9 0.1.5.1 Unable to find files
10 0.1.5.2 Slow path searching
11 0.1.5.3 Unable to generate fonts
12 0.1.5.4 TeX or Metafont failing
13 0.1.5.5 Empty Makefiles
16 0.1.5.8 `ShellWidgetClass'
17 0.1.5.9 Pointer combination warnings
23 If you have problems or suggestions, please report them to
24 <tex-k@mail.tug.org> using the bug checklist below.
26 Please report bugs in the documentation; not only factual errors or
27 inconsistent behavior, but unclear or incomplete explanations, typos,
33 Before reporting a bug, please check below to be sure it isn't already
34 known (*note Common problems::).
36 Bug reports should be sent via electronic mail to
37 <tex-k@mail.tug.org>, or by postal mail to 135 Center Hill Road /
38 Plymouth, MA 02360 / USA.
40 The general principle is that a good bug report includes all the
41 information necessary for reproduction. Therefore, to enable
42 investigation, your report should include the following:
44 * The version number(s) of the program(s) involved, and of Kpathsea
45 itself. You can get the former by giving a sole option `--version'
46 to the program, and the latter by running `kpsewhich --version'.
47 The `NEWS' and `ChangeLog' files also contain the version number.
49 * The hardware, operating system (including version number),
50 compiler, and `make' program you are using (the output of `uname
51 -a' is a start on the first two, though often incomplete). If the
52 bug involves the X window system, include X version and supplier
53 information as well (examples: X11R6 from MIT; X11R4 from HP;
54 OpenWindows 3.3 bundled with SunOS 4.1.4).
56 * Any options you gave to `configure'. This is recorded in the
57 `config.status' files.
59 If you are reporting a bug in `configure' itself, it's probably
60 system-dependent, and it will be unlikely the maintainers can do
61 anything useful if you merely report that thus-and-such is broken.
62 Therefore, you need to do some additional work: for some bugs, you
63 can look in the file `config.log' where the test that failed should
64 appear, along with the compiler invocation and source program in
65 question. You can then compile it yourself by hand, and discover
66 why the test failed. Other `configure' bugs do not involve the
67 compiler; in that case, the only recourse is to inspect the
68 `configure' shell script itself, or the Autoconf macros that
69 generated `configure'.
71 * The log of all debugging output, if the bug is in path searching.
72 You can get this by setting the environment variable
73 `KPATHSEA_DEBUG' to `-1' before running the program. Please look
74 at the log yourself to make sure the behavior is really a bug
75 before reporting it; perhaps "old" environment variable settings
76 are causing files not to be found, for example.
78 * The contents of any input files necessary to reproduce the bug.
79 For bugs in DVI-reading programs, for example, this generally
80 means a DVI file (and any EPS or other files it uses)--TeX source
81 files are helpful, but the DVI file is necessary, because that's
82 the actual program input.
84 GNU `shar', available from `ftp://prep.ai.mit.edu/pub/gnu' is a
85 convenient way of packaging multiple (possibly binary) files for
86 electronic mail. If you feel your input files are too big to send
87 by email, you can ftp them to `ftp://ftp.tug.org/incoming' (that
88 directory is writable, but not readable).
90 * If you are sending a patch (do so if you can!), please do so in
91 the form of a context diff (`diff -c') against the original
92 distribution source. Any other form of diff is either not as
93 complete or harder for me to understand. Please also include a
96 * If the bug involved is an actual crash (i.e., core dump), it is
97 easy and useful to include a stack trace from a debugger (I
98 recommend the GNU debugger GDB, available from
99 `ftp://prep.ai.mit.edu/pub/gnu'). If the cause is apparent (a
100 `NULL' value being dereferenced, for example), please send the
101 details along. If the program involved is TeX or Metafont, and
102 the crash is happening at apparently-sound code, however, the bug
103 may well be in the compiler, rather than in the program or the
104 library (*note TeX or Metafont failing: TeX or Metafont failing.).
106 * Any additional information that will be helpful in reproducing,
107 diagnosing, or fixing the bug.
112 Web2c and Kpathsea in general are discussed on the mailing list
113 <tex-k@mail.tug.org>. To join, email <tex-k-request@mail.tug.org> with
116 subscribe YOU@YOUR.PREFERRED.EMAIL.ADDRESS
118 in the body of the message.
120 You do not need to join to submit a report, nor will it affect whether
121 you get a response. There is no Usenet newsgroup equivalent (if you can
122 be the one to set this up, email `tex-k-request'). Traffic on the list
123 is fairly light, and is mainly bug reports and enhancement requests to
124 the software. The best way to decide if you want to join or not is
125 read some of the archives from `ftp://ftp.tug.org/mail/archives/tex-k/'.
127 Be aware that large data files are sometimes included in bug reports.
128 If this is a problem for you, do not join the list.
130 If you only want announcements of new releases, not bug reports and
131 discussion, join <tex-archive@math.utah.edu> (via mail to
132 <tex-archive-request@math.utah.edu>).
134 If you are looking for general TeX help, such as how to use LaTeX,
135 please use the mailing list <info-tex@shsu.edu> mailing list, which is
136 gatewayed to the `comp.text.tex' Usenet newsgroup (or post to the
137 newsgroup; the gateway is bidirectional).
142 Kpathsea provides a number of runtime debugging options, detailed below
143 by their names and corresponding numeric values. When the files you
144 expect aren't being found, the thing to do is enable these options and
147 You can set these with some runtime argument (e.g., `-d') to the
148 program; in that case, you should use the numeric values described in
149 the program's documentation (which, for Dvipsk and Xdvik, are different
150 than those below). It's best to give the `-d' (or whatever) option
151 first, for maximal output. Dvipsk and Xdvik have additional
152 program-specific debugging options as well.
154 You can also set the environment variable `KPATHSEA_DEBUG'; in this
155 case, you should use the numbers below. If you run the program under a
156 debugger and set the variable `kpathsea_debug', also use the numbers
159 In any case, by far the simplest value to use is `-1', which will
160 turn on all debugging output. This is usually better than guessing
161 which particular values will yield the output you need.
163 Debugging output always goes to standard error, so you can redirect it
164 easily. For example, in Bourne-compatible shells:
165 dvips -d -1 ... 2>/tmp/debug
167 It is sometimes helpful to run the standalone Kpsewhich utility
168 (*note Invoking kpsewhich::), instead of the original program.
170 In any case, you can _not_ use the _names_ below; you must always use
171 somebody's numbers. (Sorry.) To set more than one option, just sum
172 the corresponding numbers.
174 `KPSE_DEBUG_STAT (1)'
175 Report `stat'(2) calls. This is useful for verifying that your
176 directory structure is not forcing Kpathsea to do many additional
177 file tests (*note Slow path searching::, and *note Subdirectory
178 expansion::). If you are using an up-to-date `ls-R' database
179 (*note Filename database::), this should produce no output unless a
180 nonexistent file that must exist is searched for.
182 `KPSE_DEBUG_HASH (2)'
183 Report lookups in all hash tables: `ls-R' and `aliases' (*note
184 Filename database::); font aliases (*note Fontmap::); and config
185 file values (*note Config files::). Useful when expected values
186 are not being found, e.g.., file searches are looking at the disk
187 instead of using `ls-R'.
189 `KPSE_DEBUG_FOPEN (4)'
190 Report file openings and closings. Especially useful when your
191 system's file table is full, for seeing which files have been
192 opened but never closed. In case you want to set breakpoints in a
193 debugger: this works by redefining `fopen' (`fclose') to be
194 `kpse_fopen_trace' (`kpse_fclose_trace').
196 `KPSE_DEBUG_PATHS (8)'
197 Report general path information for each file type Kpathsea is
198 asked to search. This is useful when you are trying to track down
199 how a particular path got defined--from `texmf.cnf', `config.ps',
200 an environment variable, the compile-time default, etc. This is
201 the contents of the `kpse_format_info_type' structure defined in
204 `KPSE_DEBUG_EXPAND (16)'
205 Report the directory list corresponding to each path element
206 Kpathsea searches. This is only relevant when Kpathsea searches
207 the disk, since `ls-R' searches don't look through directory lists
210 `KPSE_DEBUG_SEARCH (32)'
211 Report on each file search: the name of the file searched for, the
212 path searched in, whether or not the file must exist (when drivers
213 search for `cmr10.vf', it need not exist), and whether or not we
214 are collecting all occurrences of the file in the path (as with,
215 e.g., `texmf.cnf' and `texfonts.map'), or just the first (as with
216 most lookups). This can help you correlate what Kpathsea is doing
217 with what is in your input file.
219 `KPSE_DEBUG_VARS (64)'
220 Report the value of each variable Kpathsea looks up. This is
221 useful for verifying that variables do indeed obtain their correct
224 `GSFTOPK_DEBUG (128)'
225 Activates debugging printout specific to `gsftopk' program.
227 `MAKETEX_DEBUG (512)'
228 If you use the optional `mktex' programs instead of the
229 traditional shell scripts, this will report the name of the site
230 file (`mktex.cnf' by default) which is read, directories created by
231 `mktexdir', the full path of the `ls-R' database built by
232 `mktexlsr', font map searches, `MT_FEATURES' in effect, parameters
233 from `mktexnam', filenames added by `mktexupd', and some
234 subsidiary commands run by the programs.
236 `MAKETEX_FINE_DEBUG (1024)'
237 When the optional `mktex' programs are used, this will print
238 additional debugging info from functions internal to these
241 Debugging output from Kpathsea is always written to standard error,
242 and begins with the string `kdebug:'. (Except for hash table buckets,
243 which just start with the number, but you can only get that output
244 running under a debugger. See comments at the `hash_summary_only'
245 variable in `kpathsea/db.c'.)
250 Kpathsea can record the time and filename found for each successful
251 search. This may be useful in finding good candidates for deletion when
252 your filesystem is full, or in discovering usage patterns at your site.
254 To do this, define the environment or config file variable
255 `TEXMFLOG'. The value is the name of the file to append the
256 information to. The file is created if it doesn't exist, and appended
259 Each successful search turns into one line in the log file: two words
260 separated by a space. The first word is the time of the search, as the
261 integer number of seconds since "the epoch", i.e., UTC midnight 1
262 January 1970 (more precisely, the result of the `time' system call).
263 The second word is the filename.
265 For example, after `setenv TEXMFLOG /tmp/log', running Dvips on
266 `story.dvi' appends the following lines:
268 774455887 /usr/local/share/texmf/dvips/config.ps
269 774455887 /usr/local/share/texmf/dvips/psfonts.map
270 774455888 /usr/local/share/texmf/dvips/texc.pro
271 774455888 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmbx10.600pk
272 774455889 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmsl10.600pk
273 774455889 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmr10.600pk
274 774455889 /usr/local/share/texmf/dvips/texc.pro
276 Only filenames that are absolute are recorded, to preserve some
277 semblance of privacy.
279 0.1.5 Common problems
280 ---------------------
282 Here are some common problems with configuration, compilation, linking,
285 0.1.5.1 Unable to find files
286 ............................
288 If a program complains it cannot find fonts (or other input files), any
289 of several things might be wrong. In any case, you may find the
290 debugging options helpful. *Note Debugging::.
292 * Perhaps you simply haven't installed all the necessary files; the
293 basic fonts and input files are distributed separately from the
294 programs. *Note unixtex.ftp::.
296 * You have (perhaps unknowingly) told Kpathsea to use search paths
297 that don't reflect where the files actually are. One common cause
298 is having environment variables set from a previous installation,
299 thus overriding what you carefully set in `texmf.cnf' (*note
300 Supported file formats::). System `/etc/profile' or other files
301 such may be the culprit.
303 * Your files reside in a directory that is only pointed to via a
304 symbolic link, in a leaf directory and is not listed in `ls-R'.
306 Unfortunately, Kpathsea's subdirectory searching has an
307 irremediable deficiency: If a directory D being searched for
308 subdirectories contains plain files and symbolic links to other
309 directories, but no true subdirectories, D will be considered a
310 leaf directory, i.e., the symbolic links will not be followed.
311 *Note Subdirectory expansion::.
313 You can work around this problem by creating an empty dummy
314 subdirectory in D. Then D will no longer be a leaf, and the
315 symlinks will be followed.
317 The directory immediately followed by the `//' in the path
318 specification, however, is always searched for subdirectories,
319 even if it is a leaf. Presumably you would not have asked for the
320 directory to be searched for subdirectories if you didn't want it
323 * If the fonts (or whatever) don't already exist, `mktexpk' (or
324 `mktexmf' or `mktextfm') will try to create them. If these rather
325 complicated shell scripts fail, you'll eventually get an error
326 message saying something like `Can't find font FONTNAME'. The best
327 solution is to fix (or at least report) the bug in `mktexpk'; the
328 workaround is to generate the necessary fonts by hand with
329 Metafont, or to grab them from a CTAN site (*note unixtex.ftp::).
331 * There is a bug in the library. *Note Reporting bugs::.
333 0.1.5.2 Slow path searching
334 ...........................
336 If your program takes an excessively long time to find fonts or other
337 input files, but does eventually succeed, here are some possible
340 * Most likely, you just have a lot of directories to search, and that
341 takes a noticeable time. The solution is to create and maintain a
342 separate `ls-R' file that lists all the files in your main TeX
343 hierarchy. *Note Filename database::. Kpathsea always uses `ls-R'
344 if it's present; there's no need to recompile or reconfigure any
347 * Your recursively-searched directories (e.g.,
348 `/usr/local/share/texmf/fonts//'), contain a mixture of files and
349 directories. This prevents Kpathsea from using a useful
350 optimization (*note Subdirectory expansion::).
352 It is best to have only directories (and perhaps a `README') in the
353 upper levels of the directory structure, and it's very important
354 to have _only_ files, and no subdirectories, in the leaf
355 directories where the dozens of TFM, PK, or whatever files reside.
357 In any case, you may find the debugging options helpful in determining
358 precisely when the disk or network is being pounded. *Note Debugging::.
360 0.1.5.3 Unable to generate fonts
361 ................................
363 Metafont outputs fonts in bitmap format, tuned for a particular device
364 at a particular resolution, in order to allow for the highest-possible
365 quality of output. Some DVI-to-whatever programs, such as Dvips, try
366 to generate these on the fly when they are needed, but this generation
367 may fail in several cases.
369 If `mktexpk' runs, but fails with this error:
370 mktexpk: Can't guess mode for NNN dpi devices.
371 mktexpk: Use a config file to specify the mode, or update me.
372 you need to ensure the resolution and mode match; just specifying the
373 resolution, as in `-D 360', is not enough.
375 You can specify the mode name with the `-mode' option on the Dvips
376 command line, or in a Dvips configuration file (*note Config files:
377 (dvips)Config files.), such as `config.ps' in your document directory,
378 `~/.dvipsrc' in your home directory, or in a system directory (again
379 named `config.ps'). (Other drivers use other files, naturally.)
381 For example, if you need 360dpi fonts, you could include this in a
386 If Metafont runs, but generates fonts at the wrong resolution or for
387 the wrong device, most likely `mktexpk''s built-in guess for the mode
388 is wrong, and you should override it as above.
390 See `ftp://ftp.tug.org/tex/modes.mf' for a list of resolutions and
391 mode names for most devices (additional submissions are welcome).
393 If Metafont runs but generates fonts at a resolution of 2602dpi (and
394 prints out the name of each character as well as just a character
395 number, and maybe tries to display the characters), then your Metafont
396 base file probably hasn't been made properly. (It's using the default
397 `proof' mode, instead of an actual device mode.) To make a proper
398 `plain.base', assuming the local mode definitions are contained in a
399 file `modes.mf', run the following command (assuming Unix):
401 inimf "plain; input modes; dump"
403 Then copy the `plain.base' file from the current directory to where the
404 base files are stored on your system (`/usr/local/share/texmf/web2c' by
405 default), and make a link (either hard or soft) from `plain.base' to
406 `mf.base' in that directory. *Note inimf invocation: (web2c)inimf
409 If `mf' is a command not found at all by `mktexpk', then you need to
410 install Metafont (*note unixtex.ftp::).
412 0.1.5.4 TeX or Metafont failing
413 ...............................
415 If TeX or Metafont get a segmentation fault or otherwise fail while
416 running a normal input file, the problem is usually a compiler bug
417 (unlikely as that may sound). Even if the trip and trap tests are
418 passed, problems may lurk. Optimization occasionally causes trouble in
419 programs other than TeX and Metafont themselves, too.
421 Insufficient swap space may also cause core dumps or other erratic
424 For a workaround, if you enabled any optimization flags, it's best to
425 omit optimization entirely. In any case, the way to find the facts is
426 to run the program under the debugger and see where it's failing.
428 Also, if you have trouble with a system C compiler, I advise trying
429 the GNU C compiler. And vice versa, unfortunately; but in that case I
430 also recommend reporting a bug to the GCC mailing list; see *Note Bugs:
433 To report compiler bugs effectively requires perseverance and
434 perspicacity: you must find the miscompiled line, and that usually
435 involves delving backwards in time from the point of error, checking
436 through TeX's (or whatever program's) data structures. Things are not
437 helped by all-too-common bugs in the debugger itself. Good luck.
439 One known cause of trouble is the way arrays are handled. Some of the
440 Pascal arrays have a lower index other than 0, and the C code will take
441 the pointer to the allocated memory, subtract the lower index, and use
442 the resulting pointer for the array. While this trick often works, ANSI
443 C doesn't guarantee that it will. It it known to fail on HP-UX 10
444 mchines when the native compiler is used, unless the `+u' compiler
445 switch was specified. Using GCC will work on this platform as well.
447 0.1.5.5 Empty Makefiles
448 .......................
450 On some systems (NetBSD, FreeBSD, AIX 4.1, and Mach10), `configure' may
451 fail to properly create the Makefiles. Instead, you get an error which
452 looks something like this:
457 sed: 1: "\\@^ac_include make/pat ...": \ can not be used as a string delimiter
459 So far as I know, the bug here is in `/bin/sh' on these systems. I
460 don't have access to a machine running any of them, so if someone can
461 find a workaround that avoids the quoting bug, I'd be most grateful.
462 (Search for `ac_include' in the `configure' script to get to the
465 It should work to run `bash configure', instead of using `/bin/sh'.
466 You can get Bash from `ftp://prep.ai.mit.edu/pub/gnu' and mirrors.
468 Another possible cause (reported for NeXT) is a bug in the `sed'
469 command. In that case the error may look like this:
471 Unrecognized command: \@^ac_include make/paths.make@r make/paths.make
473 In this case, installing GNU `sed' should solve the problem. You can
474 get GNU `sed' from the same places as Bash.
479 You may find that linking X programs results in an error from the linker
480 that `XtStrings' is undefined, something like this:
483 .../x11.c:130: undefined reference to `XtStrings'
485 This generally happens because of a mismatch between the X include
486 files with which you compiled and the X libraries with which you linked;
487 often, the include files are from MIT and the libraries from Sun.
489 The solution is to use the same X distribution for compilation and
490 linking. Probably `configure' was unable to guess the proper
491 directories from your installation. You can use the `configure'
492 options `--x-includes=PATH' and `--x-libraries=PATH' to explicitly
498 (This section adapted from the file `dlsym.c' in the X distribution.)
500 The `Xlib' library uses the standard C function `wcstombs'. Under
501 SunOS 4.1, `wcstombs' uses the `dlsym' interface defined in `libdl.so'.
502 Unfortunately, the SunOS 4.1 distribution does not include a static
505 As a result, if you try to link an X program statically under SunOS,
506 you may get undefined references to `dlopen', `dlsym', and `dlclose'.
507 One workaround is to include these definitions when you link:
509 void *dlopen() { return 0; }
510 void *dlsym() { return 0; }
511 int dlclose() { return -1; }
513 These are contained in the `dlsym.c' file in the MIT X distribution.
515 0.1.5.8 `ShellWidgetClass'
516 ..........................
518 (This section adapted from the comp.sys.sun.admin FAQ.)
520 If you are linking with Sun's OpenWindows libraries in SunOS 4.1.x,
521 you may get undefined symbols `_get_wmShellWidgetClass' and
522 `_get_applicationShellWidgetClass' when linking. This problem does not
523 arise using the standard MIT X libraries under SunOS.
525 The cause is bugs in the `Xmu' shared library as shipped from Sun.
526 There are several fixes:
528 * Install the free MIT distribution from `ftp.x.org' and mirrors.
530 * Get the OpenWindows patches listed below.
532 * Statically link the `Xmu' library into the executable.
534 * Avoid using `Xmu' at all. If you are compiling Metafont, see *Note
535 Online Metafont graphics: (web2c)Online Metafont graphics. If you
536 are compiling Xdvi, see the `-DNOTOOL' option in `xdvik/INSTALL'.
538 * Ignore the errors. The binary runs fine regardless.
541 Here is the information for getting the two patches:
544 Bug ID's: 1086793, 1086912, 1074766
545 Description: 4.1.x OpenWindows 3.0 `libXt' jumbo patch
549 Description: 4.1.x OpenWindows 3.0 undefined symbols when using shared `libXmu'.
551 The way to statically link with `libXmu' depends on whether you are
552 using a Sun compiler (e.g., `cc') or `gcc'. If the latter, alter the
553 `x_libs' Make variable to include
555 -static -lXmu -dynamic
557 If you are using the Sun compiler, use `-Bstatic' and `-Bdynamic'.
559 0.1.5.9 Pointer combination warnings
560 ....................................
562 When compiling with old C compilers, you may get some warnings about
563 "illegal pointer combinations". These are spurious; just ignore them.
564 I decline to clutter up the source with casts to get rid of them.