Mercurial > emacs
view etc/PROBLEMS @ 56791:a0d8e0ecba72
*** empty log message ***
| author | John Paul Wallington <jpw@pobox.com> |
|---|---|
| date | Wed, 25 Aug 2004 18:46:04 +0000 |
| parents | 5ea587a67aae |
| children | ce4c0cb926e3 d8411455de48 |
line wrap: on
line source
This file describes various problems that have been encountered in compiling, installing and running GNU Emacs. Try doing Ctl t and browsing through the outline headers. * Emacs startup failures ** Emacs fails to start, complaining about missing fonts. A typical error message might be something like No fonts match `-*-fixed-medium-r-*--6-*-*-*-*-*-iso8859-1' This happens because some X resource specifies a bad font family for Emacs to use. The possible places where this specification might be are: - in your ~/.Xdefaults file - client-side X resource file, such as ~/Emacs or /usr/X11R6/lib/app-defaults/Emacs or /usr/X11R6/lib/X11/app-defaults/Emacs One of these files might have bad or malformed specification of a fontset that Emacs should use. To fix the problem, you need to find the problematic line(s) and correct them. ** Emacs aborts while starting up, only when run without X. This problem often results from compiling Emacs with GCC when GCC was installed incorrectly. The usual error in installing GCC is to specify --includedir=/usr/include. Installation of GCC makes corrected copies of the system header files. GCC is supposed to use the corrected copies in preference to the original system headers. Specifying --includedir=/usr/include causes the original system header files to be used. On some systems, the definition of ioctl in the original system header files is invalid for ANSI C and causes Emacs not to work. The fix is to reinstall GCC, and this time do not specify --includedir when you configure it. Then recompile Emacs. Specifying --includedir is appropriate only in very special cases and it should *never* be the same directory where system header files are kept. ** Emacs does not start, complaining that it cannot open termcap database file. If your system uses Terminfo rather than termcap (most modern systems do), this could happen if the proper version of ncurses is not visible to the Emacs configure script (i.e. it cannot be found along the usual path the linker looks for libraries). It can happen because your version of ncurses is obsolete, or is available only in form of binaries. The solution is to install an up-to-date version of ncurses in the developer's form (header files, static libraries and symbolic links); in some GNU/Linux distributions (e.g. Debian) it constitutes a separate package. ** Emacs 20 and later fails to load Lisp files at startup. The typical error message might be like this: "Cannot open load file: fontset" This could happen if you compress the file lisp/subdirs.el. That file tells Emacs what are the directories where it should look for Lisp files. Emacs cannot work with subdirs.el compressed, since the Auto-compress mode it needs for this will not be loaded until later, when your .emacs file is processed. (The package `fontset.el' is required to set up fonts used to display text on window systems, and it's loaded very early in the startup procedure.) Similarly, any other .el file for which there's no corresponding .elc file could fail to load if it is compressed. The solution is to uncompress all .el files which don't have a .elc file. Another possible reason for such failures is stale *.elc files lurking somewhere on your load-path. The following command will print any duplicate Lisp files that are present in load-path: emacs -q -batch -f list-load-path-shadows If this command prints any file names, some of these files are stale, and should be deleted or their directories removed from your load-path. ** Emacs prints an error at startup after upgrading from an earlier version. An example of such an error is: x-complement-fontset-spec: "Wrong type argument: stringp, nil" This can be another symptom of stale *.elc files in your load-path. The following command will print any duplicate Lisp files that are present in load-path: emacs -q -batch -f list-load-path-shadows If this command prints any file names, some of these files are stale, and should be deleted or their directories removed from your load-path. ** With X11R6.4, public-patch-3, Emacs crashes at startup. Reportedly this patch in X fixes the problem. --- xc/lib/X11/imInt.c~ Wed Jun 30 13:31:56 1999 +++ xc/lib/X11/imInt.c Thu Jul 1 15:10:27 1999 @@ -1,4 +1,4 @@ -/* $TOG: imInt.c /main/5 1998/05/30 21:11:16 kaleb $ */ +/* $TOG: imInt.c /main/5 1998/05/30 21:11:16 kaleb $ */ /****************************************************************** Copyright 1992, 1993, 1994 by FUJITSU LIMITED @@ -166,8 +166,8 @@ _XimMakeImName(lcd) XLCd lcd; { - char* begin; - char* end; + char* begin = NULL; + char* end = NULL; char* ret; int i = 0; char* ximmodifier = XIMMODIFIER; @@ -182,7 +182,11 @@ } ret = Xmalloc(end - begin + 2); if (ret != NULL) { - (void)strncpy(ret, begin, end - begin + 1); + if (begin != NULL) { + (void)strncpy(ret, begin, end - begin + 1); + } else { + ret[0] = '\0'; + } ret[end - begin + 1] = '\0'; } return ret; * Crash bugs ** Emacs crashes in x-popup-dialog. This can happen if the dialog widget cannot find the font it wants to use. You can work around the problem by specifying another font with an X resource--for example, `Emacs.dialog*.font: 9x15' (or any font that happens to exist on your X server). ** Emacs crashes when you use Bibtex mode. This happens if your system puts a small limit on stack size. You can prevent the problem by using a suitable shell command (often `ulimit') to raise the stack size limit before you run Emacs. Patches to raise the stack size limit automatically in `main' (src/emacs.c) on various systems would be greatly appreciated. ** Emacs crashes with SIGBUS or SIGSEGV on HPUX 9 after you delete a frame. We think this is due to a bug in the X libraries provided by HP. With the alternative X libraries in /usr/contrib/mitX11R5/lib, the problem does not happen. ** Emacs crashes with SIGBUS or SIGSEGV on Solaris after you delete a frame. We suspect that this is a similar bug in the X libraries provided by Sun. There is a report that one of these patches fixes the bug and makes the problem stop: 105216-01 105393-01 105518-01 105621-01 105665-01 105615-02 105216-02 105667-01 105401-08 105615-03 105621-02 105686-02 105736-01 105755-03 106033-01 105379-01 105786-01 105181-04 105379-03 105786-04 105845-01 105284-05 105669-02 105837-01 105837-02 105558-01 106125-02 105407-01 Another person using a newer system (kernel patch level Generic_105181-06) suspects that the bug was fixed by one of these more recent patches: 106040-07 SunOS 5.6: X Input & Output Method patch 106222-01 OpenWindows 3.6: filemgr (ff.core) fixes 105284-12 Motif 1.2.7: sparc Runtime library patch ** Error message `Symbol's value as variable is void: x', followed by a segmentation fault and core dump. This has been tracked to a bug in tar! People report that tar erroneously added a line like this at the beginning of files of Lisp code: x FILENAME, N bytes, B tape blocks If your tar has this problem, install GNU tar--if you can manage to untar it :-). ** Crashes when displaying GIF images in Emacs built with version libungif-4.1.0 are resolved by using version libungif-4.1.0b1. Configure checks for the correct version, but this problem could occur if a binary built against a shared libungif is run on a system with an older version. ** Emacs aborts inside the function `tparam1'. This can happen if Emacs was built without terminfo support, but the terminal's capabilities use format that is only supported by terminfo. If your system has ncurses installed, this might happen if your version of ncurses is broken; upgrading to a newer version of ncurses and reconfiguring and rebuilding Emacs should solve this. All modern systems support terminfo, so even if ncurses is not the problem, you should look for a way to configure Emacs so that it uses terminfo when built. ** Emacs crashes when using the Exceed 6.0 X server. If you are using Exceed 6.1, upgrade to a later version. This was reported to prevent the crashes. ** Emacs crashes with SIGSEGV in XtInitializeWidgetClass. It crashes on X, but runs fine when called with option "-nw". This has been observed when Emacs is linked with GNU ld but without passing the -z nocombreloc flag. Emacs normally knows to pass the -z nocombreloc flag when needed, so if you come across a situation where the flag is necessary but missing, please report it via M-x report-emacs-bug. On platforms such as Solaris, you can also work around this problem by configuring your compiler to use the native linker instead of GNU ld. * General runtime problems ** Lisp problems *** Changes made to .el files do not take effect. You may have forgotten to recompile them into .elc files. Then the old .elc files will be loaded, and your changes will not be seen. To fix this, do M-x byte-recompile-directory and specify the directory that contains the Lisp files. Emacs should print a warning when loading a .elc file which is older than the corresponding .el file. *** Watch out for .emacs files and EMACSLOADPATH environment vars. These control the actions of Emacs. ~/.emacs is your Emacs init file. EMACSLOADPATH overrides which directories the function "load" will search. If you observe strange problems, check for these and get rid of them, then try again. *** Using epop3.el package causes Emacs to signal an error. The error message might be something like this: "Lisp nesting exceeds max-lisp-eval-depth" This happens because epop3 redefines the function gethash, which is a built-in primitive beginning with Emacs 21.1. We don't have a patch for epop3 that fixes this, but perhaps a newer version of epop3 corrects that. *** Buffers from `with-output-to-temp-buffer' get set up in Help mode. Changes in Emacs 20.4 to the hooks used by that function cause problems for some packages, specifically BBDB. See the function's documentation for the hooks involved. BBDB 2.00.06 fixes the problem. *** The Hyperbole package causes *Help* buffers not to be displayed in Help mode due to setting `temp-buffer-show-hook' rather than using `add-hook'. Using `(add-hook 'temp-buffer-show-hook 'help-mode-maybe)' after loading Hyperbole should fix this. ** Keyboard problems *** "Compose Character" key does strange things when used as a Meta key. If you define one key to serve as both Meta and Compose Character, you will get strange results. In previous Emacs versions, this "worked" in that the key acted as Meta--that's because the older Emacs versions did not try to support Compose Character. Now Emacs tries to do character composition in the standard X way. This means that you must pick one meaning or the other for any given key. You can use both functions (Meta, and Compose Character) if you assign them to two different keys. *** C-z just refreshes the screen instead of suspending Emacs. You are probably using a shell that doesn't support job control, even though the system itself is capable of it. Either use a different shell, or set the variable `cannot-suspend' to a non-nil value. *** With M-x enable-flow-control, you need to type C-\ twice to do incremental search--a single C-\ gets no response. This has been traced to communicating with your machine via kermit, with C-\ as the kermit escape character. One solution is to use another escape character in kermit. One user did set escape-character 17 in his .kermrc file, to make C-q the kermit escape character. ** Mailers and other helper programs *** movemail compiled with POP support can't connect to the POP server. Make sure that the `pop' entry in /etc/services, or in the services NIS map if your machine uses NIS, has the same port number as the entry on the POP server. A common error is for the POP server to be listening on port 110, the assigned port for the POP3 protocol, while the client is trying to connect on port 109, the assigned port for the old POP protocol. *** RMAIL gets error getting new mail. RMAIL gets new mail from /usr/spool/mail/$USER using a program called `movemail'. This program interlocks with /bin/mail using the protocol defined by /bin/mail. There are two different protocols in general use. One of them uses the `flock' system call. The other involves creating a lock file; `movemail' must be able to write in /usr/spool/mail in order to do this. You control which one is used by defining, or not defining, the macro MAIL_USE_FLOCK in config.h or the m- or s- file it includes. IF YOU DON'T USE THE FORM OF INTERLOCKING THAT IS NORMAL ON YOUR SYSTEM, YOU CAN LOSE MAIL! If your system uses the lock file protocol, and fascist restrictions prevent ordinary users from writing the lock files in /usr/spool/mail, you may need to make `movemail' setgid to a suitable group such as `mail'. You can use these commands (as root): chgrp mail movemail chmod 2755 movemail If your system uses the lock file protocol, and fascist restrictions prevent ordinary users from writing the lock files in /usr/spool/mail, you may need to make `movemail' setgid to a suitable group such as `mail'. To do this, use the following commands (as root) after doing the make install. chgrp mail movemail chmod 2755 movemail Installation normally copies movemail from the build directory to an installation directory which is usually under /usr/local/lib. The installed copy of movemail is usually in the directory /usr/local/lib/emacs/VERSION/TARGET. You must change the group and mode of the installed copy; changing the group and mode of the build directory copy is ineffective. *** rcs2log gives you the awk error message "too many fields". This is due to an arbitrary limit in certain versions of awk. The solution is to use gawk (GNU awk). ** Problems with hostname resolution *** Emacs fails to understand most Internet host names, even though the names work properly with other programs on the same system. *** Emacs won't work with X-windows if the value of DISPLAY is HOSTNAME:0. *** GNUs can't make contact with the specified host for nntp. This typically happens on Suns and other systems that use shared libraries. The cause is that the site has installed a version of the shared library which uses a name server--but has not installed a similar version of the unshared library which Emacs uses. The result is that most programs, using the shared library, work with the nameserver, but Emacs does not. The fix is to install an unshared library that corresponds to what you installed in the shared library, and then relink Emacs. On SunOS 4.1, simply define HAVE_RES_INIT. If you have already installed the name resolver in the file libresolv.a, then you need to compile Emacs to use that library. The easiest way to do this is to add to config.h a definition of LIBS_SYSTEM, LIBS_MACHINE or LIB_STANDARD which uses -lresolv. Watch out! If you redefine a macro that is already in use in your configuration to supply some other libraries, be careful not to lose the others. Thus, you could start by adding this to config.h: #define LIBS_SYSTEM -lresolv Then if this gives you an error for redefining a macro, and you see that the s- file defines LIBS_SYSTEM as -lfoo -lbar, you could change config.h again to say this: #define LIBS_SYSTEM -lresolv -lfoo -lbar *** Emacs does not know your host's fully-qualified domain name. You need to configure your machine with a fully qualified domain name, either in /etc/hosts, /etc/hostname, the NIS, or wherever your system calls for specifying this. If you cannot fix the configuration, you can set the Lisp variable mail-host-address to the value you want. ** NFS and RFS *** Emacs says it has saved a file, but the file does not actually appear on disk. This can happen on certain systems when you are using NFS, if the remote disk is full. It is due to a bug in NFS (or certain NFS implementations), and there is apparently nothing Emacs can do to detect the problem. Emacs checks the failure codes of all the system calls involved in writing a file, including `close'; but in the case where the problem occurs, none of those system calls fails. *** Editing files through RFS gives spurious "file has changed" warnings. It is possible that a change in Emacs 18.37 gets around this problem, but in case not, here is a description of how to fix the RFS bug that causes it. There was a serious pair of bugs in the handling of the fsync() system call in the RFS server. The first is that the fsync() call is handled as another name for the close() system call (!!). It appears that fsync() is not used by very many programs; Emacs version 18 does an fsync() before closing files to make sure that the bits are on the disk. This is fixed by the enclosed patch to the RFS server. The second, more serious problem, is that fsync() is treated as a non-blocking system call (i.e., it's implemented as a message that gets sent to the remote system without waiting for a reply). Fsync is a useful tool for building atomic file transactions. Implementing it as a non-blocking RPC call (when the local call blocks until the sync is done) is a bad idea; unfortunately, changing it will break the RFS protocol. No fix was supplied for this problem. (as always, your line numbers may vary) % rcsdiff -c -r1.2 serversyscall.c RCS file: RCS/serversyscall.c,v retrieving revision 1.2 diff -c -r1.2 serversyscall.c *** /tmp/,RCSt1003677 Wed Jan 28 15:15:02 1987 --- serversyscall.c Wed Jan 28 15:14:48 1987 *************** *** 163,169 **** /* * No return sent for close or fsync! */ ! if (syscall == RSYS_close || syscall == RSYS_fsync) proc->p_returnval = deallocate_fd(proc, msg->m_args[0]); else { --- 166,172 ---- /* * No return sent for close or fsync! */ ! if (syscall == RSYS_close) proc->p_returnval = deallocate_fd(proc, msg->m_args[0]); else { ** PSGML *** Old versions of the PSGML package use the obsolete variables `before-change-function' and `after-change-function', which are no longer used by Emacs. Please use PSGML 1.2.3 or later. *** PSGML conflicts with sgml-mode. PSGML package uses the same names of some variables (like keymap) as built-in sgml-mode.el because it was created as a replacement of that package. The conflict will be shown if you load sgml-mode.el before psgml.el. E.g. this could happen if you edit HTML page and then start to work with SGML or XML file. html-mode (from sgml-mode.el) is used for HTML file and loading of psgml.el (for sgml-mode or xml-mode) will cause an error. *** Versions of the PSGML package earlier than 1.0.3 (stable) or 1.1.2 (alpha) fail to parse DTD files correctly in Emacs 20.3 and later. Here is a patch for psgml-parse.el from PSGML 1.0.1 and, probably, earlier versions. --- psgml-parse.el 1998/08/21 19:18:18 1.1 +++ psgml-parse.el 1998/08/21 19:20:00 @@ -2383,7 +2383,7 @@ (defun sgml-push-to-entity (entity &opti (setq sgml-buffer-parse-state nil)) (cond ((stringp entity) ; a file name - (save-excursion (insert-file-contents entity)) + (insert-file-contents entity) (setq default-directory (file-name-directory entity))) ((consp (sgml-entity-text entity)) ; external id? (let* ((extid (sgml-entity-text entity)) ** AUCTeX You should not be using a version older than 11.52 if you can avoid it. *** Emacs 21 freezes when visiting a TeX file with AUCTeX installed. Emacs 21 needs version 10 or later of AUCTeX; upgrading should solve these problems. *** No colors in AUCTeX with Emacs 21. Upgrade to AUC TeX version 10 or later, and make sure it is byte-compiled with Emacs 21. ** Miscellaneous problems *** Self-documentation messages are garbled. This means that the file `etc/DOC-...' doesn't properly correspond with the Emacs executable. Redumping Emacs and then installing the corresponding pair of files should fix the problem. *** Programs running under terminal emulator do not recognize `emacs' terminal type. The cause of this is a shell startup file that sets the TERMCAP environment variable. The terminal emulator uses that variable to provide the information on the special terminal type that Emacs emulates. Rewrite your shell startup file so that it does not change TERMCAP in such a case. You could use the following conditional which sets it only if it is undefined. if ( ! ${?TERMCAP} ) setenv TERMCAP ~/my-termcap-file Or you could set TERMCAP only when you set TERM--which should not happen in a non-login shell. *** In Shell mode, you get a ^M at the end of every line. This happens to people who use tcsh, because it is trying to be too smart. It sees that the Shell uses terminal type `unknown' and turns on the flag to output ^M at the end of each line. You can fix the problem by adding this to your .cshrc file: if ($?EMACS) then if ($EMACS == "t") then unset edit stty -icrnl -onlcr -echo susp ^Z endif endif *** Emacs startup on GNU/Linux systems (and possibly other systems) is slow. This can happen if the system is misconfigured and Emacs can't get the full qualified domain name, FQDN. You should have your FQDN in the /etc/hosts file, something like this: 127.0.0.1 localhost 129.187.137.82 nuc04.t30.physik.tu-muenchen.de nuc04 The way to set this up may vary on non-GNU systems. *** Attempting to visit remote files via ange-ftp fails. If the error message is "ange-ftp-file-modtime: Specified time is not representable", then this could happen when `lukemftp' is used as the ftp client. This was reported to happen on Debian GNU/Linux, kernel version 2.4.3, with `lukemftp' 1.5-5, but might happen on other systems as well. To avoid this problem, switch to using the standard ftp client. On a Debian system, type update-alternatives --config ftp and then choose /usr/bin/netkit-ftp. *** JPEG images aren't displayed. This has been reported when Emacs is built with jpeg-6a library. Upgrading to jpeg-6b solves the problem. Configure checks for the correct version, but this problem could occur if a binary built against a shared libjpeg is run on a system with an older version. *** Dired is very slow. This could happen if invocation of the `df' program takes a long time. Possible reasons for this include: - ClearCase mounted filesystems (VOBs) that sometimes make `df' response time extremely slow (dozens of seconds); - slow automounters on some old versions of Unix; - slow operation of some versions of `df'. To work around the problem, you could either (a) set the variable `directory-free-space-program' to nil, and thus prevent Emacs from invoking `df'; (b) use `df' from the GNU Fileutils package; or (c) use CVS, which is Free Software, instead of ClearCase. *** Versions of the W3 package released before Emacs 21.1 don't run under Emacs 21. This fixed in W3 version 4.0pre.47. *** The LDAP support rely on ldapsearch program from OpenLDAP version 2. It can fail to work with ldapsearch program from OpenLDAP version 1. Version 1 of OpenLDAP is now deprecated. If you are still using it, please upgrade to version 2. As a temporary workaround, remove argument "-x" from the variable `ldap-ldapsearch-args'. *** ps-print commands fail to find prologue files ps-prin*.ps. This can happen if you use an old version of X-Symbol package: it defines compatibility functions which trick ps-print into thinking it runs in XEmacs, and look for the prologue files in a wrong directory. The solution is to upgrade X-Symbol to a later version. *** On systems with shared libraries you might encounter run-time errors from the dynamic linker telling you that it is unable to find some shared libraries, for instance those for Xaw3d or image support. These errors mean Emacs has been linked with a library whose shared library is not in the default search path of the dynamic linker. Similar problems could prevent Emacs from building, since the build process invokes Emacs several times. On many systems, it is possible to set LD_LIBRARY_PATH in your environment to specify additional directories where shared libraries can be found. Other systems allow to set LD_RUN_PATH in a similar way, but before Emacs is linked. With LD_RUN_PATH set, the linker will include a specified run-time search path in the executable. On some systems, Emacs can crash due to problems with dynamic linking. Specifically, on SGI Irix 6.5, crashes were reported with backtraces like this: (dbx) where 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] 1 general_find_symbol(0xf49239d, 0x0, 0x0, 0x0, 0xf0000000, 0xf4923aa, 0x0, 0x492ddb2) ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":2140, 0xfb65a98] 2 resolve_symbol(0xf49239d, 0x4031184, 0x0, 0xfbdd438, 0x0, 0xf4923aa, 0x0, 0x492ddb2) ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":1947, 0xfb657e4] 3 lazy_text_resolve(0xd18, 0x1a3, 0x40302b4, 0x12, 0xf0000000, 0xf4923aa, 0x0, 0x492ddb2) ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld.c":997, 0xfb64d44] 4 _rld_text_resolve(0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0) ["/comp2/mtibuild/v73/workarea/v7.3/rld/rld_bridge.s":175, 0xfb6032c] (`rld' is the dynamic linker.) We don't know yet why this happens, but setting the environment variable LD_BIND_NOW to 1 (which forces the dynamic linker to bind all shared objects early on) seems to work around the problem. Please refer to the documentation of your dynamic linker for details. *** You request inverse video, and the first Emacs frame is in inverse video, but later frames are not in inverse video. This can happen if you have an old version of the custom library in your search path for Lisp packages. Use M-x list-load-path-shadows to check whether this is true. If it is, delete the old custom library. *** When you run Ispell from Emacs, it reports a "misalignment" error. This can happen if you compiled the Ispell program to use ASCII characters only and then try to use it from Emacs with non-ASCII characters, like Latin-1. The solution is to recompile Ispell with support for 8-bit characters. To see whether your Ispell program supports 8-bit characters, type this at your shell's prompt: ispell -vv and look in the output for the string "NO8BIT". If Ispell says "!NO8BIT (8BIT)", your speller supports 8-bit characters; otherwise it does not. To rebuild Ispell with 8-bit character support, edit the local.h file in the Ispell distribution and make sure it does _not_ define NO8BIT. Then rebuild the speller. Another possible cause for "misalignment" error messages is that the version of Ispell installed on your machine is old. Upgrade. Yet another possibility is that you are trying to spell-check a word in a language that doesn't fit the dictionary you choose for use by Ispell. (Ispell can only spell-check one language at a time, because it uses a single dictionary.) Make sure that the text you are spelling and the dictionary used by Ispell conform to each other. If your spell-checking program is Aspell, it has been reported that if you have a personal configuration file (normally ~/.aspell.conf), it can cause this error. Remove that file, execute `ispell-kill-ispell' in Emacs, and then try spell-checking again. * Runtime problems related to font handling ** Under X11, some characters appear as hollow boxes. Each X11 font covers just a fraction of the characters that Emacs supports. To display the whole range of Emacs characters requires many different fonts, collected into a fontset. If some of the fonts called for in your fontset do not exist on your X server, then the characters that have no font appear as hollow boxes. You can remedy the problem by installing additional fonts. The intlfonts distribution includes a full spectrum of fonts that can display all the characters Emacs supports. Another cause of this for specific characters is fonts which have a missing glyph and no default character. This is known to occur for character number 160 (no-break space) in some fonts, such as Lucida but Emacs sets the display table for the unibyte and Latin-1 version of this character to display a space. ** Under X11, some characters appear improperly aligned in their lines. You may have bad X11 fonts; try installing the intlfonts distribution. ** Certain fonts make each line take one pixel more than it "should". This is because these fonts contain characters a little taller than the font's nominal height. Emacs needs to make sure that lines do not overlap. ** Loading fonts is very slow. You might be getting scalable fonts instead of precomputed bitmaps. Known scalable font directories are "Type1" and "Speedo". A font directory contains scalable fonts if it contains the file "fonts.scale". If this is so, re-order your X windows font path to put the scalable font directories last. See the documentation of `xset' for details. With some X servers, it may be necessary to take the scalable font directories out of your path entirely, at least for Emacs 19.26. Changes in the future may make this unnecessary. ** Font Lock displays portions of the buffer in incorrect faces. By far the most frequent cause of this is a parenthesis `(' or a brace `{' in column zero. Font Lock assumes that such a paren is outside of any comment or string. This is of course not true in general, but the vast majority of well-formatted program source files don't have such parens, and therefore this assumption is used to allow optimizations in Font Lock's syntactical analysis. These optimizations avoid some pathological cases where jit-lock, the Just-in-Time fontification introduced with Emacs 21.1, could significantly slow down scrolling through the buffer, especially scrolling backwards, and also jumping to the end of a very large buffer. Beginning with version 21.4, a parenthesis or a brace in column zero is highlighted in bold-red face if it is inside a string or a comment, to indicate that it could interfere with Font Lock (and also with indentation) and should be moved or escaped with a backslash. If you don't use large buffers, or have a very fast machine which makes the delays insignificant, you can avoid the incorrect fontification by setting the variable `font-lock-beginning-of-syntax-function' to a nil value. (This must be done _after_ turning on Font Lock.) Another alternative is to avoid a paren in column zero. For example, in a Lisp string you could precede the paren with a backslash. ** With certain fonts, when the cursor appears on a character, the character doesn't appear--you get a solid box instead. One user on a Linux-based GNU system reported that this problem went away with installation of a new X server. The failing server was XFree86 3.1.1. XFree86 3.1.2 works. ** Characters are displayed as empty boxes or with wrong font under X. This can occur when two different versions of FontConfig are used. For example, XFree86 4.3.0 has one version and Gnome usually comes with a newer version. Emacs compiled with --with-gtk will then use the newer version. In most cases the problem can be temporarily fixed by stopping the application that has the error (it can be Emacs or any other application), removing ~/.fonts.cache-1, and then start the application again. If removing ~/.fonts.cache-1 and restarting doesn't help, the application with problem must be recompiled with the same version of FontConfig as the rest of the system uses. For KDE, it is sufficient to recompile Qt. ** Emacs pauses for several seconds when changing the default font. This has been reported for fvwm 2.2.5 and the window manager of KDE 2.1. The reason for the pause is Xt waiting for a ConfigureNotify event from the window manager, which the window manager doesn't send. Xt stops waiting after a default timeout of usually 5 seconds. A workaround for this is to add something like emacs.waitForWM: false to your X resources. Alternatively, add `(wait-for-wm . nil)' to a frame's parameter list, like this: (modify-frame-parameters nil '((wait-for-wm . nil))) (this should go into your `.emacs' file). ** Underlines appear at the wrong position. This is caused by fonts having a wrong UNDERLINE_POSITION property. Examples are the font 7x13 on XFree prior to version 4.1, or the jmk neep font from the Debian xfonts-jmk package. To circumvent this problem, set x-use-underline-position-properties to nil in your `.emacs'. To see what is the value of UNDERLINE_POSITION defined by the font, type `xlsfonts -lll FONT' and look at the font's UNDERLINE_POSITION property. ** When using Exceed, fonts sometimes appear too tall. When the display is set to an Exceed X-server and fonts are specified (either explicitly with the -fn option or implicitly with X resources) then the fonts may appear "too tall". The actual character sizes are correct but there is too much vertical spacing between rows, which gives the appearance of "double spacing". To prevent this, turn off the Exceed's "automatic font substitution" feature (in the font part of the configuration window). * Internationalization problems ** Characters from the mule-unicode charsets aren't displayed under X. XFree86 4 contains many fonts in iso10646-1 encoding which have minimal character repertoires (whereas the encoding part of the font name is meant to be a reasonable indication of the repertoire according to the XLFD spec). Emacs may choose one of these to display characters from the mule-unicode charsets and then typically won't be able to find the glyphs to display many characters. (Check with C-u C-x = .) To avoid this, you may need to use a fontset which sets the font for the mule-unicode sets explicitly. E.g. to use GNU unifont, include in the fontset spec: mule-unicode-2500-33ff:-gnu-unifont-*-iso10646-1,\ mule-unicode-e000-ffff:-gnu-unifont-*-iso10646-1,\ mule-unicode-0100-24ff:-gnu-unifont-*-iso10646-1 ** The UTF-8/16/7 coding systems don't encode CJK (Far Eastern) characters. Emacs by default only supports the parts of the Unicode BMP whose code points are in the ranges 0000-33ff and e000-ffff. This excludes: most of CJK, Yi and Hangul, as well as everything outside the BMP. If you read UTF-8 data with code points outside these ranges, the characters appear in the buffer as raw bytes of the original UTF-8 (composed into a single quasi-character) and they will be written back correctly as UTF-8, assuming you don't break the composed sequences. If you read such characters from UTF-16 or UTF-7 data, they are substituted with the Unicode `replacement character', and you lose information. To edit such UTF data, turn on Utf-Translate-Cjk mode, which makes many common CJK characters available for encoding and decoding and can be extended by updating the tables it uses. This also allows you to save as UTF buffers containing characters decoded by the chinese-, japanese- and korean- coding systems, e.g. cut and pasted from elsewhere. ** Mule-UCS loads very slowly. Changes to Emacs internals interact badly with Mule-UCS's `un-define' library, which is the usual interface to Mule-UCS. Apply the following patch to Mule-UCS 0.84 and rebuild it. That will help, though loading will still be slower than in Emacs 20. (Some distributions, such as Debian, may already have applied such a patch.) --- lisp/un-define.el 6 Mar 2001 22:41:38 -0000 1.30 +++ lisp/un-define.el 19 Apr 2002 18:34:26 -0000 @@ -610,13 +624,21 @@ by calling post-read-conversion and pre- (mapcar (lambda (x) - (mapcar - (lambda (y) - (mucs-define-coding-system - (nth 0 y) (nth 1 y) (nth 2 y) - (nth 3 y) (nth 4 y) (nth 5 y) (nth 6 y)) - (coding-system-put (car y) 'alias-coding-systems (list (car x)))) - (cdr x))) + (if (fboundp 'register-char-codings) + ;; Mule 5, where we don't need the eol-type specified and + ;; register-char-codings may be very slow for these coding + ;; system definitions. + (let ((y (cadr x))) + (mucs-define-coding-system + (car x) (nth 1 y) (nth 2 y) + (nth 3 y) (nth 4 y) (nth 5 y))) + (mapcar + (lambda (y) + (mucs-define-coding-system + (nth 0 y) (nth 1 y) (nth 2 y) + (nth 3 y) (nth 4 y) (nth 5 y) (nth 6 y)) + (coding-system-put (car y) 'alias-coding-systems (list (car x))))) + (cdr x))) `((utf-8 (utf-8-unix ?u "UTF-8 coding system" Note that Emacs has native support for Unicode, roughly equivalent to Mule-UCS's, so you may not need it. ** Accented ISO-8859-1 characters are displayed as | or _. Try other font set sizes (S-mouse-1). If the problem persists with other sizes as well, your text is corrupted, probably through software that is not 8-bit clean. If the problem goes away with another font size, it's probably because some fonts pretend to be ISO-8859-1 fonts when they are really ASCII fonts. In particular the schumacher-clean fonts have this bug in some versions of X. To see what glyphs are included in a font, use `xfd', like this: xfd -fn -schumacher-clean-medium-r-normal--12-120-75-75-c-60-iso8859-1 If this shows only ASCII glyphs, the font is indeed the source of the problem. The solution is to remove the corresponding lines from the appropriate `fonts.alias' file, then run `mkfontdir' in that directory, and then run `xset fp rehash'. ** The `oc-unicode' package doesn't work with Emacs 21. This package tries to define more private charsets than there are free slots now. The current built-in Unicode support is actually more flexible. (Use option `utf-translate-cjk-mode' if you need CJK support.) Files encoded as emacs-mule using oc-unicode aren't generally read correctly by Emacs 21. ** After a while, Emacs slips into unibyte mode. The VM mail package, which is not part of Emacs, sometimes does (standard-display-european t) That should be changed to (standard-display-european 1 t) * X runtime problems ** X keyboard problems *** You "lose characters" after typing Compose Character key. This is because the Compose Character key is defined as the keysym Multi_key, and Emacs (seeing that) does the proper X11 character-composition processing. If you don't want your Compose key to do that, you can redefine it with xmodmap. For example, here's one way to turn it into a Meta key: xmodmap -e "keysym Multi_key = Meta_L" If all users at your site of a particular keyboard prefer Meta to Compose, you can make the remapping happen automatically by adding the xmodmap command to the xdm setup script for that display. *** Using X Windows, control-shift-leftbutton makes Emacs hang. Use the shell command `xset bc' to make the old X Menu package work. *** M-SPC seems to be ignored as input. See if your X server is set up to use this as a command for character composition. *** The S-C-t key combination doesn't get passed to Emacs on X. This happens because some X configurations assign the Ctrl-Shift-t combination the same meaning as the Multi_key. The offending definition is in the file `...lib/X11/locale/iso8859-1/Compose'; there might be other similar combinations which are grabbed by X for similar purposes. We think that this can be countermanded with the `xmodmap' utility, if you want to be able to bind one of these key sequences within Emacs. *** Under X, C-v and/or other keys don't work. These may have been intercepted by your window manager. In particular, AfterStep 1.6 is reported to steal C-v in its default configuration. Various Meta keys are also likely to be taken by the configuration of the `feel'. See the WM's documentation for how to change this. *** Clicking C-mouse-2 in the scroll bar doesn't split the window. This currently doesn't work with scroll-bar widgets (and we don't know a good way of implementing it with widgets). If Emacs is configured --without-toolkit-scroll-bars, C-mouse-2 on the scroll bar does work. *** Inability to send an Alt-modified key, when Emacs is communicating directly with an X server. If you have tried to bind an Alt-modified key as a command, and it does not work to type the command, the first thing you should check is whether the key is getting through to Emacs. To do this, type C-h c followed by the Alt-modified key. C-h c should say what kind of event it read. If it says it read an Alt-modified key, then make sure you have made the key binding correctly. If C-h c reports an event that doesn't have the Alt modifier, it may be because your X server has no key for the Alt modifier. The X server that comes from MIT does not set up the Alt modifier by default. If your keyboard has keys named Alt, you can enable them as follows: xmodmap -e 'add mod2 = Alt_L' xmodmap -e 'add mod2 = Alt_R' If the keyboard has just one key named Alt, then only one of those commands is needed. The modifier `mod2' is a reasonable choice if you are using an unmodified MIT version of X. Otherwise, choose any modifier bit not otherwise used. If your keyboard does not have keys named Alt, you can use some other keys. Use the keysym command in xmodmap to turn a function key (or some other 'spare' key) into Alt_L or into Alt_R, and then use the commands show above to make them modifier keys. Note that if you have Alt keys but no Meta keys, Emacs translates Alt into Meta. This is because of the great importance of Meta in Emacs. ** Window-manager and toolkit-related problems *** Gnome: Emacs' xterm-mouse-mode doesn't work on the Gnome terminal. A symptom of this bug is that double-clicks insert a control sequence into the buffer. The reason this happens is an apparent incompatibility of the Gnome terminal with Xterm, which also affects other programs using the Xterm mouse interface. A problem report has been filed. *** KDE: When running on KDE, colors or fonts are not as specified for Emacs, or messed up. For example, you could see background you set for Emacs only in the empty portions of the Emacs display, while characters have some other background. This happens because KDE's defaults apply its color and font definitions even to applications that weren't compiled for KDE. The solution is to uncheck the "Apply fonts and colors to non-KDE apps" option in Preferences->Look&Feel->Style (KDE 2). In KDE 3, this option is in the "Colors" section, rather than "Style". Alternatively, if you do want the KDE defaults to apply to other applications, but not to Emacs, you could modify the file `Emacs.ad' (should be in the `/usr/share/apps/kdisplay/app-defaults/' directory) so that it doesn't set the default background and foreground only for Emacs. For example, make sure the following resources are either not present or commented out: Emacs.default.attributeForeground Emacs.default.attributeBackground Emacs*Foreground Emacs*Background *** KDE: Emacs hangs on KDE when a large portion of text is killed. This is caused by a bug in the KDE applet `klipper' which periodically requests the X clipboard contents from applications. Early versions of klipper don't implement the ICCM protocol for large selections, which leads to Emacs being flooded with selection requests. After a while, Emacs will print a message: Timed out waiting for property-notify event A workaround is to not use `klipper'. *** CDE: Frames may cover dialogs they created when using CDE. This can happen if you have "Allow Primary Windows On Top" enabled which seems to be the default in the Common Desktop Environment. To change, go in to "Desktop Controls" -> "Window Style Manager" and uncheck "Allow Primary Windows On Top". *** Xaw3d : When using Xaw3d scroll bars without arrows, the very first mouse click in a scroll bar might be ignored by the scroll bar widget. This is probably a bug in Xaw3d; when Xaw3d is compiled with arrows, the problem disappears. *** Xaw: There are known binary incompatibilities between Xaw, Xaw3d, neXtaw, XawM and the few other derivatives of Xaw. So when you compile with one of these, it may not work to dynamically link with another one. For example, strange problems, such as Emacs exiting when you type "C-x 1", were reported when Emacs compiled with Xaw3d and libXaw was used with neXtaw at run time. The solution is to rebuild Emacs with the toolkit version you actually want to use, or set LD_PRELOAD to preload the same toolkit version you built Emacs with. *** Open Motif: Problems with file dialogs in Emacs built with Open Motif. When Emacs 21 is built with Open Motif 2.1, it can happen that the graphical file dialog boxes do not work properly. The "OK", "Filter" and "Cancel" buttons do not respond to mouse clicks. Dragging the file dialog window usually causes the buttons to work again. The solution is to use LessTif instead. LessTif is a free replacement for Motif. See the file INSTALL for information on how to do this. Another workaround is not to use the mouse to trigger file prompts, but to use the keyboard. This way, you will be prompted for a file in the minibuffer instead of a graphical file dialog. *** LessTif: Problems in Emacs built with LessTif. The problems seem to depend on the version of LessTif and the Motif emulation for which it is set up. Only the Motif 1.2 emulation seems to be stable enough in LessTif. Lesstif 0.92-17's Motif 1.2 emulation seems to work okay on FreeBSD. On GNU/Linux systems, lesstif-0.92.6 configured with "./configure --enable-build-12 --enable-default-12" is reported to be the most successful. The binary GNU/Linux package lesstif-devel-0.92.0-1.i386.rpm was reported to have problems with menu placement. On some systems, even with Motif 1.2 emulation, Emacs occasionally locks up, grabbing all mouse and keyboard events. We still don't know what causes these problems; they are not reproducible by Emacs developers. *** Motif: The Motif version of Emacs paints the screen a solid color. This has been observed to result from the following X resource: Emacs*default.attributeFont: -*-courier-medium-r-*-*-*-140-*-*-*-*-iso8859-* That the resource has this effect indicates a bug in something, but we do not yet know what. If it is an Emacs bug, we hope someone can explain what the bug is so we can fix it. In the mean time, removing the resource prevents the problem. ** General X problems *** Redisplay using X11 is much slower than previous Emacs versions. We've noticed that certain X servers draw the text much slower when scroll bars are on the left. We don't know why this happens. If this happens to you, you can work around it by putting the scroll bars on the right (as they were in Emacs 19). Here's how to do this: (set-scroll-bar-mode 'right) If you're not sure whether (or how much) this problem affects you, try that and see how much difference it makes. To set things back to normal, do (set-scroll-bar-mode 'left) *** Error messages about undefined colors on X. The messages might say something like this: Unable to load color "grey95" (typically, in the `*Messages*' buffer), or something like this: Error while displaying tooltip: (error Undefined color lightyellow) These problems could happen if some other X program has used up too many colors of the X palette, leaving Emacs with insufficient system resources to load all the colors it needs. A solution is to exit the offending X programs before starting Emacs. *** Improving performance with slow X connections. There are several ways to improve this performance, any subset of which can be carried out at the same time: 1) If you don't need X Input Methods (XIM) for entering text in some language you use, you can improve performance on WAN links by using the X resource useXIM to turn off use of XIM. This does not affect the use of Emacs' own input methods, which are part of the Leim package. 2) If the connection is very slow, you might also want to consider switching off scroll bars, menu bar, and tool bar. 3) Use ssh to forward the X connection, and enable compression on this forwarded X connection (ssh -XC remotehostname emacs ...). 4) Use lbxproxy on the remote end of the connection. This is an interface to the low bandwidth X extension in most modern X servers, which improves performance dramatically, at the slight expense of correctness of the X protocol. lbxproxy acheives the performance gain by grouping several X requests in one TCP packet and sending them off together, instead of requiring a round-trip for each X request in a seperate packet. The switches that seem to work best for emacs are: -noatomsfile -nowinattr -cheaterrors -cheatevents Note that the -nograbcmap option is known to cause problems. For more about lbxproxy, see: http://www.xfree86.org/4.3.0/lbxproxy.1.html *** Emacs gives the error, Couldn't find per display information. This can result if the X server runs out of memory because Emacs uses a large number of fonts. On systems where this happens, C-h h is likely to cause it. We do not know of a way to prevent the problem. *** Emacs does not notice when you release the mouse. There are reports that this happened with (some) Microsoft mice and that replacing the mouse made it stop. *** You can't select from submenus (in the X toolkit version). On certain systems, mouse-tracking and selection in top-level menus works properly with the X toolkit, but neither of them works when you bring up a submenu (such as Bookmarks or Compare or Apply Patch, in the Files menu). This works on most systems. There is speculation that the failure is due to bugs in old versions of X toolkit libraries, but no one really knows. If someone debugs this and finds the precise cause, perhaps a workaround can be found. *** An error message such as `X protocol error: BadMatch (invalid parameter attributes) on protocol request 93'. This comes from having an invalid X resource, such as emacs*Cursor: black (which is invalid because it specifies a color name for something that isn't a color.) The fix is to correct your X resources. *** Slow startup on X11R6 with X windows. If Emacs takes two minutes to start up on X11R6, see if your X resources specify any Adobe fonts. That causes the type-1 font renderer to start up, even if the font you asked for is not a type-1 font. One way to avoid this problem is to eliminate the type-1 fonts from your font path, like this: xset -fp /usr/X11R6/lib/X11/fonts/Type1/ *** Pull-down menus appear in the wrong place, in the toolkit version of Emacs. An X resource of this form can cause the problem: Emacs*geometry: 80x55+0+0 This resource is supposed to apply, and does apply, to the menus individually as well as to Emacs frames. If that is not what you want, rewrite the resource. To check thoroughly for such resource specifications, use `xrdb -query' to see what resources the X server records, and also look at the user's ~/.Xdefaults and ~/.Xdefaults-* files. *** --with-x-toolkit version crashes when used with shared libraries. On some systems, including Sunos 4 and DGUX 5.4.2 and perhaps others, unexec doesn't work properly with the shared library for the X toolkit. You might be able to work around this by using a nonshared libXt.a library. The real fix is to upgrade the various versions of unexec and/or ralloc. We think this has been fixed on Sunos 4 and Solaris in version 19.29. *** Emacs running under X Windows does not handle mouse clicks. *** `emacs -geometry 80x20' finds a file named `80x20'. One cause of such problems is having (setq term-file-prefix nil) in your .emacs file. Another cause is a bad value of EMACSLOADPATH in the environment. *** Emacs fails to get default settings from X Windows server. The X library in X11R4 has a bug; it interchanges the 2nd and 3rd arguments to XGetDefaults. Define the macro XBACKWARDS in config.h to tell Emacs to compensate for this. I don't believe there is any way Emacs can determine for itself whether this problem is present on a given system. *** X Windows doesn't work if DISPLAY uses a hostname. People have reported kernel bugs in certain systems that cause Emacs not to work with X Windows if DISPLAY is set using a host name. But the problem does not occur if DISPLAY is set to `unix:0.0'. I think the bug has to do with SIGIO or FIONREAD. You may be able to compensate for the bug by doing (set-input-mode nil nil). However, that has the disadvantage of turning off interrupts, so that you are unable to quit out of a Lisp program by typing C-g. The easy way to do this is to put (setq x-sigio-bug t) in your site-init.el file. * Runtime problems on character termunals ** Emacs spontaneously displays "I-search: " at the bottom of the screen. This means that Control-S/Control-Q (XON/XOFF) "flow control" is being used. C-s/C-q flow control is bad for Emacs editors because it takes away C-s and C-q as user commands. Since editors do not output long streams of text without user commands, there is no need for a user-issuable "stop output" command in an editor; therefore, a properly designed flow control mechanism would transmit all possible input characters without interference. Designing such a mechanism is easy, for a person with at least half a brain. There are three possible reasons why flow control could be taking place: 1) Terminal has not been told to disable flow control 2) Insufficient padding for the terminal in use 3) Some sort of terminal concentrator or line switch is responsible First of all, many terminals have a set-up mode which controls whether they generate XON/XOFF flow control characters. This must be set to "no XON/XOFF" in order for Emacs to work. Sometimes there is an escape sequence that the computer can send to turn flow control off and on. If so, perhaps the termcap `ti' string should turn flow control off, and the `te' string should turn it on. Once the terminal has been told "no flow control", you may find it needs more padding. The amount of padding Emacs sends is controlled by the termcap entry for the terminal in use, and by the output baud rate as known by the kernel. The shell command `stty' will print your output baud rate; `stty' with suitable arguments will set it if it is wrong. Setting to a higher speed causes increased padding. If the results are wrong for the correct speed, there is probably a problem in the termcap entry. You must speak to a local Unix wizard to fix this. Perhaps you are just using the wrong terminal type. For terminals that lack a "no flow control" mode, sometimes just giving lots of padding will prevent actual generation of flow control codes. You might as well try it. If you are really unlucky, your terminal is connected to the computer through a concentrator which sends XON/XOFF flow control to the computer, or it insists on sending flow control itself no matter how much padding you give it. Unless you can figure out how to turn flow control off on this concentrator (again, refer to your local wizard), you are screwed! You should have the terminal or concentrator replaced with a properly designed one. In the mean time, some drastic measures can make Emacs semi-work. You can make Emacs ignore C-s and C-q and let the operating system handle them. To do this on a per-session basis, just type M-x enable-flow-control RET. You will see a message that C-\ and C-^ are now translated to C-s and C-q. (Use the same command M-x enable-flow-control to turn *off* this special mode. It toggles flow control handling.) If C-\ and C-^ are inconvenient for you (for example, if one of them is the escape character of your terminal concentrator), you can choose other characters by setting the variables flow-control-c-s-replacement and flow-control-c-q-replacement. But choose carefully, since all other control characters are already used by emacs. IMPORTANT: if you type C-s by accident while flow control is enabled, Emacs output will freeze, and you will have to remember to type C-q in order to continue. If you work in an environment where a majority of terminals of a certain type are flow control hobbled, you can use the function `enable-flow-control-on' to turn on this flow control avoidance scheme automatically. Here is an example: (enable-flow-control-on "vt200" "vt300" "vt101" "vt131") If this isn't quite correct (e.g. you have a mixture of flow-control hobbled and good vt200 terminals), you can still run enable-flow-control manually. I have no intention of ever redesigning the Emacs command set for the assumption that terminals use C-s/C-q flow control. XON/XOFF flow control technique is a bad design, and terminals that need it are bad merchandise and should not be purchased. Now that X is becoming widespread, XON/XOFF seems to be on the way out. If you can get some use out of GNU Emacs on inferior terminals, more power to you, but I will not make Emacs worse for properly designed systems for the sake of inferior systems. ** Control-S and Control-Q commands are ignored completely. For some reason, your system is using brain-damaged C-s/C-q flow control despite Emacs's attempts to turn it off. Perhaps your terminal is connected to the computer through a concentrator that wants to use flow control. You should first try to tell the concentrator not to use flow control. If you succeed in this, try making the terminal work without flow control, as described in the preceding section. If that line of approach is not successful, map some other characters into C-s and C-q using keyboard-translate-table. The example above shows how to do this with C-^ and C-\. ** Screen is updated wrong, but only on one kind of terminal. This could mean that the termcap entry you are using for that terminal is wrong, or it could mean that Emacs has a bug handing the combination of features specified for that terminal. The first step in tracking this down is to record what characters Emacs is sending to the terminal. Execute the Lisp expression (open-termscript "./emacs-script") to make Emacs write all terminal output into the file ~/emacs-script as well; then do what makes the screen update wrong, and look at the file and decode the characters using the manual for the terminal. There are several possibilities: 1) The characters sent are correct, according to the terminal manual. In this case, there is no obvious bug in Emacs, and most likely you need more padding, or possibly the terminal manual is wrong. 2) The characters sent are incorrect, due to an obscure aspect of the terminal behavior not described in an obvious way by termcap. This case is hard. It will be necessary to think of a way for Emacs to distinguish between terminals with this kind of behavior and other terminals that behave subtly differently but are classified the same by termcap; or else find an algorithm for Emacs to use that avoids the difference. Such changes must be tested on many kinds of terminals. 3) The termcap entry is wrong. See the file etc/TERMS for information on changes that are known to be needed in commonly used termcap entries for certain terminals. 4) The characters sent are incorrect, and clearly cannot be right for any terminal with the termcap entry you were using. This is unambiguously an Emacs bug, and can probably be fixed in termcap.c, tparam.c, term.c, scroll.c, cm.c or dispnew.c. ** Control-S and Control-Q commands are ignored completely on a net connection. Some versions of rlogin (and possibly telnet) do not pass flow control characters to the remote system to which they connect. On such systems, emacs on the remote system cannot disable flow control on the local system. One way to cure this is to disable flow control on the local host (the one running rlogin, not the one running rlogind) using the stty command, before starting the rlogin process. On many systems, "stty start u stop u" will do this. Some versions of tcsh will prevent even this from working. One way around this is to start another shell before starting rlogin, and issue the stty command to disable flow control from that shell. If none of these methods work, the best solution is to type M-x enable-flow-control at the beginning of your emacs session, or if you expect the problem to continue, add a line such as the following to your .emacs (on the host running rlogind): (enable-flow-control-on "vt200" "vt300" "vt101" "vt131") See the entry about spontaneous display of I-search (above) for more info. ** Output from Control-V is slow. On many bit-map terminals, scrolling operations are fairly slow. Often the termcap entry for the type of terminal in use fails to inform Emacs of this. The two lines at the bottom of the screen before a Control-V command are supposed to appear at the top after the Control-V command. If Emacs thinks scrolling the lines is fast, it will scroll them to the top of the screen. If scrolling is slow but Emacs thinks it is fast, the usual reason is that the termcap entry for the terminal you are using does not specify any padding time for the `al' and `dl' strings. Emacs concludes that these operations take only as much time as it takes to send the commands at whatever line speed you are using. You must fix the termcap entry to specify, for the `al' and `dl', as much time as the operations really take. Currently Emacs thinks in terms of serial lines which send characters at a fixed rate, so that any operation which takes time for the terminal to execute must also be padded. With bit-map terminals operated across networks, often the network provides some sort of flow control so that padding is never needed no matter how slow an operation is. You must still specify a padding time if you want Emacs to realize that the operation takes a long time. This will cause padding characters to be sent unnecessarily, but they do not really cost much. They will be transmitted while the scrolling is happening and then discarded quickly by the terminal. Most bit-map terminals provide commands for inserting or deleting multiple lines at once. Define the `AL' and `DL' strings in the termcap entry to say how to do these things, and you will have fast output without wasted padding characters. These strings should each contain a single %-spec saying how to send the number of lines to be scrolled. These %-specs are like those in the termcap `cm' string. You should also define the `IC' and `DC' strings if your terminal has a command to insert or delete multiple characters. These take the number of positions to insert or delete as an argument. A `cs' string to set the scrolling region will reduce the amount of motion you see on the screen when part of the screen is scrolled. ** You type Control-H (Backspace) expecting to delete characters. Put `stty dec' in your .login file and your problems will disappear after a day or two. The choice of Backspace for erasure was based on confusion, caused by the fact that backspacing causes erasure (later, when you type another character) on most display terminals. But it is a mistake. Deletion of text is not the same thing as backspacing followed by failure to overprint. I do not wish to propagate this confusion by conforming to it. For this reason, I believe `stty dec' is the right mode to use, and I have designed Emacs to go with that. If there were a thousand other control characters, I would define Control-h to delete as well; but there are not very many other control characters, and I think that providing the most mnemonic possible Help character is more important than adapting to people who don't use `stty dec'. If you are obstinate about confusing buggy overprinting with deletion, you can redefine Backspace in your .emacs file: (global-set-key "\b" 'delete-backward-char) You can probably access help-command via f1. ** Colors are not available on a tty or in xterm. Emacs 21 supports colors on character terminals and terminal emulators, but this support relies on the terminfo or termcap database entry to specify that the display supports color. Emacs looks at the "Co" capability for the terminal to find out how many colors are supported; it should be non-zero to activate the color support within Emacs. (Most color terminals support 8 or 16 colors.) If your system uses terminfo, the name of the capability equivalent to "Co" is "colors". In addition to the "Co" capability, Emacs needs the "op" (for ``original pair'') capability, which tells how to switch the terminal back to the default foreground and background colors. Emacs will not use colors if this capability is not defined. If your terminal entry doesn't provide such a capability, try using the ANSI standard escape sequence \E[00m (that is, define a new termcap/terminfo entry and make it use your current terminal's entry plus \E[00m for the "op" capability). Finally, the "NC" capability (terminfo name: "ncv") tells Emacs which attributes cannot be used with colors. Setting this capability incorrectly might have the effect of disabling colors; try setting this capability to `0' (zero) and see if that helps. Emacs uses the database entry for the terminal whose name is the value of the environment variable TERM. With `xterm', a common terminal entry that supports color is `xterm-color', so setting TERM's value to `xterm-color' might activate the color support on an xterm-compatible emulator. Beginning with version 21.4, Emacs supports the --color command-line option which may be used to force Emacs to use one of a few popular modes for getting colors on a tty. For example, --color=ansi8 sets up for using the ANSI-standard escape sequences that support 8 colors. Some modes do not use colors unless you turn on the Font-lock mode. Some people have long ago set their `~/.emacs' files to turn on Font-lock on X only, so they won't see colors on a tty. The recommended way of turning on Font-lock is by typing "M-x global-font-lock-mode RET" or by customizing the variable `global-font-lock-mode'. * Runtime problems specific to individual Unix variants ** GNU/Linux *** GNU/Linux: On Linux-based GNU systems using libc versions 5.4.19 through 5.4.22, Emacs crashes at startup with a segmentation fault. This problem happens if libc defines the symbol __malloc_initialized. One known solution is to upgrade to a newer libc version. 5.4.33 is known to work. *** GNU/Linux: After upgrading to a newer version of Emacs, the Meta key stops working. This was reported to happen on a GNU/Linux system distributed by Mandrake. The reason is that the previous version of Emacs was modified by Mandrake to make the Alt key act as the Meta key, on a keyboard where the Windows key is the one which produces the Meta modifier. A user who started using a newer version of Emacs, which was not hacked by Mandrake, expected the Alt key to continue to act as Meta, and was astonished when that didn't happen. The solution is to find out what key on your keyboard produces the Meta modifier, and use that key instead. Try all of the keys to the left and to the right of the space bar, together with the `x' key, and see which combination produces "M-x" in the echo area. You can also use the `xmodmap' utility to show all the keys which produce a Meta modifier: xmodmap -pk | egrep -i "meta|alt" A more convenient way of finding out which keys produce a Meta modifier is to use the `xkbprint' utility, if it's available on your system: xkbprint 0:0 /tmp/k.ps This produces a PostScript file `/tmp/k.ps' with a picture of your keyboard; printing that file on a PostScript printer will show what keys can serve as Meta. The `xkeycaps' also shows a visual representation of the current keyboard settings. It also allows to modify them. *** GNU/Linux: low startup on Linux-based GNU systems. People using systems based on the Linux kernel sometimes report that startup takes 10 to 15 seconds longer than `usual'. This is because Emacs looks up the host name when it starts. Normally, this takes negligible time; the extra delay is due to improper system configuration. This problem can occur for both networked and non-networked machines. Here is how to fix the configuration. It requires being root. **** Networked Case. First, make sure the files `/etc/hosts' and `/etc/host.conf' both exist. The first line in the `/etc/hosts' file should look like this (replace HOSTNAME with your host name): 127.0.0.1 HOSTNAME Also make sure that the `/etc/host.conf' files contains the following lines: order hosts, bind multi on Any changes, permanent and temporary, to the host name should be indicated in the `/etc/hosts' file, since it acts a limited local database of addresses and names (e.g., some SLIP connections dynamically allocate ip addresses). **** Non-Networked Case. The solution described in the networked case applies here as well. However, if you never intend to network your machine, you can use a simpler solution: create an empty `/etc/host.conf' file. The command `touch /etc/host.conf' suffices to create the file. The `/etc/hosts' file is not necessary with this approach. *** GNU/Linux: Emacs on a tty switches the cursor to large blinking block. This was reported to happen on some GNU/Linux systems which use ncurses version 5.0, but could be relevant for other versions as well. These versions of ncurses come with a `linux' terminfo entry, where the "cvvis" capability (termcap "vs") is defined as "\E[?25h\E[?8c" (show cursor, change size). This escape sequence switches on a blinking hardware text-mode cursor whose size is a full character cell. This blinking cannot be stopped, since a hardware cursor always blinks. A work-around is to redefine the "cvvis" capability so that it enables a *software* cursor. The software cursor works by inverting the colors of the character at point, so what you see is a block cursor that doesn't blink. For this to work, you need to redefine the "cnorm" capability as well, so that it operates on the software cursor instead of the hardware cursor. To this end, run "infocmp linux > linux-term", edit the file `linux-term' to make both the "cnorm" and "cvvis" capabilities send the sequence "\E[?25h\E[?17;0;64c", and then run "tic linux-term" to produce a modified terminfo entry. Alternatively, if you want a blinking underscore as your Emacs cursor, change the "cvvis" capability to send the "\E[?25h\E[?0c" command. *** GNU/Linux: Error messages `internal facep []' happen on GNU/Linux systems. There is a report that replacing libc.so.5.0.9 with libc.so.5.2.16 caused this to start happening. People are not sure why, but the problem seems unlikely to be in Emacs itself. Some suspect that it is actually Xlib which won't work with libc.so.5.2.16. Using the old library version is a workaround. ** Mac OS X *** Mac OS X (Carbon): Environment Variables from dotfiles are ignored. When starting Emacs from the Dock or the Finder on Mac OS X, the environment variables that are set up in dotfiles, such as .cshrc or .profile, are ignored. This is because the Finder and Dock are not started from a shell, but instead from the Window Manager itself. The workaround for this is to create a .MacOSX/environment.plist file to setup these environment variables. These environment variables will apply to all processes regardless of where they are started. For me information, see http://developer.apple.com/qa/qa2001/qa1067.html. *** Mac OS X (Carbon): Process output truncated when using ptys. There appears to be a problem with the implementation of pty's on the Mac OS X that causes process output to be truncated. To avoid this, leave process-connection-type set to its default value of nil. ** FreeBSD *** FreeBSD 2.1.5: useless symbolic links remain in /tmp or other directories that have the +t bit. This is because of a kernel bug in FreeBSD 2.1.5 (fixed in 2.2). Emacs uses symbolic links to implement file locks. In a directory with +t bit, the directory owner becomes the owner of the symbolic link, so that it cannot be removed by anyone else. If you don't like those useless links, you can let Emacs not to using file lock by adding #undef CLASH_DETECTION to config.h. *** FreeBSD: Getting a Meta key on the console. By default, neither Alt nor any other key acts as a Meta key on FreeBSD, but this can be changed using kbdcontrol(1). Dump the current keymap to a file with the command $ kbdcontrol -d >emacs.kbd Edit emacs.kbd, and give the key you want to be the Meta key the definition `meta'. For instance, if your keyboard has a ``Windows'' key with scan code 105, change the line for scan code 105 in emacs.kbd to look like this 105 meta meta meta meta meta meta meta meta O to make the Windows key the Meta key. Load the new keymap with $ kbdcontrol -l emacs.kbd ** HP-UX *** HP/UX : Shell mode gives the message, "`tty`: Ambiguous". christos@theory.tn.cornell.edu says: The problem is that in your .cshrc you have something that tries to execute `tty`. If you are not running the shell on a real tty then tty will print "not a tty". Csh expects one word in some places, but tty is giving it back 3. The solution is to add a pair of quotes around `tty` to make it a single word: if (`tty` == "/dev/console") should be changed to: if ("`tty`" == "/dev/console") Even better, move things that set up terminal sections out of .cshrc and into .login. *** HP/UX: `Pid xxx killed due to text modification or page I/O error'. On HP/UX, you can get that error when the Emacs executable is on an NFS file system. HP/UX responds this way if it tries to swap in a page and does not get a response from the server within a timeout whose default value is just ten seconds. If this happens to you, extend the timeout period. *** HP/UX: Emacs is slow using X11R5. This happens if you use the MIT versions of the X libraries--it doesn't run as fast as HP's version. People sometimes use the version because they see the HP version doesn't have the libraries libXaw.a, libXmu.a, libXext.a and others. HP/UX normally doesn't come with those libraries installed. To get good performance, you need to install them and rebuild Emacs. *** HP/UX: The right Alt key works wrong on German HP keyboards (and perhaps other non-English HP keyboards too). This is because HP-UX defines the modifiers wrong in X. Here is a shell script to fix the problem; be sure that it is run after VUE configures the X server. xmodmap 2> /dev/null - << EOF keysym Alt_L = Meta_L keysym Alt_R = Meta_R EOF xmodmap - << EOF clear mod1 keysym Mode_switch = NoSymbol add mod1 = Meta_L keysym Meta_R = Mode_switch add mod2 = Mode_switch EOF *** HP/UX: "Cannot find callback list" messages from dialog boxes in Emacs built with Motif. This problem resulted from a bug in GCC 2.4.5. Newer GCC versions such as 2.7.0 fix the problem. *** HP/UX: Emacs does not recognize the AltGr key. To fix this, set up a file ~/.dt/sessions/sessionetc with executable rights, containing this text: -------------------------------- xmodmap 2> /dev/null - << EOF keysym Alt_L = Meta_L keysym Alt_R = Meta_R EOF xmodmap - << EOF clear mod1 keysym Mode_switch = NoSymbol add mod1 = Meta_L keysym Meta_R = Mode_switch add mod2 = Mode_switch EOF -------------------------------- *** HP/UX: Large file support is disabled. See the comments in src/s/hpux10.h. *** HP/UX 11.0: Emacs makes HP/UX 11.0 crash. This is a bug in HPUX; HPUX patch PHKL_16260 is said to fix it. ** AIX *** AIX: Trouble using ptys. People often install the pty devices on AIX incorrectly. Use `smit pty' to reinstall them properly. *** AIXterm: Your Delete key sends a Backspace to the terminal. The solution is to include in your .Xdefaults the lines: *aixterm.Translations: #override <Key>BackSpace: string(0x7f) aixterm*ttyModes: erase ^? This makes your Backspace key send DEL (ASCII 127). *** AIX: You get this message when running Emacs: Could not load program emacs Symbol smtcheckinit in csh is undefined Error was: Exec format error or this one: Could not load program .emacs Symbol _system_con in csh is undefined Symbol _fp_trapsta in csh is undefined Error was: Exec format error These can happen when you try to run on AIX 3.2.5 a program that was compiled with 3.2.4. The fix is to recompile. *** AIX 3.2.4: Releasing Ctrl/Act key has no effect, if Shift is down. Due to a feature of AIX, pressing or releasing the Ctrl/Act key is ignored when the Shift, Alt or AltGr keys are held down. This can lead to the keyboard being "control-locked"--ordinary letters are treated as control characters. You can get out of this "control-locked" state by pressing and releasing Ctrl/Act while not pressing or holding any other keys. *** AIX 4.2: Emacs gets a segmentation fault at startup. If you are using IBM's xlc compiler, compile emacs.c without optimization; that should avoid the problem. *** AIX: If linking fails because libXbsd isn't found, check if you are compiling with the system's `cc' and CFLAGS containing `-O5'. If so, you have hit a compiler bug. Please make sure to re-configure Emacs so that it isn't compiled with `-O5'. *** AIX 4.3.x or 4.4: Compiling fails. This could happen if you use /bin/c89 as your compiler, instead of the default `cc'. /bin/c89 treats certain warnings, such as benign redefinitions of macros, as errors, and fails the build. A solution is to use the default compiler `cc'. *** AIX 4: Some programs fail when run in a Shell buffer with an error message like No terminfo entry for "unknown". On AIX, many terminal type definitions are not installed by default. `unknown' is one of them. Install the "Special Generic Terminal Definitions" to make them defined. ** Solaris We list bugs in current versions here. Solaris 2.x and 4.x are covered in the section on legacy systems. *** On Solaris, C-x doesn't get through to Emacs when you use the console. This is a Solaris feature (at least on Intel x86 cpus). Type C-r C-r C-t, to toggle whether C-x gets through to Emacs. *** Problem with remote X server on Suns. On a Sun, running Emacs on one machine with the X server on another may not work if you have used the unshared system libraries. This is because the unshared libraries fail to use YP for host name lookup. As a result, the host name you specify may not be recognized. *** Emacs reports a BadAtom error (from X) running on Solaris 7 or 8. This happens when Emacs was built on some other version of Solaris. Rebuild it on Solaris 8. *** On Solaris, CTRL-t is ignored by Emacs when you use the fr.ISO-8859-15 locale (and maybe other related locales). You can fix this by editing the file: /usr/openwin/lib/locale/iso8859-15/Compose Near the bottom there is a line that reads: Ctrl<t> <quotedbl> <Y> : "\276" threequarters that should read: Ctrl<T> <quotedbl> <Y> : "\276" threequarters Note the lower case <t>. Changing this line should make C-t work. *** When using M-x dbx with the SparcWorks debugger, the `up' and `down' commands do not move the arrow in Emacs. You can fix this by adding the following line to `~/.dbxinit': dbxenv output_short_file_name off ** Irix *** Irix 5.2: unexelfsgi.c can't find cmplrs/stsupport.h. The file cmplrs/stsupport.h was included in the wrong file set in the Irix 5.2 distribution. You can find it in the optional fileset compiler_dev, or copy it from some other Irix 5.2 system. A kludgy workaround is to change unexelfsgi.c to include sym.h instead of syms.h. *** Irix 5.3: "out of virtual swap space". This message occurs when the system runs out of swap space due to too many large programs running. The solution is either to provide more swap space or to reduce the number of large programs being run. You can check the current status of the swap space by executing the command `swap -l'. You can increase swap space by changing the file /etc/fstab. Adding a line like this: /usr/swap/swap.more swap swap pri=3 0 0 where /usr/swap/swap.more is a file previously created (for instance by using /etc/mkfile), will increase the swap space by the size of that file. Execute `swap -m' or reboot the machine to activate the new swap area. See the manpages for `swap' and `fstab' for further information. The objectserver daemon can use up lots of memory because it can be swamped with NIS information. It collects information about all users on the network that can log on to the host. If you want to disable the objectserver completely, you can execute the command `chkconfig objectserver off' and reboot. That may disable some of the window system functionality, such as responding CDROM icons. You can also remove NIS support from the objectserver. The SGI `admin' FAQ has a detailed description on how to do that; see question 35 ("Why isn't the objectserver working?"). The admin FAQ can be found at ftp://viz.tamu.edu/pub/sgi/faq/. *** Irix 5.3: Emacs crashes in utmpname. This problem is fixed in Patch 3175 for Irix 5.3. It is also fixed in Irix versions 6.2 and up. *** Irix 6.0: Make tries (and fails) to build a program named unexelfsgi. A compiler bug inserts spaces into the string "unexelfsgi . o" in src/Makefile. Edit src/Makefile, after configure is run, find that string, and take out the spaces. Compiler fixes in Irix 6.0.1 should eliminate this problem. *** Irix 6.5: Emacs crashes on the SGI R10K, when compiled with GCC. This seems to be fixed in GCC 2.95. *** Trouble using ptys on IRIX, or running out of ptys. The program mkpts (which may be in `/usr/adm' or `/usr/sbin') needs to be set-UID to root, or non-root programs like Emacs will not be able to allocate ptys reliably. ** SCO Unix and UnixWare *** SCO 3.2v4: Unusable default font. The Open Desktop environment comes with default X resource settings that tell Emacs to use a variable-width font. Emacs cannot use such fonts, so it does not work. This is caused by the file /usr/lib/X11/app-defaults/ScoTerm, which is the application-specific resource file for the `scoterm' terminal emulator program. It contains several extremely general X resources that affect other programs besides `scoterm'. In particular, these resources affect Emacs also: *Font: -*-helvetica-medium-r-*--12-*-p-* *Background: scoBackground *Foreground: scoForeground The best solution is to create an application-specific resource file for Emacs, /usr/lib/X11/sco/startup/Emacs, with the following contents: Emacs*Font: -*-courier-medium-r-*-*-*-120-*-*-*-*-iso8859-1 Emacs*Background: white Emacs*Foreground: black (These settings mimic the Emacs defaults, but you can change them to suit your needs.) This resource file is only read when the X server starts up, so you should restart it by logging out of the Open Desktop environment or by running `scologin stop; scologin start` from the shell as root. Alternatively, you can put these settings in the /usr/lib/X11/app-defaults/Emacs resource file and simply restart Emacs, but then they will not affect remote invocations of Emacs that use the Open Desktop display. These resource files are not normally shared across a network of SCO machines; you must create the file on each machine individually. *** Regular expressions matching bugs on SCO systems. On SCO, there are problems in regexp matching when Emacs is compiled with the system compiler. The compiler version is "Microsoft C version 6", SCO 4.2.0h Dev Sys Maintenance Supplement 01/06/93; Quick C Compiler Version 1.00.46 (Beta). The solution is to compile with GCC. *** UnixWare 2.1: Error 12 (virtual memory exceeded) when dumping Emacs. Paul Abrahams (abrahams@acm.org) reports that with the installed virtual memory settings for UnixWare 2.1.2, an Error 12 occurs during the "make" that builds Emacs, when running temacs to dump emacs. That error indicates that the per-process virtual memory limit has been exceeded. The default limit is probably 32MB. Raising the virtual memory limit to 40MB should make it possible to finish building Emacs. You can do this with the command `ulimit' (sh) or `limit' (csh). But you have to be root to do it. According to Martin Sohnius, you can also retune this in the kernel: # /etc/conf/bin/idtune SDATLIM 33554432 ## soft data size limit # /etc/conf/bin/idtune HDATLIM 33554432 ## hard " # /etc/conf/bin/idtune SVMMSIZE unlimited ## soft process size limit # /etc/conf/bin/idtune HVMMSIZE unlimited ## hard " # /etc/conf/bin/idbuild -B (He recommends you not change the stack limit, though.) These changes take effect when you reboot. * Runtime problems specific to MS-Windows ** Emacs exits with "X protocol error" when run with an X server for MS-Windows. A certain X server for Windows had a bug which caused this. Supposedly the newer 32-bit version of this server doesn't have the problem. ** Known problems with the MS-Windows port of Emacs 21.2. Frames are not refreshed while the File or Font dialog or a pop-up menu is displayed. This also means help text for pop-up menus is not displayed at all. This is because message handling under Windows is synchronous, so we cannot handle repaint (or any other) messages while waiting for a system function to return the result of the dialog or pop-up menu interaction. Windows 95 and Windows NT up to version 4.0 do not support help text for menus. Help text is only available in later versions of Windows. There are problems with display if mouse-tracking is enabled and the mouse is moved off a frame, over another frame then back over the first frame. A workaround is to click the left mouse button inside the frame after moving back into it. Some minor flickering still persists during mouse-tracking, although not as severely as in 21.1. Emacs can sometimes abort when non-ASCII text, possibly with null characters, is copied and pasted into a buffer. An inactive cursor remains in an active window after the Windows Manager driven switch of the focus, until a key is pressed. Windows input methods are not recognized by Emacs (as of v21.2). Some of these input methods cause the keyboard to send characters encoded in the appropriate coding system (e.g., ISO 8859-1 for Latin-1 characters, ISO 8859-8 for Hebrew characters, etc.). To make this work, set the keyboard coding system to the appropriate value after you activate the Windows input method. For example, if you activate the Hebrew input method, type "C-x RET k iso-8859-8 RET". (Emacs ought to recognize the Windows language-change event and set up the appropriate keyboard encoding automatically, but it doesn't do that yet.) The %b specifier for format-time-string does not produce abbreviated month names with consistent widths for some locales on some versions of Windows. This is caused by a deficiency in the underlying system library function. ** Problems running Perl under Emacs on MS-Windows NT/95. `perl -de 0' just hangs when executed in an Emacs subshell. The fault lies with Perl (indirectly with Windows NT/95). The problem is that the Perl debugger explicitly opens a connection to "CON", which is the DOS/NT equivalent of "/dev/tty", for interacting with the user. On Unix, this is okay, because Emacs (or the shell?) creates a pseudo-tty so that /dev/tty is really the pipe Emacs is using to communicate with the subprocess. On NT, this fails because CON always refers to the handle for the relevant console (approximately equivalent to a tty), and cannot be redirected to refer to the pipe Emacs assigned to the subprocess as stdin. A workaround is to modify perldb.pl to use STDIN/STDOUT instead of CON. For Perl 4: *** PERL/LIB/PERLDB.PL.orig Wed May 26 08:24:18 1993 --- PERL/LIB/PERLDB.PL Mon Jul 01 15:28:16 1996 *************** *** 68,74 **** $rcfile=".perldb"; } else { ! $console = "con"; $rcfile="perldb.ini"; } --- 68,74 ---- $rcfile=".perldb"; } else { ! $console = ""; $rcfile="perldb.ini"; } For Perl 5: *** perl/5.001/lib/perl5db.pl.orig Sun Jun 04 21:13:40 1995 --- perl/5.001/lib/perl5db.pl Mon Jul 01 17:00:08 1996 *************** *** 22,28 **** $rcfile=".perldb"; } elsif (-e "con") { ! $console = "con"; $rcfile="perldb.ini"; } else { --- 22,28 ---- $rcfile=".perldb"; } elsif (-e "con") { ! $console = ""; $rcfile="perldb.ini"; } else { ** On MS-Windows 95, Alt-f6 does not get through to Emacs. This character seems to be trapped by the kernel in Windows 95. You can enter M-f6 by typing ESC f6. ** Typing Alt-Shift has strange effects on MS-Windows. This combination of keys is a command to change keyboard layout. If you proceed to type another non-modifier key before you let go of Alt and Shift, the Alt and Shift act as modifiers in the usual way. A more permanent work around is to change it to another key combination, or disable it in the keyboard control panel. ** Interrupting Cygwin port of Bash from Emacs doesn't work. Cygwin 1.x builds of the ported Bash cannot be interrupted from the MS-Windows version of Emacs. This is due to some change in the Bash port or in the Cygwin library which apparently make Bash ignore the keyboard interrupt event sent by Emacs to Bash. (Older Cygwin ports of Bash, up to b20.1, did receive SIGINT from Emacs.) ** Accessing remote files with ange-ftp hangs the MS-Windows version of Emacs. If the FTP client is the Cygwin port of GNU `ftp', this appears to be due to some bug in the Cygwin DLL or some incompatibility between it and the implementation of asynchronous subprocesses in the Windows port of Emacs. Specifically, some parts of the FTP server responses are not flushed out, apparently due to buffering issues, which confuses ange-ftp. The solution is to downgrade to an older version of the Cygwin DLL (version 1.3.2 was reported to solve the problem), or use the stock Windows FTP client, usually found in the `C:\WINDOWS' or 'C:\WINNT' directory. To force ange-ftp use the stock Windows client, set the variable `ange-ftp-ftp-program-name' to the absolute file name of the client's executable. For example: (setq ange-ftp-ftp-program-name "c:/windows/ftp.exe") If you want to stick with the Cygwin FTP client, you can work around this problem by putting this in your `.emacs' file: (setq ange-ftp-ftp-program-args '("-i" "-n" "-g" "-v" "--prompt" "") ** lpr commands don't work on MS-Windows with some cheap printers. This problem may also strike other platforms, but the solution is likely to be a global one, and not Emacs specific. Many cheap inkjet, and even some cheap laser printers, do not print plain text anymore, they will only print through graphical printer drivers. A workaround on MS-Windows is to use Windows' basic built in editor to print (this is possibly the only useful purpose it has): (setq printer-name "") ;; notepad takes the default (setq lpr-command "notepad") ;; notepad (setq lpr-switches nil) ;; not needed (setq lpr-printer-switch "/P") ;; run notepad as batch printer ** Antivirus software interacts badly with the MS-Windows version of Emacs. The usual manifestation of these problems is that subprocesses don't work or even wedge the entire system. In particular, "M-x shell RET" was reported to fail to work. But other commands also sometimes don't work when an antivirus package is installed. The solution is to switch the antivirus software to a less aggressive mode (e.g., disable the ``auto-protect'' feature), or even uninstall or disable it entirely. ** On MS-Windows 95/98/ME, subprocesses do not terminate properly. This is a limitation of the Operating System, and can cause problems when shutting down Windows. Ensure that all subprocesses are exited cleanly before exiting Emacs. For more details, see the FAQ at http://www.gnu.org/software/emacs/windows/. ** MS-Windows 95/98/ME crashes when Emacs invokes non-existent programs. When a program you are trying to run is not found on the PATH, Windows might respond by crashing or locking up your system. In particular, this has been reported when trying to compile a Java program in JDEE when javac.exe is installed, but not on the system PATH. ** Pressing the mouse button on MS-Windows does not give a mouse-2 event. This is usually a problem with the mouse driver. Because most Windows programs do not do anything useful with the middle mouse button, many mouse drivers allow you to define the wheel press to do something different. Some drivers do not even have the option to generate a middle button press. In such cases, setting the wheel press to "scroll" sometimes works if you press the button twice. Trying a generic mouse driver might help. ** Scrolling the mouse wheel on MS-Windows always scrolls the top window. This is another common problem with mouse drivers. Instead of generating scroll events, some mouse drivers try to fake scroll bar movement. But they are not intelligent enough to handle multiple scroll bars within a frame. Trying a generic mouse driver might help. ** Mail sent through Microsoft Exchange in some encodings appears to be mangled and is not seen correctly in Rmail or Gnus. We don't know exactly what happens, but it isn't an Emacs problem in cases we've seen. ** On MS-Windows, you cannot use the right-hand ALT key and the left-hand CTRL key together to type a Control-Meta character. This is a consequence of a misfeature beyond Emacs's control. Under Windows, the AltGr key on international keyboards generates key events with the modifiers Right-Alt and Left-Ctrl. Since Emacs cannot distinguish AltGr from an explicit Right-Alt and Left-Ctrl combination, whenever it sees Right-Alt and Left-Ctrl it assumes that AltGr has been pressed. The variable `w32-recognize-altgr' can be set to nil to tell Emacs that AltGr is really Ctrl and Alt. ** Under some X-servers running on MS-Windows, Emacs' display is incorrect. The symptoms are that Emacs does not completely erase blank areas of the screen during scrolling or some other screen operations (e.g., selective display or when killing a region). M-x recenter will cause the screen to be completely redisplayed and the "extra" characters will disappear. This is known to occur under Exceed 6, and possibly earlier versions as well; it is reportedly solved in version 6.2.0.16 and later. The problem lies in the X-server settings. There are reports that you can solve the problem with Exceed by running `Xconfig' from within NT, choosing "X selection", then un-checking the boxes "auto-copy X selection" and "auto-paste to X selection". Of this does not work, please inform bug-gnu-emacs@gnu.org. Then please call support for your X-server and see if you can get a fix. If you do, please send it to bug-gnu-emacs@gnu.org so we can list it here. * Build-time problems ** Configuration *** The `configure' script doesn't find the jpeg library. There are reports that this happens on some systems because the linker by default only looks for shared libraries, but jpeg distribution by default only installs a nonshared version of the library, `libjpeg.a'. If this is the problem, you can configure the jpeg library with the `--enable-shared' option and then rebuild libjpeg. This produces a shared version of libjpeg, which you need to install. Finally, rerun the Emacs configure script, which should now find the jpeg library. Alternatively, modify the generated src/Makefile to link the .a file explicitly, and edit src/config.h to define HAVE_JPEG. *** AIX: You get this compiler error message: Processing include file ./XMenuInt.h 1501-106: (S) Include file X11/Xlib.h not found. This means your system was installed with only the X11 runtime i.d libraries. You have to find your sipo (bootable tape) and install X11Dev... with smit. ** Compilation *** Building Emacs over NFS fails with ``Text file busy''. This was reported to happen when building Emacs on a GNU/Linux system (RedHat Linux 6.2) using a build directory automounted from Solaris (SunOS 5.6) file server, but it might not be limited to that configuration alone. Presumably, the NFS server doesn't commit the files' data to disk quickly enough, and the Emacs executable file is left ``busy'' for several seconds after Emacs has finished dumping itself. This causes the subsequent commands which invoke the dumped Emacs executable to fail with the above message. In some of these cases, a time skew between the NFS server and the machine where Emacs is built is detected and reported by GNU Make (it says that some of the files have modification time in the future). This might be a symptom of NFS-related problems. If the NFS server runs on Solaris, apply the Solaris patch 105379-05 (Sunos 5.6: /kernel/misc/nfssrv patch). If that doesn't work, or if you have a different version of the OS or the NFS server, you can force the NFS server to use 1KB blocks, which was reported to fix the problem albeit at a price of slowing down file I/O. You can force 1KB blocks by specifying the "-o rsize=1024,wsize=1024" options to the `mount' command, or by adding ",rsize=1024,wsize=1024" to the mount options in the appropriate system configuration file, such as `/etc/auto.home'. Alternatively, when Make fails due to this problem, you could wait for a few seconds and then invoke Make again. In one particular case, waiting for 10 or more seconds between the two Make invocations seemed to work around the problem. Similar problems can happen if your machine NFS-mounts a directory onto itself. Suppose the Emacs sources live in `/usr/local/src' and you are working on the host called `marvin'. Then an entry in the `/etc/fstab' file like the following is asking for trouble: marvin:/usr/local/src /usr/local/src ...options.omitted... The solution is to remove this line from `etc/fstab'. *** Building Emacs with GCC 2.9x fails in the `src' directory. This may happen if you use a development version of GNU `cpp' from one of the GCC snapshots between Oct 2000 and Feb 2001, or from a released version of GCC newer than 2.95.2 which was prepared around those dates; similar problems were reported with some snapshots of GCC 3.1 around Sep 30 2001. The preprocessor in those versions is incompatible with a traditional Unix cpp (e.g., it expands ".." into ". .", which breaks relative file names that reference the parent directory; or inserts TAB characters before lines that set Make variables). The solution is to make sure the preprocessor is run with the `-traditional' option. The `configure' script does that automatically when it detects the known problems in your cpp, but you might hit some unknown ones. To force the `configure' script to use `-traditional', run the script like this: CPP='gcc -E -traditional' ./configure ... (replace the ellipsis "..." with any additional arguments you pass to the script). Note that this problem does not pertain to the MS-Windows port of Emacs, since it doesn't use the preprocessor to generate Makefiles. *** src/Makefile and lib-src/Makefile are truncated--most of the file missing. *** Compiling wakeup, in lib-src, says it can't make wakeup.c. This can happen if configure uses GNU sed version 2.03. That version had a bug. GNU sed version 2.05 works properly.To solve the problem, install the current version of GNU Sed, then rerun Emacs's configure script. *** Compiling lib-src says there is no rule to make test-distrib.c. This results from a bug in a VERY old version of GNU Sed. To solve the problem, install the current version of GNU Sed, then rerun Emacs's configure script. *** Building the MS-Windows port with Cygwin GCC can fail. Emacs may not build using recent Cygwin builds of GCC, such as Cygwin version 1.1.8, using the default configure settings. It appears to be necessary to specify the -mwin32 flag when compiling, and define __MSVCRT__, like so: configure --with-gcc --cflags -mwin32 --cflags -D__MSVCRT__ *** Building the MS-Windows port fails with a CreateProcess failure. Some versions of mingw32 make on some versions of Windows do not seem to detect the shell correctly. Try "make SHELL=cmd.exe", or if that fails, try running make from Cygwin bash instead. *** Building the MS-Windows port with Leim fails in the `leim' directory. The error message might be something like this: Converting d:/emacs-21.3/leim/CXTERM-DIC/4Corner.tit to quail-package... Invalid ENCODE: value in TIT dictionary NMAKE : fatal error U1077: '"../src/obj-spd/i386/emacs.exe"' : return code '0xffffffff' Stop. This can happen if the Leim distribution is unpacked with a program which converts the `*.tit' files to DOS-style CR-LF text format. The `*.tit' files in the leim/CXTERM-DIC directory require Unix-style line endings to compile properly, because Emacs reads them without any code or EOL conversions. The solution is to make sure the program used to unpack Leim does not change the files' line endings behind your back. The GNU FTP site has in the `/gnu/emacs/windows' directory a program called `djtarnt.exe' which can be used to unpack `.tar.gz' and `.zip' archives without mangling them. *** Building `ctags' for MS-Windows with the MinGW port of GCC fails. This might happen due to a bug in the MinGW header assert.h, which defines the `assert' macro with a trailing semi-colon. The following patch to assert.h should solve this: *** include/assert.h.orig Sun Nov 7 02:41:36 1999 --- include/assert.h Mon Jan 29 11:49:10 2001 *************** *** 41,47 **** /* * If not debugging, assert does nothing. */ ! #define assert(x) ((void)0); #else /* debugging enabled */ --- 41,47 ---- /* * If not debugging, assert does nothing. */ ! #define assert(x) ((void)0) #else /* debugging enabled */ ** Linking *** Building Emacs with a system compiler fails to link because of an undefined symbol such as __eprintf which does not appear in Emacs. This can happen if some of the libraries linked into Emacs were built with GCC, but Emacs itself is being linked with a compiler other than GCC. Object files compiled with GCC might need some helper functions from libgcc.a, the library which comes with GCC, but the system compiler does not instruct the linker to search libgcc.a during the link stage. A solution is to link with GCC, like this: make CC=gcc Since the .o object files already exist, this will not recompile Emacs with GCC, but just restart by trying again to link temacs. *** AIX 1.3 ptf 0013: Link failure. There is a real duplicate definition of the function `_slibc_free' in the library /lib/libc_s.a (just do nm on it to verify). The workaround/fix is: cd /lib ar xv libc_s.a NLtmtime.o ar dv libc_s.a NLtmtime.o *** AIX 4.1.2: Linker error messages such as ld: 0711-212 SEVERE ERROR: Symbol .__quous, found in the global symbol table of archive /usr/lib/libIM.a, was not defined in archive member shr.o. This is a problem in libIM.a. You can work around it by executing these shell commands in the src subdirectory of the directory where you build Emacs: cp /usr/lib/libIM.a . chmod 664 libIM.a ranlib libIM.a Then change -lIM to ./libIM.a in the command to link temacs (in Makefile). *** Sun with acc: Link failure when using acc on a Sun. To use acc, you need additional options just before the libraries, such as /usr/lang/SC2.0.1/values-Xt.o -L/usr/lang/SC2.0.1/cg87 -L/usr/lang/SC2.0.1 and you need to add -lansi just before -lc. The precise file names depend on the compiler version, so we cannot easily arrange to supply them. *** Linking says that the functions insque and remque are undefined. Change oldXMenu/Makefile by adding insque.o to the variable OBJS. *** `tparam' reported as a multiply-defined symbol when linking with ncurses. This problem results from an incompatible change in ncurses, in version 1.9.9e approximately. This version is unable to provide a definition of tparm without also defining tparam. This is also incompatible with Terminfo; as a result, the Emacs Terminfo support does not work with this version of ncurses. The fix is to install a newer version of ncurses, such as version 4.2. ** Dumping *** Linux: Segfault during `make bootstrap' under certain recent versions of the Linux kernel. With certain recent Linux kernels (like the one of Redhat Fedora Core 1), the new "Exec-shield" functionality is enabled by default, which creates a different memory layout that breaks the emacs dumper. You can check the Exec-shield state like this: cat /proc/sys/kernel/exec-shield It returns 1 or 2 when Exec-shield is enabled, 0 otherwise. Please read your system documentation for more details on Exec-shield and associated commands. When Exec-shield is enabled, building Emacs will segfault during the execution of this command: temacs --batch --load loadup [dump|bootstrap] To work around this problem, it is necessary to temporarily disable Exec-shield while building Emacs, using the `setarch' command like this: setarch i386 ./configure <configure parameters> setarch i386 make <make parameters> *** Fatal signal in the command temacs -l loadup inc dump. This command is the final stage of building Emacs. It is run by the Makefile in the src subdirectory, or by build.com on VMS. It has been known to get fatal errors due to insufficient swapping space available on the machine. On 68000s, it has also happened because of bugs in the subroutine `alloca'. Verify that `alloca' works right, even for large blocks (many pages). *** test-distrib says that the distribution has been clobbered. *** or, temacs prints "Command key out of range 0-127". *** or, temacs runs and dumps emacs, but emacs totally fails to work. *** or, temacs gets errors dumping emacs. This can be because the .elc files have been garbled. Do not be fooled by the fact that most of a .elc file is text: these are binary files and can contain all 256 byte values. In particular `shar' cannot be used for transmitting GNU Emacs. It typically truncates "lines". What appear to be "lines" in a binary file can of course be of any length. Even once `shar' itself is made to work correctly, `sh' discards null characters when unpacking the shell archive. I have also seen character \177 changed into \377. I do not know what transfer means caused this problem. Various network file transfer programs are suspected of clobbering the high bit. If you have a copy of Emacs that has been damaged in its nonprinting characters, you can fix them: 1) Record the names of all the .elc files. 2) Delete all the .elc files. 3) Recompile alloc.c with a value of PURESIZE twice as large. (See puresize.h.) You might as well save the old alloc.o. 4) Remake emacs. It should work now. 5) Running emacs, do Meta-x byte-compile-file repeatedly to recreate all the .elc files that used to exist. You may need to increase the value of the variable max-lisp-eval-depth to succeed in running the compiler interpreted on certain .el files. 400 was sufficient as of last report. 6) Reinstall the old alloc.o (undoing changes to alloc.c if any) and remake temacs. 7) Remake emacs. It should work now, with valid .elc files. *** temacs prints "Pure Lisp storage exhausted". This means that the Lisp code loaded from the .elc and .el files during temacs -l loadup inc dump took up more space than was allocated. This could be caused by 1) adding code to the preloaded Lisp files 2) adding more preloaded files in loadup.el 3) having a site-init.el or site-load.el which loads files. Note that ANY site-init.el or site-load.el is nonstandard; if you have received Emacs from some other site and it contains a site-init.el or site-load.el file, consider deleting that file. 4) getting the wrong .el or .elc files (not from the directory you expected). 5) deleting some .elc files that are supposed to exist. This would cause the source files (.el files) to be loaded instead. They take up more room, so you lose. 6) a bug in the Emacs distribution which underestimates the space required. If the need for more space is legitimate, change the definition of PURESIZE in puresize.h. But in some of the cases listed above, this problem is a consequence of something else that is wrong. Be sure to check and fix the real problem. *** Linux: Emacs crashes when dumping itself on Mac PPC running Yellow Dog GNU/Linux. The crashes happen inside the function Fmake_symbol; here's a typical C backtrace printed by GDB: 0x190c0c0 in Fmake_symbol () (gdb) where #0 0x190c0c0 in Fmake_symbol () #1 0x1942ca4 in init_obarray () #2 0x18b3500 in main () #3 0x114371c in __libc_start_main (argc=5, argv=0x7ffff5b4, envp=0x7ffff5cc, This could happen because GCC version 2.95 and later changed the base of the load address to 0x10000000. Emacs needs to be told about this, but we currently cannot do that automatically, because that breaks other versions of GNU/Linux on the MacPPC. Until we find a way to distinguish between the Yellow Dog and the other varieties of GNU/Linux systems on the PPC, you will have to manually uncomment the following section near the end of the file src/m/macppc.h in the Emacs distribution: #if 0 /* This breaks things on PPC GNU/Linux except for Yellowdog, even with identical GCC, as, ld. Let's take it out until we know what's really going on here. */ /* GCC 2.95 and newer on GNU/Linux PPC changed the load address to 0x10000000. */ #if defined __linux__ #if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 95) #define DATA_SEG_BITS 0x10000000 #endif #endif #endif /* 0 */ Remove the "#if 0" and "#endif" directives which surround this, save the file, and then reconfigure and rebuild Emacs. The dumping process should now succeed. *** HPUX 10.20: Emacs crashes during dumping on the HPPA machine. This seems to be due to a GCC bug; it is fixed in GCC 2.8.1. ** Installation *** Installing Emacs gets an error running `install-info'. You need to install a recent version of Texinfo; that package supplies the `install-info' command. ** First execution *** Emacs binary is not in executable format, and cannot be run. This was reported to happen when Emacs is built in a directory mounted via NFS, for some combinations of NFS client and NFS server. Usually, the file `emacs' produced in these cases is full of binary null characters, and the `file' utility says: emacs: ASCII text, with no line terminators We don't know what exactly causes this failure. A work-around is to build Emacs in a directory on a local disk. *** The dumped Emacs crashes when run, trying to write pure data. Two causes have been seen for such problems. 1) On a system where getpagesize is not a system call, it is defined as a macro. If the definition (in both unexec.c and malloc.c) is wrong, it can cause problems like this. You might be able to find the correct value in the man page for a.out (5). 2) Some systems allocate variables declared static among the initialized variables. Emacs makes all initialized variables in most of its files pure after dumping, but the variables declared static and not initialized are not supposed to be pure. On these systems you may need to add "#define static" to the m- or the s- file. * Emacs 19 problems ** Error messages `Wrong number of arguments: #<subr where-is-internal>, 5'. This typically results from having the powerkey library loaded. Powerkey was designed for Emacs 19.22. It is obsolete now because Emacs 19 now has this feature built in; and powerkey also calls where-is-internal in an obsolete way. So the fix is to arrange not to load powerkey. * Runtime problems on legacy systems This section covers bugs reported on very old hardware or software. If you are using hardware and an operating system shipped after 2000, it is unlikely you will see any of these. ** Ancient operating systems *** ISC Unix **** ISC: display-time causes kernel problems on ISC systems. Under Interactive Unix versions 3.0.1 and 4.0 (and probably other versions), display-time causes the loss of large numbers of STREVENT cells. Eventually the kernel's supply of these cells is exhausted. This makes emacs and the whole system run slow, and can make other processes die, in particular pcnfsd. Other emacs functions that communicate with remote processes may have the same problem. Display-time seems to be far the worst. The only known fix: Don't run display-time. *** SunOS **** Sun 4.0.x: M-x shell persistently reports "Process shell exited abnormally with code 1". This happened on Suns as a result of what is said to be a bug in Sunos version 4.0.x. The only fix was to reboot the machine. **** SunOS4.1.1 and SunOS4.1.3: Mail is lost when sent to local aliases. Many emacs mail user agents (VM and rmail, for instance) use the sendmail.el library. This library can arrange for mail to be delivered by passing messages to the /usr/lib/sendmail (usually) program . In doing so, it passes the '-t' flag to sendmail, which means that the name of the recipient of the message is not on the command line and, therefore, that sendmail must parse the message to obtain the destination address. There is a bug in the SunOS4.1.1 and SunOS4.1.3 versions of sendmail. In short, when given the -t flag, the SunOS sendmail won't recognize non-local (i.e. NIS) aliases. It has been reported that the Solaris 2.x versions of sendmail do not have this bug. For those using SunOS 4.1, the best fix is to install sendmail V8 or IDA sendmail (which have other advantages over the regular sendmail as well). At the time of this writing, these official versions are available: Sendmail V8 on ftp.cs.berkeley.edu in /ucb/sendmail: sendmail.8.6.9.base.tar.Z (the base system source & documentation) sendmail.8.6.9.cf.tar.Z (configuration files) sendmail.8.6.9.misc.tar.Z (miscellaneous support programs) sendmail.8.6.9.xdoc.tar.Z (extended documentation, with postscript) IDA sendmail on vixen.cso.uiuc.edu in /pub: sendmail-5.67b+IDA-1.5.tar.gz **** Sunos 5.3: Subprocesses remain, hanging but not zombies. A bug in Sunos 5.3 causes Emacs subprocesses to remain after Emacs exits. Sun patch # 101415-02 is part of the fix for this, but it only applies to ptys, and doesn't fix the problem with subprocesses communicating through pipes. **** Sunos 4: You get the error ld: Undefined symbol __lib_version. This is the result of using cc or gcc with the shared library meant for acc (the Sunpro compiler). Check your LD_LIBRARY_PATH and delete /usr/lang/SC2.0.1 or some similar directory. **** SunOS 4.1.3: Emacs unpredictably crashes in _yp_dobind_soft. This happens if you configure Emacs specifying just `sparc-sun-sunos4' on a system that is version 4.1.3. You must specify the precise version number (or let configure figure out the configuration, which it can do perfectly well for SunOS). **** Sunos 4.1.3: Emacs gets hung shortly after startup. We think this is due to a bug in Sunos. The word is that one of these Sunos patches fixes the bug: 100075-11 100224-06 100347-03 100482-05 100557-02 100623-03 100804-03 101080-01 100103-12 100249-09 100496-02 100564-07 100630-02 100891-10 101134-01 100170-09 100296-04 100377-09 100507-04 100567-04 100650-02 101070-01 101145-01 100173-10 100305-15 100383-06 100513-04 100570-05 100689-01 101071-03 101200-02 100178-09 100338-05 100421-03 100536-02 100584-05 100784-01 101072-01 101207-01 We don't know which of these patches really matter. If you find out which ones, please inform bug-gnu-emacs@gnu.org. **** SunOS 4: Emacs processes keep going after you kill the X server (or log out, if you logged in using X). Someone reported that recompiling with GCC 2.7.0 fixed this problem. **** SunOS: You get linker errors ld: Undefined symbol _get_wmShellWidgetClass _get_applicationShellWidgetClass The fix to this is to install patch 100573 for OpenWindows 3.0 or link libXmu statically. *** Apollo Domain **** Shell mode ignores interrupts on Apollo Domain. You may find that M-x shell prints the following message: Warning: no access to tty; thus no job control in this shell... This can happen if there are not enough ptys on your system. Here is how to make more of them. % cd /dev % ls pty* # shows how many pty's you have. I had 8, named pty0 to pty7) % /etc/crpty 8 # creates eight new pty's *** Irix *** Irix 6.2: No visible display on mips-sgi-irix6.2 when compiling with GCC 2.8.1. This problem went away after installing the latest IRIX patches as of 8 Dec 1998. The same problem has been reported on Irix 6.3. *** Irix 6.3: substituting environment variables in file names in the minibuffer gives peculiar error messages such as Substituting nonexistent environment variable "" This is not an Emacs bug; it is caused by something in SGI patch 003082 August 11, 1998. *** OPENSTEP **** OPENSTEP 4.2: Compiling syntax.c with gcc 2.7.2.1 fails. The compiler was reported to crash while compiling syntax.c with the following message: cc: Internal compiler error: program cc1obj got fatal signal 11 To work around this, replace the macros UPDATE_SYNTAX_TABLE_FORWARD, INC_BOTH, and INC_FROM with functions. To this end, first define 3 functions, one each for every macro. Here's an example: static int update_syntax_table_forward(int from) { return(UPDATE_SYNTAX_TABLE_FORWARD(from)); }/*update_syntax_table_forward*/ Then replace all references to UPDATE_SYNTAX_TABLE_FORWARD in syntax.c with a call to the function update_syntax_table_forward. *** Solaris 2.x **** Strange results from format %d in a few cases, on a Sun. Sun compiler version SC3.0 has been found to miscompile part of editfns.c. The workaround is to compile with some other compiler such as GCC. **** On Solaris, Emacs dumps core if lisp-complete-symbol is called. If you compile Emacs with the -fast or -xO4 option with version 3.0.2 of the Sun C compiler, Emacs dumps core when lisp-complete-symbol is called. The problem does not happen if you compile with GCC. **** On Solaris, Emacs crashes if you use (display-time). This can happen if you configure Emacs without specifying the precise version of Solaris that you are using. **** Solaris 2.3 and 2.4: Unpredictable segmentation faults. A user reported that this happened in 19.29 when it was compiled with the Sun compiler, but not when he recompiled with GCC 2.7.0. We do not know whether something in Emacs is partly to blame for this. **** Solaris 2.4: Emacs dumps core on startup. Bill Sebok says that the cause of this is Solaris 2.4 vendor patch 102303-05, which extends the Solaris linker to deal with the Solaris Common Desktop Environment's linking needs. You can fix the problem by removing this patch and installing patch 102049-02 instead. However, that linker version won't work with CDE. Solaris 2.5 comes with a linker that has this bug. It is reported that if you install all the latest patches (as of June 1996), the bug is fixed. We suspect the crucial patch is one of these, but we don't know for certain. 103093-03: [README] SunOS 5.5: kernel patch (2140557 bytes) 102832-01: [README] OpenWindows 3.5: Xview Jumbo Patch (4181613 bytes) 103242-04: [README] SunOS 5.5: linker patch (595363 bytes) (One user reports that the bug was fixed by those patches together with patches 102980-04, 103279-01, 103300-02, and 103468-01.) If you can determine which patch does fix the bug, please tell bug-gnu-emacs@gnu.org. Meanwhile, the GNU linker links Emacs properly on both Solaris 2.4 and Solaris 2.5. **** Solaris 2.4: Dired hangs and C-g does not work. Or Emacs hangs forever waiting for termination of a subprocess that is a zombie. casper@fwi.uva.nl says the problem is in X11R6. Rebuild libX11.so after changing the file xc/config/cf/sunLib.tmpl. Change the lines #if ThreadedX #define SharedX11Reqs -lthread #endif to: #if OSMinorVersion < 4 #if ThreadedX #define SharedX11Reqs -lthread #endif #endif Be sure also to edit x/config/cf/sun.cf so that OSMinorVersion is 4 (as it should be for Solaris 2.4). The file has three definitions for OSMinorVersion: the first is for x86, the second for SPARC under Solaris, and the third for SunOS 4. Make sure to update the definition for your type of machine and system. Then do `make Everything' in the top directory of X11R6, to rebuild the makefiles and rebuild X. The X built this way work only on Solaris 2.4, not on 2.3. For multithreaded X to work it is necessary to install patch 101925-02 to fix problems in header files [2.4]. You need to reinstall gcc or re-run just-fixinc after installing that patch. However, Frank Rust <frust@iti.cs.tu-bs.de> used a simpler solution: he changed #define ThreadedX YES to #define ThreadedX NO in sun.cf and did `make World' to rebuild X11R6. Removing all `-DXTHREAD*' flags and `-lthread' entries from lib/X11/Makefile and typing 'make install' in that directory also seemed to work. **** Solaris 2.x: GCC complains "64 bit integer types not supported". This suggests that GCC is not installed correctly. Most likely you are using GCC 2.7.2.3 (or earlier) on Solaris 2.6 (or later); this does not work without patching. To run GCC 2.7.2.3 on Solaris 2.6 or later, you must patch fixinc.svr4 and reinstall GCC from scratch as described in the Solaris FAQ <http://www.wins.uva.nl/pub/solaris/solaris2.html>. A better fix is to upgrade to GCC 2.8.1 or later. **** Solaris 2.7: Building Emacs with WorkShop Compilers 5.0 98/12/15 C 5.0 failed, apparently with non-default CFLAGS, most probably due to compiler bugs. Using Sun Solaris 2.7 Sun WorkShop 6 update 1 C release was reported to work without problems. It worked OK on another system with Solaris 8 using apparently the same 5.0 compiler and the default CFLAGS. **** Solaris 2.x: Emacs dumps core when built with Motif. The Solaris Motif libraries are buggy, at least up through Solaris 2.5.1. Install the current Motif runtime library patch appropriate for your host. (Make sure the patch is current; some older patch versions still have the bug.) You should install the other patches recommended by Sun for your host, too. You can obtain Sun patches from ftp://sunsolve.sun.com/pub/patches/; look for files with names ending in `.PatchReport' to see which patches are currently recommended for your host. On Solaris 2.6, Emacs is said to work with Motif when Solaris patch 105284-12 is installed, but fail when 105284-15 is installed. 105284-18 might fix it again. *** Solaris 2.6 and 7: the Compose key does not work. This is a bug in Motif in Solaris. Supposedly it has been fixed for the next major release of Solaris. However, if someone with Sun support complains to Sun about the bug, they may release a patch. If you do this, mention Sun bug #4188711. One workaround is to use a locale that allows non-ASCII characters. For example, before invoking emacs, set the LC_ALL environment variable to "en_US" (American English). The directory /usr/lib/locale lists the supported locales; any locale other than "C" or "POSIX" should do. pen@lysator.liu.se says (Feb 1998) that the Compose key does work if you link with the MIT X11 libraries instead of the Solaris X11 libraries. *** Ultrix and Digital Unix **** Ultrix 4.2: `make install' fails on install-doc with `Error 141'. This happens on Ultrix 4.2 due to failure of a pipeline of tar commands. We don't know why they fail, but the bug seems not to be in Emacs. The workaround is to run the shell command in install-doc by hand. **** Digital Unix 4.0: Garbled display on non-X terminals when Emacs runs. So far it appears that running `tset' triggers this problem (when TERM is vt100, at least). If you do not run `tset', then Emacs displays properly. If someone can tell us precisely which effect of running `tset' actually causes the problem, we may be able to implement a fix in Emacs. **** Ultrix: `expand-file-name' fails to work on any but the machine you dumped Emacs on. On Ultrix, if you use any of the functions which look up information in the passwd database before dumping Emacs (say, by using expand-file-name in site-init.el), then those functions will not work in the dumped Emacs on any host but the one Emacs was dumped on. The solution? Don't use expand-file-name in site-init.el, or in anything it loads. Yuck - some solution. I'm not sure why this happens; if you can find out exactly what is going on, and perhaps find a fix or a workaround, please let us know. Perhaps the YP functions cache some information, the cache is included in the dumped Emacs, and is then inaccurate on any other host. *** SVr4 **** SVr4: On some variants of SVR4, Emacs does not work at all with X. Try defining BROKEN_FIONREAD in your config.h file. If this solves the problem, please send a bug report to tell us this is needed; be sure to say exactly what type of machine and system you are using. **** SVr4: After running emacs once, subsequent invocations crash. Some versions of SVR4 have a serious bug in the implementation of the mmap () system call in the kernel; this causes emacs to run correctly the first time, and then crash when run a second time. Contact your vendor and ask for the mmap bug fix; in the mean time, you may be able to work around the problem by adding a line to your operating system description file (whose name is reported by the configure script) that reads: #define SYSTEM_MALLOC This makes Emacs use memory less efficiently, but seems to work around the kernel bug. *** Linux 1.x **** Linux 1.0-1.04: Typing C-c C-c in Shell mode kills your X server. This happens with Linux kernel 1.0 thru 1.04, approximately. The workaround is to define SIGNALS_VIA_CHARACTERS in config.h and recompile Emacs. Newer Linux kernel versions don't have this problem. **** Linux 1.3: Output from subprocess (such as man or diff) is randomly truncated on GNU/Linux systems. This is due to a kernel bug which seems to be fixed in Linux version 1.3.75. ** MS-DOS *** When compiling with DJGPP on MS-Windows NT, "config msdos" fails. If the error message is "VDM has been already loaded", this is because Windows has a program called `redir.exe' that is incompatible with a program by the same name supplied with DJGPP, which is used by config.bat. To resolve this, move the DJGPP's `bin' subdirectory to the front of your PATH environment variable. *** When compiling with DJGPP on MS-Windows 95, Make fails for some targets like make-docfile. This can happen if long file name support (the setting of environment variable LFN) when Emacs distribution was unpacked and during compilation are not the same. See the MSDOG section of INSTALL for the explanation of how to avoid this problem. *** Emacs compiled with DJGPP complains at startup: "Wrong type of argument: internal-facep, msdos-menu-active-face" This can happen if you define an environment variable `TERM'. Emacs on MSDOS uses an internal terminal emulator which is disabled if the value of `TERM' is anything but the string "internal". Emacs then works as if its terminal were a dumb glass teletype that doesn't support faces. To work around this, arrange for `TERM' to be undefined when Emacs runs. The best way to do that is to add an [emacs] section to the DJGPP.ENV file which defines an empty value for `TERM'; this way, only Emacs gets the empty value, while the rest of your system works as before. *** MS-DOS: Emacs crashes at startup. Some users report that Emacs 19.29 requires dpmi memory management, and crashes on startup if the system does not have it. We don't yet know why this happens--perhaps these machines don't have enough real memory, or perhaps something is wrong in Emacs or the compiler. However, arranging to use dpmi support is a workaround. You can find out if you have a dpmi host by running go32 without arguments; it will tell you if it uses dpmi memory. For more information about dpmi memory, consult the djgpp FAQ. (djgpp is the GNU C compiler as packaged for MSDOS.) Compiling Emacs under MSDOS is extremely sensitive for proper memory configuration. If you experience problems during compilation, consider removing some or all memory resident programs (notably disk caches) and make sure that your memory managers are properly configured. See the djgpp faq for configuration hints. *** Emacs compiled with DJGPP for MS-DOS/MS-Windows cannot access files in the directory with the special name `dev' under the root of any drive, e.g. `c:/dev'. This is an unfortunate side-effect of the support for Unix-style device names such as /dev/null in the DJGPP runtime library. A work-around is to rename the problem directory to another name. *** MS-DOS+DJGPP: Problems on MS-DOG if DJGPP v2.0 is used to compile Emacs. There are two DJGPP library bugs which cause problems: * Running `shell-command' (or `compile', or `grep') you get `Searching for program: permission denied (EACCES), c:/command.com'; * After you shell to DOS, Ctrl-Break kills Emacs. To work around these bugs, you can use two files in the msdos subdirectory: `is_exec.c' and `sigaction.c'. Compile them and link them into the Emacs executable `temacs'; then they will replace the incorrect library functions. *** MS-DOS: Emacs compiled for MSDOS cannot find some Lisp files, or other run-time support files, when long filename support is enabled. Usually, this problem will manifest itself when Emacs exits immediately after flashing the startup screen, because it cannot find the Lisp files it needs to load at startup. Redirect Emacs stdout and stderr to a file to see the error message printed by Emacs. Another manifestation of this problem is that Emacs is unable to load the support for editing program sources in languages such as C and Lisp. This can happen if the Emacs distribution was unzipped without LFN support, thus causing long filenames to be truncated to the first 6 characters and a numeric tail that Windows 95 normally attaches to it. You should unzip the files again with a utility that supports long filenames (such as djtar from DJGPP or InfoZip's UnZip program compiled with DJGPP v2). The MSDOG section of the file INSTALL explains this issue in more detail. Another possible reason for such failures is that Emacs compiled for MSDOS is used on Windows NT, where long file names are not supported by this version of Emacs, but the distribution was unpacked by an unzip program that preserved the long file names instead of truncating them to DOS 8+3 limits. To be useful on NT, the MSDOS port of Emacs must be unzipped by a DOS utility, so that long file names are properly truncated. ** Archaic window managers and toolkits *** OpenLook: Under OpenLook, the Emacs window disappears when you type M-q. Some versions of the Open Look window manager interpret M-q as a quit command for whatever window you are typing at. If you want to use Emacs with that window manager, you should try to configure the window manager to use some other command. You can disable the shortcut keys entirely by adding this line to ~/.OWdefaults: OpenWindows.WindowMenuAccelerators: False **** twm: A position you specified in .Xdefaults is ignored, using twm. twm normally ignores "program-specified" positions. You can tell it to obey them with this command in your `.twmrc' file: UsePPosition "on" #allow clients to request a position ** Bugs related to old DEC hardware *** The Compose key on a DEC keyboard does not work as Meta key. This shell command should fix it: xmodmap -e 'keycode 0xb1 = Meta_L' *** Keyboard input gets confused after a beep when using a DECserver as a concentrator. This problem seems to be a matter of configuring the DECserver to use 7 bit characters rather than 8 bit characters. * Build problems on legacy systems ** BSD/386 1.0: --with-x-toolkit option configures wrong. This problem is due to bugs in the shell in version 1.0 of BSD/386. The workaround is to edit the configure file to use some other shell, such as bash. ** Digital Unix 4.0: Emacs fails to build, giving error message Invalid dimension for the charset-ID 160 This is due to a bug or an installation problem in GCC 2.8.0. Installing a more recent version of GCC fixes the problem. ** Digital Unix 4.0: Failure in unexec while dumping emacs. This problem manifests itself as an error message unexec: Bad address, writing data section to ... The user suspects that this happened because his X libraries were built for an older system version, ./configure --x-includes=/usr/include --x-libraries=/usr/shlib made the problem go away. ** Sunos 4.1.1: there are errors compiling sysdep.c. If you get errors such as "sysdep.c", line 2017: undefined structure or union "sysdep.c", line 2017: undefined structure or union "sysdep.c", line 2019: nodename undefined This can result from defining LD_LIBRARY_PATH. It is very tricky to use that environment variable with Emacs. The Emacs configure script links many test programs with the system libraries; you must make sure that the libraries available to configure are the same ones available when you build Emacs. ** SunOS 4.1.1: You get this error message from GNU ld: /lib/libc.a(_Q_sub.o): Undefined symbol __Q_get_rp_rd referenced from text segment The problem is in the Sun shared C library, not in GNU ld. The solution is to install Patch-ID# 100267-03 from Sun. ** Sunos 4.1: Undefined symbols when linking using --with-x-toolkit. If you get the undefined symbols _atowc _wcslen, _iswprint, _iswspace, _iswcntrl, _wcscpy, and _wcsncpy, then you need to add -lXwchar after -lXaw in the command that links temacs. This problem seems to arise only when the international language extensions to X11R5 are installed. ** SunOS: Emacs gets error message from linker on Sun. If the error message says that a symbol such as `f68881_used' or `ffpa_used' or `start_float' is undefined, this probably indicates that you have compiled some libraries, such as the X libraries, with a floating point option other than the default. It's not terribly hard to make this work with small changes in crt0.c together with linking with Fcrt1.o, Wcrt1.o or Mcrt1.o. However, the easiest approach is to build Xlib with the default floating point option: -fsoft. ** SunOS: Undefined symbols _dlopen, _dlsym and/or _dlclose. If you see undefined symbols _dlopen, _dlsym, or _dlclose when linking with -lX11, compile and link against the file mit/util/misc/dlsym.c in the MIT X11R5 distribution. Alternatively, link temacs using shared libraries with s/sunos4shr.h. (This doesn't work if you use the X toolkit.) If you get the additional error that the linker could not find lib_version.o, try extracting it from X11/usr/lib/X11/libvim.a in X11R4, then use it in the link. ** VMS: Compilation errors on VMS. You will get warnings when compiling on VMS because there are variable names longer than 32 (or whatever it is) characters. This is not an error. Ignore it. VAX C does not support #if defined(foo). Uses of this construct were removed, but some may have crept back in. They must be rewritten. There is a bug in the C compiler which fails to sign extend characters in conditional expressions. The bug is: char c = -1, d = 1; int i; i = d ? c : d; The result is i == 255; the fix is to typecast the char in the conditional expression as an (int). Known occurrences of such constructs in Emacs have been fixed. ** Vax C compiler bugs affecting Emacs. You may get one of these problems compiling Emacs: foo.c line nnn: compiler error: no table entry for op STASG foo.c: fatal error in /lib/ccom These are due to bugs in the C compiler; the code is valid C. Unfortunately, the bugs are unpredictable: the same construct may compile properly or trigger one of these bugs, depending on what else is in the source file being compiled. Even changes in header files that should not affect the file being compiled can affect whether the bug happens. In addition, sometimes files that compile correctly on one machine get this bug on another machine. As a result, it is hard for me to make sure this bug will not affect you. I have attempted to find and alter these constructs, but more can always appear. However, I can tell you how to deal with it if it should happen. The bug comes from having an indexed reference to an array of Lisp_Objects, as an argument in a function call: Lisp_Object *args; ... ... foo (5, args[i], ...)... putting the argument into a temporary variable first, as in Lisp_Object *args; Lisp_Object tem; ... tem = args[i]; ... foo (r, tem, ...)... causes the problem to go away. The `contents' field of a Lisp vector is an array of Lisp_Objects, so you may see the problem happening with indexed references to that. ** 68000 C compiler problems Various 68000 compilers have different problems. These are some that have been observed. *** Using value of assignment expression on union type loses. This means that x = y = z; or foo (x = z); does not work if x is of type Lisp_Object. *** "cannot reclaim" error. This means that an expression is too complicated. You get the correct line number in the error message. The code must be rewritten with simpler expressions. *** XCONS, XSTRING, etc macros produce incorrect code. If temacs fails to run at all, this may be the cause. Compile this test program and look at the assembler code: struct foo { char x; unsigned int y : 24; }; lose (arg) struct foo arg; { test ((int *) arg.y); } If the code is incorrect, your compiler has this problem. In the XCONS, etc., macros in lisp.h you must replace (a).u.val with ((a).u.val + coercedummy) where coercedummy is declared as int. This problem will not happen if the m-...h file for your type of machine defines NO_UNION_TYPE. That is the recommended setting now. *** C compilers lose on returning unions. I hear that some C compilers cannot handle returning a union type. Most of the functions in GNU Emacs return type Lisp_Object, which is defined as a union on some rare architectures. This problem will not happen if the m-...h file for your type of machine defines NO_UNION_TYPE. Copyright 1987,88,89,93,94,95,96,97,98,1999,2001,2002,2004 Free Software Foundation, Inc. Copying and redistribution of this file with or without modification are permitted without royalty provided this notice is preserved. Local variables: mode: outline paragraph-separate: "[ ]*$" end: arch-tag: 49fc0d95-88cb-4715-b21c-f27fb5a4764a