 105216-02 |
| 2035 | 105667-01 105401-08 105615-03 105621-02 105686-02 105736-01 105755-03 |
| 2036 | 106033-01 105379-01 105786-01 105181-04 105379-03 105786-04 105845-01 |
| 2037 | 105284-05 105669-02 105837-01 105837-02 105558-01 106125-02 105407-01 |
| 2038 | |
| 2039 | Another person using a newer system (kernel patch level Generic_105181-06) |
| 2040 | suspects that the bug was fixed by one of these more recent patches: |
| 2041 | |
| 2042 | 106040-07 SunOS 5.6: X Input & Output Method patch |
| 2043 | 106222-01 OpenWindows 3.6: filemgr (ff.core) fixes |
| 2044 | 105284-12 Motif 1.2.7: sparc Runtime library patch |
| 2045 | |
| 2046 | *** Solaris 7 or 8: Emacs reports a BadAtom error (from X) |
| 2047 | |
| 2048 | This happens when Emacs was built on some other version of Solaris. |
| 2049 | Rebuild it on Solaris 8. |
| 2050 | |
| 2051 | *** When using M-x dbx with the SparcWorks debugger, the `up' and `down' |
| 2052 | commands do not move the arrow in Emacs. |
| 2053 | |
| 2054 | You can fix this by adding the following line to `~/.dbxinit': |
| 2055 | |
| 2056 | dbxenv output_short_file_name off |
| 2057 | |
| 2058 | *** On Solaris, CTRL-t is ignored by Emacs when you use |
| 2059 | the fr.ISO-8859-15 locale (and maybe other related locales). |
| 2060 | |
| 2061 | You can fix this by editing the file: |
| 2062 | |
| 2063 | /usr/openwin/lib/locale/iso8859-15/Compose |
| 2064 | |
| 2065 | Near the bottom there is a line that reads: |
| 2066 | |
| 2067 | Ctrl<t> <quotedbl> <Y> : "\276" threequarters |
| 2068 | |
| 2069 | that should read: |
| 2070 | |
| 2071 | Ctrl<T> <quotedbl> <Y> : "\276" threequarters |
| 2072 | |
| 2073 | Note the lower case <t>. Changing this line should make C-t work. |
| 2074 | |
| 2075 | *** On Solaris, Emacs fails to set menu-bar-update-hook on startup, with error |
| 2076 | "Error in menu-bar-update-hook: (error Point before start of properties)". |
| 2077 | This seems to be a GCC optimization bug that occurs for GCC 4.1.2 (-g |
| 2078 | and -g -O2) and GCC 4.2.3 (-g -O and -g -O2). You can fix this by |
| 2079 | compiling with GCC 4.2.3 or CC 5.7, with no optimizations. |
| 2080 | |
| 2081 | ** Irix |
| 2082 | |
| 2083 | *** Irix 6.5: Emacs crashes on the SGI R10K, when compiled with GCC. |
| 2084 | |
| 2085 | This seems to be fixed in GCC 2.95. |
| 2086 | |
| 2087 | *** Irix: Trouble using ptys, or running out of ptys. |
| 2088 | |
| 2089 | The program mkpts (which may be in `/usr/adm' or `/usr/sbin') needs to |
| 2090 | be set-UID to root, or non-root programs like Emacs will not be able |
| 2091 | to allocate ptys reliably. |
| 2092 | |
| 2093 | * Runtime problems specific to MS-Windows |
| 2094 | |
| 2095 | ** PATH can contain unexpanded environment variables |
| 2096 | |
| 2097 | Old releases of TCC (version 9) and 4NT (up to version 8) do not correctly |
| 2098 | expand App Paths entries of type REG_EXPAND_SZ. When Emacs is run from TCC |
| 2099 | and such an entry exists for emacs.exe, exec-path will contain the |
| 2100 | unexpanded entry. This has been fixed in TCC 10. For more information, |
| 2101 | see bug#2062. |
| 2102 | |
| 2103 | ** Setting w32-pass-rwindow-to-system and w32-pass-lwindow-to-system to nil |
| 2104 | does not prevent the Start menu from popping up when the left or right |
| 2105 | ``Windows'' key is pressed. |
| 2106 | |
| 2107 | This was reported to happen when XKeymacs is installed. At least with |
| 2108 | XKeymacs Version 3.47, deactivating XKeymacs when Emacs is active is |
| 2109 | not enough to avoid its messing with the keyboard input. Exiting |
| 2110 | XKeymacs completely is reported to solve the problem. |
| 2111 | |
| 2112 | ** Windows 95 and networking. |
| 2113 | |
| 2114 | To support server sockets, Emacs 22.1 loads ws2_32.dll. If this file |
| 2115 | is missing, all Emacs networking features are disabled. |
| 2116 | |
| 2117 | Old versions of Windows 95 may not have the required DLL. To use |
| 2118 | Emacs' networking features on Windows 95, you must install the |
| 2119 | "Windows Socket 2" update available from MicroSoft's support Web. |
| 2120 | |
| 2121 | ** Emacs exits with "X protocol error" when run with an X server for MS-Windows. |
| 2122 | |
| 2123 | A certain X server for Windows had a bug which caused this. |
| 2124 | Supposedly the newer 32-bit version of this server doesn't have the |
| 2125 | problem. |
| 2126 | |
| 2127 | ** Emacs crashes when opening a file with a UNC path and rails-mode is loaded. |
| 2128 | |
| 2129 | Loading rails-mode seems to interfere with UNC path handling. This has been |
| 2130 | reported as a bug against both Emacs and rails-mode, so look for an updated |
| 2131 | rails-mode that avoids this crash, or avoid using UNC paths if using |
| 2132 | rails-mode. |
| 2133 | |
| 2134 | ** Known problems with the MS-Windows port of Emacs 22.3 |
| 2135 | |
| 2136 | M-x term does not work on MS-Windows. TTY emulation on Windows is |
| 2137 | undocumented, and programs such as stty which are used on posix platforms |
| 2138 | to control tty emulation do not exist for native windows terminals. |
| 2139 | |
| 2140 | Using create-fontset-from-ascii-font or the --font startup parameter |
| 2141 | with a Chinese, Japanese or Korean font leads to display problems. |
| 2142 | Use a Latin-only font as your default font. If you want control over |
| 2143 | which font is used to display Chinese, Japanese or Korean character, |
| 2144 | use create-fontset-from-fontset-spec to define a fontset. |
| 2145 | |
| 2146 | Frames are not refreshed while the File or Font dialog or a pop-up menu |
| 2147 | is displayed. This also means help text for pop-up menus is not |
| 2148 | displayed at all. This is because message handling under Windows is |
| 2149 | synchronous, so we cannot handle repaint (or any other) messages while |
| 2150 | waiting for a system function to return the result of the dialog or |
| 2151 | pop-up menu interaction. |
| 2152 | |
| 2153 | Windows 95 and Windows NT up to version 4.0 do not support help text |
| 2154 | for menus. Help text is only available in later versions of Windows. |
| 2155 | |
| 2156 | When "ClearType" method is selected as the "method to smooth edges of |
| 2157 | screen fonts" (in Display Properties, Appearance tab, under |
| 2158 | "Effects"), there are various problems related to display of |
| 2159 | characters: Bold fonts can be hard to read, small portions of some |
| 2160 | characters could appear chopped, etc. This happens because, under |
| 2161 | ClearType, characters are drawn outside their advertised bounding box. |
| 2162 | Emacs 21 disabled the use of ClearType, whereas Emacs 22 allows it and |
| 2163 | has some code to enlarge the width of the bounding box. Apparently, |
| 2164 | this display feature needs more changes to get it 100% right. A |
| 2165 | workaround is to disable ClearType. |
| 2166 | |
| 2167 | There are problems with display if mouse-tracking is enabled and the |
| 2168 | mouse is moved off a frame, over another frame then back over the first |
| 2169 | frame. A workaround is to click the left mouse button inside the frame |
| 2170 | after moving back into it. |
| 2171 | |
| 2172 | Some minor flickering still persists during mouse-tracking, although |
| 2173 | not as severely as in 21.1. |
| 2174 | |
| 2175 | An inactive cursor remains in an active window after the Windows |
| 2176 | Manager driven switch of the focus, until a key is pressed. |
| 2177 | |
| 2178 | Windows input methods are not recognized by Emacs. However, some |
| 2179 | of these input methods cause the keyboard to send characters encoded |
| 2180 | in the appropriate coding system (e.g., ISO 8859-1 for Latin-1 |
| 2181 | characters, ISO 8859-8 for Hebrew characters, etc.). To make these |
| 2182 | input methods work with Emacs, set the keyboard coding system to the |
| 2183 | appropriate value after you activate the Windows input method. For |
| 2184 | example, if you activate the Hebrew input method, type this: |
| 2185 | |
| 2186 | C-x RET k hebrew-iso-8bit RET |
| 2187 | |
| 2188 | (Emacs ought to recognize the Windows language-change event and set up |
| 2189 | the appropriate keyboard encoding automatically, but it doesn't do |
| 2190 | that yet.) In addition, to use these Windows input methods, you |
| 2191 | should set your "Language for non-Unicode programs" (on Windows XP, |
| 2192 | this is on the Advanced tab of Regional Settings) to the language of |
| 2193 | the input method. |
| 2194 | |
| 2195 | To bind keys that produce non-ASCII characters with modifiers, you |
| 2196 | must specify raw byte codes. For instance, if you want to bind |
| 2197 | META-a-grave to a command, you need to specify this in your `~/.emacs': |
| 2198 | |
| 2199 | (global-set-key [?\M-\340] ...) |
| 2200 | |
| 2201 | The above example is for the Latin-1 environment where the byte code |
| 2202 | of the encoded a-grave is 340 octal. For other environments, use the |
| 2203 | encoding appropriate to that environment. |
| 2204 | |
| 2205 | The %b specifier for format-time-string does not produce abbreviated |
| 2206 | month names with consistent widths for some locales on some versions |
| 2207 | of Windows. This is caused by a deficiency in the underlying system |
| 2208 | library function. |
| 2209 | |
| 2210 | The function set-time-zone-rule gives incorrect results for many |
| 2211 | non-US timezones. This is due to over-simplistic handling of |
| 2212 | daylight savings switchovers by the Windows libraries. |
| 2213 | |
| 2214 | Files larger than 4GB cause overflow in the size (represented as a |
| 2215 | 32-bit integer) reported by `file-attributes'. This affects Dired as |
| 2216 | well, since the Windows port uses a Lisp emulation of `ls' that relies |
| 2217 | on `file-attributes'. |
| 2218 | |
| 2219 | Sound playing is not supported with the `:data DATA' key-value pair. |
| 2220 | You _must_ use the `:file FILE' method. |
| 2221 | |
| 2222 | ** Typing Alt-Shift has strange effects on MS-Windows. |
| 2223 | |
| 2224 | This combination of keys is a command to change keyboard layout. If |
| 2225 | you proceed to type another non-modifier key before you let go of Alt |
| 2226 | and Shift, the Alt and Shift act as modifiers in the usual way. A |
| 2227 | more permanent work around is to change it to another key combination, |
| 2228 | or disable it in the "Regional and Language Options" applet of the |
| 2229 | Control Panel. (The exact sequence of mouse clicks in the "Regional |
| 2230 | and Language Options" applet needed to find the key combination that |
| 2231 | changes the keyboard layout depends on your Windows version; for XP, |
| 2232 | in the Languages tab, click "Details" and then "Key Settings".) |
| 2233 | |
| 2234 | ** Interrupting Cygwin port of Bash from Emacs doesn't work. |
| 2235 | |
| 2236 | Cygwin 1.x builds of the ported Bash cannot be interrupted from the |
| 2237 | MS-Windows version of Emacs. This is due to some change in the Bash |
| 2238 | port or in the Cygwin library which apparently make Bash ignore the |
| 2239 | keyboard interrupt event sent by Emacs to Bash. (Older Cygwin ports |
| 2240 | of Bash, up to b20.1, did receive SIGINT from Emacs.) |
| 2241 | |
| 2242 | ** Accessing remote files with ange-ftp hangs the MS-Windows version of Emacs. |
| 2243 | |
| 2244 | If the FTP client is the Cygwin port of GNU `ftp', this appears to be |
| 2245 | due to some bug in the Cygwin DLL or some incompatibility between it |
| 2246 | and the implementation of asynchronous subprocesses in the Windows |
| 2247 | port of Emacs. Specifically, some parts of the FTP server responses |
| 2248 | are not flushed out, apparently due to buffering issues, which |
| 2249 | confuses ange-ftp. |
| 2250 | |
| 2251 | The solution is to downgrade to an older version of the Cygwin DLL |
| 2252 | (version 1.3.2 was reported to solve the problem), or use the stock |
| 2253 | Windows FTP client, usually found in the `C:\WINDOWS' or 'C:\WINNT' |
| 2254 | directory. To force ange-ftp use the stock Windows client, set the |
| 2255 | variable `ange-ftp-ftp-program-name' to the absolute file name of the |
| 2256 | client's executable. For example: |
| 2257 | |
| 2258 | (setq ange-ftp-ftp-program-name "c:/windows/ftp.exe") |
| 2259 | |
| 2260 | If you want to stick with the Cygwin FTP client, you can work around |
| 2261 | this problem by putting this in your `.emacs' file: |
| 2262 | |
| 2263 | (setq ange-ftp-ftp-program-args '("-i" "-n" "-g" "-v" "--prompt" "") |
| 2264 | |
| 2265 | ** lpr commands don't work on MS-Windows with some cheap printers. |
| 2266 | |
| 2267 | This problem may also strike other platforms, but the solution is |
| 2268 | likely to be a global one, and not Emacs specific. |
| 2269 | |
| 2270 | Many cheap inkjet, and even some cheap laser printers, do not |
| 2271 | print plain text anymore, they will only print through graphical |
| 2272 | printer drivers. A workaround on MS-Windows is to use Windows' basic |
| 2273 | built in editor to print (this is possibly the only useful purpose it |
| 2274 | has): |
| 2275 | |
| 2276 | (setq printer-name "") ;; notepad takes the default |
| 2277 | (setq lpr-command "notepad") ;; notepad |
| 2278 | (setq lpr-switches nil) ;; not needed |
| 2279 | (setq lpr-printer-switch "/P") ;; run notepad as batch printer |
| 2280 | |
| 2281 | ** Antivirus software interacts badly with the MS-Windows version of Emacs. |
| 2282 | |
| 2283 | The usual manifestation of these problems is that subprocesses don't |
| 2284 | work or even wedge the entire system. In particular, "M-x shell RET" |
| 2285 | was reported to fail to work. But other commands also sometimes don't |
| 2286 | work when an antivirus package is installed. |
| 2287 | |
| 2288 | The solution is to switch the antivirus software to a less aggressive |
| 2289 | mode (e.g., disable the ``auto-protect'' feature), or even uninstall |
| 2290 | or disable it entirely. |
| 2291 | |
| 2292 | ** Pressing the mouse button on MS-Windows does not give a mouse-2 event. |
| 2293 | |
| 2294 | This is usually a problem with the mouse driver. Because most Windows |
| 2295 | programs do not do anything useful with the middle mouse button, many |
| 2296 | mouse drivers allow you to define the wheel press to do something |
| 2297 | different. Some drivers do not even have the option to generate a |
| 2298 | middle button press. In such cases, setting the wheel press to |
| 2299 | "scroll" sometimes works if you press the button twice. Trying a |
| 2300 | generic mouse driver might help. |
| 2301 | |
| 2302 | ** Scrolling the mouse wheel on MS-Windows always scrolls the top window. |
| 2303 | |
| 2304 | This is another common problem with mouse drivers. Instead of |
| 2305 | generating scroll events, some mouse drivers try to fake scroll bar |
| 2306 | movement. But they are not intelligent enough to handle multiple |
| 2307 | scroll bars within a frame. Trying a generic mouse driver might help. |
| 2308 | |
| 2309 | ** Mail sent through Microsoft Exchange in some encodings appears to be |
| 2310 | mangled and is not seen correctly in Rmail or Gnus. We don't know |
| 2311 | exactly what happens, but it isn't an Emacs problem in cases we've |
| 2312 | seen. |
| 2313 | |
| 2314 | ** On MS-Windows, you cannot use the right-hand ALT key and the left-hand |
| 2315 | CTRL key together to type a Control-Meta character. |
| 2316 | |
| 2317 | This is a consequence of a misfeature beyond Emacs's control. |
| 2318 | |
| 2319 | Under Windows, the AltGr key on international keyboards generates key |
| 2320 | events with the modifiers Right-Alt and Left-Ctrl. Since Emacs cannot |
| 2321 | distinguish AltGr from an explicit Right-Alt and Left-Ctrl |
| 2322 | combination, whenever it sees Right-Alt and Left-Ctrl it assumes that |
| 2323 | AltGr has been pressed. The variable `w32-recognize-altgr' can be set |
| 2324 | to nil to tell Emacs that AltGr is really Ctrl and Alt. |
| 2325 | |
| 2326 | ** Under some X-servers running on MS-Windows, Emacs' display is incorrect. |
| 2327 | |
| 2328 | The symptoms are that Emacs does not completely erase blank areas of the |
| 2329 | screen during scrolling or some other screen operations (e.g., selective |
| 2330 | display or when killing a region). M-x recenter will cause the screen |
| 2331 | to be completely redisplayed and the "extra" characters will disappear. |
| 2332 | |
| 2333 | This is known to occur under Exceed 6, and possibly earlier versions |
| 2334 | as well; it is reportedly solved in version 6.2.0.16 and later. The |
| 2335 | problem lies in the X-server settings. |
| 2336 | |
| 2337 | There are reports that you can solve the problem with Exceed by |
| 2338 | running `Xconfig' from within NT, choosing "X selection", then |
| 2339 | un-checking the boxes "auto-copy X selection" and "auto-paste to X |
| 2340 | selection". |
| 2341 | |
| 2342 | Of this does not work, please inform bug-gnu-emacs@gnu.org. Then |
| 2343 | please call support for your X-server and see if you can get a fix. |
| 2344 | If you do, please send it to bug-gnu-emacs@gnu.org so we can list it |
| 2345 | here. |
| 2346 | |
| 2347 | * Build-time problems |
| 2348 | |
| 2349 | ** Configuration |
| 2350 | |
| 2351 | *** The `configure' script doesn't find the jpeg library. |
| 2352 | |
| 2353 | There are reports that this happens on some systems because the linker |
| 2354 | by default only looks for shared libraries, but jpeg distribution by |
| 2355 | default only installs a nonshared version of the library, `libjpeg.a'. |
| 2356 | |
| 2357 | If this is the problem, you can configure the jpeg library with the |
| 2358 | `--enable-shared' option and then rebuild libjpeg. This produces a |
| 2359 | shared version of libjpeg, which you need to install. Finally, rerun |
| 2360 | the Emacs configure script, which should now find the jpeg library. |
| 2361 | Alternatively, modify the generated src/Makefile to link the .a file |
| 2362 | explicitly, and edit src/config.h to define HAVE_JPEG. |
| 2363 | |
| 2364 | *** `configure' warns ``accepted by the compiler, rejected by the preprocessor''. |
| 2365 | |
| 2366 | This indicates a mismatch between the C compiler and preprocessor that |
| 2367 | configure is using. For example, on Solaris 10 trying to use |
| 2368 | CC=/opt/SUNWspro/bin/cc (the Sun Studio compiler) together with |
| 2369 | CPP=/usr/ccs/lib/cpp can result in errors of this form (you may also |
| 2370 | see the error ``"/usr/include/sys/isa_defs.h", line 500: undefined control''). |
| 2371 | |
| 2372 | The solution is to tell configure to use the correct C preprocessor |
| 2373 | for your C compiler (CPP="/opt/SUNWspro/bin/cc -E" in the above |
| 2374 | example). |
| 2375 | |
| 2376 | *** `configure' fails with ``"junk.c", line 660: invalid input token: 8.elc'' |
| 2377 | |
| 2378 | The final stage of the Emacs configure process uses the C preprocessor |
| 2379 | to generate the Makefiles. Errors of this form can occur if the C |
| 2380 | preprocessor inserts extra whitespace into its output. The solution |
| 2381 | is to find the switches that stop your preprocessor from inserting extra |
| 2382 | whitespace, add them to CPPFLAGS, and re-run configure. For example, |
| 2383 | this error can occur on Solaris 10 when using the Sun Studio compiler |
| 2384 | ``Sun C 5.8'' with its preprocessor CPP="/opt/SUNWspro/bin/cc -E". |
| 2385 | The relevant switch in this case is "-Xs" (``compile assuming |
| 2386 | (pre-ANSI) K & R C style code''). |
| 2387 | |
| 2388 | ** Compilation |
| 2389 | |
| 2390 | *** Building Emacs over NFS fails with ``Text file busy''. |
| 2391 | |
| 2392 | This was reported to happen when building Emacs on a GNU/Linux system |
| 2393 | (Red Hat Linux 6.2) using a build directory automounted from Solaris |
| 2394 | (SunOS 5.6) file server, but it might not be limited to that |
| 2395 | configuration alone. Presumably, the NFS server doesn't commit the |
| 2396 | files' data to disk quickly enough, and the Emacs executable file is |
| 2397 | left ``busy'' for several seconds after Emacs has finished dumping |
| 2398 | itself. This causes the subsequent commands which invoke the dumped |
| 2399 | Emacs executable to fail with the above message. |
| 2400 | |
| 2401 | In some of these cases, a time skew between the NFS server and the |
| 2402 | machine where Emacs is built is detected and reported by GNU Make |
| 2403 | (it says that some of the files have modification time in the future). |
| 2404 | This might be a symptom of NFS-related problems. |
| 2405 | |
| 2406 | If the NFS server runs on Solaris, apply the Solaris patch 105379-05 |
| 2407 | (Sunos 5.6: /kernel/misc/nfssrv patch). If that doesn't work, or if |
| 2408 | you have a different version of the OS or the NFS server, you can |
| 2409 | force the NFS server to use 1KB blocks, which was reported to fix the |
| 2410 | problem albeit at a price of slowing down file I/O. You can force 1KB |
| 2411 | blocks by specifying the "-o rsize=1024,wsize=1024" options to the |
| 2412 | `mount' command, or by adding ",rsize=1024,wsize=1024" to the mount |
| 2413 | options in the appropriate system configuration file, such as |
| 2414 | `/etc/auto.home'. |
| 2415 | |
| 2416 | Alternatively, when Make fails due to this problem, you could wait for |
| 2417 | a few seconds and then invoke Make again. In one particular case, |
| 2418 | waiting for 10 or more seconds between the two Make invocations seemed |
| 2419 | to work around the problem. |
| 2420 | |
| 2421 | Similar problems can happen if your machine NFS-mounts a directory |
| 2422 | onto itself. Suppose the Emacs sources live in `/usr/local/src' and |
| 2423 | you are working on the host called `marvin'. Then an entry in the |
| 2424 | `/etc/fstab' file like the following is asking for trouble: |
| 2425 | |
| 2426 | marvin:/usr/local/src /usr/local/src ...options.omitted... |
| 2427 | |
| 2428 | The solution is to remove this line from `etc/fstab'. |
| 2429 | |
| 2430 | *** Building Emacs with GCC 2.9x fails in the `src' directory. |
| 2431 | |
| 2432 | This may happen if you use a development version of GNU `cpp' from one |
| 2433 | of the GCC snapshots between Oct 2000 and Feb 2001, or from a released |
| 2434 | version of GCC newer than 2.95.2 which was prepared around those |
| 2435 | dates; similar problems were reported with some snapshots of GCC 3.1 |
| 2436 | around Sep 30 2001. The preprocessor in those versions is |
| 2437 | incompatible with a traditional Unix cpp (e.g., it expands ".." into |
| 2438 | ". .", which breaks relative file names that reference the parent |
| 2439 | directory; or inserts TAB characters before lines that set Make |
| 2440 | variables). |
| 2441 | |
| 2442 | The solution is to make sure the preprocessor is run with the |
| 2443 | `-traditional' option. The `configure' script does that automatically |
| 2444 | when it detects the known problems in your cpp, but you might hit some |
| 2445 | unknown ones. To force the `configure' script to use `-traditional', |
| 2446 | run the script like this: |
| 2447 | |
| 2448 | CPP='gcc -E -traditional' ./configure ... |
| 2449 | |
| 2450 | (replace the ellipsis "..." with any additional arguments you pass to |
| 2451 | the script). |
| 2452 | |
| 2453 | Note that this problem does not pertain to the MS-Windows port of |
| 2454 | Emacs, since it doesn't use the preprocessor to generate Makefiles. |
| 2455 | |
| 2456 | *** src/Makefile and lib-src/Makefile are truncated--most of the file missing. |
| 2457 | *** Compiling wakeup, in lib-src, says it can't make wakeup.c. |
| 2458 | |
| 2459 | This can happen if configure uses GNU sed version 2.03. That version |
| 2460 | had a bug. GNU sed version 2.05 works properly.To solve the |
| 2461 | problem, install the current version of GNU Sed, then rerun Emacs's |
| 2462 | configure script. |
| 2463 | |
| 2464 | *** Compiling lib-src says there is no rule to make test-distrib.c. |
| 2465 | |
| 2466 | This results from a bug in a VERY old version of GNU Sed. To solve |
| 2467 | the problem, install the current version of GNU Sed, then rerun |
| 2468 | Emacs's configure script. |
| 2469 | |
| 2470 | *** Building a 32-bit executable on a 64-bit GNU/Linux architecture. |
| 2471 | |
| 2472 | First ensure that the necessary 32-bit system libraries and include |
| 2473 | files are installed. Then use: |
| 2474 | |
| 2475 | env CC="gcc -m32" ./configure --build=i386-linux-gnu \ |
| 2476 | --x-libraries=/usr/X11R6/lib |
| 2477 | |
| 2478 | (using the location of the 32-bit X libraries on your system). |
| 2479 | |
| 2480 | *** Building Emacs for Cygwin can fail with GCC 3 |
| 2481 | |
| 2482 | As of Emacs 22.1, there have been stability problems with Cygwin |
| 2483 | builds of Emacs using GCC 3. Cygwin users are advised to use GCC 4. |
| 2484 | |
| 2485 | *** Building the native MS-Windows port fails due to unresolved externals |
| 2486 | |
| 2487 | The linker error messages look like this: |
| 2488 | |
| 2489 | oo-spd/i386/ctags.o:ctags.c:(.text+0x156e): undefined reference to `_imp__re_set_syntax' |
| 2490 | collect2: ld returned 1 exit status |
| 2491 | |
| 2492 | This happens because GCC finds an incompatible header regex.h |
| 2493 | somewhere on the include path, before the version of regex.h supplied |
| 2494 | with Emacs. One such incompatible version of regex.h is part of the |
| 2495 | GnuWin32 Regex package. |
| 2496 | |
| 2497 | The solution is to remove the incompatible regex.h from the include |
| 2498 | path, when compiling Emacs. Alternatively, re-run the configure.bat |
| 2499 | script with the "-isystem C:/GnuWin32/include" switch (adapt for your |
| 2500 | system's place where you keep the GnuWin32 include files) -- this will |
| 2501 | cause the compiler to search headers in the directories specified by |
| 2502 | the Emacs Makefile _before_ it looks in the GnuWin32 include |
| 2503 | directories. |
| 2504 | |
| 2505 | *** Building the native MS-Windows port with Cygwin GCC can fail. |
| 2506 | |
| 2507 | Emacs may not build using some Cygwin builds of GCC, such as Cygwin |
| 2508 | version 1.1.8, using the default configure settings. It appears to be |
| 2509 | necessary to specify the -mwin32 flag when compiling, and define |
| 2510 | __MSVCRT__, like so: |
| 2511 | |
| 2512 | configure --with-gcc --cflags -mwin32 --cflags -D__MSVCRT__ |
| 2513 | |
| 2514 | *** Building the MS-Windows port fails with a CreateProcess failure. |
| 2515 | |
| 2516 | Some versions of mingw32 make on some versions of Windows do not seem |
| 2517 | to detect the shell correctly. Try "make SHELL=cmd.exe", or if that |
| 2518 | fails, try running make from Cygwin bash instead. |
| 2519 | |
| 2520 | *** Building `ctags' for MS-Windows with the MinGW port of GCC fails. |
| 2521 | |
| 2522 | This might happen due to a bug in the MinGW header assert.h, which |
| 2523 | defines the `assert' macro with a trailing semi-colon. The following |
| 2524 | patch to assert.h should solve this: |
| 2525 | |
| 2526 | *** include/assert.h.orig Sun Nov 7 02:41:36 1999 |
| 2527 | --- include/assert.h Mon Jan 29 11:49:10 2001 |
| 2528 | *************** |
| 2529 | *** 41,47 **** |
| 2530 | /* |
| 2531 | * If not debugging, assert does nothing. |
| 2532 | */ |
| 2533 | ! #define assert(x) ((void)0); |
| 2534 | |
| 2535 | #else /* debugging enabled */ |
| 2536 | |
| 2537 | --- 41,47 ---- |
| 2538 | /* |
| 2539 | * If not debugging, assert does nothing. |
| 2540 | */ |
| 2541 | ! #define assert(x) ((void)0) |
| 2542 | |
| 2543 | #else /* debugging enabled */ |
| 2544 | |
| 2545 | |
| 2546 | *** Building the MS-Windows port with Visual Studio 2005 fails. |
| 2547 | |
| 2548 | Microsoft no longer ships the single threaded version of the C library |
| 2549 | with their compiler, and the multithreaded static library is missing |
| 2550 | some functions that Microsoft have deemed non-threadsafe. The |
| 2551 | dynamically linked C library has all the functions, but there is a |
| 2552 | conflict between the versions of malloc in the DLL and in Emacs, which |
| 2553 | is not resolvable due to the way Windows does dynamic linking. |
| 2554 | |
| 2555 | We recommend the use of the MinGW port of GCC for compiling Emacs, as |
| 2556 | not only does it not suffer these problems, but it is also Free |
| 2557 | software like Emacs. |
| 2558 | |
| 2559 | *** Building the MS-Windows port with Visual Studio fails compiling emacs.rc |
| 2560 | |
| 2561 | If the build fails with the following message then the problem |
| 2562 | described here most likely applies: |
| 2563 | |
| 2564 | ../nt/emacs.rc(1) : error RC2176 : old DIB in icons\emacs.ico; pass it |
| 2565 | through SDKPAINT |
| 2566 | |
| 2567 | The Emacs icon contains a high resolution PNG icon for Vista, which is |
| 2568 | not recognized by older versions of the resource compiler. There are |
| 2569 | several workarounds for this problem: |
| 2570 | 1. Use Free MinGW tools to compile, which do not have this problem. |
| 2571 | 2. Install the latest Windows SDK. |
| 2572 | 3. Replace emacs.ico with an older or edited icon. |
| 2573 | |
| 2574 | ** Linking |
| 2575 | |
| 2576 | *** Building Emacs with a system compiler fails to link because of an |
| 2577 | undefined symbol such as __eprintf which does not appear in Emacs. |
| 2578 | |
| 2579 | This can happen if some of the libraries linked into Emacs were built |
| 2580 | with GCC, but Emacs itself is being linked with a compiler other than |
| 2581 | GCC. Object files compiled with GCC might need some helper functions |
| 2582 | from libgcc.a, the library which comes with GCC, but the system |
| 2583 | compiler does not instruct the linker to search libgcc.a during the |
| 2584 | link stage. |
| 2585 | |
| 2586 | A solution is to link with GCC, like this: |
| 2587 | |
| 2588 | make CC=gcc |
| 2589 | |
| 2590 | Since the .o object files already exist, this will not recompile Emacs |
| 2591 | with GCC, but just restart by trying again to link temacs. |
| 2592 | |
| 2593 | *** AIX 1.3 ptf 0013: Link failure. |
| 2594 | |
| 2595 | There is a real duplicate definition of the function `_slibc_free' in |
| 2596 | the library /lib/libc_s.a (just do nm on it to verify). The |
| 2597 | workaround/fix is: |
| 2598 | |
| 2599 | cd /lib |
| 2600 | ar xv libc_s.a NLtmtime.o |
| 2601 | ar dv libc_s.a NLtmtime.o |
| 2602 | |
| 2603 | *** AIX 4.1.2: Linker error messages such as |
| 2604 | ld: 0711-212 SEVERE ERROR: Symbol .__quous, found in the global symbol table |
| 2605 | of archive /usr/lib/libIM.a, was not defined in archive member shr.o. |
| 2606 | |
| 2607 | This is a problem in libIM.a. You can work around it by executing |
| 2608 | these shell commands in the src subdirectory of the directory where |
| 2609 | you build Emacs: |
| 2610 | |
| 2611 | cp /usr/lib/libIM.a . |
| 2612 | chmod 664 libIM.a |
| 2613 | ranlib libIM.a |
| 2614 | |
| 2615 | Then change -lIM to ./libIM.a in the command to link temacs (in |
| 2616 | Makefile). |
| 2617 | |
| 2618 | *** Sun with acc: Link failure when using acc on a Sun. |
| 2619 | |
| 2620 | To use acc, you need additional options just before the libraries, such as |
| 2621 | |
| 2622 | /usr/lang/SC2.0.1/values-Xt.o -L/usr/lang/SC2.0.1/cg87 -L/usr/lang/SC2.0.1 |
| 2623 | |
| 2624 | and you need to add -lansi just before -lc. |
| 2625 | |
| 2626 | The precise file names depend on the compiler version, so we |
| 2627 | cannot easily arrange to supply them. |
| 2628 | |
| 2629 | *** Linking says that the functions insque and remque are undefined. |
| 2630 | |
| 2631 | Change oldXMenu/Makefile by adding insque.o to the variable OBJS. |
| 2632 | |
| 2633 | *** `tparam' reported as a multiply-defined symbol when linking with ncurses. |
| 2634 | |
| 2635 | This problem results from an incompatible change in ncurses, in |
| 2636 | version 1.9.9e approximately. This version is unable to provide a |
| 2637 | definition of tparm without also defining tparam. This is also |
| 2638 | incompatible with Terminfo; as a result, the Emacs Terminfo support |
| 2639 | does not work with this version of ncurses. |
| 2640 | |
| 2641 | The fix is to install a newer version of ncurses, such as version 4.2. |
| 2642 | |
| 2643 | ** Bootstrapping |
| 2644 | |
| 2645 | Bootstrapping (compiling the .el files) is normally only necessary |
| 2646 | with CVS builds, since the .elc files are pre-compiled in releases. |
| 2647 | |
| 2648 | *** "No rule to make target" with Ubuntu 8.04 make 3.81-3build1 |
| 2649 | |
| 2650 | Compiling the lisp files fails at random places, complaining: |
| 2651 | "No rule to make target `/path/to/some/lisp.elc'". |
| 2652 | The causes of this problem are not understood. Using GNU make 3.81 compiled |
| 2653 | from source, rather than the Ubuntu version, worked. See Bug#327,821. |
| 2654 | |
| 2655 | ** Dumping |
| 2656 | |
| 2657 | *** Linux: Segfault during `make bootstrap' under certain recent versions of the Linux kernel. |
| 2658 | |
| 2659 | With certain recent Linux kernels (like the one of Red Hat Fedora Core |
| 2660 | 1 and newer), the new "Exec-shield" functionality is enabled by default, which |
| 2661 | creates a different memory layout that breaks the emacs dumper. Emacs tries |
| 2662 | to handle this at build time, but if the workaround used fails, these |
| 2663 | instructions can be useful. |
| 2664 | The work-around explained here is not enough on Fedora Core 4 (and possible |
| 2665 | newer). Read the next item. |
| 2666 | |
| 2667 | Configure can overcome the problem of exec-shield if the architecture is |
| 2668 | x86 and the program setarch is present. On other architectures no |
| 2669 | workaround is known. |
| 2670 | |
| 2671 | You can check the Exec-shield state like this: |
| 2672 | |
| 2673 | cat /proc/sys/kernel/exec-shield |
| 2674 | |
| 2675 | It returns non-zero when Exec-shield is enabled, 0 otherwise. Please |
| 2676 | read your system documentation for more details on Exec-shield and |
| 2677 | associated commands. Exec-shield can be turned off with this command: |
| 2678 | |
| 2679 | echo "0" > /proc/sys/kernel/exec-shield |
| 2680 | |
| 2681 | When Exec-shield is enabled, building Emacs will segfault during the |
| 2682 | execution of this command: |
| 2683 | |
| 2684 | ./temacs --batch --load loadup [dump|bootstrap] |
| 2685 | |
| 2686 | To work around this problem, it is necessary to temporarily disable |
| 2687 | Exec-shield while building Emacs, or, on x86, by using the `setarch' |
| 2688 | command when running temacs like this: |
| 2689 | |
| 2690 | setarch i386 ./temacs --batch --load loadup [dump|bootstrap] |
| 2691 | |
| 2692 | |
| 2693 | *** Fedora Core 4 GNU/Linux: Segfault during dumping. |
| 2694 | |
| 2695 | In addition to exec-shield explained above "Linux: Segfault during |
| 2696 | `make bootstrap' under certain recent versions of the Linux kernel" |
| 2697 | item, Linux kernel shipped with Fedora Core 4 randomizes the virtual |
| 2698 | address space of a process. As the result dumping may fail even if |
| 2699 | you turn off exec-shield. In this case, use the -R option to the setarch |
| 2700 | command: |
| 2701 | |
| 2702 | setarch i386 -R ./temacs --batch --load loadup [dump|bootstrap] |
| 2703 | |
| 2704 | or |
| 2705 | |
| 2706 | setarch i386 -R make bootstrap |
| 2707 | |
| 2708 | *** Fatal signal in the command temacs -l loadup inc dump. |
| 2709 | |
| 2710 | This command is the final stage of building Emacs. It is run by the |
| 2711 | Makefile in the src subdirectory. |
| 2712 | |
| 2713 | It has been known to get fatal errors due to insufficient swapping |
| 2714 | space available on the machine. |
| 2715 | |
| 2716 | On 68000s, it has also happened because of bugs in the |
| 2717 | subroutine `alloca'. Verify that `alloca' works right, even |
| 2718 | for large blocks (many pages). |
| 2719 | |
| 2720 | *** test-distrib says that the distribution has been clobbered. |
| 2721 | *** or, temacs prints "Command key out of range 0-127". |
| 2722 | *** or, temacs runs and dumps emacs, but emacs totally fails to work. |
| 2723 | *** or, temacs gets errors dumping emacs. |
| 2724 | |
| 2725 | This can be because the .elc files have been garbled. Do not be |
| 2726 | fooled by the fact that most of a .elc file is text: these are |
| 2727 | binary files and can contain all 256 byte values. |
| 2728 | |
| 2729 | In particular `shar' cannot be used for transmitting GNU Emacs. |
| 2730 | It typically truncates "lines". What appear to be "lines" in |
| 2731 | a binary file can of course be of any length. Even once `shar' |
| 2732 | itself is made to work correctly, `sh' discards null characters |
| 2733 | when unpacking the shell archive. |
| 2734 | |
| 2735 | I have also seen character \177 changed into \377. I do not know |
| 2736 | what transfer means caused this problem. Various network |
| 2737 | file transfer programs are suspected of clobbering the high bit. |
| 2738 | |
| 2739 | If you have a copy of Emacs that has been damaged in its |
| 2740 | nonprinting characters, you can fix them: |
| 2741 | |
| 2742 | 1) Record the names of all the .elc files. |
| 2743 | 2) Delete all the .elc files. |
| 2744 | 3) Recompile alloc.c with a value of PURESIZE twice as large. |
| 2745 | (See puresize.h.) You might as well save the old alloc.o. |
| 2746 | 4) Remake emacs. It should work now. |
| 2747 | 5) Running emacs, do Meta-x byte-compile-file repeatedly |
| 2748 | to recreate all the .elc files that used to exist. |
| 2749 | You may need to increase the value of the variable |
| 2750 | max-lisp-eval-depth to succeed in running the compiler interpreted |
| 2751 | on certain .el files. 400 was sufficient as of last report. |
| 2752 | 6) Reinstall the old alloc.o (undoing changes to alloc.c if any) |
| 2753 | and remake temacs. |
| 2754 | 7) Remake emacs. It should work now, with valid .elc files. |
| 2755 | |
| 2756 | *** temacs prints "Pure Lisp storage exhausted". |
| 2757 | |
| 2758 | This means that the Lisp code loaded from the .elc and .el |
| 2759 | files during temacs -l loadup inc dump took up more |
| 2760 | space than was allocated. |
| 2761 | |
| 2762 | This could be caused by |
| 2763 | 1) adding code to the preloaded Lisp files |
| 2764 | 2) adding more preloaded files in loadup.el |
| 2765 | 3) having a site-init.el or site-load.el which loads files. |
| 2766 | Note that ANY site-init.el or site-load.el is nonstandard; |
| 2767 | if you have received Emacs from some other site |
| 2768 | and it contains a site-init.el or site-load.el file, consider |
| 2769 | deleting that file. |
| 2770 | 4) getting the wrong .el or .elc files |
| 2771 | (not from the directory you expected). |
| 2772 | 5) deleting some .elc files that are supposed to exist. |
| 2773 | This would cause the source files (.el files) to be |
| 2774 | loaded instead. They take up more room, so you lose. |
| 2775 | 6) a bug in the Emacs distribution which underestimates |
| 2776 | the space required. |
| 2777 | |
| 2778 | If the need for more space is legitimate, change the definition |
| 2779 | of PURESIZE in puresize.h. |
| 2780 | |
| 2781 | But in some of the cases listed above, this problem is a consequence |
| 2782 | of something else that is wrong. Be sure to check and fix the real |
| 2783 | problem. |
| 2784 | |
| 2785 | *** Linux: Emacs crashes when dumping itself on Mac PPC running Yellow Dog GNU/Linux. |
| 2786 | |
| 2787 | The crashes happen inside the function Fmake_symbol; here's a typical |
| 2788 | C backtrace printed by GDB: |
| 2789 | |
| 2790 | 0x190c0c0 in Fmake_symbol () |
| 2791 | (gdb) where |
| 2792 | #0 0x190c0c0 in Fmake_symbol () |
| 2793 | #1 0x1942ca4 in init_obarray () |
| 2794 | #2 0x18b3500 in main () |
| 2795 | #3 0x114371c in __libc_start_main (argc=5, argv=0x7ffff5b4, envp=0x7ffff5cc, |
| 2796 | |
| 2797 | This could happen because GCC version 2.95 and later changed the base |
| 2798 | of the load address to 0x10000000. Emacs needs to be told about this, |
| 2799 | but we currently cannot do that automatically, because that breaks |
| 2800 | other versions of GNU/Linux on the MacPPC. Until we find a way to |
| 2801 | distinguish between the Yellow Dog and the other varieties of |
| 2802 | GNU/Linux systems on the PPC, you will have to manually uncomment the |
| 2803 | following section near the end of the file src/m/macppc.h in the Emacs |
| 2804 | distribution: |
| 2805 | |
| 2806 | #if 0 /* This breaks things on PPC GNU/Linux except for Yellowdog, |
| 2807 | even with identical GCC, as, ld. Let's take it out until we |
| 2808 | know what's really going on here. */ |
| 2809 | /* GCC 2.95 and newer on GNU/Linux PPC changed the load address to |
| 2810 | 0x10000000. */ |
| 2811 | #if defined __linux__ |
| 2812 | #if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 95) |
| 2813 | #define DATA_SEG_BITS 0x10000000 |
| 2814 | #endif |
| 2815 | #endif |
| 2816 | #endif /* 0 */ |
| 2817 | |
| 2818 | Remove the "#if 0" and "#endif" directives which surround this, save |
| 2819 | the file, and then reconfigure and rebuild Emacs. The dumping process |
| 2820 | should now succeed. |
| 2821 | |
| 2822 | *** OpenBSD 4.0 macppc: Segfault during dumping. |
| 2823 | |
| 2824 | The build aborts with signal 11 when the command `./temacs --batch |
| 2825 | --load loadup bootstrap' tries to load files.el. A workaround seems |
| 2826 | to be to reduce the level of compiler optimization used during the |
| 2827 | build (from -O2 to -O1). It is possible this is an OpenBSD |
| 2828 | GCC problem specific to the macppc architecture, possibly only |
| 2829 | occurring with older versions of GCC (e.g. 3.3.5). |
| 2830 | |
| 2831 | *** openSUSE 10.3: Segfault in bcopy during dumping. |
| 2832 | |
| 2833 | This is due to a bug in the bcopy implementation in openSUSE 10.3. |
| 2834 | It is/will be fixed in an openSUSE update. |
| 2835 | |
| 2836 | ** Installation |
| 2837 | |
| 2838 | *** Installing Emacs gets an error running `install-info'. |
| 2839 | |
| 2840 | You need to install a recent version of Texinfo; that package |
| 2841 | supplies the `install-info' command. |
| 2842 | |
| 2843 | *** Installing to a directory with spaces in the name fails. |
| 2844 | |
| 2845 | For example, if you call configure with a directory-related option |
| 2846 | with spaces in the value, eg --enable-locallisppath='/path/with\ spaces'. |
| 2847 | Using directory paths with spaces is not supported at this time: you |
| 2848 | must re-configure without using spaces. |
| 2849 | |
| 2850 | *** Installing to a directory with non-ASCII characters in the name fails. |
| 2851 | |
| 2852 | Installation may fail, or the Emacs executable may not start |
| 2853 | correctly, if a directory name containing non-ASCII characters is used |
| 2854 | as a `configure' argument (e.g. `--prefix'). The problem can also |
| 2855 | occur if a non-ASCII directory is specified in the EMACSLOADPATH |
| 2856 | envvar. |
| 2857 | |
| 2858 | *** On Solaris, use GNU Make when installing an out-of-tree build |
| 2859 | |
| 2860 | The Emacs configuration process allows you to configure the |
| 2861 | build environment so that you can build emacs in a directory |
| 2862 | outside of the distribution tree. When installing Emacs from an |
| 2863 | out-of-tree build directory on Solaris, you may need to use GNU |
| 2864 | make. The make programs bundled with Solaris support the VPATH |
| 2865 | macro but use it differently from the way the VPATH macro is |
| 2866 | used by GNU make. The differences will cause the "make install" |
| 2867 | step to fail, leaving you with an incomplete emacs |
| 2868 | installation. GNU make is available in /usr/sfw/bin on Solaris |
| 2869 | 10 and can be installed as /opt/sfw/bin/gmake from the Solaris 9 |
| 2870 | Software Companion CDROM. |
| 2871 | |
| 2872 | The problems due to the VPATH processing differences affect only |
| 2873 | out of tree builds so, if you are on a Solaris installation |
| 2874 | without GNU make, you can install Emacs completely by installing |
| 2875 | from a build environment using the original emacs distribution tree. |
| 2876 | |
| 2877 | ** First execution |
| 2878 | |
| 2879 | *** Emacs binary is not in executable format, and cannot be run. |
| 2880 | |
| 2881 | This was reported to happen when Emacs is built in a directory mounted |
| 2882 | via NFS, for some combinations of NFS client and NFS server. |
| 2883 | Usually, the file `emacs' produced in these cases is full of |
| 2884 | binary null characters, and the `file' utility says: |
| 2885 | |
| 2886 | emacs: ASCII text, with no line terminators |
| 2887 | |
| 2888 | We don't know what exactly causes this failure. A work-around is to |
| 2889 | build Emacs in a directory on a local disk. |
| 2890 | |
| 2891 | *** The dumped Emacs crashes when run, trying to write pure data. |
| 2892 | |
| 2893 | Two causes have been seen for such problems. |
| 2894 | |
| 2895 | 1) On a system where getpagesize is not a system call, it is defined |
| 2896 | as a macro. If the definition (in both unexec.c and malloc.c) is wrong, |
| 2897 | it can cause problems like this. You might be able to find the correct |
| 2898 | value in the man page for a.out (5). |
| 2899 | |
| 2900 | 2) Some systems allocate variables declared static among the |
| 2901 | initialized variables. Emacs makes all initialized variables in most |
| 2902 | of its files pure after dumping, but the variables declared static and |
| 2903 | not initialized are not supposed to be pure. On these systems you |
| 2904 | may need to add "#define static" to the m- or the s- file. |
| 2905 | |
| 2906 | * Emacs 19 problems |
| 2907 | |
| 2908 | ** Error messages `Wrong number of arguments: #<subr where-is-internal>, 5'. |
| 2909 | |
| 2910 | This typically results from having the powerkey library loaded. |
| 2911 | Powerkey was designed for Emacs 19.22. It is obsolete now because |
| 2912 | Emacs 19 now has this feature built in; and powerkey also calls |
| 2913 | where-is-internal in an obsolete way. |
| 2914 | |
| 2915 | So the fix is to arrange not to load powerkey. |
| 2916 | |
| 2917 | * Runtime problems on legacy systems |
| 2918 | |
| 2919 | This section covers bugs reported on very old hardware or software. |
| 2920 | If you are using hardware and an operating system shipped after 2000, |
| 2921 | it is unlikely you will see any of these. |
| 2922 | |
| 2923 | ** Ancient operating systems |
| 2924 | |
| 2925 | AIX 4.2 was end-of-lifed on Dec 31st, 1999. |
| 2926 | |
| 2927 | *** AIX: You get this compiler error message: |
| 2928 | |
| 2929 | Processing include file ./XMenuInt.h |
| 2930 | 1501-106: (S) Include file X11/Xlib.h not found. |
| 2931 | |
| 2932 | This means your system was installed with only the X11 runtime i.d |
| 2933 | libraries. You have to find your sipo (bootable tape) and install |
| 2934 | X11Dev... with smit. |
| 2935 | |
| 2936 | (This report must be ancient. Bootable tapes are long dead.) |
| 2937 | |
| 2938 | *** AIX 3.2.4: Releasing Ctrl/Act key has no effect, if Shift is down. |
| 2939 | |
| 2940 | Due to a feature of AIX, pressing or releasing the Ctrl/Act key is |
| 2941 | ignored when the Shift, Alt or AltGr keys are held down. This can |
| 2942 | lead to the keyboard being "control-locked"--ordinary letters are |
| 2943 | treated as control characters. |
| 2944 | |
| 2945 | You can get out of this "control-locked" state by pressing and |
| 2946 | releasing Ctrl/Act while not pressing or holding any other keys. |
| 2947 | |
| 2948 | *** AIX 3.2.5: You get this message when running Emacs: |
| 2949 | |
| 2950 | Could not load program emacs |
| 2951 | Symbol smtcheckinit in csh is undefined |
| 2952 | Error was: Exec format error |
| 2953 | |
| 2954 | or this one: |
| 2955 | |
| 2956 | Could not load program .emacs |
| 2957 | Symbol _system_con in csh is undefined |
| 2958 | Symbol _fp_trapsta in csh is undefined |
| 2959 | Error was: Exec format error |
| 2960 | |
| 2961 | These can happen when you try to run on AIX 3.2.5 a program that was |
| 2962 | compiled with 3.2.4. The fix is to recompile. |
| 2963 | |
| 2964 | *** AIX 4.2: Emacs gets a segmentation fault at startup. |
| 2965 | |
| 2966 | If you are using IBM's xlc compiler, compile emacs.c |
| 2967 | without optimization; that should avoid the problem. |
| 2968 | |
| 2969 | *** ISC Unix |
| 2970 | |
| 2971 | **** ISC: display-time causes kernel problems on ISC systems. |
| 2972 | |
| 2973 | Under Interactive Unix versions 3.0.1 and 4.0 (and probably other |
| 2974 | versions), display-time causes the loss of large numbers of STREVENT |
| 2975 | cells. Eventually the kernel's supply of these cells is exhausted. |
| 2976 | This makes emacs and the whole system run slow, and can make other |
| 2977 | processes die, in particular pcnfsd. |
| 2978 | |
| 2979 | Other emacs functions that communicate with remote processes may have |
| 2980 | the same problem. Display-time seems to be far the worst. |
| 2981 | |
| 2982 | The only known fix: Don't run display-time. |
| 2983 | |
| 2984 | **** Sunos 5.3: Subprocesses remain, hanging but not zombies. |
| 2985 | |
| 2986 | A bug in Sunos 5.3 causes Emacs subprocesses to remain after Emacs |
| 2987 | exits. Sun patch # 101415-02 is part of the fix for this, but it only |
| 2988 | applies to ptys, and doesn't fix the problem with subprocesses |
| 2989 | communicating through pipes. |
| 2990 | |
| 2991 | *** Irix |
| 2992 | |
| 2993 | *** Irix 6.2: No visible display on mips-sgi-irix6.2 when compiling with GCC 2.8.1. |
| 2994 | |
| 2995 | This problem went away after installing the latest IRIX patches |
| 2996 | as of 8 Dec 1998. |
| 2997 | |
| 2998 | The same problem has been reported on Irix 6.3. |
| 2999 | |
| 3000 | *** Irix 6.3: substituting environment variables in file names |
| 3001 | in the minibuffer gives peculiar error messages such as |
| 3002 | |
| 3003 | Substituting nonexistent environment variable "" |
| 3004 | |
| 3005 | This is not an Emacs bug; it is caused by something in SGI patch |
| 3006 | 003082 August 11, 1998. |
| 3007 | |
| 3008 | *** OPENSTEP |
| 3009 | |
| 3010 | **** OPENSTEP 4.2: Compiling syntax.c with gcc 2.7.2.1 fails. |
| 3011 | |
| 3012 | The compiler was reported to crash while compiling syntax.c with the |
| 3013 | following message: |
| 3014 | |
| 3015 | cc: Internal compiler error: program cc1obj got fatal signal 11 |
| 3016 | |
| 3017 | To work around this, replace the macros UPDATE_SYNTAX_TABLE_FORWARD, |
| 3018 | INC_BOTH, and INC_FROM with functions. To this end, first define 3 |
| 3019 | functions, one each for every macro. Here's an example: |
| 3020 | |
| 3021 | static int update_syntax_table_forward(int from) |
| 3022 | { |
| 3023 | return(UPDATE_SYNTAX_TABLE_FORWARD(from)); |
| 3024 | }/*update_syntax_table_forward*/ |
| 3025 | |
| 3026 | Then replace all references to UPDATE_SYNTAX_TABLE_FORWARD in syntax.c |
| 3027 | with a call to the function update_syntax_table_forward. |
| 3028 | |
| 3029 | *** Solaris 2.x |
| 3030 | |
| 3031 | **** Strange results from format %d in a few cases, on a Sun. |
| 3032 | |
| 3033 | Sun compiler version SC3.0 has been found to miscompile part of |
| 3034 | editfns.c. The workaround is to compile with some other compiler such |
| 3035 | as GCC. |
| 3036 | |
| 3037 | **** On Solaris, Emacs dumps core if lisp-complete-symbol is called. |
| 3038 | |
| 3039 | If you compile Emacs with the -fast or -xO4 option with version 3.0.2 |
| 3040 | of the Sun C compiler, Emacs dumps core when lisp-complete-symbol is |
| 3041 | called. The problem does not happen if you compile with GCC. |
| 3042 | |
| 3043 | **** On Solaris, Emacs crashes if you use (display-time). |
| 3044 | |
| 3045 | This can happen if you configure Emacs without specifying the precise |
| 3046 | version of Solaris that you are using. |
| 3047 | |
| 3048 | **** Solaris 2.3 and 2.4: Unpredictable segmentation faults. |
| 3049 | |
| 3050 | A user reported that this happened in 19.29 when it was compiled with |
| 3051 | the Sun compiler, but not when he recompiled with GCC 2.7.0. |
| 3052 | |
| 3053 | We do not know whether something in Emacs is partly to blame for this. |
| 3054 | |
| 3055 | **** Solaris 2.4: Emacs dumps core on startup. |
| 3056 | |
| 3057 | Bill Sebok says that the cause of this is Solaris 2.4 vendor patch |
| 3058 | 102303-05, which extends the Solaris linker to deal with the Solaris |
| 3059 | Common Desktop Environment's linking needs. You can fix the problem |
| 3060 | by removing this patch and installing patch 102049-02 instead. |
| 3061 | However, that linker version won't work with CDE. |
| 3062 | |
| 3063 | Solaris 2.5 comes with a linker that has this bug. It is reported that if |
| 3064 | you install all the latest patches (as of June 1996), the bug is fixed. |
| 3065 | We suspect the crucial patch is one of these, but we don't know |
| 3066 | for certain. |
| 3067 | |
| 3068 | 103093-03: [README] SunOS 5.5: kernel patch (2140557 bytes) |
| 3069 | 102832-01: [README] OpenWindows 3.5: Xview Jumbo Patch (4181613 bytes) |
| 3070 | 103242-04: [README] SunOS 5.5: linker patch (595363 bytes) |
| 3071 | |
| 3072 | (One user reports that the bug was fixed by those patches together |
| 3073 | with patches 102980-04, 103279-01, 103300-02, and 103468-01.) |
| 3074 | |
| 3075 | If you can determine which patch does fix the bug, please tell |
| 3076 | bug-gnu-emacs@gnu.org. |
| 3077 | |
| 3078 | Meanwhile, the GNU linker links Emacs properly on both Solaris 2.4 and |
| 3079 | Solaris 2.5. |
| 3080 | |
| 3081 | **** Solaris 2.4: Dired hangs and C-g does not work. Or Emacs hangs |
| 3082 | forever waiting for termination of a subprocess that is a zombie. |
| 3083 | |
| 3084 | casper@fwi.uva.nl says the problem is in X11R6. Rebuild libX11.so |
| 3085 | after changing the file xc/config/cf/sunLib.tmpl. Change the lines |
| 3086 | |
| 3087 | #if ThreadedX |
| 3088 | #define SharedX11Reqs -lthread |
| 3089 | #endif |
| 3090 | |
| 3091 | to: |
| 3092 | |
| 3093 | #if OSMinorVersion < 4 |
| 3094 | #if ThreadedX |
| 3095 | #define SharedX11Reqs -lthread |
| 3096 | #endif |
| 3097 | #endif |
| 3098 | |
| 3099 | Be sure also to edit x/config/cf/sun.cf so that OSMinorVersion is 4 |
| 3100 | (as it should be for Solaris 2.4). The file has three definitions for |
| 3101 | OSMinorVersion: the first is for x86, the second for SPARC under |
| 3102 | Solaris, and the third for SunOS 4. Make sure to update the |
| 3103 | definition for your type of machine and system. |
| 3104 | |
| 3105 | Then do `make Everything' in the top directory of X11R6, to rebuild |
| 3106 | the makefiles and rebuild X. The X built this way work only on |
| 3107 | Solaris 2.4, not on 2.3. |
| 3108 | |
| 3109 | For multithreaded X to work it is necessary to install patch |
| 3110 | 101925-02 to fix problems in header files [2.4]. You need |
| 3111 | to reinstall gcc or re-run just-fixinc after installing that |
| 3112 | patch. |
| 3113 | |
| 3114 | However, Frank Rust <frust@iti.cs.tu-bs.de> used a simpler solution: |
| 3115 | he changed |
| 3116 | #define ThreadedX YES |
| 3117 | to |
| 3118 | #define ThreadedX NO |
| 3119 | in sun.cf and did `make World' to rebuild X11R6. Removing all |
| 3120 | `-DXTHREAD*' flags and `-lthread' entries from lib/X11/Makefile and |
| 3121 | typing 'make install' in that directory also seemed to work. |
| 3122 | |
| 3123 | **** Solaris 2.x: GCC complains "64 bit integer types not supported". |
| 3124 | |
| 3125 | This suggests that GCC is not installed correctly. Most likely you |
| 3126 | are using GCC 2.7.2.3 (or earlier) on Solaris 2.6 (or later); this |
| 3127 | does not work without patching. To run GCC 2.7.2.3 on Solaris 2.6 or |
| 3128 | later, you must patch fixinc.svr4 and reinstall GCC from scratch as |
| 3129 | described in the Solaris FAQ |
| 3130 | <http://www.wins.uva.nl/pub/solaris/solaris2.html>. A better fix is |
| 3131 | to upgrade to GCC 2.8.1 or later. |
| 3132 | |
| 3133 | **** Solaris 2.7: Building Emacs with WorkShop Compilers 5.0 98/12/15 |
| 3134 | C 5.0 failed, apparently with non-default CFLAGS, most probably due to |
| 3135 | compiler bugs. Using Sun Solaris 2.7 Sun WorkShop 6 update 1 C |
| 3136 | release was reported to work without problems. It worked OK on |
| 3137 | another system with Solaris 8 using apparently the same 5.0 compiler |
| 3138 | and the default CFLAGS. |
| 3139 | |
| 3140 | **** Solaris 2.x: Emacs dumps core when built with Motif. |
| 3141 | |
| 3142 | The Solaris Motif libraries are buggy, at least up through Solaris 2.5.1. |
| 3143 | Install the current Motif runtime library patch appropriate for your host. |
| 3144 | (Make sure the patch is current; some older patch versions still have the bug.) |
| 3145 | You should install the other patches recommended by Sun for your host, too. |
| 3146 | You can obtain Sun patches from ftp://sunsolve.sun.com/pub/patches/; |
| 3147 | look for files with names ending in `.PatchReport' to see which patches |
| 3148 | are currently recommended for your host. |
| 3149 | |
| 3150 | On Solaris 2.6, Emacs is said to work with Motif when Solaris patch |
| 3151 | 105284-12 is installed, but fail when 105284-15 is installed. |
| 3152 | 105284-18 might fix it again. |
| 3153 | |
| 3154 | **** Solaris 2.6 and 7: the Compose key does not work. |
| 3155 | |
| 3156 | This is a bug in Motif in Solaris. Supposedly it has been fixed for |
| 3157 | the next major release of Solaris. However, if someone with Sun |
| 3158 | support complains to Sun about the bug, they may release a patch. |
| 3159 | If you do this, mention Sun bug #4188711. |
| 3160 | |
| 3161 | One workaround is to use a locale that allows non-ASCII characters. |
| 3162 | For example, before invoking emacs, set the LC_ALL environment |
| 3163 | variable to "en_US" (American English). The directory /usr/lib/locale |
| 3164 | lists the supported locales; any locale other than "C" or "POSIX" |
| 3165 | should do. |
| 3166 | |
| 3167 | pen@lysator.liu.se says (Feb 1998) that the Compose key does work |
| 3168 | if you link with the MIT X11 libraries instead of the Solaris X11 |
| 3169 | libraries. |
| 3170 | |
| 3171 | *** HP/UX versions before 11.0 |
| 3172 | |
| 3173 | HP/UX 9 was end-of-lifed in December 1998. |
| 3174 | HP/UX 10 was end-of-lifed in May 1999. |
| 3175 | |
| 3176 | **** HP/UX 9: Emacs crashes with SIGBUS or SIGSEGV after you delete a frame. |
| 3177 | |
| 3178 | We think this is due to a bug in the X libraries provided by HP. With |
| 3179 | the alternative X libraries in /usr/contrib/mitX11R5/lib, the problem |
| 3180 | does not happen. |
| 3181 | |
| 3182 | *** HP/UX 10: Large file support is disabled. |
| 3183 | |
| 3184 | See the comments in src/s/hpux10.h. |
| 3185 | |
| 3186 | *** HP/UX: Emacs is slow using X11R5. |
| 3187 | |
| 3188 | This happens if you use the MIT versions of the X libraries--it |
| 3189 | doesn't run as fast as HP's version. People sometimes use the version |
| 3190 | because they see the HP version doesn't have the libraries libXaw.a, |
| 3191 | libXmu.a, libXext.a and others. HP/UX normally doesn't come with |
| 3192 | those libraries installed. To get good performance, you need to |
| 3193 | install them and rebuild Emacs. |
| 3194 | |
| 3195 | *** Ultrix and Digital Unix |
| 3196 | |
| 3197 | **** Ultrix 4.2: `make install' fails on install-doc with `Error 141'. |
| 3198 | |
| 3199 | This happens on Ultrix 4.2 due to failure of a pipeline of tar |
| 3200 | commands. We don't know why they fail, but the bug seems not to be in |
| 3201 | Emacs. The workaround is to run the shell command in install-doc by |
| 3202 | hand. |
| 3203 | |
| 3204 | **** Digital Unix 4.0: Garbled display on non-X terminals when Emacs runs. |
| 3205 | |
| 3206 | So far it appears that running `tset' triggers this problem (when TERM |
| 3207 | is vt100, at least). If you do not run `tset', then Emacs displays |
| 3208 | properly. If someone can tell us precisely which effect of running |
| 3209 | `tset' actually causes the problem, we may be able to implement a fix |
| 3210 | in Emacs. |
| 3211 | |
| 3212 | **** Ultrix: `expand-file-name' fails to work on any but the machine you dumped Emacs on. |
| 3213 | |
| 3214 | On Ultrix, if you use any of the functions which look up information |
| 3215 | in the passwd database before dumping Emacs (say, by using |
| 3216 | expand-file-name in site-init.el), then those functions will not work |
| 3217 | in the dumped Emacs on any host but the one Emacs was dumped on. |
| 3218 | |
| 3219 | The solution? Don't use expand-file-name in site-init.el, or in |
| 3220 | anything it loads. Yuck - some solution. |
| 3221 | |
| 3222 | I'm not sure why this happens; if you can find out exactly what is |
| 3223 | going on, and perhaps find a fix or a workaround, please let us know. |
| 3224 | Perhaps the YP functions cache some information, the cache is included |
| 3225 | in the dumped Emacs, and is then inaccurate on any other host. |
| 3226 | |
| 3227 | *** SVr4 |
| 3228 | |
| 3229 | **** SVr4: On some variants of SVR4, Emacs does not work at all with X. |
| 3230 | |
| 3231 | Try defining BROKEN_FIONREAD in your config.h file. If this solves |
| 3232 | the problem, please send a bug report to tell us this is needed; be |
| 3233 | sure to say exactly what type of machine and system you are using. |
| 3234 | |
| 3235 | **** SVr4: After running emacs once, subsequent invocations crash. |
| 3236 | |
| 3237 | Some versions of SVR4 have a serious bug in the implementation of the |
| 3238 | mmap () system call in the kernel; this causes emacs to run correctly |
| 3239 | the first time, and then crash when run a second time. |
| 3240 | |
| 3241 | Contact your vendor and ask for the mmap bug fix; in the mean time, |
| 3242 | you may be able to work around the problem by adding a line to your |
| 3243 | operating system description file (whose name is reported by the |
| 3244 | configure script) that reads: |
| 3245 | #define SYSTEM_MALLOC |
| 3246 | This makes Emacs use memory less efficiently, but seems to work around |
| 3247 | the kernel bug. |
| 3248 | |
| 3249 | *** Irix 5 and earlier |
| 3250 | |
| 3251 | Exactly when Irix-5 end-of-lifed is obscure. But since Irix 6.0 |
| 3252 | shipped in 1994, it has been some years. |
| 3253 | |
| 3254 | **** Irix 5.2: unexelfsgi.c can't find cmplrs/stsupport.h. |
| 3255 | |
| 3256 | The file cmplrs/stsupport.h was included in the wrong file set in the |
| 3257 | Irix 5.2 distribution. You can find it in the optional fileset |
| 3258 | compiler_dev, or copy it from some other Irix 5.2 system. A kludgy |
| 3259 | workaround is to change unexelfsgi.c to include sym.h instead of |
| 3260 | syms.h. |
| 3261 | |
| 3262 | **** Irix 5.3: "out of virtual swap space". |
| 3263 | |
| 3264 | This message occurs when the system runs out of swap space due to too |
| 3265 | many large programs running. The solution is either to provide more |
| 3266 | swap space or to reduce the number of large programs being run. You |
| 3267 | can check the current status of the swap space by executing the |
| 3268 | command `swap -l'. |
| 3269 | |
| 3270 | You can increase swap space by changing the file /etc/fstab. Adding a |
| 3271 | line like this: |
| 3272 | |
| 3273 | /usr/swap/swap.more swap swap pri=3 0 0 |
| 3274 | |
| 3275 | where /usr/swap/swap.more is a file previously created (for instance |
| 3276 | by using /etc/mkfile), will increase the swap space by the size of |
| 3277 | that file. Execute `swap -m' or reboot the machine to activate the |
| 3278 | new swap area. See the manpages for `swap' and `fstab' for further |
| 3279 | information. |
| 3280 | |
| 3281 | The objectserver daemon can use up lots of memory because it can be |
| 3282 | swamped with NIS information. It collects information about all users |
| 3283 | on the network that can log on to the host. |
| 3284 | |
| 3285 | If you want to disable the objectserver completely, you can execute |
| 3286 | the command `chkconfig objectserver off' and reboot. That may disable |
| 3287 | some of the window system functionality, such as responding CDROM |
| 3288 | icons. |
| 3289 | |
| 3290 | You can also remove NIS support from the objectserver. The SGI `admin' |
| 3291 | FAQ has a detailed description on how to do that; see question 35 |
| 3292 | ("Why isn't the objectserver working?"). The admin FAQ can be found at |
| 3293 | ftp://viz.tamu.edu/pub/sgi/faq/. |
| 3294 | |
| 3295 | **** Irix 5.3: Emacs crashes in utmpname. |
| 3296 | |
| 3297 | This problem is fixed in Patch 3175 for Irix 5.3. |
| 3298 | It is also fixed in Irix versions 6.2 and up. |
| 3299 | |
| 3300 | **** Irix 6.0: Make tries (and fails) to build a program named unexelfsgi. |
| 3301 | |
| 3302 | A compiler bug inserts spaces into the string "unexelfsgi . o" |
| 3303 | in src/Makefile. Edit src/Makefile, after configure is run, |
| 3304 | find that string, and take out the spaces. |
| 3305 | |
| 3306 | Compiler fixes in Irix 6.0.1 should eliminate this problem. |
| 3307 | |
| 3308 | *** SCO Unix and UnixWare |
| 3309 | |
| 3310 | **** SCO 3.2v4: Unusable default font. |
| 3311 | |
| 3312 | The Open Desktop environment comes with default X resource settings |
| 3313 | that tell Emacs to use a variable-width font. Emacs cannot use such |
| 3314 | fonts, so it does not work. |
| 3315 | |
| 3316 | This is caused by the file /usr/lib/X11/app-defaults/ScoTerm, which is |
| 3317 | the application-specific resource file for the `scoterm' terminal |
| 3318 | emulator program. It contains several extremely general X resources |
| 3319 | that affect other programs besides `scoterm'. In particular, these |
| 3320 | resources affect Emacs also: |
| 3321 | |
| 3322 | *Font: -*-helvetica-medium-r-*--12-*-p-* |
| 3323 | *Background: scoBackground |
| 3324 | *Foreground: scoForeground |
| 3325 | |
| 3326 | The best solution is to create an application-specific resource file for |
| 3327 | Emacs, /usr/lib/X11/sco/startup/Emacs, with the following contents: |
| 3328 | |
| 3329 | Emacs*Font: -*-courier-medium-r-*-*-*-120-*-*-*-*-iso8859-1 |
| 3330 | Emacs*Background: white |
| 3331 | Emacs*Foreground: black |
| 3332 | |
| 3333 | (These settings mimic the Emacs defaults, but you can change them to |
| 3334 | suit your needs.) This resource file is only read when the X server |
| 3335 | starts up, so you should restart it by logging out of the Open Desktop |
| 3336 | environment or by running `scologin stop; scologin start` from the shell |
| 3337 | as root. Alternatively, you can put these settings in the |
| 3338 | /usr/lib/X11/app-defaults/Emacs resource file and simply restart Emacs, |
| 3339 | but then they will not affect remote invocations of Emacs that use the |
| 3340 | Open Desktop display. |
| 3341 | |
| 3342 | These resource files are not normally shared across a network of SCO |
| 3343 | machines; you must create the file on each machine individually. |
| 3344 | |
| 3345 | **** SCO 4.2.0: Regular expressions matching bugs on SCO systems. |
| 3346 | |
| 3347 | On SCO, there are problems in regexp matching when Emacs is compiled |
| 3348 | with the system compiler. The compiler version is "Microsoft C |
| 3349 | version 6", SCO 4.2.0h Dev Sys Maintenance Supplement 01/06/93; Quick |
| 3350 | C Compiler Version 1.00.46 (Beta). The solution is to compile with |
| 3351 | GCC. |
| 3352 | |
| 3353 | **** UnixWare 2.1: Error 12 (virtual memory exceeded) when dumping Emacs. |
| 3354 | |
| 3355 | Paul Abrahams (abrahams@acm.org) reports that with the installed |
| 3356 | virtual memory settings for UnixWare 2.1.2, an Error 12 occurs during |
| 3357 | the "make" that builds Emacs, when running temacs to dump emacs. That |
| 3358 | error indicates that the per-process virtual memory limit has been |
| 3359 | exceeded. The default limit is probably 32MB. Raising the virtual |
| 3360 | memory limit to 40MB should make it possible to finish building Emacs. |
| 3361 | |
| 3362 | You can do this with the command `ulimit' (sh) or `limit' (csh). |
| 3363 | But you have to be root to do it. |
| 3364 | |
| 3365 | According to Martin Sohnius, you can also retune this in the kernel: |
| 3366 | |
| 3367 | # /etc/conf/bin/idtune SDATLIM 33554432 ## soft data size limit |
| 3368 | # /etc/conf/bin/idtune HDATLIM 33554432 ## hard " |
| 3369 | # /etc/conf/bin/idtune SVMMSIZE unlimited ## soft process size limit |
| 3370 | # /etc/conf/bin/idtune HVMMSIZE unlimited ## hard " |
| 3371 | # /etc/conf/bin/idbuild -B |
| 3372 | |
| 3373 | (He recommends you not change the stack limit, though.) |
| 3374 | These changes take effect when you reboot. |
| 3375 | |
| 3376 | *** Linux 1.x |
| 3377 | |
| 3378 | **** Linux 1.0-1.04: Typing C-c C-c in Shell mode kills your X server. |
| 3379 | |
| 3380 | This happens with Linux kernel 1.0 thru 1.04, approximately. The workaround is |
| 3381 | to define SIGNALS_VIA_CHARACTERS in config.h and recompile Emacs. |
| 3382 | Newer Linux kernel versions don't have this problem. |
| 3383 | |
| 3384 | **** Linux 1.3: Output from subprocess (such as man or diff) is randomly |
| 3385 | truncated on GNU/Linux systems. |
| 3386 | |
| 3387 | This is due to a kernel bug which seems to be fixed in Linux version |
| 3388 | 1.3.75. |
| 3389 | |
| 3390 | ** Windows 3.1, 95, 98, and ME |
| 3391 | |
| 3392 | *** MS-Windows NT/95: Problems running Perl under Emacs |
| 3393 | |
| 3394 | `perl -de 0' just hangs when executed in an Emacs subshell. |
| 3395 | The fault lies with Perl (indirectly with Windows NT/95). |
| 3396 | |
| 3397 | The problem is that the Perl debugger explicitly opens a connection to |
| 3398 | "CON", which is the DOS/NT equivalent of "/dev/tty", for interacting |
| 3399 | with the user. |
| 3400 | |
| 3401 | On Unix, this is okay, because Emacs (or the shell?) creates a |
| 3402 | pseudo-tty so that /dev/tty is really the pipe Emacs is using to |
| 3403 | communicate with the subprocess. |
| 3404 | |
| 3405 | On NT, this fails because CON always refers to the handle for the |
| 3406 | relevant console (approximately equivalent to a tty), and cannot be |
| 3407 | redirected to refer to the pipe Emacs assigned to the subprocess as |
| 3408 | stdin. |
| 3409 | |
| 3410 | A workaround is to modify perldb.pl to use STDIN/STDOUT instead of CON. |
| 3411 | |
| 3412 | For Perl 4: |
| 3413 | |
| 3414 | *** PERL/LIB/PERLDB.PL.orig Wed May 26 08:24:18 1993 |
| 3415 | --- PERL/LIB/PERLDB.PL Mon Jul 01 15:28:16 1996 |
| 3416 | *************** |
| 3417 | *** 68,74 **** |
| 3418 | $rcfile=".perldb"; |
| 3419 | } |
| 3420 | else { |
| 3421 | ! $console = "con"; |
| 3422 | $rcfile="perldb.ini"; |
| 3423 | } |
| 3424 | |
| 3425 | --- 68,74 ---- |
| 3426 | $rcfile=".perldb"; |
| 3427 | } |
| 3428 | else { |
| 3429 | ! $console = ""; |
| 3430 | $rcfile="perldb.ini"; |
| 3431 | } |
| 3432 | |
| 3433 | |
| 3434 | For Perl 5: |
| 3435 | *** perl/5.001/lib/perl5db.pl.orig Sun Jun 04 21:13:40 1995 |
| 3436 | --- perl/5.001/lib/perl5db.pl Mon Jul 01 17:00:08 1996 |
| 3437 | *************** |
| 3438 | *** 22,28 **** |
| 3439 | $rcfile=".perldb"; |
| 3440 | } |
| 3441 | elsif (-e "con") { |
| 3442 | ! $console = "con"; |
| 3443 | $rcfile="perldb.ini"; |
| 3444 | } |
| 3445 | else { |
| 3446 | --- 22,28 ---- |
| 3447 | $rcfile=".perldb"; |
| 3448 | } |
| 3449 | elsif (-e "con") { |
| 3450 | ! $console = ""; |
| 3451 | $rcfile="perldb.ini"; |
| 3452 | } |
| 3453 | else { |
| 3454 | |
| 3455 | *** MS-Windows 95: Alt-f6 does not get through to Emacs. |
| 3456 | |
| 3457 | This character seems to be trapped by the kernel in Windows 95. |
| 3458 | You can enter M-f6 by typing ESC f6. |
| 3459 | |
| 3460 | *** MS-Windows 95/98/ME: subprocesses do not terminate properly. |
| 3461 | |
| 3462 | This is a limitation of the Operating System, and can cause problems |
| 3463 | when shutting down Windows. Ensure that all subprocesses are exited |
| 3464 | cleanly before exiting Emacs. For more details, see the FAQ at |
| 3465 | http://www.gnu.org/software/emacs/windows/. |
| 3466 | |
| 3467 | *** MS-Windows 95/98/ME: crashes when Emacs invokes non-existent programs. |
| 3468 | |
| 3469 | When a program you are trying to run is not found on the PATH, |
| 3470 | Windows might respond by crashing or locking up your system. In |
| 3471 | particular, this has been reported when trying to compile a Java |
| 3472 | program in JDEE when javac.exe is installed, but not on the system |
| 3473 | PATH. |
| 3474 | |
| 3475 | ** MS-DOS |
| 3476 | |
| 3477 | *** When compiling with DJGPP on MS-Windows NT or later, "config msdos" fails. |
| 3478 | |
| 3479 | If the error message is "VDM has been already loaded", this is because |
| 3480 | Windows has a program called `redir.exe' that is incompatible with a |
| 3481 | program by the same name supplied with DJGPP, which is used by |
| 3482 | config.bat. To resolve this, move the DJGPP's `bin' subdirectory to |
| 3483 | the front of your PATH environment variable. |
| 3484 | |
| 3485 | *** When Emacs compiled with DJGPP runs on Windows 2000 and later, it cannot |
| 3486 | find your HOME directory. |
| 3487 | |
| 3488 | This was reported to happen when you click on "Save for future |
| 3489 | sessions" button in a Customize buffer. You might see an error |
| 3490 | message like this one: |
| 3491 | |
| 3492 | basic-save-buffer-2: c:/FOO/BAR/~dosuser/: no such directory |
| 3493 | |
| 3494 | (The telltale sign is the "~USER" part at the end of the directory |
| 3495 | Emacs complains about, where USER is your username or the literal |
| 3496 | string "dosuser", which is the default username set up by the DJGPP |
| 3497 | startup file DJGPP.ENV.) |
| 3498 | |
| 3499 | This happens when the functions `user-login-name' and |
| 3500 | `user-real-login-name' return different strings for your username as |
| 3501 | Emacs sees it. To correct this, make sure both USER and USERNAME |
| 3502 | environment variables are set to the same value. Windows 2000 and |
| 3503 | later sets USERNAME, so if you want to keep that, make sure USER is |
| 3504 | set to the same value. If you don't want to set USER globally, you |
| 3505 | can do it in the [emacs] section of your DJGPP.ENV file. |
| 3506 | |
| 3507 | *** When Emacs compiled with DJGPP runs on Vista, it runs out of memory. |
| 3508 | |
| 3509 | If Emacs running on Vista displays "!MEM FULL!" in the mode line, you |
| 3510 | are hitting the memory allocation bugs in the Vista DPMI server. See |
| 3511 | msdos/INSTALL for how to work around these bugs (search for "Vista"). |
| 3512 | |
| 3513 | *** When compiling with DJGPP on MS-Windows 95, Make fails for some targets |
| 3514 | like make-docfile. |
| 3515 | |
| 3516 | This can happen if long file name support (the setting of environment |
| 3517 | variable LFN) when Emacs distribution was unpacked and during |
| 3518 | compilation are not the same. See msdos/INSTALL for the explanation |
| 3519 | of how to avoid this problem. |
| 3520 | |
| 3521 | *** Emacs compiled with DJGPP complains at startup: |
| 3522 | |
| 3523 | "Wrong type of argument: internal-facep, msdos-menu-active-face" |
| 3524 | |
| 3525 | This can happen if you define an environment variable `TERM'. Emacs |
| 3526 | on MSDOS uses an internal terminal emulator which is disabled if the |
| 3527 | value of `TERM' is anything but the string "internal". Emacs then |
| 3528 | works as if its terminal were a dumb glass teletype that doesn't |
| 3529 | support faces. To work around this, arrange for `TERM' to be |
| 3530 | undefined when Emacs runs. The best way to do that is to add an |
| 3531 | [emacs] section to the DJGPP.ENV file which defines an empty value for |
| 3532 | `TERM'; this way, only Emacs gets the empty value, while the rest of |
| 3533 | your system works as before. |
| 3534 | |
| 3535 | *** MS-DOS: Emacs crashes at startup. |
| 3536 | |
| 3537 | Some users report that Emacs 19.29 requires dpmi memory management, |
| 3538 | and crashes on startup if the system does not have it. We don't yet |
| 3539 | know why this happens--perhaps these machines don't have enough real |
| 3540 | memory, or perhaps something is wrong in Emacs or the compiler. |
| 3541 | However, arranging to use dpmi support is a workaround. |
| 3542 | |
| 3543 | You can find out if you have a dpmi host by running go32 without |
| 3544 | arguments; it will tell you if it uses dpmi memory. For more |
| 3545 | information about dpmi memory, consult the djgpp FAQ. (djgpp |
| 3546 | is the GNU C compiler as packaged for MSDOS.) |
| 3547 | |
| 3548 | Compiling Emacs under MSDOS is extremely sensitive for proper memory |
| 3549 | configuration. If you experience problems during compilation, consider |
| 3550 | removing some or all memory resident programs (notably disk caches) |
| 3551 | and make sure that your memory managers are properly configured. See |
| 3552 | the djgpp faq for configuration hints. |
| 3553 | |
| 3554 | *** Emacs compiled with DJGPP for MS-DOS/MS-Windows cannot access files |
| 3555 | in the directory with the special name `dev' under the root of any |
| 3556 | drive, e.g. `c:/dev'. |
| 3557 | |
| 3558 | This is an unfortunate side-effect of the support for Unix-style |
| 3559 | device names such as /dev/null in the DJGPP runtime library. A |
| 3560 | work-around is to rename the problem directory to another name. |
| 3561 | |
| 3562 | *** MS-DOS+DJGPP: Problems on MS-DOG if DJGPP v2.0 is used to compile Emacs. |
| 3563 | |
| 3564 | There are two DJGPP library bugs which cause problems: |
| 3565 | |
| 3566 | * Running `shell-command' (or `compile', or `grep') you get |
| 3567 | `Searching for program: permission denied (EACCES), c:/command.com'; |
| 3568 | * After you shell to DOS, Ctrl-Break kills Emacs. |
| 3569 | |
| 3570 | To work around these bugs, you can use two files in the msdos |
| 3571 | subdirectory: `is_exec.c' and `sigaction.c'. Compile them and link |
| 3572 | them into the Emacs executable `temacs'; then they will replace the |
| 3573 | incorrect library functions. |
| 3574 | |
| 3575 | *** MS-DOS: Emacs compiled for MSDOS cannot find some Lisp files, or other |
| 3576 | run-time support files, when long filename support is enabled. |
| 3577 | |
| 3578 | Usually, this problem will manifest itself when Emacs exits |
| 3579 | immediately after flashing the startup screen, because it cannot find |
| 3580 | the Lisp files it needs to load at startup. Redirect Emacs stdout |
| 3581 | and stderr to a file to see the error message printed by Emacs. |
| 3582 | |
| 3583 | Another manifestation of this problem is that Emacs is unable to load |
| 3584 | the support for editing program sources in languages such as C and |
| 3585 | Lisp. |
| 3586 | |
| 3587 | This can happen if the Emacs distribution was unzipped without LFN |
| 3588 | support, thus causing long filenames to be truncated to the first 6 |
| 3589 | characters and a numeric tail that Windows 95 normally attaches to it. |
| 3590 | You should unzip the files again with a utility that supports long |
| 3591 | filenames (such as djtar from DJGPP or InfoZip's UnZip program |
| 3592 | compiled with DJGPP v2). The file msdos/INSTALL explains this issue |
| 3593 | in more detail. |
| 3594 | |
| 3595 | Another possible reason for such failures is that Emacs compiled for |
| 3596 | MSDOS is used on Windows NT, where long file names are not supported |
| 3597 | by this version of Emacs, but the distribution was unpacked by an |
| 3598 | unzip program that preserved the long file names instead of truncating |
| 3599 | them to DOS 8+3 limits. To be useful on NT, the MSDOS port of Emacs |
| 3600 | must be unzipped by a DOS utility, so that long file names are |
| 3601 | properly truncated. |
| 3602 | |
| 3603 | ** Archaic window managers and toolkits |
| 3604 | |
| 3605 | *** OpenLook: Under OpenLook, the Emacs window disappears when you type M-q. |
| 3606 | |
| 3607 | Some versions of the Open Look window manager interpret M-q as a quit |
| 3608 | command for whatever window you are typing at. If you want to use |
| 3609 | Emacs with that window manager, you should try to configure the window |
| 3610 | manager to use some other command. You can disable the |
| 3611 | shortcut keys entirely by adding this line to ~/.OWdefaults: |
| 3612 | |
| 3613 | OpenWindows.WindowMenuAccelerators: False |
| 3614 | |
| 3615 | **** twm: A position you specified in .Xdefaults is ignored, using twm. |
| 3616 | |
| 3617 | twm normally ignores "program-specified" positions. |
| 3618 | You can tell it to obey them with this command in your `.twmrc' file: |
| 3619 | |
| 3620 | UsePPosition "on" #allow clients to request a position |
| 3621 | |
| 3622 | ** Bugs related to old DEC hardware |
| 3623 | |
| 3624 | *** The Compose key on a DEC keyboard does not work as Meta key. |
| 3625 | |
| 3626 | This shell command should fix it: |
| 3627 | |
| 3628 | xmodmap -e 'keycode 0xb1 = Meta_L' |
| 3629 | |
| 3630 | *** Keyboard input gets confused after a beep when using a DECserver |
| 3631 | as a concentrator. |
| 3632 | |
| 3633 | This problem seems to be a matter of configuring the DECserver to use |
| 3634 | 7 bit characters rather than 8 bit characters. |
| 3635 | |
| 3636 | * Build problems on legacy systems |
| 3637 | |
| 3638 | ** BSD/386 1.0: --with-x-toolkit option configures wrong. |
| 3639 | |
| 3640 | This problem is due to bugs in the shell in version 1.0 of BSD/386. |
| 3641 | The workaround is to edit the configure file to use some other shell, |
| 3642 | such as bash. |
| 3643 | |
| 3644 | ** Digital Unix 4.0: Emacs fails to build, giving error message |
| 3645 | Invalid dimension for the charset-ID 160 |
| 3646 | |
| 3647 | This is due to a bug or an installation problem in GCC 2.8.0. |
| 3648 | Installing a more recent version of GCC fixes the problem. |
| 3649 | |
| 3650 | ** Digital Unix 4.0: Failure in unexec while dumping emacs. |
| 3651 | |
| 3652 | This problem manifests itself as an error message |
| 3653 | |
| 3654 | unexec: Bad address, writing data section to ... |
| 3655 | |
| 3656 | The user suspects that this happened because his X libraries |
| 3657 | were built for an older system version, |
| 3658 | |
| 3659 | ./configure --x-includes=/usr/include --x-libraries=/usr/shlib |
| 3660 | |
| 3661 | made the problem go away. |
| 3662 | |
| 3663 | ** Sunos 4.1.1: there are errors compiling sysdep.c. |
| 3664 | |
| 3665 | If you get errors such as |
| 3666 | |
| 3667 | "sysdep.c", line 2017: undefined structure or union |
| 3668 | "sysdep.c", line 2017: undefined structure or union |
| 3669 | "sysdep.c", line 2019: nodename undefined |
| 3670 | |
| 3671 | This can result from defining LD_LIBRARY_PATH. It is very tricky |
| 3672 | to use that environment variable with Emacs. The Emacs configure |
| 3673 | script links many test programs with the system libraries; you must |
| 3674 | make sure that the libraries available to configure are the same |
| 3675 | ones available when you build Emacs. |
| 3676 | |
| 3677 | ** SunOS 4.1.1: You get this error message from GNU ld: |
| 3678 | |
| 3679 | /lib/libc.a(_Q_sub.o): Undefined symbol __Q_get_rp_rd referenced from text segment |
| 3680 | |
| 3681 | The problem is in the Sun shared C library, not in GNU ld. |
| 3682 | |
| 3683 | The solution is to install Patch-ID# 100267-03 from Sun. |
| 3684 | |
| 3685 | ** Sunos 4.1: Undefined symbols when linking using --with-x-toolkit. |
| 3686 | |
| 3687 | If you get the undefined symbols _atowc _wcslen, _iswprint, _iswspace, |
| 3688 | _iswcntrl, _wcscpy, and _wcsncpy, then you need to add -lXwchar after |
| 3689 | -lXaw in the command that links temacs. |
| 3690 | |
| 3691 | This problem seems to arise only when the international language |
| 3692 | extensions to X11R5 are installed. |
| 3693 | |
| 3694 | ** SunOS: Emacs gets error message from linker on Sun. |
| 3695 | |
| 3696 | If the error message says that a symbol such as `f68881_used' or |
| 3697 | `ffpa_used' or `start_float' is undefined, this probably indicates |
| 3698 | that you have compiled some libraries, such as the X libraries, |
| 3699 | with a floating point option other than the default. |
| 3700 | |
| 3701 | It's not terribly hard to make this work with small changes in |
| 3702 | crt0.c together with linking with Fcrt1.o, Wcrt1.o or Mcrt1.o. |
| 3703 | However, the easiest approach is to build Xlib with the default |
| 3704 | floating point option: -fsoft. |
| 3705 | |
| 3706 | ** HPUX 10.20: Emacs crashes during dumping on the HPPA machine. |
| 3707 | |
| 3708 | This seems to be due to a GCC bug; it is fixed in GCC 2.8.1. |
| 3709 | |
| 3710 | ** Vax C compiler bugs affecting Emacs. |
| 3711 | |
| 3712 | You may get one of these problems compiling Emacs: |
| 3713 | |
| 3714 | foo.c line nnn: compiler error: no table entry for op STASG |
| 3715 | foo.c: fatal error in /lib/ccom |
| 3716 | |
| 3717 | These are due to bugs in the C compiler; the code is valid C. |
| 3718 | Unfortunately, the bugs are unpredictable: the same construct |
| 3719 | may compile properly or trigger one of these bugs, depending |
| 3720 | on what else is in the source file being compiled. Even changes |
| 3721 | in header files that should not affect the file being compiled |
| 3722 | can affect whether the bug happens. In addition, sometimes files |
| 3723 | that compile correctly on one machine get this bug on another machine. |
| 3724 | |
| 3725 | As a result, it is hard for me to make sure this bug will not affect |
| 3726 | you. I have attempted to find and alter these constructs, but more |
| 3727 | can always appear. However, I can tell you how to deal with it if it |
| 3728 | should happen. The bug comes from having an indexed reference to an |
| 3729 | array of Lisp_Objects, as an argument in a function call: |
| 3730 | Lisp_Object *args; |
| 3731 | ... |
| 3732 | ... foo (5, args[i], ...)... |
| 3733 | putting the argument into a temporary variable first, as in |
| 3734 | Lisp_Object *args; |
| 3735 | Lisp_Object tem; |
| 3736 | ... |
| 3737 | tem = args[i]; |
| 3738 | ... foo (r, tem, ...)... |
| 3739 | causes the problem to go away. |
| 3740 | The `contents' field of a Lisp vector is an array of Lisp_Objects, |
| 3741 | so you may see the problem happening with indexed references to that. |
| 3742 | |
| 3743 | ** 68000 C compiler problems |
| 3744 | |
| 3745 | Various 68000 compilers have different problems. |
| 3746 | These are some that have been observed. |
| 3747 | |
| 3748 | *** Using value of assignment expression on union type loses. |
| 3749 | This means that x = y = z; or foo (x = z); does not work |
| 3750 | if x is of type Lisp_Object. |
| 3751 | |
| 3752 | *** "cannot reclaim" error. |
| 3753 | |
| 3754 | This means that an expression is too complicated. You get the correct |
| 3755 | line number in the error message. The code must be rewritten with |
| 3756 | simpler expressions. |
| 3757 | |
| 3758 | *** XCONS, XSTRING, etc macros produce incorrect code. |
| 3759 | |
| 3760 | If temacs fails to run at all, this may be the cause. |
| 3761 | Compile this test program and look at the assembler code: |
| 3762 | |
| 3763 | struct foo { char x; unsigned int y : 24; }; |
| 3764 | |
| 3765 | lose (arg) |
| 3766 | struct foo arg; |
| 3767 | { |
| 3768 | test ((int *) arg.y); |
| 3769 | } |
| 3770 | |
| 3771 | If the code is incorrect, your compiler has this problem. |
| 3772 | In the XCONS, etc., macros in lisp.h you must replace (a).u.val with |
| 3773 | ((a).u.val + coercedummy) where coercedummy is declared as int. |
| 3774 | |
| 3775 | This problem will only happen if USE_LISP_UNION_TYPE is manually |
| 3776 | defined in lisp.h. |
| 3777 | |
| 3778 | *** C compilers lose on returning unions. |
| 3779 | |
| 3780 | I hear that some C compilers cannot handle returning a union type. |
| 3781 | Most of the functions in GNU Emacs return type Lisp_Object, which is |
| 3782 | defined as a union on some rare architectures. |
| 3783 | |
| 3784 | This problem will only happen if USE_LISP_UNION_TYPE is manually |
| 3785 | defined in lisp.h. |
| 3786 | |
| 3787 | \f |
| 3788 | This file is part of GNU Emacs. |
| 3789 | |
| 3790 | GNU Emacs is free software: you can redistribute it and/or modify |
| 3791 | it under the terms of the GNU General Public License as published by |
| 3792 | the Free Software Foundation, either version 3 of the License, or |
| 3793 | (at your option) any later version. |
| 3794 | |
| 3795 | GNU Emacs is distributed in the hope that it will be useful, |
| 3796 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 3797 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 3798 | GNU General Public License for more details. |
| 3799 | |
| 3800 | You should have received a copy of the GNU General Public License |
| 3801 | along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. |
| 3802 | |
| 3803 | \f |
| 3804 | Local variables: |
| 3805 | mode: outline |
| 3806 | paragraph-separate: "[ \f]*$" |
| 3807 | end: |
| 3808 | |
| 3809 | arch-tag: 49fc0d95-88cb-4715-b21c-f27fb5a4764a |