Mercurial > hg > xemacs-beta

comparison PROBLEMS @ 312:d1b52dcaa789 r21-0b54

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Import from CVS: tag r21-0b54
author cvs
date Mon, 13 Aug 2007 10:43:55 +0200
parents 33bdb3d4b97f
children 512e409c26a2
comparison
equal deleted inserted replaced
311:b4ad76366919 312:d1b52dcaa789
27 27
28 28
29 * Problems with building XEmacs 29 * Problems with building XEmacs
30 =============================== 30 ===============================
31 31
32 ** Don't use -O2 with gcc 2.7.2 under Intel/XXX without also using 32 ** General
33 *** Don't use -O2 with gcc 2.7.2 under Intel/XXX without also using
33 `-fno-strength-reduce'. 34 `-fno-strength-reduce'.
34 35
35 gcc will generate incorrect code otherwise. This bug is present in at 36 gcc will generate incorrect code otherwise. This bug is present in at
36 least 2.6.x and 2.7.[0-2]. This bug has been fixed in GCC 2.7.2.1 and 37 least 2.6.x and 2.7.[0-2]. This bug has been fixed in GCC 2.7.2.1 and
37 later. This bug is O/S independent, but is limited to x86 architectures. 38 later. This bug is O/S independent, but is limited to x86 architectures.
38 39
39 This problem is known to be fixed in egcs (or pgcc) 1.0 or later. 40 This problem is known to be fixed in egcs (or pgcc) 1.0 or later.
40 41
41 ** Don't use -O2 with gcc 2.7.2 under Intel architectures without also 42 *** Don't use -O2 with gcc 2.7.2 under Intel architectures without also
42 using `-fno-caller-saves'. 43 using `-fno-caller-saves'.
43 44
44 gcc will generate incorrect code otherwise. This bug is still 45 gcc will generate incorrect code otherwise. This bug is still
45 present in gcc 2.7.2.3. There have been no reports to indicate the 46 present in gcc 2.7.2.3. There have been no reports to indicate the
46 bug is present in egcs 1.0 (or pgcc 1.0) or later. This bug is O/S 47 bug is present in egcs 1.0 (or pgcc 1.0) or later. This bug is O/S
47 independent, but limited to x86 architectures. 48 independent, but limited to x86 architectures.
48 49
49 This problem is known to be fixed in egcs (or pgcc) 1.0 or later. 50 This problem is known to be fixed in egcs (or pgcc) 1.0 or later.
50 51
51 ** Excessive optimization with pgcc can break XEmacs 52 *** When using gcc, you get the error message "undefined symbol __fixunsdfsi".
53 When using gcc, you get the error message "undefined symbol __main".
54
55 This means that you need to link with the gcc library. It may be called
56 "gcc-gnulib" or "libgcc.a"; figure out where it is, and define LIB_GCC in
57 config.h to point to it.
58
59 It may also work to use the GCC version of `ld' instead of the standard one.
60
61 *** Excessive optimization with pgcc can break XEmacs
52 62
53 It has been reported on some systems that compiling with -O6 can lead 63 It has been reported on some systems that compiling with -O6 can lead
54 to XEmacs failures. The workaround is to use a lower optimization 64 to XEmacs failures. The workaround is to use a lower optimization
55 level. -O2 and -O4 have been tested extensively. 65 level. -O2 and -O4 have been tested extensively.
56 66
57 All of this depends heavily on the version of pgcc and the version 67 All of this depends heavily on the version of pgcc and the version
58 of libc. Snapshots near the release of pgcc-1.0 have been tested 68 of libc. Snapshots near the release of pgcc-1.0 have been tested
59 extensively and no sign of breakage has been seen on systems using 69 extensively and no sign of breakage has been seen on systems using
60 glibc-2. 70 glibc-2.
61 71
62 ** `compress' and `uncompress' not found and XFree86 72 *** src/Makefile and lib-src/Makefile are truncated--most of the file missing.
73
74 This can happen if configure uses GNU sed version 2.03. That version
75 had a bug. GNU sed version 2.05 works properly.
76
77 *** When compiling with X11, you get "undefined symbol _XtStrings".
78
79 This means that you are trying to link emacs against the X11r4 version of
80 libXt.a, but you have compiled either Emacs or the code in the lwlib
81 subdirectory with the X11r5 header files. That doesn't work.
82
83 Remember, you can't compile lwlib for r4 and emacs for r5, or vice versa.
84 They must be in sync.
85
86 *** test-distrib says that the distribution has been clobbered
87 or, temacs prints "Command key out of range 0-127"
88 or, temacs runs and dumps xemacs, but xemacs totally fails to work.
89 or, temacs gets errors dumping xemacs
90
91 This can be because the .elc files have been garbled. Do not be
92 fooled by the fact that most of a .elc file is text: these are binary
93 files and can contain all 256 byte values.
94
95 In particular `shar' cannot be used for transmitting GNU Emacs. It
96 typically truncates "lines". (this does not apply to GNU shar, which
97 uses uuencode to encode binary files.)
98
99 If you have a copy of Emacs that has been damaged in its nonprinting
100 characters, you can fix them by running:
101
102 make all-elc
103
104 This will rebuild all the needed .elc files.
105
106 *** `compress' and `uncompress' not found and XFree86
63 107
64 XFree86 installs a very old version of libz.a by default ahead of where 108 XFree86 installs a very old version of libz.a by default ahead of where
65 more modern version of libz might be installed. This will cause problems 109 more modern version of libz might be installed. This will cause problems
66 when attempting to link against libMagick. The fix is to remove the old 110 when attempting to link against libMagick. The fix is to remove the old
67 libz.a in the X11 binary directory. 111 libz.a in the X11 binary directory.
68 112
69 ** Excessive optimization on AIX 4.2 can lead to compiler failure. 113
114 ** AIX
115 *** On AIX, you get this compiler error message:
116
117 Processing include file ./XMenuInt.h
118 1501-106: (S) Include file X11/Xlib.h not found.
119
120 This means your system was installed with only the X11 runtime i.d
121 libraries. You have to find your sipo (bootable tape) and install
122 X11Dev... with smit.
123
124 *** On AIX 4.1.2, linker error messages such as
125 ld: 0711-212 SEVERE ERROR: Symbol .__quous, found in the global symbol table
126 of archive /usr/lib/libIM.a, was not defined in archive member shr.o.
127
128 This is a problem in libIM.a. You can work around it by executing
129 these shell commands in the src subdirectory of the directory where
130 you build Emacs:
131
132 cp /usr/lib/libIM.a .
133 chmod 664 libIM.a
134 ranlib libIM.a
135
136 Then change -lIM to ./libIM.a in the command to link temacs (in
137 Makefile).
138
139 *** Link failure on IBM AIX 1.3 ptf 0013.
140
141 There is a real duplicate definition of the function `_slibc_free' in
142 the library /lib/libc_s.a (just do nm on it to verify). The
143 workaround/fix is:
144
145 cd /lib
146 ar xv libc_s.a NLtmtime.o
147 ar dv libc_s.a NLtmtime.o
148
149 *** Excessive optimization on AIX 4.2 can lead to compiler failure.
70 150
71 Valdis.Kletnieks@vt.edu writes: 151 Valdis.Kletnieks@vt.edu writes:
72 At least at the b34 level, and the latest-and-greatest IBM xlc 152 At least at the b34 level, and the latest-and-greatest IBM xlc
73 (3.1.4.4), there are problems with -O3. I haven't investigated 153 (3.1.4.4), there are problems with -O3. I haven't investigated
74 further. 154 further.
75 155
76 ** Sed problems on Solaris 2.5 156
157 ** SunOS/Solaris
158 *** Link failure when using acc on a Sun.
159
160 To use acc, you need additional options just before the libraries, such as
161
162 /usr/lang/SC2.0.1/values-Xt.o -L/usr/lang/SC2.0.1/cg87 -L/usr/lang/SC2.0.1
163
164 and you need to add -lansi just before -lc.
165
166 The precise file names depend on the compiler version, so we
167 cannot easily arrange to supply them.
168
169 *** Problems finding X11 libraries on Solaris with Openwindows
170
171 Some users have reported problems in this area. The reported solution
172 is to define the environment variable OPENWINHOME, even if you must set
173 it to `/usr/openwin'.
174
175 *** Sed problems on Solaris 2.5
77 176
78 There have been reports of Sun sed truncating very lines in the 177 There have been reports of Sun sed truncating very lines in the
79 Makefile during configuration. The workaround is to use GNU sed or, 178 Makefile during configuration. The workaround is to use GNU sed or,
80 even better, think of a better way to generate Makefile, and send us a 179 even better, think of a better way to generate Makefile, and send us a
81 patch. :-) 180 patch. :-)
82 181
83 ** test-distrib says that the distribution has been clobbered 182 *** On Solaris 2 I get undefined symbols from libcurses.a.
84 or, temacs prints "Command key out of range 0-127" 183
85 or, temacs runs and dumps xemacs, but xemacs totally fails to work. 184 You probably have /usr/ucblib/ on your LD_LIBRARY_PATH. Do the link with
86 or, temacs gets errors dumping xemacs 185 LD_LIBRARY_PATH unset. Generally, avoid using any ucb* stuff when
87 186 building XEmacs.
88 This can be because the .elc files have been garbled. Do not be 187
89 fooled by the fact that most of a .elc file is text: these are binary 188 *** On Solaris 2 I cannot make alloc.o, glyphs.o or process.o.
90 files and can contain all 256 byte values. 189
91 190 The SparcWorks C compiler may have difficulty building those modules
92 In particular `shar' cannot be used for transmitting GNU Emacs. It 191 with optimization level -xO4. Try using only "-fast" optimization
93 typically truncates "lines". (this does not apply to GNU shar, which 192 for just those modules. (Or use gcc).
94 uses uuencode to encode binary files.) 193
95 194 *** Solaris 2.3 /bin/sh coredumps during configuration.
96 If you have a copy of Emacs that has been damaged in its nonprinting 195
97 characters, you can fix them by running: 196 This only occurs if you have LANG != C. This is a known bug with
98 197 /bin/sh fixed by installing Patch-ID# 101613-01. Or, you can use
99 make all-elc 198 bash, as a workaround.
100 199
101 This will rebuild all the needed .elc files. 200 *** On SunOS, you get linker errors
102 201 ld: Undefined symbol
103 ** `Error: No ExtNode to pop!' on Linux systems with Lesstif. 202 _get_wmShellWidgetClass
203 _get_applicationShellWidgetClass
204
205 The fix to this is to install patch 100573 for OpenWindows 3.0
206 or link libXmu statically.
207
208 *** On Sunos 4, you get the error ld: Undefined symbol __lib_version.
209
210 This is the result of using cc or gcc with the shared library meant
211 for acc (the Sunpro compiler). Check your LD_LIBRARY_PATH and delete
212 /usr/lang/SC2.0.1 or some similar directory.
213
214 *** Undefined symbols when linking on Sunos 4.1.
215
216 If you get the undefined symbols _atowc _wcslen, _iswprint, _iswspace,
217 _iswcntrl, _wcscpy, and _wcsncpy, then you need to add -lXwchar after
218 -lXaw in the command that links temacs.
219
220 This problem seems to arise only when the international language
221 extensions to X11R5 are installed.
222
223 *** On a Sun running SunOS 4.1.1, you get this error message from GNU ld:
224
225 /lib/libc.a(_Q_sub.o): Undefined symbol __Q_get_rp_rd referenced from text segment
226
227 The problem is in the Sun shared C library, not in GNU ld.
228
229 The solution is to install Patch-ID# 100267-03 from Sun.
230
231 *** SunOS 4.1.2: undefined symbol _get_wmShellWidgetClass
232
233 Apparently the version of libXmu.so.a that Sun ships is hosed: it's missing
234 some stuff that is in libXmu.a (the static version). Sun has a patch for
235 this, but a workaround is to use the static version of libXmu, by changing
236 the link command from "-lXmu" to "-Bstatic -lXmu -Bdynamic". If you have
237 OpenWindows 3.0, ask Sun for these patches:
238 100512-02 4.1.x OpenWindows 3.0 libXt Jumbo patch
239 100573-03 4.1.x OpenWindows 3.0 undefined symbols with shared libXmu
240
241 *** Random other SunOS 4.1.[12] link errors.
242
243 The X headers and libraries that Sun ships in /usr/{include,lib}/X11 are
244 broken. Use the ones in /usr/openwin/{include,lib} instead.
245
246 ** Linux
247 *** Under Linux, you get "too many arguments to function `getpgrp'".
248
249 You have probably installed LessTiff under `/usr/local' and `libXm.so'
250 could not be found when linking `getpgrp()' test program, making XEmacs
251 think that `getpgrp()' takes an argument. Try adding `/usr/local/lib'
252 in `/etc/ld.so.conf' and run `ldconfig'. Then run XEmacs's `configure'
253 again. As with all problems of this type, reading the config.log file
254 generated from configure and seeing the log of how the test failed can
255 prove enlightening.
256
257 *** `Error: No ExtNode to pop!' on Linux systems with Lesstif.
104 258
105 This error message has been observed with lesstif-0.75a. It does not 259 This error message has been observed with lesstif-0.75a. It does not
106 appear to cause any harm. 260 appear to cause any harm.
107 261
108 ** Linking with -rpath on IRIX. 262 *** xemacs: can't resolve symbol '__malloc_hook'
263
264 This is a Linux problem where you've compiled the XEmacs binary on a libc
265 5.4 with version higher than 5.4.19 and attempted to run the binary against
266 an earlier version. The solution is to upgrade your old library.
267
268 ** IRIX
269 *** Linking with -rpath on IRIX.
109 270
110 Darrell Kindred <dkindred@cmu.edu> writes: 271 Darrell Kindred <dkindred@cmu.edu> writes:
111 There are a couple of problems [with use of -rpath with Irix ld], though: 272 There are a couple of problems [with use of -rpath with Irix ld], though:
112 273
113 1. The ld in IRIX 5.3 ignores all but the last -rpath 274 1. The ld in IRIX 5.3 ignores all but the last -rpath
124 rather than quoting $LDFLAGS with prefix-args, like 285 rather than quoting $LDFLAGS with prefix-args, like
125 src/Makefile does. So if you specify --x-libraries 286 src/Makefile does. So if you specify --x-libraries
126 or --site-runtime-libraries, you must use --use-gcc=no, 287 or --site-runtime-libraries, you must use --use-gcc=no,
127 or configure will fail. 288 or configure will fail.
128 289
129 ** On Irix 6.3, the SGI ld quits with segmentation fault when linking temacs 290 *** On Irix 6.3, the SGI ld quits with segmentation fault when linking temacs
130 291
131 This occurs if you use the SGI linker version 7.1. Installing the 292 This occurs if you use the SGI linker version 7.1. Installing the
132 patch SG0001872 fixes this problem. 293 patch SG0001872 fixes this problem.
133 294
134 ** xemacs: can't resolve symbol '__malloc_hook' 295 *** On Irix 6.0, make tries (and fails) to build a program named unexelfsgi
135 296
136 This is a Linux problem where you've compiled the XEmacs binary on a libc 297 A compiler bug inserts spaces into the string "unexelfsgi . o"
137 5.4 with version higher than 5.4.19 and attempted to run the binary against 298 in src/Makefile. Edit src/Makefile, after configure is run,
138 an earlier version. The solution is to upgrade your old library. 299 find that string, and take out the spaces.
139 300
140 ** Compilation errors on VMS. 301 Compiler fixes in Irix 6.0.1 should eliminate this problem.
141 302
142 Sorry, XEmacs does not work under VMS. You might consider working on 303 *** On Irix 5.2, unexelfsgi.c can't find cmplrs/stsupport.h.
143 the port if you really want to have XEmacs work under VMS. 304
144 305 The file cmplrs/stsupport.h was included in the wrong file set in the
145 ** On Solaris 2 I get undefined symbols from libcurses.a. 306 Irix 5.2 distribution. You can find it in the optional fileset
146 307 compiler_dev, or copy it from some other Irix 5.2 system. A kludgy
147 You probably have /usr/ucblib/ on your LD_LIBRARY_PATH. Do the link with 308 workaround is to change unexelfsgi.c to include sym.h instead of
148 LD_LIBRARY_PATH unset. Generally, avoid using any ucb* stuff when 309 syms.h.
149 building XEmacs. 310
150 311 *** Coredumping in Irix 6.2
151 ** On Solaris 2 I cannot make alloc.o, glyphs.o or process.o. 312
152 313 Pete Forman <gsez020@compo.bedford.waii.com> writes:
153 The SparcWorks C compiler may have difficulty building those modules 314 A problem noted by myself and others (I've lost the references) was
154 with optimization level -xO4. Try using only "-fast" optimization 315 that XEmacs coredumped when the cut or copy toolbar buttons were
155 for just those modules. (Or use gcc). 316 pressed. This has been fixed by loading the SGI patchset (Feb 98)
156 317 without having to recompile XEmacs.
157 ** On Digital UNIX, the DEC C compiler might have a problem compiling 318
319 My versions are XEmacs 20.3 (problem first noted in 19.15) and IRIX
320 6.2, compiled using -n32. I'd guess that the relevant individual
321 patch was "SG0002580: multiple fixes for X libraries". SGI recommends
322 that the complete patch set be installed rather than parts of it.
323
324 ** Digital UNIX/OSF/VMS
325 *** On Digital UNIX, the DEC C compiler might have a problem compiling
158 some files. 326 some files.
159 327
160 In particular, src/extents.c and src/faces.c might cause the DEC C 328 In particular, src/extents.c and src/faces.c might cause the DEC C
161 compiler to abort. When this happens: cd src, compile the files by 329 compiler to abort. When this happens: cd src, compile the files by
162 hand, cd .., and redo the "make" command. When recompiling the files by 330 hand, cd .., and redo the "make" command. When recompiling the files by
165 - V4.n: Add "-oldc" to the compile command. 333 - V4.n: Add "-oldc" to the compile command.
166 334
167 A related compiler bug has been fixed by the DEC compiler team. The 335 A related compiler bug has been fixed by the DEC compiler team. The
168 new versions of the compiler should run fine. 336 new versions of the compiler should run fine.
169 337
170 ** On HPUX, the HP C compiler might have a problem compiling some files 338 *** Under some versions of OSF XEmacs runs fine if built without
339 optimization but will crash randomly if built with optimization.
340
341 Using 'cc -g' is not sufficient to eliminate all optimization. Try
342 'cc -g -O0' instead.
343
344 *** Compilation errors on VMS.
345
346 Sorry, XEmacs does not work under VMS. You might consider working on
347 the port if you really want to have XEmacs work under VMS.
348
349 ** HP-UX
350 *** On HPUX, the HP C compiler might have a problem compiling some files
171 with optimization. 351 with optimization.
172 352
173 Richard Cognot <cognot@ensg.u-nancy.fr> writes: 353 Richard Cognot <cognot@ensg.u-nancy.fr> writes:
174 354
175 Had to drop once again to level 2 optimization, at least to 355 Had to drop once again to level 2 optimization, at least to
179 applying patches for the C compiler). Trouble is I still 359 applying patches for the C compiler). Trouble is I still
180 haven't found the same patch for hpux 10.10, and I don't 360 haven't found the same patch for hpux 10.10, and I don't
181 remember the patch numbers. I think potential XEmacs builders 361 remember the patch numbers. I think potential XEmacs builders
182 on HP should be warned about this. 362 on HP should be warned about this.
183 363
184 ** I don't have `xmkmf' and `imake' on my HP. 364 *** I don't have `xmkmf' and `imake' on my HP.
185 365
186 You can get these standard X tools by anonymous FTP to 366 You can get these standard X tools by anonymous FTP to
187 hpcvaaz.cv.hp.com. Essentially all X programs need these. 367 hpcvaaz.cv.hp.com. Essentially all X programs need these.
188 368
189 ** On HP-UX, problems with make 369 *** On HP-UX, problems with make
190 370
191 Marcus Thiessel <marcus_thiessel@hp.com> 371 Marcus Thiessel <marcus_thiessel@hp.com>
192 372
193 Some releases of XEmacs (e.g. 20.4) require GNU make to build 373 Some releases of XEmacs (e.g. 20.4) require GNU make to build
194 successfully. You don't need GNU make when building 21.x. 374 successfully. You don't need GNU make when building 21.x.
195 375
196 ** On HP-UX 9.05 XEmacs won't compile or coredump during the build. 376 *** On HP-UX 9.05 XEmacs won't compile or coredump during the build.
197 377
198 Marcus Thiessel <marcus_thiessel@hp.com> 378 Marcus Thiessel <marcus_thiessel@hp.com>
199 379
200 This might be a sed problem. For your own safety make sure to use 380 This might be a sed problem. For your own safety make sure to use
201 GNU sed while dumping XEmacs. 381 GNU sed while dumping XEmacs.
202 382
203 ** On HP-UX 11.0 XEmacs causes excessive X11 errors when running. 383 *** On HP-UX 11.0 XEmacs causes excessive X11 errors when running.
204 384
205 Marcus Thiessel <marcus_thiessel@hp.com> 385 Marcus Thiessel <marcus_thiessel@hp.com>
206 386
207 Unfortunately, XEmacs releases don't work with Motif2.1. It 387 Unfortunately, XEmacs releases don't work with Motif2.1. It
208 will compile but you will get excessive X11 errors like 388 will compile but you will get excessive X11 errors like
216 --x-libraries="/usr/lib/Motif1.2_R6 -L/usr/lib/X11R6" 396 --x-libraries="/usr/lib/Motif1.2_R6 -L/usr/lib/X11R6"
217 397
218 Make sure /usr/lib/Motif1.2_R6/libXm.sl is a link to 398 Make sure /usr/lib/Motif1.2_R6/libXm.sl is a link to
219 /usr/lib/Motif1.2_R6/libXm.3. 399 /usr/lib/Motif1.2_R6/libXm.3.
220 400
221 ** Solaris 2.3 /bin/sh coredumps during configuration. 401 ** SCO OpenServer
222 402 *** Native cc on SCO OpenServer 5 is now OK. Icc may still throw you
223 This only occurs if you have LANG != C. This is a known bug with
224 /bin/sh fixed by installing Patch-ID# 101613-01. Or, you can use
225 bash, as a workaround.
226
227 ** On Irix 6.0, make tries (and fails) to build a program named unexelfsgi
228
229 A compiler bug inserts spaces into the string "unexelfsgi . o"
230 in src/Makefile. Edit src/Makefile, after configure is run,
231 find that string, and take out the spaces.
232
233 Compiler fixes in Irix 6.0.1 should eliminate this problem.
234
235 ** Coredumping in Irix 6.2
236
237 Pete Forman <gsez020@compo.bedford.waii.com> writes:
238 A problem noted by myself and others (I've lost the references) was
239 that XEmacs coredumped when the cut or copy toolbar buttons were
240 pressed. This has been fixed by loading the SGI patchset (Feb 98)
241 without having to recompile XEmacs.
242
243 My versions are XEmacs 20.3 (problem first noted in 19.15) and IRIX
244 6.2, compiled using -n32. I'd guess that the relevant individual
245 patch was "SG0002580: multiple fixes for X libraries". SGI recommends
246 that the complete patch set be installed rather than parts of it.
247
248 ** Native cc on SCO OpenServer 5 is now OK. Icc may still throw you
249 a curve. Here is what Robert Lipe <robertl@arnet.com> says: 403 a curve. Here is what Robert Lipe <robertl@arnet.com> says:
250 404
251 Unlike XEmacs 19.13, building with the native cc on SCO OpenServer 5 405 Unlike XEmacs 19.13, building with the native cc on SCO OpenServer 5
252 now produces a functional binary. I will typically build this 406 now produces a functional binary. I will typically build this
253 configuration for COFF with: 407 configuration for COFF with:
310 to take advantage of the keyboard map in emacskeys.sco. 464 to take advantage of the keyboard map in emacskeys.sco.
311 465
312 Note: Much of the above entry is probably not valid for XEmacs 21.0 466 Note: Much of the above entry is probably not valid for XEmacs 21.0
313 and later. 467 and later.
314 468
315 ** Under some versions of OSF XEmacs runs fine if built without 469
316 optimization but will crash randomly if built with optimization. 470
317 471 * Problems with running XEmacs
318 Using 'cc -g' is not sufficient to eliminate all optimization. Try 472 ==============================
319 'cc -g -O0' instead. 473 ** General
320 474 *** C-z just refreshes the screen instead of suspending Emacs.
321 ** On SunOS, you get linker errors
322 ld: Undefined symbol
323 _get_wmShellWidgetClass
324 _get_applicationShellWidgetClass
325
326 The fix to this is to install patch 100573 for OpenWindows 3.0
327 or link libXmu statically.
328
329 ** On Sunos 4, you get the error ld: Undefined symbol __lib_version.
330
331 This is the result of using cc or gcc with the shared library meant
332 for acc (the Sunpro compiler). Check your LD_LIBRARY_PATH and delete
333 /usr/lang/SC2.0.1 or some similar directory.
334
335 ** On AIX 4.1.2, linker error messages such as
336 ld: 0711-212 SEVERE ERROR: Symbol .__quous, found in the global symbol table
337 of archive /usr/lib/libIM.a, was not defined in archive member shr.o.
338
339 This is a problem in libIM.a. You can work around it by executing
340 these shell commands in the src subdirectory of the directory where
341 you build Emacs:
342
343 cp /usr/lib/libIM.a .
344 chmod 664 libIM.a
345 ranlib libIM.a
346
347 Then change -lIM to ./libIM.a in the command to link temacs (in
348 Makefile).
349
350 ** On Irix 5.2, unexelfsgi.c can't find cmplrs/stsupport.h.
351
352 The file cmplrs/stsupport.h was included in the wrong file set in the
353 Irix 5.2 distribution. You can find it in the optional fileset
354 compiler_dev, or copy it from some other Irix 5.2 system. A kludgy
355 workaround is to change unexelfsgi.c to include sym.h instead of
356 syms.h.
357
358 ** Link failure when using acc on a Sun.
359
360 To use acc, you need additional options just before the libraries, such as
361
362 /usr/lang/SC2.0.1/values-Xt.o -L/usr/lang/SC2.0.1/cg87 -L/usr/lang/SC2.0.1
363
364 and you need to add -lansi just before -lc.
365
366 The precise file names depend on the compiler version, so we
367 cannot easily arrange to supply them.
368
369 ** Link failure on IBM AIX 1.3 ptf 0013.
370
371 There is a real duplicate definition of the function `_slibc_free' in
372 the library /lib/libc_s.a (just do nm on it to verify). The
373 workaround/fix is:
374
375 cd /lib
376 ar xv libc_s.a NLtmtime.o
377 ar dv libc_s.a NLtmtime.o
378
379 ** Undefined symbols when linking on Sunos 4.1.
380
381 If you get the undefined symbols _atowc _wcslen, _iswprint, _iswspace,
382 _iswcntrl, _wcscpy, and _wcsncpy, then you need to add -lXwchar after
383 -lXaw in the command that links temacs.
384
385 This problem seems to arise only when the international language
386 extensions to X11R5 are installed.
387
388 ** src/Makefile and lib-src/Makefile are truncated--most of the file missing.
389
390 This can happen if configure uses GNU sed version 2.03. That version
391 had a bug. GNU sed version 2.05 works properly.
392
393 ** On AIX, you get this compiler error message:
394
395 Processing include file ./XMenuInt.h
396 1501-106: (S) Include file X11/Xlib.h not found.
397
398 This means your system was installed with only the X11 runtime i.d
399 libraries. You have to find your sipo (bootable tape) and install
400 X11Dev... with smit.
401
402 ** C-z just refreshes the screen instead of suspending Emacs.
403 475
404 You are probably using a shell that doesn't support job control, even 476 You are probably using a shell that doesn't support job control, even
405 though the system itself is capable of it. Try using a different 477 though the system itself is capable of it. Try using a different
406 shell. 478 shell.
407 479
408 ** On a Sun running SunOS 4.1.1, you get this error message from GNU ld: 480 *** You type Control-H (Backspace) expecting to delete characters.
409
410 /lib/libc.a(_Q_sub.o): Undefined symbol __Q_get_rp_rd referenced from text segment
411
412 The problem is in the Sun shared C library, not in GNU ld.
413
414 The solution is to install Patch-ID# 100267-03 from Sun.
415
416 ** SunOS 4.1.2: undefined symbol _get_wmShellWidgetClass
417
418 Apparently the version of libXmu.so.a that Sun ships is hosed: it's missing
419 some stuff that is in libXmu.a (the static version). Sun has a patch for
420 this, but a workaround is to use the static version of libXmu, by changing
421 the link command from "-lXmu" to "-Bstatic -lXmu -Bdynamic". If you have
422 OpenWindows 3.0, ask Sun for these patches:
423 100512-02 4.1.x OpenWindows 3.0 libXt Jumbo patch
424 100573-03 4.1.x OpenWindows 3.0 undefined symbols with shared libXmu
425
426 ** Random other SunOS 4.1.[12] link errors.
427
428 The X headers and libraries that Sun ships in /usr/{include,lib}/X11 are
429 broken. Use the ones in /usr/openwin/{include,lib} instead.
430
431 ** When using gcc, you get the error message "undefined symbol __fixunsdfsi".
432 When using gcc, you get the error message "undefined symbol __main".
433
434 This means that you need to link with the gcc library. It may be called
435 "gcc-gnulib" or "libgcc.a"; figure out where it is, and define LIB_GCC in
436 config.h to point to it.
437
438 It may also work to use the GCC version of `ld' instead of the standard one.
439
440 ** When compiling with X11, you get "undefined symbol _XtStrings".
441
442 This means that you are trying to link emacs against the X11r4 version of
443 libXt.a, but you have compiled either Emacs or the code in the lwlib
444 subdirectory with the X11r5 header files. That doesn't work.
445
446 Remember, you can't compile lwlib for r4 and emacs for r5, or vice versa.
447 They must be in sync.
448
449 ** Problems finding X11 libraries on Solaris with Openwindows
450
451 Some users have reported problems in this area. The reported solution
452 is to define the environment variable OPENWINHOME, even if you must set
453 it to `/usr/openwin'.
454
455 ** Under Linux, you get "too many arguments to function `getpgrp'".
456
457 You have probably installed LessTiff under `/usr/local' and `libXm.so'
458 could not be found when linking `getpgrp()' test program, making XEmacs
459 think that `getpgrp()' takes an argument. Try adding `/usr/local/lib'
460 in `/etc/ld.so.conf' and run `ldconfig'. Then run XEmacs's `configure'
461 again. As with all problems of this type, reading the config.log file
462 generated from configure and seeing the log of how the test failed can
463 prove enlightening.
464
465
466 * Problems with running XEmacs
467 ==============================
468 ** On Solaris 2.6, XEmacs dumps core when exiting.
469
470 This happens if you're XEmacs is running on the same machine as the X
471 server, and the optimized memory transport has been turned on by
472 setting the environment variable XSUNTRANSPORT. The crash occurs
473 during the call to XCloseDisplay.
474
475 If this describes your situation, you need to undefine the
476 XSUNTRANSPORT environment variable.
477
478 ** `C-z', or `M-x suspend-emacs' hangs instead of suspending.
479
480 If you build with `gpm' support on Linux, you cannot suspend XEmacs
481 because gpm installs a buggy SIGTSTP handler. Either compile with
482 `--with-gpm=no', or don't suspend XEmacs on the Linux console until
483 this bug is fixed.
484
485 ** You type Control-H (Backspace) expecting to delete characters.
486 481
487 Emacs has traditionally used Control-H for help; unfortunately this 482 Emacs has traditionally used Control-H for help; unfortunately this
488 interferes with its use as Backspace on TTY's. One way to solve this 483 interferes with its use as Backspace on TTY's. One way to solve this
489 problem is to put this in your .emacs: 484 problem is to put this in your .emacs:
490 485
495 This checks whether the TTY erase char is C-h, and if it is, makes 490 This checks whether the TTY erase char is C-h, and if it is, makes
496 Control-H (Backspace) work sensibly, and moves help to Meta-? (ESC ?). 491 Control-H (Backspace) work sensibly, and moves help to Meta-? (ESC ?).
497 492
498 Note that you can probably also access help using F1. 493 Note that you can probably also access help using F1.
499 494
500 ** Mail agents (VM, Gnus, rmail) cannot get new mail 495 *** Mail agents (VM, Gnus, rmail) cannot get new mail
501 496
502 rmail and VM get new mail from /usr/spool/mail/$USER using a program 497 rmail and VM get new mail from /usr/spool/mail/$USER using a program
503 called `movemail'. This program interlocks with /bin/mail using the 498 called `movemail'. This program interlocks with /bin/mail using the
504 protocol defined by /bin/mail. 499 protocol defined by /bin/mail.
505 500
525 installed copy of movemail is usually in the directory 520 installed copy of movemail is usually in the directory
526 /usr/local/lib/emacs/VERSION/TARGET. You must change the group and 521 /usr/local/lib/emacs/VERSION/TARGET. You must change the group and
527 mode of the installed copy; changing the group and mode of the build 522 mode of the installed copy; changing the group and mode of the build
528 directory copy is ineffective. 523 directory copy is ineffective.
529 524
530 ** XEmacs crashes on Digital Unix within font-lock, or when dealing 525 *** VM appears to hang in large folders.
531 with large compilation buffers.
532
533 The default stack size under Digital Unix is rather small (2M as
534 opposed to Solaris 8M), hosing the regexp code, which uses alloca()
535 extensively, overflowing the stack when complex regexps are used.
536 Workarounds:
537
538 1) Increase your stack size, using `ulimit -s 8192' or a (t)csh
539 equivalent;
540
541 2) Recompile regex.c with REGEX_MALLOC defined.
542
543 ** On Solaris, C-x doesn't get through to Emacs when you use the console.
544
545 This is a Solaris feature (at least on Intel x86 cpus). Type C-r
546 C-r C-t, to toggle whether C-x gets through to Emacs.
547
548 ** VM appears to hang in large folders.
549 526
550 This is normal (trust us) when upgrading to VM-6.22 from earlier 527 This is normal (trust us) when upgrading to VM-6.22 from earlier
551 versions. Let VM finish what it is doing and all will be well. 528 versions. Let VM finish what it is doing and all will be well.
552 529
553 ** Changes made to .el files do not take effect. 530 *** Changes made to .el files do not take effect.
554 531
555 You may have forgotten to recompile them into .elc files. Then the 532 You may have forgotten to recompile them into .elc files. Then the
556 old .elc files will be loaded, and your changes will not be seen. To 533 old .elc files will be loaded, and your changes will not be seen. To
557 fix this, do `M-x byte-recompile-directory' and specify the directory 534 fix this, do `M-x byte-recompile-directory' and specify the directory
558 that contains the Lisp files. 535 that contains the Lisp files.
559 536
560 Note that you will get a warning when loading a .elc file that is 537 Note that you will get a warning when loading a .elc file that is
561 older than the corresponding .el file. 538 older than the corresponding .el file.
562 539
563 ** Things which should be bold or italic (such as the initial 540 *** Things which should be bold or italic (such as the initial
564 copyright notice) are not. 541 copyright notice) are not.
565 542
566 The fonts of the "bold" and "italic" faces are generated from the font 543 The fonts of the "bold" and "italic" faces are generated from the font
567 of the "default" face; in this way, your bold and italic fonts will 544 of the "default" face; in this way, your bold and italic fonts will
568 have the appropriate size and family. However, emacs can only be 545 have the appropriate size and family. However, emacs can only be
581 then emacs won't be able to guess the names of the "bold" and "italic" 558 then emacs won't be able to guess the names of the "bold" and "italic"
582 versions. All X fonts can be referred to via XLFD-style names, so you 559 versions. All X fonts can be referred to via XLFD-style names, so you
583 should use those forms. See the man pages for X(1), xlsfonts(1), and 560 should use those forms. See the man pages for X(1), xlsfonts(1), and
584 xfontsel(1). 561 xfontsel(1).
585 562
586 ** The dumped Emacs crashes when run, trying to write pure data. 563 *** The dumped Emacs crashes when run, trying to write pure data.
587 564
588 Two causes have been seen for such problems. 565 Two causes have been seen for such problems.
589 566
590 1) On a system where getpagesize is not a system call, it is defined 567 1) On a system where getpagesize is not a system call, it is defined
591 as a macro. If the definition (in both unexec.c and malloc.c) is wrong, 568 as a macro. If the definition (in both unexec.c and malloc.c) is wrong,
596 initialized variables. Emacs makes all initialized variables in most 573 initialized variables. Emacs makes all initialized variables in most
597 of its files pure after dumping, but the variables declared static and 574 of its files pure after dumping, but the variables declared static and
598 not initialized are not supposed to be pure. On these systems you 575 not initialized are not supposed to be pure. On these systems you
599 may need to add "#define static" to the m- or the s- file. 576 may need to add "#define static" to the m- or the s- file.
600 577
601 ** Reading and writing files is very very slow. 578 *** Reading and writing files is very very slow.
602 579
603 Try evaluating the form (setq lock-directory nil) and see if that helps. 580 Try evaluating the form (setq lock-directory nil) and see if that helps.
604 There is a problem with file-locking on some systems (possibly related 581 There is a problem with file-locking on some systems (possibly related
605 to NFS) that I don't understand. Please send mail to the address 582 to NFS) that I don't understand. Please send mail to the address
606 xemacs@xemacs.org if you figure this one out. 583 xemacs@xemacs.org if you figure this one out.
607 584
608 ** The Emacs window disappears when you type M-q. 585 *** When emacs starts up, I get lots of warnings about unknown keysyms.
609
610 Some versions of the Open Look window manager interpret M-q as a quit
611 command for whatever window you are typing at. If you want to use
612 Emacs with that window manager, you should try to configure the window
613 manager to use some other command. You can disable the
614 shortcut keys entirely by adding this line to ~/.OWdefaults:
615
616 OpenWindows.WindowMenuAccelerators: False
617
618 ** The `Alt' key doesn't behave as `Meta' when running DECwindows.
619
620 The default DEC keyboard mapping has the Alt keys set up to generate the
621 keysym `Multi_key', which has a meaning to xemacs which is distinct from that
622 of the `Meta_L' and `Meta-R' keysyms. A second problem is that certain keys
623 have the Mod2 modifier attached to them for no adequately explored reason.
624 The correct fix is to pass this file to xmodmap upon starting X:
625
626 clear mod2
627 keysym Multi_key = Alt_L
628 add mod1 = Alt_L
629 add mod1 = Alt_R
630
631 ** The Compose key on a DEC keyboard does not work as Meta key.
632
633 This shell command should fix it:
634
635 xmodmap -e 'keycode 0xb1 = Meta_L'
636
637 ** When emacs starts up, I get lots of warnings about unknown keysyms.
638 586
639 If you are running the prebuilt binaries, the Motif library expects to find 587 If you are running the prebuilt binaries, the Motif library expects to find
640 certain thing in the XKeysymDB file. This file is normally in /usr/lib/X11/ 588 certain thing in the XKeysymDB file. This file is normally in /usr/lib/X11/
641 or in /usr/openwin/lib/. If you keep yours in a different place, set the 589 or in /usr/openwin/lib/. If you keep yours in a different place, set the
642 environment variable $XKEYSYMDB to point to it before starting emacs. If 590 environment variable $XKEYSYMDB to point to it before starting emacs. If
643 you still have the problem after doing that, perhaps your version of X is 591 you still have the problem after doing that, perhaps your version of X is
644 too old. There is a copy of the MIT X11R5 XKeysymDB file in the emacs `etc' 592 too old. There is a copy of the MIT X11R5 XKeysymDB file in the emacs `etc'
645 directory. Try using that one. 593 directory. Try using that one.
646 594
647 ** My X resources used to work, and now some of them are being ignored. 595 *** My X resources used to work, and now some of them are being ignored.
648 596
649 Check the resources in .../etc/Emacs.ad (which is the same as the file 597 Check the resources in .../etc/Emacs.ad (which is the same as the file
650 sample.Xdefaults). Perhaps some of the default resources built in to 598 sample.Xdefaults). Perhaps some of the default resources built in to
651 emacs are now overriding your existing resources. Copy and edit the 599 emacs are now overriding your existing resources. Copy and edit the
652 resources in Emacs.ad as necessary. 600 resources in Emacs.ad as necessary.
653 601
654 ** I get complaints about the mapping of my HP keyboard at startup, 602 *** I have focus problems when I use `M-o' to switch to another screen
655 but I haven't changed anything.
656
657 The default HP keymap is set up to have Mod1 assigned to two different keys:
658 Meta_L and Mode_switch (even though there is not actually a Mode_switch key on
659 the keyboard -- it uses an "imaginary" keycode.) There actually is a reason
660 for this, but it's not a good one. The correct fix is to execute this command
661 upon starting X:
662
663 xmodmap -e 'remove mod1 = Mode_switch'
664
665 ** I have focus problems when I use `M-o' to switch to another screen
666 without using the mouse. 603 without using the mouse.
667 604
668 The focus issues with a program like XEmacs, which has multiple 605 The focus issues with a program like XEmacs, which has multiple
669 homogeneous top-level windows, are very complicated, and as a result, 606 homogeneous top-level windows, are very complicated, and as a result,
670 most window managers don't implement them correctly. 607 most window managers don't implement them correctly.
685 (Many people suggest that XEmacs should warp the mouse when focusing 622 (Many people suggest that XEmacs should warp the mouse when focusing
686 on another screen in point-to-type mode. This is not ICCCM-compliant 623 on another screen in point-to-type mode. This is not ICCCM-compliant
687 behavior. Implementing such policy is the responsibility of the 624 behavior. Implementing such policy is the responsibility of the
688 window manager itself, it is not legal for a client to do this.) 625 window manager itself, it is not legal for a client to do this.)
689 626
690 ** Emacs spontaneously displays "I-search: " at the bottom of the screen. 627 *** Emacs spontaneously displays "I-search: " at the bottom of the screen.
691 628
692 This means that Control-S/Control-Q (XON/XOFF) "flow control" is being 629 This means that Control-S/Control-Q (XON/XOFF) "flow control" is being
693 used. C-s/C-q flow control is bad for Emacs editors because it takes 630 used. C-s/C-q flow control is bad for Emacs editors because it takes
694 away C-s and C-q as user commands. Since editors do not output long 631 away C-s and C-q as user commands. Since editors do not output long
695 streams of text without user commands, there is no need for a 632 streams of text without user commands, there is no need for a
769 widespread, XON/XOFF seems to be on the way out. If you can get some 706 widespread, XON/XOFF seems to be on the way out. If you can get some
770 use out of GNU Emacs on inferior terminals, more power to you, but I 707 use out of GNU Emacs on inferior terminals, more power to you, but I
771 will not make Emacs worse for properly designed systems for the sake 708 will not make Emacs worse for properly designed systems for the sake
772 of inferior systems. 709 of inferior systems.
773 710
774 ** Control-S and Control-Q commands are ignored completely. 711 *** Control-S and Control-Q commands are ignored completely.
775 712
776 For some reason, your system is using brain-damaged C-s/C-q flow 713 For some reason, your system is using brain-damaged C-s/C-q flow
777 control despite Emacs's attempts to turn it off. Perhaps your 714 control despite Emacs's attempts to turn it off. Perhaps your
778 terminal is connected to the computer through a concentrator 715 terminal is connected to the computer through a concentrator
779 that wants to use flow control. 716 that wants to use flow control.
784 721
785 If that line of approach is not successful, map some other characters 722 If that line of approach is not successful, map some other characters
786 into C-s and C-q using keyboard-translate-table. The example above 723 into C-s and C-q using keyboard-translate-table. The example above
787 shows how to do this with C-^ and C-\. 724 shows how to do this with C-^ and C-\.
788 725
789 ** Control-S and Control-Q commands are ignored completely on a net 726 *** Control-S and Control-Q commands are ignored completely on a net
790 connection. 727 connection.
791 728
792 Some versions of rlogin (and possibly telnet) do not pass flow 729 Some versions of rlogin (and possibly telnet) do not pass flow
793 control characters to the remote system to which they connect. 730 control characters to the remote system to which they connect.
794 On such systems, emacs on the remote system cannot disable flow 731 On such systems, emacs on the remote system cannot disable flow
811 (enable-flow-control-on "vt200" "vt300" "vt101" "vt131") 748 (enable-flow-control-on "vt200" "vt300" "vt101" "vt131")
812 749
813 See the entry about spontaneous display of I-search (above) for more 750 See the entry about spontaneous display of I-search (above) for more
814 info. 751 info.
815 752
816 ** TTY redisplay is slow. 753 *** TTY redisplay is slow.
817 754
818 XEmacs has fairly new TTY redisplay support (beginning from 19.12), 755 XEmacs has fairly new TTY redisplay support (beginning from 19.12),
819 which doesn't include some basic TTY optimizations -- like using 756 which doesn't include some basic TTY optimizations -- like using
820 scrolling regions to move around blocks of text. This is why 757 scrolling regions to move around blocks of text. This is why
821 redisplay on the traditional terminals, or over slow lines can be very 758 redisplay on the traditional terminals, or over slow lines can be very
822 slow. 759 slow.
823 760
824 If you are interested in fixing this, please let us know at 761 If you are interested in fixing this, please let us know at
825 <xemacs@xemacs.org>. 762 <xemacs@xemacs.org>.
826 763
827 ** Screen is updated wrong, but only on one kind of terminal. 764 *** Screen is updated wrong, but only on one kind of terminal.
828 765
829 This could mean that the termcap entry you are using for that terminal 766 This could mean that the termcap entry you are using for that terminal
830 is wrong, or it could mean that Emacs has a bug handing the 767 is wrong, or it could mean that Emacs has a bug handing the
831 combination of features specified for that terminal. 768 combination of features specified for that terminal.
832 769
861 798
862 This is unambiguously an Emacs bug, and can probably be fixed in 799 This is unambiguously an Emacs bug, and can probably be fixed in
863 termcap.c, terminfo.c, tparam.c, cm.c, redisplay-tty.c, 800 termcap.c, terminfo.c, tparam.c, cm.c, redisplay-tty.c,
864 redisplay-output.c, or redisplay.c. 801 redisplay-output.c, or redisplay.c.
865 802
866 ** Your Delete key sends a Backspace to the terminal, using an AIXterm. 803 *** My buffers are full of \000 characters or otherwise corrupt.
804
805 Some compilers have trouble with gmalloc.c and ralloc.c; try recompiling
806 without optimization. If that doesn't work, try recompiling with
807 SYSTEM_MALLOC defined, and/or with REL_ALLOC undefined.
808
809 *** A position you specified in .Xdefaults is ignored, using twm.
810
811 twm normally ignores "program-specified" positions.
812 You can tell it to obey them with this command in your `.twmrc' file:
813
814 UsePPosition "on" #allow clents to request a position
815
816 *** With M-x enable-flow-control, you need to type C-\ twice to do
817 incremental search--a single C-\ gets no response.
818
819 This has been traced to communicating with your machine via kermit,
820 with C-\ as the kermit escape character. One solution is to use
821 another escape character in kermit. One user did
822
823 set escape-character 17
824
825 in his .kermrc file, to make C-q the kermit escape character.
826
827 *** The Motif version of Emacs paints the screen a solid color.
828
829 This has been observed to result from the following X resource:
830
831 Emacs*default.attributeFont: -*-courier-medium-r-*-*-*-140-*-*-*-*-iso8859-*
832
833 That the resource has this effect indicates a bug in something, but we
834 do not yet know what. If it is an Emacs bug, we hope someone can
835 explain what the bug is so we can fix it. In the mean time, removing
836 the resource prevents the problem.
837
838 *** After running emacs once, subsequent invocations crash.
839
840 Some versions of SVR4 have a serious bug in the implementation of the
841 mmap () system call in the kernel; this causes emacs to run correctly
842 the first time, and then crash when run a second time.
843
844 Contact your vendor and ask for the mmap bug fix; in the mean time,
845 you may be able to work around the problem by adding a line to your
846 operating system description file (whose name is reported by the
847 configure script) that reads:
848 #define SYSTEM_MALLOC
849 This makes Emacs use memory less efficiently, but seems to work around
850 the kernel bug.
851
852 *** Inability to send an Alt-modified key, when Emacs is communicating
853 directly with an X server.
854
855 If you have tried to bind an Alt-modified key as a command, and it
856 does not work to type the command, the first thing you should check is
857 whether the key is getting through to Emacs. To do this, type C-h c
858 followed by the Alt-modified key. C-h c should say what kind of event
859 it read. If it says it read an Alt-modified key, then make sure you
860 have made the key binding correctly.
861
862 If C-h c reports an event that doesn't have the Alt modifier, it may
863 be because your X server has no key for the Alt modifier. The X
864 server that comes from MIT does not set up the Alt modifier by
865 default.
866
867 If your keyboard has keys named Alt, you can enable them as follows:
868
869 xmodmap -e 'add mod2 = Alt_L'
870 xmodmap -e 'add mod2 = Alt_R'
871
872 If the keyboard has just one key named Alt, then only one of those
873 commands is needed. The modifier `mod2' is a reasonable choice if you
874 are using an unmodified MIT version of X. Otherwise, choose any
875 modifier bit not otherwise used.
876
877 If your keyboard does not have keys named Alt, you can use some other
878 keys. Use the keysym command in xmodmap to turn a function key (or
879 some other 'spare' key) into Alt_L or into Alt_R, and then use the
880 commands show above to make them modifier keys.
881
882 Note that if you have Alt keys but no Meta keys, Emacs translates Alt
883 into Meta. This is because of the great importance of Meta in Emacs.
884
885 *** In Shell mode, you get a ^M at the end of every line.
886
887 This happens to people who use tcsh, because it is trying to be too
888 smart. It sees that the Shell uses terminal type `unknown' and turns
889 on the flag to output ^M at the end of each line. You can fix the
890 problem by adding this to your .cshrc file:
891
892 if ($?EMACS) then
893 if ($EMACS == "t") then
894 unset edit
895 stty -icrnl -onlcr -echo susp ^Z
896 endif
897 endif
898
899 *** An error message such as `X protocol error: BadMatch (invalid
900 parameter attributes) on protocol request 93'.
901
902 This comes from having an invalid X resource, such as
903 emacs*Cursor: black
904 (which is invalid because it specifies a color name for something
905 that isn't a color.)
906
907 The fix is to correct your X resources.
908
909 *** Once you pull down a menu from the menubar, it won't go away.
910
911 It has been claimed that this is caused by a bug in certain very old
912 (1990?) versions of the twm window manager. It doesn't happen with
913 recent vintages, or with other window managers.
914
915 *** Emacs ignores the "help" key when running OLWM.
916
917 OLWM grabs the help key, and retransmits it to the appropriate client
918 using XSendEvent. Allowing emacs to react to synthetic events is a
919 security hole, so this is turned off by default. You can enable it by
920 setting the variable x-allow-sendevents to t. You can also cause fix
921 this by telling OLWM to not grab the help key, with the null binding
922 "OpenWindows.KeyboardCommand.Help:".
923
924 *** Programs running under terminal emulator do not recognize `emacs'
925 terminal type.
926
927 The cause of this is a shell startup file that sets the TERMCAP
928 environment variable. The terminal emulator uses that variable to
929 provide the information on the special terminal type that Emacs
930 emulates.
931
932 Rewrite your shell startup file so that it does not change TERMCAP
933 in such a case. You could use the following conditional which sets
934 it only if it is undefined.
935
936 if ( ! ${?TERMCAP} ) setenv TERMCAP ~/my-termcap-file
937
938 Or you could set TERMCAP only when you set TERM--which should not
939 happen in a non-login shell.
940
941 *** The popup menu appears at the buttom/right of my screen.
942
943 You probably have something like the following in your ~/.Xdefaults
944
945 Emacs.geometry: 81x56--9--1
946
947 Use the following instead
948
949 Emacs*EmacsFrame.geometry: 81x56--9--1
950
951
952 ** AIX
953 *** Your Delete key sends a Backspace to the terminal, using an AIXterm.
867 954
868 The solution is to include in your .Xdefaults the lines: 955 The solution is to include in your .Xdefaults the lines:
869 956
870 *aixterm.Translations: #override <Key>BackSpace: string(0x7f) 957 *aixterm.Translations: #override <Key>BackSpace: string(0x7f)
871 aixterm*ttyModes: erase ^? 958 aixterm*ttyModes: erase ^?
872 959
873 This makes your Backspace key send DEL (ASCII 127). 960 This makes your Backspace key send DEL (ASCII 127).
874 961
875 ** With certain fonts, when the cursor appears on a character, the 962 *** On AIX 4, some programs fail when run in a Shell buffer
876 character doesn't appear--you get a solid box instead. 963 with an error message like No terminfo entry for "unknown".
877 964
878 One user on a Linux system reported that this problem went away with 965 On AIX, many terminal type definitions are not installed by default.
879 installation of a new X server. The failing server was XFree86 3.1.1. 966 `unknown' is one of them. Install the "Special Generic Terminal
880 XFree86 3.1.2 works. 967 Definitions" to make them defined.
881 968
882 ** On SunOS 4.1.3, Emacs unpredictably crashes in _yp_dobind_soft. 969 *** On AIX, you get this message when running Emacs:
970
971 Could not load program emacs
972 Symbol smtcheckinit in csh is undefined
973 Error was: Exec format error
974
975 or this one:
976
977 Could not load program .emacs
978 Symbol _system_con in csh is undefined
979 Symbol _fp_trapsta in csh is undefined
980 Error was: Exec format error
981
982 These can happen when you try to run on AIX 3.2.5 a program that was
983 compiled with 3.2.4. The fix is to recompile.
984
985 *** Trouble using ptys on AIX.
986
987 People often install the pty devices on AIX incorrectly.
988 Use `smit pty' to reinstall them properly.
989
990
991 ** SunOS/Solaris
992 *** The Emacs window disappears when you type M-q.
993
994 Some versions of the Open Look window manager interpret M-q as a quit
995 command for whatever window you are typing at. If you want to use
996 Emacs with that window manager, you should try to configure the window
997 manager to use some other command. You can disable the
998 shortcut keys entirely by adding this line to ~/.OWdefaults:
999
1000 OpenWindows.WindowMenuAccelerators: False
1001
1002 *** When Emacs tries to ring the bell, you get an error like
1003
1004 audio: sst_open: SETQSIZE" Invalid argument
1005 audio: sst_close: SETREG MMR2, Invalid argument
1006
1007 you have probably compiled using an ANSI C compiler, but with non-ANSI
1008 include files. In particular, on Suns, the file
1009 /usr/include/sun/audioio.h uses the _IOW macro to define the constant
1010 AUDIOSETQSIZE. _IOW in turn uses a K&R preprocessor feature that is
1011 now explicitly forbidden in ANSI preprocessors, namely substitution
1012 inside character constants. All ANSI C compilers must provide a
1013 workaround for this problem. Lucid's C compiler is shipped with a new
1014 set of system include files. If you are using GCC, there is a script
1015 called fixincludes that creates new versions of some system include
1016 files that use this obsolete feature.
1017
1018 *** On Solaris 2.6, XEmacs dumps core when exiting.
1019
1020 This happens if you're XEmacs is running on the same machine as the X
1021 server, and the optimized memory transport has been turned on by
1022 setting the environment variable XSUNTRANSPORT. The crash occurs
1023 during the call to XCloseDisplay.
1024
1025 If this describes your situation, you need to undefine the
1026 XSUNTRANSPORT environment variable.
1027
1028 *** On Solaris, C-x doesn't get through to Emacs when you use the console.
1029
1030 This is a Solaris feature (at least on Intel x86 cpus). Type C-r
1031 C-r C-t, to toggle whether C-x gets through to Emacs.
1032
1033 *** On Solaris 2.4, Dired hangs and C-g does not work. Or Emacs hangs
1034 forever waiting for termination of a subprocess that is a zombie.
1035
1036 casper@fwi.uva.nl says the problem is in X11R6. Rebuild libX11.so
1037 after changing the file xc/config/cf/sunLib.tmpl. Change the lines
1038
1039 #if ThreadedX
1040 #define SharedX11Reqs -lthread
1041 #endif
1042
1043 to:
1044
1045 #if OSMinorVersion < 4
1046 #if ThreadedX
1047 #define SharedX11Reqs -lthread
1048 #endif
1049 #endif
1050
1051 Be sure also to edit x/config/cf/sun.cf so that OSMinorVersion is 4
1052 (as it should be for Solaris 2.4). The file has three definitions for
1053 OSMinorVersion: the first is for x86, the second for SPARC under
1054 Solaris, and the third for SunOS 4. Make sure to update the
1055 definition for your type of machine and system.
1056
1057 Then do `make Everything' in the top directory of X11R6, to rebuild
1058 the makefiles and rebuild X. The X built this way work only on
1059 Solaris 2.4, not on 2.3.
1060
1061 For multithreaded X to work it necessary to install patch
1062 101925-02 to fix problems in header files [2.4]. You need
1063 to reinstall gcc or re-run just-fixinc after installing that
1064 patch.
1065
1066 However, Frank Rust <frust@iti.cs.tu-bs.de> used a simpler solution:
1067 he changed
1068 #define ThreadedX YES
1069 to
1070 #define ThreadedX NO
1071 in sun.cf and did `make World' to rebuild X11R6. Removing all
1072 `-DXTHREAD*' flags and `-lthread' entries from lib/X11/Makefile and
1073 typing 'make install' in that directory also seemed to work.
1074
1075 *** On SunOS 4.1.3, Emacs unpredictably crashes in _yp_dobind_soft.
883 1076
884 This happens if you configure Emacs specifying just `sparc-sun-sunos4' 1077 This happens if you configure Emacs specifying just `sparc-sun-sunos4'
885 on a system that is version 4.1.3. You must specify the precise 1078 on a system that is version 4.1.3. You must specify the precise
886 version number (or let configure figure out the configuration, which 1079 version number (or let configure figure out the configuration, which
887 it can do perfectly well for SunOS). 1080 it can do perfectly well for SunOS).
888 1081
889 ** On Irix, I don't see the toolbar icons and I'm getting lots of 1082 *** Mail is lost when sent to local aliases.
1083
1084 Many emacs mail user agents (VM and rmail, for instance) use the
1085 sendmail.el library. This library can arrange for mail to be
1086 delivered by passing messages to the /usr/lib/sendmail (usually)
1087 program . In doing so, it passes the '-t' flag to sendmail, which
1088 means that the name of the recipient of the message is not on the
1089 command line and, therefore, that sendmail must parse the message to
1090 obtain the destination address.
1091
1092 There is a bug in the SunOS4.1.1 and SunOS4.1.3 versions of sendmail.
1093 In short, when given the -t flag, the SunOS sendmail won't recognize
1094 non-local (i.e. NIS) aliases. It has been reported that the Solaris
1095 2.x versions of sendmail do not have this bug. For those using SunOS
1096 4.1, the best fix is to install sendmail V8 or IDA sendmail (which
1097 have other advantages over the regular sendmail as well). At the time
1098 of this writing, these official versions are available:
1099
1100 Sendmail V8 on ftp.cs.berkeley.edu in /ucb/sendmail:
1101 sendmail.8.6.9.base.tar.Z (the base system source & documentation)
1102 sendmail.8.6.9.cf.tar.Z (configuration files)
1103 sendmail.8.6.9.misc.tar.Z (miscellaneous support programs)
1104 sendmail.8.6.9.xdoc.tar.Z (extended documentation, with postscript)
1105
1106 IDA sendmail on vixen.cso.uiuc.edu in /pub:
1107 sendmail-5.67b+IDA-1.5.tar.gz
1108
1109 *** Emacs fails to understand most Internet host names, even though
1110 the names work properly with other programs on the same system.
1111 Emacs won't work with X-windows if the value of DISPLAY is HOSTNAME:0.
1112 Gnus can't make contact with the specified host for nntp.
1113
1114 This typically happens on Suns and other systems that use shared
1115 libraries. The cause is that the site has installed a version of the
1116 shared library which uses a name server--but has not installed a
1117 similar version of the unshared library which Emacs uses.
1118
1119 The result is that most programs, using the shared library, work with
1120 the nameserver, but Emacs does not.
1121
1122 The fix is to install an unshared library that corresponds to what you
1123 installed in the shared library, and then relink Emacs.
1124
1125 On SunOS 4.1, simply define HAVE_RES_INIT.
1126
1127 If you have already installed the name resolver in the file libresolv.a,
1128 then you need to compile Emacs to use that library. The easiest way to
1129 do this is to add to config.h a definition of LIBS_SYSTEM, LIBS_MACHINE
1130 or LIB_STANDARD which uses -lresolv. Watch out! If you redefine a macro
1131 that is already in use in your configuration to supply some other libraries,
1132 be careful not to lose the others.
1133
1134 Thus, you could start by adding this to config.h:
1135
1136 #define LIBS_SYSTEM -lresolv
1137
1138 Then if this gives you an error for redefining a macro, and you see that
1139 the s- file defines LIBS_SYSTEM as -lfoo -lbar, you could change config.h
1140 again to say this:
1141
1142 #define LIBS_SYSTEM -lresolv -lfoo -lbar
1143
1144 *** With process-connection-type set to t, each line of subprocess
1145 output is terminated with a ^M, making ange-ftp and GNUS not work.
1146
1147 On SunOS systems, this problem has been seen to be a result of an
1148 incomplete installation of gcc 2.2 which allowed some non-ANSI
1149 compatible include files into the compilation. In particular this
1150 affected virtually all ioctl() calls.
1151
1152
1153 ** Linux
1154 *** `C-z', or `M-x suspend-emacs' hangs instead of suspending.
1155
1156 If you build with `gpm' support on Linux, you cannot suspend XEmacs
1157 because gpm installs a buggy SIGTSTP handler. Either compile with
1158 `--with-gpm=no', or don't suspend XEmacs on the Linux console until
1159 this bug is fixed.
1160
1161 *** With certain fonts, when the cursor appears on a character, the
1162 character doesn't appear--you get a solid box instead.
1163
1164 One user on a Linux system reported that this problem went away with
1165 installation of a new X server. The failing server was XFree86 3.1.1.
1166 XFree86 3.1.2 works.
1167
1168 *** Slow startup on Linux.
1169
1170 People using systems based on the Linux kernel sometimes report that
1171 startup takes 10 to 15 seconds longer than `usual'.
1172
1173 This is because Emacs looks up the host name when it starts.
1174 Normally, this takes negligible time; the extra delay is due to
1175 improper system configuration. This problem can occur for both
1176 networked and non-networked machines.
1177
1178 Here is how to fix the configuration. It requires being root.
1179
1180 **** Networked Case
1181
1182 First, make sure the files `/etc/hosts' and `/etc/host.conf' both
1183 exist. The first line in the `/etc/hosts' file should look like this
1184 (replace HOSTNAME with your host name):
1185
1186 127.0.0.1 localhost HOSTNAME
1187
1188 Also make sure that the `/etc/host.conf' files contains the following
1189 lines:
1190
1191 order hosts, bind
1192 multi on
1193
1194 Any changes, permanent and temporary, to the host name should be
1195 indicated in the `/etc/hosts' file, since it acts a limited local
1196 database of addresses and names (e.g., some SLIP connections
1197 dynamically allocate ip addresses).
1198
1199 **** Non-Networked Case
1200
1201 The solution described in the networked case applies here as well.
1202 However, if you never intend to network your machine, you can use a
1203 simpler solution: create an empty `/etc/host.conf' file. The command
1204 `touch /etc/host.conf' suffices to create the file. The `/etc/hosts'
1205 file is not necessary with this approach.
1206
1207
1208 ** IRIX
1209 *** On Irix, I don't see the toolbar icons and I'm getting lots of
890 entries in the warnings buffer. 1210 entries in the warnings buffer.
891 1211
892 SGI ships a really old Xpm library in /usr/lib which does not work at 1212 SGI ships a really old Xpm library in /usr/lib which does not work at
893 all well with XEmacs. The solution is to install your own copy of the 1213 all well with XEmacs. The solution is to install your own copy of the
894 latest version of Xpm somewhere and then use the --site-includes and 1214 latest version of Xpm somewhere and then use the --site-includes and
895 --site-libraries flags to tell configure where to find it. 1215 --site-libraries flags to tell configure where to find it.
896 1216
897 ** On HPUX, you get "poll: Interrupted system call" message in the 1217 *** Trouble using ptys on IRIX, or running out of ptys.
1218
1219 The program mkpts (which may be in `/usr/adm' or `/usr/sbin') needs to
1220 be set-UID to root, or non-root programs like Emacs will not be able
1221 to allocate ptys reliably.
1222
1223 *** Motif dialog boxes lose on Irix.
1224
1225 Larry Auton <lda@control.att.com> writes:
1226 Beware of not specifying
1227
1228 --with-dialogs=athena
1229
1230 if it builds with the motif dialogs [boom!] you're a dead man.
1231
1232 *** Beware of the default image & graphics library on Irix
1233
1234 Richard Cognot <cognot@ensg.u-nancy.fr> writes:
1235
1236 You *have* to compile your own jpeg lib. The one delivered with SGI
1237 systems is a C++ lib, which apparently XEmacs cannot cope with.
1238
1239
1240 ** Digital UNIX/OSF/VMS/Ultrix
1241 *** XEmacs crashes on Digital Unix within font-lock, or when dealing
1242 with large compilation buffers.
1243
1244 The default stack size under Digital Unix is rather small (2M as
1245 opposed to Solaris 8M), hosing the regexp code, which uses alloca()
1246 extensively, overflowing the stack when complex regexps are used.
1247 Workarounds:
1248
1249 1) Increase your stack size, using `ulimit -s 8192' or a (t)csh
1250 equivalent;
1251
1252 2) Recompile regex.c with REGEX_MALLOC defined.
1253
1254 *** The `Alt' key doesn't behave as `Meta' when running DECwindows.
1255
1256 The default DEC keyboard mapping has the Alt keys set up to generate the
1257 keysym `Multi_key', which has a meaning to xemacs which is distinct from that
1258 of the `Meta_L' and `Meta-R' keysyms. A second problem is that certain keys
1259 have the Mod2 modifier attached to them for no adequately explored reason.
1260 The correct fix is to pass this file to xmodmap upon starting X:
1261
1262 clear mod2
1263 keysym Multi_key = Alt_L
1264 add mod1 = Alt_L
1265 add mod1 = Alt_R
1266
1267 *** The Compose key on a DEC keyboard does not work as Meta key.
1268
1269 This shell command should fix it:
1270
1271 xmodmap -e 'keycode 0xb1 = Meta_L'
1272
1273 *** `expand-file-name' fails to work on any but the machine you dumped
1274 Emacs on.
1275
1276 On Ultrix, if you use any of the functions which look up information
1277 in the passwd database before dumping Emacs (say, by using
1278 expand-file-name in site-init.el), then those functions will not work
1279 in the dumped Emacs on any host but the one Emacs was dumped on.
1280
1281 The solution? Don't use expand-file-name in site-init.el, or in
1282 anything it loads. Yuck - some solution.
1283
1284 I'm not sure why this happens; if you can find out exactly what is
1285 going on, and perhaps find a fix or a workaround, please let us know.
1286 Perhaps the YP functions cache some information, the cache is included
1287 in the dumped Emacs, and is then inaccurate on any other host.
1288
1289
1290 ** HP-UX
1291 *** I get complaints about the mapping of my HP keyboard at startup,
1292 but I haven't changed anything.
1293
1294 The default HP keymap is set up to have Mod1 assigned to two different keys:
1295 Meta_L and Mode_switch (even though there is not actually a Mode_switch key on
1296 the keyboard -- it uses an "imaginary" keycode.) There actually is a reason
1297 for this, but it's not a good one. The correct fix is to execute this command
1298 upon starting X:
1299
1300 xmodmap -e 'remove mod1 = Mode_switch'
1301
1302 *** On HP-UX, you get "poll: Interrupted system call" message in the
898 window where XEmacs was launched. 1303 window where XEmacs was launched.
899 1304
900 Richard Cognot <cognot@ensg.u-nancy.fr> writes: 1305 Richard Cognot <cognot@ensg.u-nancy.fr> writes:
901 1306
902 I get a very strange problem when linking libc.a dynamically: every 1307 I get a very strange problem when linking libc.a dynamically: every
914 after the release otherwise. 1319 after the release otherwise.
915 1320
916 Note: The above entry is probably not valid for XEmacs 21.0 and 1321 Note: The above entry is probably not valid for XEmacs 21.0 and
917 later. 1322 later.
918 1323
919 ** When Emacs tries to ring the bell, you get an error like 1324 *** The right Alt key works wrong on German HP keyboards (and perhaps
920
921 audio: sst_open: SETQSIZE" Invalid argument
922 audio: sst_close: SETREG MMR2, Invalid argument
923
924 you have probably compiled using an ANSI C compiler, but with non-ANSI
925 include files. In particular, on Suns, the file
926 /usr/include/sun/audioio.h uses the _IOW macro to define the constant
927 AUDIOSETQSIZE. _IOW in turn uses a K&R preprocessor feature that is
928 now explicitly forbidden in ANSI preprocessors, namely substitution
929 inside character constants. All ANSI C compilers must provide a
930 workaround for this problem. Lucid's C compiler is shipped with a new
931 set of system include files. If you are using GCC, there is a script
932 called fixincludes that creates new versions of some system include
933 files that use this obsolete feature.
934
935 ** My buffers are full of \000 characters or otherwise corrupt.
936
937 Some compilers have trouble with gmalloc.c and ralloc.c; try recompiling
938 without optimization. If that doesn't work, try recompiling with
939 SYSTEM_MALLOC defined, and/or with REL_ALLOC undefined.
940
941 ** On AIX 4, some programs fail when run in a Shell buffer
942 with an error message like No terminfo entry for "unknown".
943
944 On AIX, many terminal type definitions are not installed by default.
945 `unknown' is one of them. Install the "Special Generic Terminal
946 Definitions" to make them defined.
947
948 ** Emacs exits with "X protocol error" when run with an X server for
949 Windows.
950
951 A certain X server for Windows had a bug which caused this.
952 Supposedly the newer 32-bit version of this server doesn't have the
953 problem.
954
955 ** A position you specified in .Xdefaults is ignored, using twm.
956
957 twm normally ignores "program-specified" positions.
958 You can tell it to obey them with this command in your `.twmrc' file:
959
960 UsePPosition "on" #allow clents to request a position
961
962 ** The right Alt key works wrong on German HP keyboards (and perhaps
963 other non-English HP keyboards too). 1325 other non-English HP keyboards too).
964 1326
965 This is because HPUX defines the modifiers wrong in X. Here is a 1327 This is because HP-UX defines the modifiers wrong in X. Here is a
966 shell script to fix the problem; be sure that it is run after VUE 1328 shell script to fix the problem; be sure that it is run after VUE
967 configures the X server. 1329 configures the X server.
968 1330
969 xmodmap 2> /dev/null - << EOF 1331 xmodmap 2> /dev/null - << EOF
970 keysym Alt_L = Meta_L 1332 keysym Alt_L = Meta_L
977 add mod1 = Meta_L 1339 add mod1 = Meta_L
978 keysym Meta_R = Mode_switch 1340 keysym Meta_R = Mode_switch
979 add mod2 = Mode_switch 1341 add mod2 = Mode_switch
980 EOF 1342 EOF
981 1343
982 ** Trouble using ptys on IRIX, or running out of ptys. 1344 *** `Pid xxx killed due to text modification or page I/O error'
983 1345
984 The program mkpts (which may be in `/usr/adm' or `/usr/sbin') needs to 1346 On HP-UX, you can get that error when the Emacs executable is on an NFS
985 be set-UID to root, or non-root programs like Emacs will not be able 1347 file system. HP-UX responds this way if it tries to swap in a page and
986 to allocate ptys reliably. 1348 does not get a response from the server within a timeout whose default
987 1349 value is just ten seconds.
988 ** Motif dialog boxes lose on Irix. 1350
989 1351 If this happens to you, extend the timeout period.
990 Larry Auton <lda@control.att.com> writes: 1352
991 Beware of not specifying 1353 *** Shell mode on HP-UX gives the message, "`tty`: Ambiguous".
992 1354
993 --with-dialogs=athena 1355 christos@theory.tn.cornell.edu says:
994 1356
995 if it builds with the motif dialogs [boom!] you're a dead man. 1357 The problem is that in your .cshrc you have something that tries to
996 1358 execute `tty`. If you are not running the shell on a real tty then tty
997 ** Beware of the default image & graphics library on Irix 1359 will print "not a tty". Csh expects one word in some places, but tty
998 1360 is giving it back 3.
999 Richard Cognot <cognot@ensg.u-nancy.fr> writes: 1361
1000 You *have* to compile your own jpeg lib. The one delivered with SGI 1362 The solution is to add a pair of quotes around `tty` to make it a
1001 systems is a C++ lib, which apparently XEmacs cannot cope with. 1363 single word:
1002 1364
1003 ** Slow startup on Linux. 1365 if (`tty` == "/dev/console")
1004 1366
1005 People using systems based on the Linux kernel sometimes report that 1367 should be changed to:
1006 startup takes 10 to 15 seconds longer than `usual'. 1368
1007 1369 if ("`tty`" == "/dev/console")
1008 This is because Emacs looks up the host name when it starts. 1370
1009 Normally, this takes negligible time; the extra delay is due to 1371 Even better, move things that set up terminal sections out of .cshrc
1010 improper system configuration. This problem can occur for both 1372 and into .login.
1011 networked and non-networked machines. 1373
1012 1374
1013 Here is how to fix the configuration. It requires being root. 1375 ** SCO
1014 1376 *** Regular expressions matching bugs on SCO systems.
1015 *** Networked Case
1016
1017 First, make sure the files `/etc/hosts' and `/etc/host.conf' both
1018 exist. The first line in the `/etc/hosts' file should look like this
1019 (replace HOSTNAME with your host name):
1020
1021 127.0.0.1 localhost HOSTNAME
1022
1023 Also make sure that the `/etc/host.conf' files contains the following
1024 lines:
1025
1026 order hosts, bind
1027 multi on
1028
1029 Any changes, permanent and temporary, to the host name should be
1030 indicated in the `/etc/hosts' file, since it acts a limited local
1031 database of addresses and names (e.g., some SLIP connections
1032 dynamically allocate ip addresses).
1033
1034 *** Non-Networked Case
1035
1036 The solution described in the networked case applies here as well.
1037 However, if you never intend to network your machine, you can use a
1038 simpler solution: create an empty `/etc/host.conf' file. The command
1039 `touch /etc/host.conf' suffices to create the file. The `/etc/hosts'
1040 file is not necessary with this approach.
1041
1042 ** On Solaris 2.4, Dired hangs and C-g does not work. Or Emacs hangs
1043 forever waiting for termination of a subprocess that is a zombie.
1044
1045 casper@fwi.uva.nl says the problem is in X11R6. Rebuild libX11.so
1046 after changing the file xc/config/cf/sunLib.tmpl. Change the lines
1047
1048 #if ThreadedX
1049 #define SharedX11Reqs -lthread
1050 #endif
1051
1052 to:
1053
1054 #if OSMinorVersion < 4
1055 #if ThreadedX
1056 #define SharedX11Reqs -lthread
1057 #endif
1058 #endif
1059
1060 Be sure also to edit x/config/cf/sun.cf so that OSMinorVersion is 4
1061 (as it should be for Solaris 2.4). The file has three definitions for
1062 OSMinorVersion: the first is for x86, the second for SPARC under
1063 Solaris, and the third for SunOS 4. Make sure to update the
1064 definition for your type of machine and system.
1065
1066 Then do `make Everything' in the top directory of X11R6, to rebuild
1067 the makefiles and rebuild X. The X built this way work only on
1068 Solaris 2.4, not on 2.3.
1069
1070 For multithreaded X to work it necessary to install patch
1071 101925-02 to fix problems in header files [2.4]. You need
1072 to reinstall gcc or re-run just-fixinc after installing that
1073 patch.
1074
1075 However, Frank Rust <frust@iti.cs.tu-bs.de> used a simpler solution:
1076 he changed
1077 #define ThreadedX YES
1078 to
1079 #define ThreadedX NO
1080 in sun.cf and did `make World' to rebuild X11R6. Removing all
1081 `-DXTHREAD*' flags and `-lthread' entries from lib/X11/Makefile and
1082 typing 'make install' in that directory also seemed to work.
1083
1084 ** With M-x enable-flow-control, you need to type C-\ twice to do
1085 incremental search--a single C-\ gets no response.
1086
1087 This has been traced to communicating with your machine via kermit,
1088 with C-\ as the kermit escape character. One solution is to use
1089 another escape character in kermit. One user did
1090
1091 set escape-character 17
1092
1093 in his .kermrc file, to make C-q the kermit escape character.
1094
1095 ** The Motif version of Emacs paints the screen a solid color.
1096
1097 This has been observed to result from the following X resource:
1098
1099 Emacs*default.attributeFont: -*-courier-medium-r-*-*-*-140-*-*-*-*-iso8859-*
1100
1101 That the resource has this effect indicates a bug in something, but we
1102 do not yet know what. If it is an Emacs bug, we hope someone can
1103 explain what the bug is so we can fix it. In the mean time, removing
1104 the resource prevents the problem.
1105
1106 ** Regular expressions matching bugs on SCO systems.
1107 1377
1108 On SCO, there are problems in regexp matching when Emacs is compiled 1378 On SCO, there are problems in regexp matching when Emacs is compiled
1109 with the system compiler. The compiler version is "Microsoft C 1379 with the system compiler. The compiler version is "Microsoft C
1110 version 6", SCO 4.2.0h Dev Sys Maintenance Supplement 01/06/93; Quick 1380 version 6", SCO 4.2.0h Dev Sys Maintenance Supplement 01/06/93; Quick
1111 C Compiler Version 1.00.46 (Beta). The solution is to compile with 1381 C Compiler Version 1.00.46 (Beta). The solution is to compile with
1112 GCC. 1382 GCC.
1113 1383
1114 ** In Shell mode, you get a ^M at the end of every line. 1384
1115 1385 ** Windows
1116 This happens to people who use tcsh, because it is trying to be too 1386 *** Emacs exits with "X protocol error" when run with an X server for
1117 smart. It sees that the Shell uses terminal type `unknown' and turns 1387 Windows.
1118 on the flag to output ^M at the end of each line. You can fix the 1388
1119 problem by adding this to your .cshrc file: 1389 A certain X server for Windows had a bug which caused this.
1120 1390 Supposedly the newer 32-bit version of this server doesn't have the
1121 if ($?EMACS) then 1391 problem.
1122 if ($EMACS == "t") then 1392
1123 unset edit
1124 stty -icrnl -onlcr -echo susp ^Z
1125 endif
1126 endif
1127
1128 ** An error message such as `X protocol error: BadMatch (invalid
1129 parameter attributes) on protocol request 93'.
1130
1131 This comes from having an invalid X resource, such as
1132 emacs*Cursor: black
1133 (which is invalid because it specifies a color name for something
1134 that isn't a color.)
1135
1136 The fix is to correct your X resources.
1137
1138 ** Mail is lost when sent to local aliases.
1139
1140 Many emacs mail user agents (VM and rmail, for instance) use the
1141 sendmail.el library. This library can arrange for mail to be
1142 delivered by passing messages to the /usr/lib/sendmail (usually)
1143 program . In doing so, it passes the '-t' flag to sendmail, which
1144 means that the name of the recipient of the message is not on the
1145 command line and, therefore, that sendmail must parse the message to
1146 obtain the destination address.
1147
1148 There is a bug in the SunOS4.1.1 and SunOS4.1.3 versions of sendmail.
1149 In short, when given the -t flag, the SunOS sendmail won't recognize
1150 non-local (i.e. NIS) aliases. It has been reported that the Solaris
1151 2.x versions of sendmail do not have this bug. For those using SunOS
1152 4.1, the best fix is to install sendmail V8 or IDA sendmail (which
1153 have other advantages over the regular sendmail as well). At the time
1154 of this writing, these official versions are available:
1155
1156 Sendmail V8 on ftp.cs.berkeley.edu in /ucb/sendmail:
1157 sendmail.8.6.9.base.tar.Z (the base system source & documentation)
1158 sendmail.8.6.9.cf.tar.Z (configuration files)
1159 sendmail.8.6.9.misc.tar.Z (miscellaneous support programs)
1160 sendmail.8.6.9.xdoc.tar.Z (extended documentation, with postscript)
1161
1162 IDA sendmail on vixen.cso.uiuc.edu in /pub:
1163 sendmail-5.67b+IDA-1.5.tar.gz
1164
1165 ** On AIX, you get this message when running Emacs:
1166
1167 Could not load program emacs
1168 Symbol smtcheckinit in csh is undefined
1169 Error was: Exec format error
1170
1171 or this one:
1172
1173 Could not load program .emacs
1174 Symbol _system_con in csh is undefined
1175 Symbol _fp_trapsta in csh is undefined
1176 Error was: Exec format error
1177
1178 These can happen when you try to run on AIX 3.2.5 a program that was
1179 compiled with 3.2.4. The fix is to recompile.
1180
1181 ** After running emacs once, subsequent invocations crash.
1182
1183 Some versions of SVR4 have a serious bug in the implementation of the
1184 mmap () system call in the kernel; this causes emacs to run correctly
1185 the first time, and then crash when run a second time.
1186
1187 Contact your vendor and ask for the mmap bug fix; in the mean time,
1188 you may be able to work around the problem by adding a line to your
1189 operating system description file (whose name is reported by the
1190 configure script) that reads:
1191 #define SYSTEM_MALLOC
1192 This makes Emacs use memory less efficiently, but seems to work around
1193 the kernel bug.
1194
1195 ** Inability to send an Alt-modified key, when Emacs is communicating
1196 directly with an X server.
1197
1198 If you have tried to bind an Alt-modified key as a command, and it
1199 does not work to type the command, the first thing you should check is
1200 whether the key is getting through to Emacs. To do this, type C-h c
1201 followed by the Alt-modified key. C-h c should say what kind of event
1202 it read. If it says it read an Alt-modified key, then make sure you
1203 have made the key binding correctly.
1204
1205 If C-h c reports an event that doesn't have the Alt modifier, it may
1206 be because your X server has no key for the Alt modifier. The X
1207 server that comes from MIT does not set up the Alt modifier by
1208 default.
1209
1210 If your keyboard has keys named Alt, you can enable them as follows:
1211
1212 xmodmap -e 'add mod2 = Alt_L'
1213 xmodmap -e 'add mod2 = Alt_R'
1214
1215 If the keyboard has just one key named Alt, then only one of those
1216 commands is needed. The modifier `mod2' is a reasonable choice if you
1217 are using an unmodified MIT version of X. Otherwise, choose any
1218 modifier bit not otherwise used.
1219
1220 If your keyboard does not have keys named Alt, you can use some other
1221 keys. Use the keysym command in xmodmap to turn a function key (or
1222 some other 'spare' key) into Alt_L or into Alt_R, and then use the
1223 commands show above to make them modifier keys.
1224
1225 Note that if you have Alt keys but no Meta keys, Emacs translates Alt
1226 into Meta. This is because of the great importance of Meta in Emacs.
1227
1228 ** `Pid xxx killed due to text modification or page I/O error'
1229
1230 On HP/UX, you can get that error when the Emacs executable is on an NFS
1231 file system. HP/UX responds this way if it tries to swap in a page and
1232 does not get a response from the server within a timeout whose default
1233 value is just ten seconds.
1234
1235 If this happens to you, extend the timeout period.
1236
1237 ** `expand-file-name' fails to work on any but the machine you dumped
1238 Emacs on.
1239
1240 On Ultrix, if you use any of the functions which look up information
1241 in the passwd database before dumping Emacs (say, by using
1242 expand-file-name in site-init.el), then those functions will not work
1243 in the dumped Emacs on any host but the one Emacs was dumped on.
1244
1245 The solution? Don't use expand-file-name in site-init.el, or in
1246 anything it loads. Yuck - some solution.
1247
1248 I'm not sure why this happens; if you can find out exactly what is
1249 going on, and perhaps find a fix or a workaround, please let us know.
1250 Perhaps the YP functions cache some information, the cache is included
1251 in the dumped Emacs, and is then inaccurate on any other host.
1252
1253 ** Emacs fails to understand most Internet host names, even though
1254 the names work properly with other programs on the same system.
1255 Emacs won't work with X-windows if the value of DISPLAY is HOSTNAME:0.
1256 Gnus can't make contact with the specified host for nntp.
1257
1258 This typically happens on Suns and other systems that use shared
1259 libraries. The cause is that the site has installed a version of the
1260 shared library which uses a name server--but has not installed a
1261 similar version of the unshared library which Emacs uses.
1262
1263 The result is that most programs, using the shared library, work with
1264 the nameserver, but Emacs does not.
1265
1266 The fix is to install an unshared library that corresponds to what you
1267 installed in the shared library, and then relink Emacs.
1268
1269 On SunOS 4.1, simply define HAVE_RES_INIT.
1270
1271 If you have already installed the name resolver in the file libresolv.a,
1272 then you need to compile Emacs to use that library. The easiest way to
1273 do this is to add to config.h a definition of LIBS_SYSTEM, LIBS_MACHINE
1274 or LIB_STANDARD which uses -lresolv. Watch out! If you redefine a macro
1275 that is already in use in your configuration to supply some other libraries,
1276 be careful not to lose the others.
1277
1278 Thus, you could start by adding this to config.h:
1279
1280 #define LIBS_SYSTEM -lresolv
1281
1282 Then if this gives you an error for redefining a macro, and you see that
1283 the s- file defines LIBS_SYSTEM as -lfoo -lbar, you could change config.h
1284 again to say this:
1285
1286 #define LIBS_SYSTEM -lresolv -lfoo -lbar
1287
1288 ** Trouble using ptys on AIX.
1289
1290 People often install the pty devices on AIX incorrectly.
1291 Use `smit pty' to reinstall them properly.
1292
1293 ** Shell mode on HP/UX gives the message, "`tty`: Ambiguous".
1294
1295 christos@theory.tn.cornell.edu says:
1296
1297 The problem is that in your .cshrc you have something that tries to
1298 execute `tty`. If you are not running the shell on a real tty then tty
1299 will print "not a tty". Csh expects one word in some places, but tty
1300 is giving it back 3.
1301
1302 The solution is to add a pair of quotes around `tty` to make it a
1303 single word:
1304
1305 if (`tty` == "/dev/console")
1306
1307 should be changed to:
1308
1309 if ("`tty`" == "/dev/console")
1310
1311 Even better, move things that set up terminal sections out of .cshrc
1312 and into .login.
1313
1314 ** With process-connection-type set to t, each line of subprocess
1315 output is terminated with a ^M, making ange-ftp and GNUS not work.
1316
1317 On SunOS systems, this problem has been seen to be a result of an
1318 incomplete installation of gcc 2.2 which allowed some non-ANSI
1319 compatible include files into the compilation. In particular this
1320 affected virtually all ioctl() calls.
1321
1322 ** Once you pull down a menu from the menubar, it won't go away.
1323
1324 It has been claimed that this is caused by a bug in certain very old
1325 (1990?) versions of the twm window manager. It doesn't happen with
1326 recent vintages, or with other window managers.
1327
1328 ** Emacs ignores the "help" key when running OLWM.
1329
1330 OLWM grabs the help key, and retransmits it to the appropriate client
1331 using XSendEvent. Allowing emacs to react to synthetic events is a
1332 security hole, so this is turned off by default. You can enable it by
1333 setting the variable x-allow-sendevents to t. You can also cause fix
1334 this by telling OLWM to not grab the help key, with the null binding
1335 "OpenWindows.KeyboardCommand.Help:".
1336
1337 ** Programs running under terminal emulator do not recognize `emacs'
1338 terminal type.
1339
1340 The cause of this is a shell startup file that sets the TERMCAP
1341 environment variable. The terminal emulator uses that variable to
1342 provide the information on the special terminal type that Emacs
1343 emulates.
1344
1345 Rewrite your shell startup file so that it does not change TERMCAP
1346 in such a case. You could use the following conditional which sets
1347 it only if it is undefined.
1348
1349 if ( ! ${?TERMCAP} ) setenv TERMCAP ~/my-termcap-file
1350
1351 Or you could set TERMCAP only when you set TERM--which should not
1352 happen in a non-login shell.
1353
1354 ** The popup menu appears at the buttom/right of my screen.
1355
1356 You probably have something like the following in your ~/.Xdefaults
1357
1358 Emacs.geometry: 81x56--9--1
1359
1360 Use the following instead
1361
1362 Emacs*EmacsFrame.geometry: 81x56--9--1
1363 1393
1364 1394
1365 * Compatibility problems (with Emacs 18, GNU Emacs, or previous XEmacs/lemacs) 1395 * Compatibility problems (with Emacs 18, GNU Emacs, or previous XEmacs/lemacs)
1366 ============================================================================== 1396 ==============================================================================
1367 1397
1368 ** "Symbol's value as variable is void: unread-command-char". 1398 *** "Symbol's value as variable is void: unread-command-char".
1369 "Wrong type argument: arrayp, #<keymap 143 entries>" 1399 "Wrong type argument: arrayp, #<keymap 143 entries>"
1370 "Wrong type argument: stringp, [#<keypress-event return>]" 1400 "Wrong type argument: stringp, [#<keypress-event return>]"
1371 1401
1372 There are a few incompatible changes in XEmacs, and these are the 1402 There are a few incompatible changes in XEmacs, and these are the
1373 symptoms. Some of the emacs-lisp code you are running needs to be 1403 symptoms. Some of the emacs-lisp code you are running needs to be