Kpathsea: A library for path searching

Table of Contents

Next: , Up: (dir)

Kpathsea library

This manual documents how to install and use the Kpathsea library for filename lookup. It corresponds to version 6.1.0, released in May 2012.

Next: , Previous: Top, Up: Top

1 Introduction

This manual corresponds to version 6.1.0 of the Kpathsea library, released in May 2012.

The library's fundamental purpose is to return a filename from a list of directories specified by the user, similar to what shells do when looking up program names to execute.

The following software, all of which we maintain, uses this library:

Other software that we do not maintain also uses it.

We are still actively maintaining the library (and probably always will be, despite our hopes). If you have comments or suggestions, please send them to us (see Reporting bugs).

We distribute the library under the GNU Library General Public License (LGPL). In short, this means if you write a program using the library, you must (offer to) distribute the source to the library, along with any changes you have made, and allow anyone to modify the library source and distribute their modifications. It does not mean you have to distribute the source to your program, although we hope you will. See the accompanying files for the text of the GNU licenses.

If you know enough about TeX to be reading this manual, then you (or your institution) should consider joining the TeX Users Group (if you're already a member, thanks!). TUG produces the periodical TUGboat, sponsors an annual meeting and publishes the proceedings, and arranges courses on TeX for all levels of users throughout the world. See for information.

Up: Introduction

1.1 History

(This section is for those people who are curious about how the library came about.) (If you like to read historical accounts of software, we urge you to seek out the GNU Autoconf manual and the “Errors of TeX” paper by Don Knuth, published in Software—Practice and Experience 19(7), July 1989.)

[Karl writes.] My first ChangeLog entry for Web2c seems to be February 1990, but I may have done some work before then. In any case, Tim Morgan and I were jointly maintaining it for a time. (I should mention here that Tim had made Web2c into a real distribution long before I had ever used it or even heard of it, and Tom Rokicki did the original implementation. I was using pxp and pc on VAX 11/750's and the hot new Sun 2 machines.)

It must have been later in 1990 and 1991 that I started working on TeX for the Impatient. Dvips, Xdvi, Web2c, and the GNU fontutils (which I was also writing at the time) all used different environment variables, and, more importantly, had different bugs in their path searching. This became extremely painful, as I was stressing everything to the limit working on the book. I also desperately wanted to implement subdirectory searching, since I couldn't stand putting everything in one big directory, and also couldn't stand having to explicitly specify cm, pandora, ... in a path.

In the first incarnation, I just hacked separately on each program—that was the original subdirectory searching code in both Xdvi and Dvips, though I think Paul Vojta has completely rewritten Xdvi's support by now. That is, I tried to go with the flow in each program, rather than changing the program's calling sequences to conform to common routines.

Then, as bugs inevitably appeared, I found I was fixing the same thing three times (Web2c and fontutils were always sharing code, since I maintained those—there was no Dvipsk or Xdvik or Dviljk at this point). After a while, I finally started sharing source files. They weren't yet a library, though. I just kept things up to date with shell scripts. (I was developing on a 386 running ISC 2.2 at the time, and so didn't have symbolic links. An awful experience.)

The ChangeLogs for Xdvik and Dvipsk record initial releases of those distributions in May and June 1992. I think it was because I was tired of the different configuration strategies of each program, not so much because of the path searching. (Autoconf was being developed by David MacKenzie and others, and I was adapting it to TeX and friends.)

I started to make a separate library that other programs could link with on my birthday in April 1993, according to the ChangeLog. I don't remember exactly why I finally took the time to make it a separate library; a conversation with david zuhn that initiated it. Just seemed like it was time.

Dviljk got started in March 1994 after I bought a Laserjet 4. (Kpathsea work got suspended while Norm Walsh and I, with Gustaf Neumann's help, implemented a way for TeX to get at all those neat builtin LJ4 fonts ... such a treat to have something to typeset in besides Palatino!)

By spring of 1995, I had implemented just about all the path-searching features in Kpathsea that I plan to, driven beyond my initial goals by Thomas Esser and others. I then started to integrate Web2c with Kpathsea. After the release of a stable Web2c, I hope to be able to stop development, and turn most of my attention back to making fonts for GNU. (Always assuming Micros**t hasn't completely obliterated Unix by then, or that software patents haven't stopped software development by anybody smaller than a company with a million-dollar-a-year legal budget. Which is actually what I think is likely to happen, but that's another story...)

[Olaf writes.] At the end of 1997, UNIX is still alive and kicking, individuals still develop software, and Web2c development still continues. Karl had been looking for some time for someone to take up part of the burden, and I volunteered.

Next: , Previous: Introduction, Up: Top

2 Installation

(A copy of this chapter is in the distribution file kpathsea/INSTALL.)

The procedure for Kpathsea (and Web2c, etc.) configuration and installation follows. If you encounter trouble, see Common problems, a copy of which is in the file kpathsea/BUGS.

Next: , Up: Installation

2.1 Simple installation

Installing TeX and friends for the first time can be a daunting experience. Thus, you may prefer to skip this whole thing and just get precompiled executables: see unixtex.ftp.

This section explains what to do if you wish to take the defaults for everything, and generally to install in the simplest possible way. Most steps here refer to corresponding subsection in the next section which explains how to override defaults and generally gives more details.

By default everything will be installed under /usr/local and the following discussion assumes this. However, if you already have TeX installed, its location is used to derive the directory under which everything is to be installed.

  1. Be sure you have enough disk space: approximately 8 megabytes for the compressed archives, 15MB for sources, 50MB for compilation, 40MB for the (initial) installed system (including library files). See Disk space.
  2. Retrieve these distribution archives:
    These are the sources, which you will be compiling.
    This is a basic set of input files. You should unpack it in the directory /usr/local/share; doing so will create a texmf subdirectory there.

    These archives are mirrored on the CTAN hosts, in the systems/web2c directory.

    See Kpathsea application distributions.

  3. When using the default search paths, there is no need to edit any distribution files. See Changing search paths.
  4. At the top level of the distribution, run ‘sh configure’. (If you have the GNU Bash shell installed, run ‘bash configure’.) See Running configure.
  5. make’. See Running make. If you are using a BSD 4.4 system such as FreeBSD or NetBSD, you may have to use GNU make (often installed in /usr/local/bin), not the BSD make.
  6. make install’. See Installing files.
  7. make distclean’. See Cleaning up.
  8. Set up a cron job to rebuild the filename database that makes searching faster. This line will rebuild it every midnight:
              0 0 * * * cd /usr/local/share/texmf && /bindir/mktexlsr

    See Filename database generation, and Filename database.

  9. If you're installing Dvips, you also need to set up configuration files for your printers and make any additional PostScript fonts available. See Installation. If you have any color printers, see Color device configuration.
  10. The first time you run a DVI driver, a bunch of PK fonts will be built by Metafont via mktexpk (and added to the filename database). This will take some time. Don't be alarmed; they will created only this first time (unless something is wrong with your path definitions).

    By default, mktexpk will create these fonts in a hierarchy under /var/tmp/texfonts; it simply assumes that /var/tmp exists and is globally writable. If you need a different arrangement, see mktex configuration.

    See mktex scripts.

  11. For some simple tests, try ‘tex story \\bye’ and ‘latex sample2e’. Then run xdvi story or dvips sample2e on the resulting DVI files to preview/print the documents. See Installation testing.

Next: , Previous: Simple installation, Up: Installation

2.2 Custom installation

Most sites need to modify the default installation procedure in some way, perhaps merely changing the prefix from ‘/usr/local’, perhaps adding extra compiler or loader options to work around configure bugs. This section explains how to override default choices. For additional distribution-specific information:

These instructions are for Unix systems. Other operating-system specific distributions have their own instructions. The code base itself supports Amiga, DOS, OS/2, and VMS.

Following are the same steps as in the previous section (which describes the simplest installation), but with much more detail.

Next: , Up: Custom installation

2.2.1 Disk space

Here is a table showing the disk space needed for each distribution (described in the next section). The `(totals)' line reflects the ‘texk’ source distribution and ‘texklib’; the individual distributions don't enter into it. Sizes are in megabytes. All numbers are approximate.

Distribution .tar.gz Unpacked Compiled Installed
dviljk .9 3.8
dvipsk .9 3.2
xdvik .7 2.5
web2c 1.3 5.0
web 1.9 6.5 - -
texk 7.5 32.1 95.3 33.5
texklib 6.3 15.0 - 15.0
(totals) 14.6 47.1 95.3 48.5

Next: , Previous: Disk space, Up: Custom installation

2.2.2 Kpathsea application distributions

The archive contains all of the Kpathsea applications I maintain, and the library itself. For example, since NeXT does not generally support X11, you'd probably want to skip ‘xdvik’ (or simply remove it after unpacking texk.tar.gz. If you are not interested in all of them, you can also retrieve them separately:

DVI to PCL, for LaserJet printers.
DVI to PostScript, for previewers, printers, or PDF generation.
The software needed to compile TeX and friends.
The original WEB source files, also used in compilation.
DVI previewing under the X window system.

If you want to use the Babel LaTeX package for support of non-English typesetting, you may need to retrieve additional files. See the file install.txt in the Babel distribution.

Next: , Previous: Kpathsea application distributions, Up: Custom installation

2.2.3 Changing search paths

If the search paths for your installation differ from the standard TeX directory structure (see Introduction), edit the file kpathsea/ as desired, before running configure. For example, if you have all your fonts or macros in one big directory.

You may also wish to edit the file mktex.cnf, either before or after installation, to control various aspects of mktexpk and friends. See mktex configuration.

You do not need to edit to change the default top-level or other installation directories (only the paths). You can and should do that when you run configure (next step).

You also do not need to edit if you are willing to rely on texmf.cnf at runtime to define the paths, and let the compile-time default paths be incorrect. Usually there is no harm in doing this.

The section below explains default generation in more detail.

Next: , Up: Changing search paths Default path features

The purpose of having all the different files described in the section above is to avoid having the same information in more than one place. If you change the installation directories or top-level prefix at configure-time, those changes will propagate through the whole sequence. And if you change the default paths in, those changes are propagated to the compile-time defaults.

The Make definitions are all repeated in several Makefile's; but changing the top-level Makefile should suffice, as it passes down all the variable definitions, thus overriding the submakes. (The definitions are repeated so you can run Make in the subdirectories, if you should have occasion to.)

By default, the bitmap font paths end with ‘/$MAKETEX_MODE’, thus including the device name (usually a Metafont mode name such as ‘ljfour’). This distinguishes two different devices with the same resolution—a write/white from a write/black 300dpi printer, for example.

However, since most sites don't have this complication, Kpathsea (specifically, the kpse_init_prog function in kpathsea/proginit.c) has a special case: if the mode has not been explicitly set by the user (or in a configuration file), it sets MAKETEX_MODE to /. This makes the default PK path, for example, expand into .../pk//, so fonts will be found even if there is no subdirectory for the mode (if you arranged things that way because your site has only one printer, for example) or if the program is mode-independent (e.g., pktype).

To make the paths independent of the mode, simply edit before installation, or the installed texmf.cnf, and remove the ‘$MAKETEX_MODE’.

See mktex script arguments, for how this interacts with mktexpk.

See TeX directory structure, for a description of the default arrangement of the input files that comprise the TeX system.

Previous: Default path features, Up: Changing search paths Default path generation

This section describes how the default paths are constructed.

You may wish to ignore the whole mess and simply edit texmf.cnf after it is installed, perhaps even copying it into place beforehand so you can complete the installation, if it seems necessary.

To summarize the chain of events that go into defining the default paths:

  1. configure’ creates a Makefile from each
  2. When Make runs in the kpathsea directory, it creates a file texmf.sed that substitutes the Make value of $(var) for a string @var@. The variables in question are the one that define the installation directories.
  3. texmf.sed (together with a little extra magic—see kpathsea/Makefile) is applied to to generate texmf.cnf. This is the file that will eventually be installed and used.
  4. The definitions in texmf.cnf are recast as C #define's in paths.h. These values will be the compile-time defaults; they are not used at runtime unless no texmf.cnf file can be found.

    (That's a lie: the compile-time defaults are what any extra :'s in texmf.cnf expand into; but the paths as distributed have no extra :'s, and there's no particular reason for them to.)

Next: , Previous: Changing search paths, Up: Custom installation

2.2.4 Running configure

Run sh configure options (in the top-level directory, the one containing kpathsea/), possibly using a shell other than sh (see configure shells).

configure adapts the source distribution to the present system via #define's in */c-auto.h, which are created from the corresponding It also creates a Makefile from the corresponding, doing ‘@var@’ and ‘ac_include’ substitutions).

configure is the best place to control the configuration, compilation, and installed location of the software, either via command-line options, or by setting environment variables before invoking it. For example, you can disable mktexpk by default with the option ‘--disable-mktexpk’. See configure options.

Next: , Up: Running configure configure shells

Considerable effort has gone into trying to ensure that the configure scripts can be run by most Bourne shell variants. If sh runs into trouble, your best bet is to use Bash, the GNU Bourne-again shell (see Top).

Bourne shell variants for which problems have been reported in the past are:

Old versions of the Korn shell may fail to handle the scripts. The Korn shell may be installed as /bin/sh on AIX, in which case /bin/bsh may serve instead.
Old versions of ash are unable to handle the scripts. Ash is sometimes installed as /bin/sh on NetBSD, FreeBSD, and Linux systems. /bin/bash should be available for those systems, but might not be part of a default installation.
Ultrix /bin/sh
/bin/sh under Ultrix is a DEC-grown shell that is notably deficient in many ways. /bin/sh5 may be necessary.

Next: , Previous: configure shells, Up: Running configure configure options

For a complete list of all configure options, run ‘configure --help’ or see Running configure scripts, (a copy is in the file kpathsea/README.CONFIGURE). The generic options are listed first in the ‘--help’ output, and the package-specific options come last. The environment variables configure pays attention to are listed below.

Options particularly likely to be useful are ‘--prefix’, ‘--datadir’, and the like; see configure scenarios.

This section gives pointers to descriptions of the ‘--with’ and ‘--enable’ options to configure that Kpathsea-using programs accept.

Enable or disable the dynamic generation programs. See mktex configuration.
Build Kpathsea as a shared library, and link against it. Also build the usual static library. See Shared library.
Build only the shared library. Implies ‘--enable-shared’.
Enables make targets that are useful for the maintainer and likely to be a pain for anyone else; the makefiles created when this option is enabled may not work at all for you. You have been warned.

Next: , Previous: configure options, Up: Running configure configure environment

configure uses the value of the following environment variables in determining your system's characteristics, and substitutes for them in Makefile's:

The compiler to use: default is gcc if it's installed, otherwise cc.
Options to give the compiler: default is ‘-g -O2’ for gcc, ‘-g’ otherwise. CFLAGS comes after any other options. You may need to include -w here if your compilations commonly have useless warnings (e.g., NULL redefined), or configure may fail to detect the presence of header files (it takes the messages on standard error to mean the header file doesn't exist).
Options to pass to the compiler preprocessor; this matters most for configuration, not the actual source compilation. The configure script often does only preprocessing (e.g., to check for the existence of #include files), and CFLAGS is not used for this. You may need to set this to something like ‘-I/usr/local/include/wwwhatever’ if you have the libwww library installed for hyper-xdvik (see xdvik/INSTALL).
Additional preprocessor options, but not used by configure. Provided for enabling or disabling program features, as documented in the various program-specific installation instructions. DEFS comes before any compiler options included by the distribution Makefiles or by configure.
Additional options to give to the loader. LDFLAGS comes before any other linker options.
Additional libraries to link with.

Next: , Previous: configure environment, Up: Running configure configure scenarios

Here are some common installation scenarios:

Previous: configure scenarios, Up: Running configure Shared library

You can compile Kpathsea as a shared library on a few systems, by specifying the option ‘--enable-shared’ when you run ‘configure’.

The main advantage in doing this is that the executables can then share the code, thus decreasing memory and disk space requirements.

On some systems, you can record the location of shared libraries in a binary, usually by giving certain options to the linker. Then individual users do not need to set their system's environment variable (e.g., LD_LIBRARY_PATH) to find shared libraries. If you want to do this, you will need to add the necessary options to LDFLAGS yourself; for example, on Solaris, include something like ‘-R${prefix}/lib’, on IRIX or Linux, use ‘-rpath${prefix}/lib’. (Unfortunately, making this happen by default is very difficult, because of interactions with an existing installed shared library.)

Next: , Previous: Running configure, Up: Custom installation

2.2.5 Running make

make (still in the top-level directory). This also creates the texmf.cnf and paths.h files that define the default search paths, and (by default) the ‘plain’ and ‘latex’ TeX formats.

You can override directory names and other values at make-time. make/paths.make lists the variables most commonly reset. For example, ‘make default_texsizes=600’ changes the list of fallback resolutions.

You can also override each of configure's environment variables (see configure environment). The Make variables have the same names.

Finally, you can supply additional options via the following variables. (configure does not use these.)

Preprocessor options.
Compiler options.
Loader options (included at beginning of link commands).
More loader options (included at end of link commands).
Additional Make arguments passed to all sub-make's. You may need to include assignments to the other variables here via XMAKEARGS; for example: ‘make XMAKEARGS="CFLAGS=-O XDEFS=-DA4"’.

It's generally a bad idea to use a different compiler (‘CC’) or libraries (LIBS) for compilation than you did for configuration, since the values configure determined may then be incorrect.

Adding compiler options to change the “universe” you are using (typically BSD vs. system V) is generally a cause of trouble. It's best to use the native environment, whatever that is; configure and the software usually adapt best to that. In particular, under Solaris 2.x, you should not use the BSD-compatibility library (libucb) or include files (ucbinclude).

If you want to use the Babel LaTeX package for support of non-English typesetting, you need to modify some files before making the LaTeX format. See the file install.txt in the Babel distribution.

Next: , Previous: Running make, Up: Custom installation

2.2.6 Installing files

The basic command is the usual make install. For security issues, see Security.

The first time you install any manual in the GNU Info system, you should add a line (you choose where) to the file dir in your ‘$(infodir)’ directory. Sample text for this is given near the top of the Texinfo source files (kpathsea/kpathsea.texi, dvipsk/dvips.texi, and web2c/doc/web2c.texi). If you have a recent version of the GNU Texinfo distribution installed ( or later), this should happen automatically.

On the offchance that this is your first Info installation, the dir file I use is included in the distribution as etc/dir-example.

You may wish to use one of the following targets, especially if you are installing on multiple architectures:

If you use the Andrew File System, the normal path (e.g., prefix/bin) only gets you to a read-only copy of the files, and you must specify a different path for installation. The best way to do this is by setting the ‘prefix’ variable on the make command line. The sequence becomes something like this:

     configure --prefix=/whatever
     make install prefix=/afs/

With AFS, you will definitely want to use relative filenames in ls-R (see Filename database), not absolute filenames. This is done by default, but check anyway.

Next: , Previous: Installing files, Up: Custom installation

2.2.7 Cleaning up

The basic command is make distclean. This removes all files created by the build.


Next: , Previous: Cleaning up, Up: Custom installation

2.2.8 Filename database generation

You will probably want to set up a cron entry on the appropriate machine(s) to rebuild the filename database nightly or so, as in:

     0 0 * * * cd texmf && /bindir/mktexlsr

See Filename database.

Although the mktex... scripts make every effort to add newly-created files on the fly, it can't hurt to make sure you get a fresh version every so often.

Next: , Previous: Filename database generation, Up: Custom installation

2.2.9 mktex scripts

If Kpathsea cannot otherwise find a file, for some file types it is configured by default to invoke an external program to create it dynamically (see mktex configuration). These are collectively known as mktex scripts, since most of them are named mktex....

For example, this is useful for fonts (bitmaps, TFM's, and arbitrarily-sizable Metafont sources such as the Sauter and EC fonts), since any given document can use fonts never before referenced. Building all fonts in advance is therefore impractical, if not impossible.

It is also useful for the TeX ‘.fmt’ (and Metafont ‘.base’ and Metapost ‘.mem’ files, see Memory dumps), where pre-generating every format consumes a lot of both time and space.

The script is passed the name of the file to create and possibly other arguments, as explained below. It must echo the full pathname of the file it created (and nothing else) to standard output; it can write diagnostics to standard error.

Next: , Up: mktex scripts mktex configuration

The list of file types and program names that can run an external program to create missing files is listed in the next section. In the absence of configure options specifying otherwise, everything but mktextex will be enabled by default. The configure options to change the defaults are:


The configure setting is overridden if the environment variable or configuration file value named for the script is set; e.g., MKTEXPK (see mktex script arguments).

mktexfmt reads a file fmtutil.cnf, typically located in texmf/web2c/ to glean its configuration information. The rest of the files and features in this section are primarily intended for the font generation scripts.

As distributed, all the scripts source a file texmf/web2c/mktex.cnf if it exists, so you can override various defaults. See mktex.opt, for instance, which defines the default mode, resolution, some special directory names, etc. If you prefer not to change the distributed scripts, you can simply create mktex.cnf with the appropriate definitions (you do not need to create it if you have nothing to put in it). mktex.cnf has no special syntax; it's an arbitrary Bourne shell script. The distribution contains a sample mktex.cnf for you to copy and modify as you please (it is not installed anywhere).

In addition, you can configure a number of features with the MT_FEATURES variable, which you can define:

If none of the options below are enabled, mktexpk, mktextfm, and mktexmf follow the following procedure to decide where fonts should be installed. Find the tree where the font's sources are, and test the permissions of the ‘fonts’ directory of that tree to determine whether it is writable. If it is, put the files in the tree in appropriate locations. If it isn't writable, see whether the tree is a system tree (named in SYSTEXMF). If so, the VARTEXFONTS tree is used. In all other cases the working directory is used.

The ‘appendonlydir’ option is enabled by default.

Tell mktexdir to create directories append-only, i.e., set their sticky bit (see Mode Structure). This feature is silently ignored on non-Unix platforms (e.g. Windows/NT and MS-DOS) which don't support similar functionality. This feature is enabled by default.
Use 8.3 names; e.g., dpi600/ instead of cmr10.600pk. Note that this feature only affects filenames that would otherwise clash with other TeX-related filenames; mktex scripts do nothing about filenames which exceed the 8+3 MS-DOS limits but remain unique when truncated (by the OS) to these limits, and nether do the scripts care about possible clashes with files which aren't related with TeX. For example, cmr10.600pk would clash with cmr10.600gf and is therefore changed when ‘dosnames’ is in effect, but mf.pool and mp.base don't clash with any TeX-related files and are therefore unchanged.

This feature is turned on by default on MS-DOS. If you do not wish ‘dosnames’ to be set on an MS-DOS platform, you need to set the MT_FEATURES environment variable to a value that doesn't include ‘dosnames’. You can also change the default setting by editing mktex.opt, but only if you use the mktex shell scripts; the emulation programs don't consult mktex.opt.

Instead of deriving the location of a font in the destination tree from the location of the sources, the aliases and directory names from the Fontname distribution are used. (see Introduction).
Let mktexpk and mktextfm create metafont driver files in a temporary directory. These will be used for just one metafont run and not installed permanently.
Omit the directory level for the mode name; this is fine as long as you generate fonts for only one mode.
Omit the font supplier name directory level.
Omit the font typeface name directory level.
Omit the font supplier and typeface name directory levels. This feature is deprecated in favour of ‘stripsupplier’ and ‘striptypeface’.
When this option is enabled, fonts that would otherwise be written in system texmf tree go to the VARTEXFONTS tree instead. The default value in kpathsea/ is /var/tmp/texfonts. The Linux File System Standard recommends /var/tex/fonts.

The ‘varfonts’ setting in MT_FEATURES is overridden by the USE_VARTEXFONTS environment variable: if set to ‘1’, the feature is enabled, and if set to ‘0’, the feature is disabled.

Force generated files that would go into a system tree (as defined by SYSTEXMF) into TEXMFVAR. Starting with teTeX-3.0, the variable TEXMFVAR is always set. The ‘varfonts’ feature takes precedence if also set.

The ‘texmfvar’ setting in MT_FEATURES is overridden by the USE_TEXMFVAR environment variable: if set to ‘1’, the feature is enabled, and if set to ‘0’, the feature is disabled.

Next: , Previous: mktex configuration, Up: mktex scripts mktex script names

The following table shows the default name of the script for each of the file types which support runtime generation.

(‘.fmt’, ‘.base’, ‘.mem’) TeX/Metafont/MetaPost formats. This script is also named fmtutil, and reads fmtutil.cnf for configuration information.
(‘.mf’) Metafont input files.
(‘.ocp’) Omega compiled process files.
(‘.ofm’) Omega font metric files.
(‘pk’) Glyph fonts.
(‘.tex’) TeX input files (disabled by default).
(‘.tfm’) TFM files.

These names can be overridden by an environment variable specific to the program—for example, DVIPSMAKEPK for Dvipsk.

If a mktex... script fails, the invocation is appended to a file missfont.log (by default) in the current directory. You can then execute the log file to create the missing files after fixing the problem.

If the current directory is not writable and the environment variable or configuration file value TEXMFOUTPUT is set, its value is used. Otherwise, nothing is written. The name ‘missfont.log’ is overridden by the MISSFONT_LOG environment variable or configuration file value.

Previous: mktex script names, Up: mktex scripts mktex script arguments

The first argument to a mktex script is always the name of the file to be created.

In the default mktexpk implementation, additional arguments may also be passed:

--dpi num
Sets the resolution of the generated font to num.
--mfmode name
Sets the Metafont mode to name.
--bdpi num
Sets the “base dpi” for the font. This must match the mode being used.
--mag string
A “magstep” string suitable for the Metafont mag variable. This must match the combination of bdpi and dpi being used.
--destdir string
A directory name. If the directory is absolute, it is used as-is. Otherwise, it is appended to the root destination directory set in the script.

Previous: mktex scripts, Up: Custom installation

2.2.10 Installation testing

Besides the tests listed in Simple installation, you can try running ‘make check’. This includes the torture tests (trip, trap, and mptrap) that come with Web2c (see Triptrap).

Next: , Previous: Custom installation, Up: Installation

2.3 Security

None of the programs in the TeX system require any special system privileges, so there's no first-level security concern of people gaining illegitimate root access.

A TeX document, however, can write to arbitrary files, e.g., ~/.rhosts, and thus an unwitting user who runs TeX on a random document is vulnerable to a trojan horse attack. This loophole is closed by default, but you can be permissive if you so desire in texmf.cnf. See tex invocation. MetaPost has the same issue.

Dvips, Xdvi, and TeX can also execute shell commands under some circumstances. To disable this, see the ‘-R’ option in Option details, the xdvi man page, and tex invocation, respectively.

Another security issue arises because it's very useful—almost necessary—to make arbitrary fonts on user demand with mktexpk and friends. Where do these files get installed? By default, the mktexpk distributed with Kpathsea assumes a world-writable /var/tmp directory; this is a simple and convenient approach, but it may not suit your situation because it means that a local cache of fonts is created on every machine.

To avoid this duplication, many people consider a shared, globally writable font tree desirable, in spite of the potential security problems. To do this you should change the value of VARTEXFONTS in texmf.cnf to refer to some globally known directory. See mktex configuration.

The first restriction you can apply is to make newly-created directories under texmf be append-only with an option in mktex.cnf. See mktex configuration.

Another approach is to establish a group (or user) for TeX files, make the texmf tree writable only to that group (or user), and make mktexpk et al. setgid to that group (or setuid to that user). Then users must invoke the scripts to install things. (If you're worried about the inevitable security holes in scripts, then you could write a C wrapper to exec the script.)

The mktex... scripts install files with the same read and write permissions as the directory they are installed in. The executable, sgid, suid, and sticky bits are always cleared.

Any directories created by the mktex... scripts have the same permissions as their parent directory, unless the appendonlydir feature is used, in which case the sticky bit is always set.

Next: , Previous: Security, Up: Installation

2.4 TeX directory structure

This section describes the default installation hierarchy of the distribution. It conforms to both the GNU coding standards and the TeX directory structure (TDS) standard. For rationale and further explanation, please see those documents. The GNU standard is available as and mirrors. The TDS document is available from CTAN:/tex-archive/tds (see unixtex.ftp).

You can change the default paths in many ways (see Changing search paths). One common desire is to put everything (binaries and all) under a single top-level directory such as /usr/local/texmf or /opt/texmf—in the terms used below, make prefix and texmf the same. For specific instructions on doing that, see configure scenarios.

Here is a skeleton of the default directory structure, extracted from the TDS document:

     prefix/      installation root (/usr/local by default)
      bin/         executables
      man/         man pages
      include/     C header files
      info/        GNU info files
      lib/         libraries (libkpathsea.*)
      share/       architecture-independent files
       texmf/      TDS root
        bibtex/     BibTeX input files
         bib/        BibTeX databases
          base/       base distribution (e.g., ‘xampl.bib’)
          misc/       single-file databases
          pkg/       name of a package
         bst/        BibTeX style files
          base/       base distribution (e.g., ‘plain.bst’, ‘acm.bst’)
          misc/       single-file styles
          pkg/       name of a package
        doc/         additional documentation
        dvips/       .pro’, ‘.ps’, ‘
        fonts/       font-related files
         type/         file type (e.g., ‘tfm’, ‘pk’)
          mode/          type of output device (types ‘pk’ and ‘gf’ only)
           supplier/       name of a font supplier (e.g., ‘public’)
            typeface/        name of a typeface (e.g., ‘cm’)
             dpinnn/           font resolution (types ‘pk’ and ‘gf’ only)
        metafont/    Metafont (non-font) input files
         base/        base distribution (e.g., ‘’)
         misc/        single-file packages (e.g., ‘’)
         pkg/           name of a package (e.g., ‘mfpic’)
        metapost/    MetaPost input files
         base/        base distribution (e.g., ‘’)
         misc/        single-file packages
         pkg/           name of a package
         support/     support files for MetaPost-related utilities (e.g., ‘’)
        mft/         MFT’ inputs (e.g., ‘plain.mft’)
        tex/         TeX input files
         format/         name of a format (e.g., ‘plain’)
          base/        base distribution for format (e.g., ‘plain.tex’)
          misc/        single-file packages (e.g., ‘webmac.tex’)
          local/       local additions to or local configuration files for format
          pkg/           name of a package (e.g., ‘graphics’, ‘mfnfss’)
         generic/     format-independent packages
          hyphen/      hyphenation patterns (e.g., ‘hyphen.tex’)
          images/      image input files (e.g., Encapsulated PostScript)
          misc/        single-file format-independent packages (e.g., ‘null.tex’).
          pkg/           name of a package (e.g., ‘babel’)
        web2c/        implementation-dependent files (.pool, .fmt, texmf.cnf, etc.)

Some concrete examples for most file types:


Next: , Previous: TeX directory structure, Up: Installation

2.5 unixtex.ftp: Obtaining TeX

This is, last updated 13 June 2010. Also available as Email with comments or questions.

The principal free TeX distribution for Unix-like systems is TeX Live, on the web at The pages there describe many ways to acquire TeX, over the Internet or on physical media, both the sources and precompiled binaries for many systems, either standalone or as part of various operating system distributions.

Web2C, Kpathsea, Dvips, and Dviljk are no longer released as a separate packages. Their sources are now maintained as part of TeX Live.

The host is the original source for the files for which Donald Knuth is directly responsible: tex.web, plain.tex, etc. However, unless you want to build your TeX library tree ab initio, it is more reliable and less work to retrieve these files as part of a larger package. In any case, that ftp site is not the canonical source for anything except what was created as part of Stanford TeX project, so do not rely on the other files available there being up-to-date.

Previous: unixtex.ftp, Up: Installation

2.6 Reporting bugs

(A copy of this chapter is in the file kpathsea/BUGS.)

If you have problems or suggestions, please report them to using the bug checklist below.

Please report bugs in the documentation; not only factual errors or inconsistent behavior, but unclear or incomplete explanations, typos, wrong fonts, ...

Next: , Up: Reporting bugs

2.6.1 Bug checklist

Before reporting a bug, please check below to be sure it isn't already known (see Common problems).

Bug reports should be sent via electronic mail to

The general principle is that a good bug report includes all the information necessary for reproduction. Therefore, to enable investigation, your report should include the following:

Next: , Previous: Bug checklist, Up: Reporting bugs

2.6.2 Mailing lists

Web2c and Kpathsea in general are discussed on the mailing list To join, email with a line consisting of


in the body of the message.

You do not need to join to submit a report, nor will it affect whether you get a response. There is no Usenet newsgroup equivalent (if you can be the one to set this up, email ‘tex-k-request’). Traffic on the list is fairly light, and is mainly bug reports and enhancement requests to the software. The best way to decide if you want to join or not is read some of the archives from

Be aware that large data files are sometimes included in bug reports. If this is a problem for you, do not join the list.

If you are looking for general TeX help, such as how to use LaTeX, please use the mailing list mailing list ( which is gatewayed to the ‘comp.text.tex’ Usenet newsgroup (or post to the newsgroup; the gateway is bidirectional).

Next: , Previous: Mailing lists, Up: Reporting bugs

2.6.3 Debugging

Kpathsea provides a number of runtime debugging options, detailed below by their names and corresponding numeric values. When the files you expect aren't being found, the thing to do is enable these options and examine the output.

You can set these with some runtime argument (e.g., ‘-d’) to the program; in that case, you should use the numeric values described in the program's documentation (which, for Dvipsk and Xdvik, are different than those below). It's best to give the ‘-d’ (or whatever) option first, for maximal output. Dvipsk and Xdvik have additional program-specific debugging options as well.

You can also set the environment variable KPATHSEA_DEBUG; in this case, you should use the numbers below. If you run the program under a debugger and set the instance variable kpse->debug, also use the numbers below.

In any case, by far the simplest value to use is ‘-1’, which will turn on all debugging output. This is usually better than guessing which particular values will yield the output you need.

Debugging output always goes to standard error, so you can redirect it easily. For example, in Bourne-compatible shells:

     dvips -d -1 ... 2>/tmp/debug

It is sometimes helpful to run the standalone Kpsewhich utility (see Invoking kpsewhich), instead of the original program.

In any case, you can not use the names below; you must always use somebody's numbers. (Sorry.) To set more than one option, just sum the corresponding numbers.

Report ‘stat’(2) calls. This is useful for verifying that your directory structure is not forcing Kpathsea to do many additional file tests (see Slow path searching, and see Subdirectory expansion). If you are using an up-to-date ls-R database (see Filename database), this should produce no output unless a nonexistent file that must exist is searched for.
Report lookups in all hash tables: ls-R and aliases (see Filename database); font aliases (see Fontmap); and config file values (see Config files). Useful when expected values are not being found, e.g.., file searches are looking at the disk instead of using ls-R.
Report file openings and closings. Especially useful when your system's file table is full, for seeing which files have been opened but never closed. In case you want to set breakpoints in a debugger: this works by redefining ‘fopen’ (‘fclose’) to be ‘kpse_fopen_trace’ (‘kpse_fclose_trace’).
Report general path information for each file type Kpathsea is asked to search. This is useful when you are trying to track down how a particular path got defined—from texmf.cnf,, an environment variable, the compile-time default, etc. This is the contents of the kpse_format_info_type structure defined in tex-file.h.
Report the directory list corresponding to each path element Kpathsea searches. This is only relevant when Kpathsea searches the disk, since ls-R searches don't look through directory lists in this way.
Report on each file search: the name of the file searched for, the path searched in, whether or not the file must exist (when drivers search for cmr10.vf, it need not exist), and whether or not we are collecting all occurrences of the file in the path (as with, e.g., texmf.cnf and, or just the first (as with most lookups). This can help you correlate what Kpathsea is doing with what is in your input file.
Report the value of each variable Kpathsea looks up. This is useful for verifying that variables do indeed obtain their correct values.
Activates debugging printout specific to gsftopk program.
If you use the optional mktex programs instead of the traditional shell scripts, this will report the name of the site file (mktex.cnf by default) which is read, directories created by mktexdir, the full path of the ls-R database built by mktexlsr, font map searches, MT_FEATURES in effect, parameters from mktexnam, filenames added by mktexupd, and some subsidiary commands run by the programs.
When the optional mktex programs are used, this will print additional debugging info from functions internal to these programs.

Debugging output from Kpathsea is always written to standard error, and begins with the string ‘kdebug:’. (Except for hash table buckets, which just start with the number, but you can only get that output running under a debugger. See comments at the hash_summary_only variable in kpathsea/db.c.)

Next: , Previous: Debugging, Up: Reporting bugs

2.6.4 Logging

Kpathsea can record the time and filename found for each successful search. This may be useful in finding good candidates for deletion when your filesystem is full, or in discovering usage patterns at your site.

To do this, define the environment or config file variable TEXMFLOG. The value is the name of the file to append the information to. The file is created if it doesn't exist, and appended to if it does.

Each successful search turns into one line in the log file: two words separated by a space. The first word is the time of the search, as the integer number of seconds since “the epoch”, i.e., UTC midnight 1 January 1970 (more precisely, the result of the time system call). The second word is the filename.

For example, after setenv TEXMFLOG /tmp/log, running Dvips on story.dvi appends the following lines:

     774455887 /usr/local/share/texmf/dvips/
     774455887 /usr/local/share/texmf/dvips/
     774455888 /usr/local/share/texmf/dvips/
     774455888 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmbx10.600pk
     774455889 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmsl10.600pk
     774455889 /usr/local/share/texmf/fonts/pk/ljfour/public/cm/cmr10.600pk
     774455889 /usr/local/share/texmf/dvips/

Only filenames that are absolute are recorded, to preserve some semblance of privacy.

In addition to this Kpathsea-specific logging, pdftex provides an option -recorder to write the names of all files accessed during a run to the file basefile.fls.

Finally, most systems provide a general tool to output each system call, thus including opening and closing files. It might be named strace, truss, struss, or something else.

Previous: Logging, Up: Reporting bugs

2.6.5 Common problems

Here are some common problems with configuration, compilation, linking, execution, ...

Next: , Up: Common problems Unable to find files

If a program complains it cannot find fonts (or other input files), any of several things might be wrong. In any case, you may find the debugging options helpful. See Debugging.

Next: , Previous: Unable to find files, Up: Common problems Slow path searching

If your program takes an excessively long time to find fonts or other input files, but does eventually succeed, here are some possible culprits:

In any case, you may find the debugging options helpful in determining precisely when the disk or network is being pounded. See Debugging.

Next: , Previous: Slow path searching, Up: Common problems Unable to generate fonts

Metafont outputs fonts in bitmap format, tuned for a particular device at a particular resolution, in order to allow for the highest-possible quality of output. Some DVI-to-whatever programs, such as Dvips, try to generate these on the fly when they are needed, but this generation may fail in several cases.

If mktexpk runs, but fails with this error:

     mktexpk: Can't guess mode for nnn dpi devices.
     mktexpk: Use a config file to specify the mode, or update me.

you need to ensure the resolution and mode match; just specifying the resolution, as in -D 360, is not enough.

You can specify the mode name with the -mode option on the Dvips command line, or in a Dvips configuration file (see Config files), such as in your document directory, ~/.dvipsrc in your home directory, or in a system directory (again named (Other drivers use other files, naturally.)

For example, if you need 360dpi fonts, you could include this in a configuration file:

     D 360
     M lqmed

If Metafont runs, but generates fonts at the wrong resolution or for the wrong device, most likely mktexpk's built-in guess for the mode is wrong, and you should override it as above.

See for a list of resolutions and mode names for most devices (additional submissions are welcome).

If Metafont runs but generates fonts at a resolution of 2602dpi (and prints out the name of each character as well as just a character number, and maybe tries to display the characters), then your Metafont base file probably hasn't been made properly. (It's using the default proof mode, instead of an actual device mode.) To make a proper plain.base, assuming the local mode definitions are contained in a file, run the following command (assuming Unix):

     inimf "plain; input modes; dump"

Then copy the plain.base file from the current directory to where the base files are stored on your system (/usr/local/share/texmf/web2c by default), and make a link (either hard or soft) from plain.base to mf.base in that directory. See inimf invocation.

If mf is a command not found at all by mktexpk, then you need to install Metafont (see unixtex.ftp).

Next: , Previous: Unable to generate fonts, Up: Common problems TeX or Metafont failing

If TeX or Metafont get a segmentation fault or otherwise fail while running a normal input file, the problem is usually a compiler bug (unlikely as that may sound). Even if the trip and trap tests are passed, problems may lurk. Optimization occasionally causes trouble in programs other than TeX and Metafont themselves, too.

Insufficient swap space may also cause core dumps or other erratic behavior.

For a workaround, if you enabled any optimization flags, it's best to omit optimization entirely. In any case, the way to find the facts is to run the program under the debugger and see where it's failing.

Also, if you have trouble with a system C compiler, I advise trying the GNU C compiler. And vice versa, unfortunately; but in that case I also recommend reporting a bug to the GCC mailing list; see Bugs.

To report compiler bugs effectively requires perseverance and perspicacity: you must find the miscompiled line, and that usually involves delving backwards in time from the point of error, checking through TeX's (or whatever program's) data structures. Things are not helped by all-too-common bugs in the debugger itself. Good luck.

One known cause of trouble is the way arrays are handled. Some of the Pascal arrays have a lower index other than 0, and the C code will take the pointer to the allocated memory, subtract the lower index, and use the resulting pointer for the array. While this trick often works, ANSI C doesn't guarantee that it will. It it known to fail on HP-UX 10 machines when the native compiler is used, unless the ‘+u’ compiler switch was specified. Using GCC will work on this platform as well.

Next: , Previous: TeX or Metafont failing, Up: Common problems Empty Makefiles

On some systems (NetBSD, FreeBSD, AIX 4.1, and Mach10), configure may fail to properly create the Makefiles. Instead, you get an error which looks something like this:

     prompt$ ./configure
     creating Makefile
     sed: 1: "\\@^ac_include make/pat ...": \ can not
       be used as a string delimiter

So far as I know, the bug here is in /bin/sh on these systems. I don't have access to a machine running any of them, so if someone can find a workaround that avoids the quoting bug, I'd be most grateful. (Search for ac_include in the configure script to get to the problematic code.)

It should work to run bash configure, instead of using /bin/sh. You can get Bash from and mirrors.

Another possible cause (reported for NeXT) is a bug in the sed command. In that case the error may look like this:

     Unrecognized command: \@^ac_include make/paths.make@r make/paths.make

In this case, installing GNU sed should solve the problem. You can get GNU sed from the same places as Bash.

Next: , Previous: Empty Makefiles, Up: Common problems XtStrings

You may find that linking X programs results in an error from the linker that ‘XtStrings’ is undefined, something like this:

     gcc -o virmf ...
     .../x11.c:130: undefined reference to `XtStrings'

This generally happens because of a mismatch between the X include files with which you compiled and the X libraries with which you linked; often, the include files are from MIT and the libraries from Sun.

The solution is to use the same X distribution for compilation and linking. Probably ‘configure’ was unable to guess the proper directories from your installation. You can use the configure options ‘--x-includes=path’ and ‘--x-libraries=path’ to explicitly specify them.

Next: , Previous: XtStrings, Up: Common problems dlopen

(This section adapted from the file dlsym.c in the X distribution.)

The Xlib library uses the standard C function wcstombs. Under SunOS 4.1, wcstombs uses the ‘dlsym’ interface defined in Unfortunately, the SunOS 4.1 distribution does not include a static ‘libdl.a’ library.

As a result, if you try to link an X program statically under SunOS, you may get undefined references to dlopen, dlsym, and dlclose. One workaround is to include these definitions when you link:

     void *dlopen() { return 0; }
     void *dlsym()  { return 0; }
     int dlclose()  { return -1; }

These are contained in the dlsym.c file in the MIT X distribution.

Next: , Previous: dlopen, Up: Common problems ShellWidgetClass

(This section adapted from the comp.sys.sun.admin FAQ.)

If you are linking with Sun's OpenWindows libraries in SunOS 4.1.x, you may get undefined symbols _get_wmShellWidgetClass and _get_applicationShellWidgetClass when linking. This problem does not arise using the standard MIT X libraries under SunOS.

The cause is bugs in the Xmu shared library as shipped from Sun. There are several fixes:

Here is the information for getting the two patches:

     Patch ID: 100512-02
     Bug ID's: 1086793, 1086912, 1074766
     Description: 4.1.x OpenWindows 3.0 libXt jumbo patch
     Patch ID: 100573-03
     Bug ID: 1087332
     Description: 4.1.x OpenWindows 3.0 undefined symbols when using shared libXmu.

The way to statically link with libXmu depends on whether you are using a Sun compiler (e.g., cc) or gcc. If the latter, alter the x_libs Make variable to include

     -static -lXmu -dynamic

If you are using the Sun compiler, use ‘-Bstatic’ and ‘-Bdynamic’.

Previous: ShellWidgetClass, Up: Common problems Pointer combination warnings

When compiling with old C compilers, you may get some warnings about “illegal pointer combinations”. These are spurious; just ignore them. I decline to clutter up the source with casts to get rid of them.

Next: , Previous: Installation, Up: Top

3 Path searching

This chapter describes the generic path searching mechanism Kpathsea provides. For information about searching for particular file types (e.g., TeX fonts), see the next chapter.

Next: , Up: Path searching

3.1 Searching overview

A search path is a colon-separated list of path elements, which are directory names with a few extra frills. A search path can come from (a combination of) many sources; see below. To look up a file ‘foo’ along a path ‘.:/dir’, Kpathsea checks each element of the path in turn: first ./foo, then /dir/foo, returning the first match (or possibly all matches).

The “colon” and “slash” mentioned here aren't necessarily ‘:’ and ‘/’ on non-Unix systems. Kpathsea tries to adapt to other operating systems' conventions.

To check a particular path element e, Kpathsea first sees if a prebuilt database (see Filename database) applies to e, i.e., if the database is in a directory that is a prefix of e. If so, the path specification is matched against the contents of the database.

If the database does not exist, or does not apply to this path element, or contains no matches, the filesystem is searched (if this was not forbidden by the specification with ‘!!’ and if the file being searched for must exist). Kpathsea constructs the list of directories that correspond to this path element, and then checks in each for the file being searched for. (To help speed future lookups of files in the same directory, the directory in which a file is found is floated to the top of the directory list.)

The “file must exist” condition comes into play with VF files and input files read by the TeX ‘\openin’ command. These files might very well not exist (consider cmr10.vf), and so it would be wrong to search the disk for them. Therefore, if you fail to update ls-R when you install a new VF file, it will not be found.

Each path element is checked in turn: first the database, then the disk. If a match is found, the search stops and the result is returned. This avoids possibly-expensive processing of path specifications that are never needed on a particular run. (Unless the search explicitly requested all matches.)

Although the simplest and most common path element is a directory name, Kpathsea supports additional features in search paths: layered default values, environment variable names, config file values, users' home directories, and recursive subdirectory searching. Thus, we say that Kpathsea expands a path element, meaning transforming all the magic specifications into the basic directory name or names. This process is described in the sections below. It happens in the same order as the sections.

Exception to all of the above: If the filename being searched for is absolute or explicitly relative, i.e., starts with ‘/’ or ‘./’ or ‘../’, Kpathsea simply checks if that file exists.

Ordinarily, if Kpathsea tries to access a file or directory that cannot be read, it gives a warning. This is so you will be alerted to directories or files that accidentally lack any read permission (for example, a lost+found directory). If you prefer not to see these warnings, include the value ‘readable’ in the TEX_HUSH environment variable or config file value.

This generic path searching algorithm is implemented in kpathsea/pathsearch.c. It is employed by a higher-level algorithm when searching for a file of a particular type (see File lookup, and Glyph lookup).

Next: , Previous: Searching overview, Up: Path searching

3.2 Path sources

A search path can come from many sources. In the order in which Kpathsea uses them:

  1. A user-set environment variable, e.g., TEXINPUTS. Environment variables with an underscore and the program name appended override; for example, TEXINPUTS_latex overrides TEXINPUTS if the program being run is named ‘latex’.
  2. A program-specific configuration file, e.g., an ‘S /a:/b’ line in Dvips' (see Config files).
  3. A line in a Kpathsea configuration file texmf.cnf, e.g., ‘TEXINPUTS=/c:/d’ (see below).
  4. The compile-time default (specified in kpathsea/paths.h).

You can see each of these values for a given search path by using the debugging options (see Debugging).

These sources may be combined via default expansion (see Default expansion).

Up: Path sources

3.2.1 Config files

As mentioned above, Kpathsea reads runtime configuration files named texmf.cnf for search path and other definitions. The search path used to look for these configuration files is named TEXMFCNF, and is constructed in the usual way, as described above, except that configuration files cannot be used to define the path, naturally; also, an ls-R database is not used to search for them.

Kpathsea reads all texmf.cnf files in the search path, not just the first one found; definitions in earlier files override those in later files. Thus, if the search path is ‘.:$TEXMF’, values from ./texmf.cnf override those from $TEXMF/texmf.cnf.

If Kpathsea cannot find any texmf.cnf file, it reports a warning including all the directories it checked. If you don't want to see this warning, set the environment variable KPATHSEA_WARNING to the single character ‘0’ (zero, not oh).

While (or instead of) reading this description, you may find it helpful to look at the distributed texmf.cnf, which uses or at least mentions most features. The format of texmf.cnf files follows:

Here is a configuration file fragment illustrating most of these points:

     % TeX input files -- i.e., anything to be found by \input or \openin ...
     latex209_inputs = .:$TEXMF/tex/latex209//:$TEXMF/tex//
     latex2e_inputs = .:$TEXMF/tex/latex//:$TEXMF/tex//
     TEXINPUTS = .:$TEXMF/tex//
     TEXINPUTS.latex209 = $latex209_inputs
     TEXINPUTS.latex2e = $latex2e_inputs
     TEXINPUTS.latex = $latex2e_inputs

This format has obvious similarities to Bourne shell scripts—change the comment character to #, disallow spaces around the =, and get rid of the .name convention, and it could be run through the shell. However, there seemed little advantage in this, since all the information would have to passed back to Kpathsea and parsed there anyway, since the sh process couldn't affect its parent's environment.

The implementation of all this is in kpathsea/cnf.c.

Next: , Previous: Path sources, Up: Path searching

3.3 Path expansion

Kpathsea recognizes certain special characters and constructions in search paths, similar to that in shells. As a general example: ‘~$USER/{foo,bar}//baz’ expands to all subdirectories under directories foo and bar in $USER's home directory that contain a directory or file baz. These expansions are explained in the sections below.

Next: , Up: Path expansion

3.3.1 Default expansion

If the highest-priority search path (see Path sources) contains an extra colon (i.e., leading, trailing, or doubled), Kpathsea inserts at that point the next-highest-priority search path that is defined. If that inserted path has an extra colon, the same happens with the next-highest. (An extra colon in the compile-time default value has unpredictable results, so installers beware.)

For example, given an environment variable setting

     setenv TEXINPUTS /home/karl:

and a TEXINPUTS value from texmf.cnf of


then the final value used for searching will be:


Put another way, default expansion works on “formats” (search paths), and not directly on environment variables. Example, showing the trailing ‘:’ ignored in the first case and expanded in the second:

     $ env TTFONTS=/tmp: kpsewhich --expand-path '$TTFONTS'
     $ env TTFONTS=/tmp: kpsewhich --show-path=.ttf

Since Kpathsea looks for multiple configuration files, it would be natural to expect that (for example) an extra colon in ./texmf.cnf would expand to the path in $TEXMF/texmf.cnf. Or, with Dvips' configuration files, that an extra colon in config.$PRINTER would expand to the path in This doesn't happen. It's not clear this would be desirable in all cases, and trying to devise a way to specify the path to which the extra colon should expand seemed truly baroque. Technicality: Since it would be useless to insert the default value in more than one place, Kpathsea changes only one extra ‘:’ and leaves any others in place (they will eventually be ignored). Kpathsea checks first for a leading ‘:’, then a trailing ‘:’, then a doubled ‘:’.

You can trace this by debugging “paths” (see Debugging). Default expansion is implemented in the source file kpathsea/kdefault.c.

Next: , Previous: Default expansion, Up: Path expansion

3.3.2 Variable expansion

$foo’ or ‘${foo}’ in a path element is replaced by (1) the value of an environment variable ‘foo’ (if defined); (2) the value of ‘foo’ from texmf.cnf (if defined); (3) the empty string.

If the character after the ‘$’ is alphanumeric or ‘_’, the variable name consists of all consecutive such characters. If the character after the ‘$’ is a ‘{’, the variable name consists of everything up to the next ‘}’ (braces may not be nested around variable names). Otherwise, Kpathsea gives a warning and ignores the ‘$’ and its following character.

You must quote the $'s and braces as necessary for your shell. Shell variable values cannot be seen by Kpathsea, i.e., ones defined by set in C shells and without export in Bourne shells.

For example, given

     setenv tex /home/texmf
     setenv TEXINPUTS .:$tex:${tex}prev

the final TEXINPUTS path is the three directories:


The ‘.progname’ suffix on variables and ‘_progname’ on environment variable names are not implemented for general variable expansions. These are only recognized when search paths are initialized (see Path sources).

Variable expansion is implemented in the source file kpathsea/variable.c.

Next: , Previous: Variable expansion, Up: Path expansion

3.3.3 Tilde expansion

A leading ‘~’ in a path element is replaced by the value of the environment variable HOME, or . if HOME is not set. On Windows, the environment variable USERPROFILE is checked instead of HOME.

A leading ‘~user’ in a path element is replaced by user's home directory from the system passwd database.

For example,

     setenv TEXINPUTS ~/mymacros:

will prepend a directory mymacros in your home directory to the default path.

As a special case, if a home directory ends in ‘/’, the trailing slash is dropped, to avoid inadvertently creating a ‘//’ construct in the path. For example, if the home directory of the user ‘root’ is ‘/’, the path element ‘~root/mymacros’ expands to just ‘/mymacros’, not ‘//mymacros’.

Tilde expansion is implemented in the source file kpathsea/tilde.c.

Next: , Previous: Tilde expansion, Up: Path expansion

3.3.4 Brace expansion

x{a,b}y’ expands to ‘xay:xby’. For example:


expands to ‘foo/1/baz:foo/2/baz’. ‘:’ is the path separator on the current system; e.g., on a DOS system, it's ‘;’.

Braces can be nested; for example, ‘x{A,B{1,2}}y’ expands to ‘xAy:xB1y:xB2y’.

Multiple non-nested braces are expanded from right to left; for example, ‘x{A,B}{1,2}y’ expands to ‘x{A,B}1y:x{A,B}2y’, which expands to ‘xA1y:xB1y:xA2y:xB2y’.

This feature can be used to implement multiple TeX hierarchies, by assigning a brace list to $TEXMF, as mentioned in

You can also use the path separator instead of the comma. The last example could have been written ‘x{A:B}{1:2}y’.

Brace expansion is implemented in the source file kpathsea/expand.c.

Next: , Previous: Brace expansion, Up: Path expansion

3.3.5 KPSE_DOT expansion

When KPSE_DOT is defined in the environment, it names a directory that should be considered the current directory for the purpose of looking up files in the search paths. This feature is needed by the ‘mktex...’ scripts mktex scripts, because these change the working directory. You should not ever define it yourself.

Previous: KPSE_DOT expansion, Up: Path expansion

3.3.6 Subdirectory expansion

Two or more consecutive slashes in a path element following a directory d is replaced by all subdirectories of d: first those subdirectories directly under d, then the subsubdirectories under those, and so on. At each level, the order in which the directories are searched is unspecified. (It's “directory order”, and definitely not alphabetical.)

If you specify any filename components after the ‘//’, only subdirectories which match those components are included. For example, ‘/a//b’ would expand into directories /a/1/b, /a/2/b, /a/1/1/b, and so on, but not /a/b/c or /a/1.

You can include multiple ‘//’ constructs in the path.

//’ at the beginning of a path is ignored; you didn't really want to search every directory on the system, did you?

I should mention one related implementation trick, which I took from GNU find. Matthew Farwell suggested it, and David MacKenzie implemented it.

The trick is that in every real Unix implementation (as opposed to the POSIX specification), a directory which contains no subdirectories will have exactly two links (namely, one for . and one for ..). That is to say, the st_nlink field in the ‘stat’ structure will be two. Thus, we don't have to stat everything in the bottom-level (leaf) directories—we can just check st_nlink, notice it's two, and do no more work.

But if you have a directory that contains a single subdirectory and 500 regular files, st_nlink will be 3, and Kpathsea has to stat every one of those 501 entries. Therein lies slowness.

You can disable the trick by undefining ST_NLINK_TRICK in kpathsea/config.h. (It is undefined by default except under Unix.)

Unfortunately, in some cases files in leaf directories are stat'd: if the path specification is, say, ‘$TEXMF/fonts//pk//’, then files in a subdirectory ‘.../pk’, even if it is a leaf, are checked. The reason cannot be explained without reference to the implementation, so read kpathsea/elt-dirs.c (search for ‘may descend’) if you are curious. And if you can find a way to solve the problem, please let me know.

Subdirectory expansion is implemented in the source file kpathsea/elt-dirs.c.

Next: , Previous: Path expansion, Up: Path searching

3.4 Filename database (ls-R)

Kpathsea goes to some lengths to minimize disk accesses for searches (see Subdirectory expansion). Nevertheless, in practice searching each possible directory in typical TeX installations takes an excessively long time.

Therefore, Kpathsea can use an externally-built filename database file named ls-R that maps files to directories, thus avoiding the need to exhaustively search the disk.

A second database file aliases allows you to give additional names to the files listed in ls-R. This can be helpful to adapt to “8.3” filename conventions in source files.

The ls-R and aliases features are implemented in the source file kpathsea/db.c.

Next: , Up: Filename database

3.4.1 ls-R

As mentioned above, you must name the main filename database ls-R. You can put one at the root of each TeX installation hierarchy you wish to search ($TEXMF by default); most sites have only one hierarchy. Kpathsea looks for ls-R files along the TEXMFDBS path, so that should presumably match the list of hierarchies.

The recommended way to create and maintain ‘ls-R’ is to run the mktexlsr script, which is installed in ‘$(bindir)’ (/usr/local/bin by default). That script goes to some trouble to follow symbolic links as necessary, etc. It's also invoked by the distributed ‘mktex...’ scripts.

At its simplest, though, you can build ls-R with the command

     cd /your/texmf/root && ls -LAR ./ >ls-R

presuming your ls produces the right output format (see the section below). GNU ls, for example, outputs in this format. Also presuming your ls hasn't been aliased in a system file (e.g., /etc/profile) to something problematic, e.g., ‘ls --color=tty’. In that case, you will have to disable the alias before generating ls-R. For the precise definition of the file format, see Database format.

Regardless of whether you use the supplied script or your own, you will almost certainly want to invoke it via cron, so when you make changes in the installed files (say if you install a new LaTeX package), ls-R will be automatically updated.

The ‘-A’ option to ls includes files beginning with ‘.’ (except for . and ..), such as the file .tex included with the LaTeX tools package. (On the other hand, directories whose names begin with ‘.’ are always ignored.)

If your system does not support symbolic links, omit the ‘-L’.

ls -LAR /your/texmf/root will also work. But using ‘./’ avoids embedding absolute pathnames, so the hierarchy can be easily transported. It also avoids possible trouble with automounters or other network filesystem conventions.

Kpathsea warns you if it finds an ls-R file, but the file does not contain any usable entries. The usual culprit is running plain ‘ls -R’ instead of ‘ls -LR ./’ or ‘ls -R /your/texmf/root’. Another possibility is some system directory name starting with a ‘.’ (perhaps if you are using AFS); Kpathsea ignores everything under such directories.

Because the database may be out-of-date for a particular run, if a file is not found in the database, by default Kpathsea goes ahead and searches the disk. If a particular path element begins with ‘!!’, however, only the database will be searched for that element, never the disk. If the database does not exist, nothing will be searched. Because this can surprise users (“I see the font foo.tfm when I do an ls; why can't Dvips find it?”), it is not in any of the default search paths.

Next: , Previous: ls-R, Up: Filename database

3.4.2 Filename aliases

In some circumstances, you may wish to find a file under several names. For example, suppose a TeX document was created using a DOS system and tries to read longtabl.sty. But now it's being run on a Unix system, and the file has its original name, longtable.sty. The file won't be found. You need to give the actual file longtable.sty an alias ‘longtabl.sty’.

You can handle this by creating a file aliases as a companion to the ls-R for the hierarchy containing the file in question. (You must have an ls-R for the alias feature to work.)

The format of aliases is simple: two whitespace-separated words per line; the first is the real name longtable.sty, and second is the alias (longtabl.sty). These must be base filenames, with no directory components. longtable.sty must be in the sibling ls-R.

Also, blank lines and lines starting with ‘%’ or ‘#’ are ignored in aliases, to allow for comments.

If a real file longtabl.sty exists, it is used regardless of any aliases.

Previous: Filename aliases, Up: Filename database

3.4.3 Database format

The “database” read by Kpathsea is a line-oriented file of plain text. The format is that generated by GNU (and most other) ls programs given the ‘-R’ option, as follows.

For example, here's the first few lines of ls-R (which totals about 30K bytes) on my system:


Previous: Filename database, Up: Path searching

3.5 kpsewhich: Standalone path searching

The Kpsewhich program exercises the path searching functionality independent of any particular application. This can also be useful as a sort of find program to locate files in your TeX hierarchies, perhaps in administrative scripts. It is used heavily in the distributed ‘mktex...’ scripts.


     kpsewhich option... filename...

The options and filename(s) to look up can be intermixed. Options can start with either ‘-’ or ‘--’, and any unambiguous abbreviation is accepted.

Next: , Up: Invoking kpsewhich

3.5.1 Path searching options

Kpsewhich looks up each non-option argument on the command line as a filename, and returns the first file found.

Various options alter the path searching behavior:

Report all matches found, one per line. By default, if there is more than one match, just one will be reported (chosen effectively at random).
Set the resolution to num; this only affects ‘gf’ and ‘pk’ lookups. ‘-D’ is a synonym, for compatibility with Dvips. Default is 600.
Set the engine name to name. By default it is not set. The engine name is used in some search paths to allow files with the same name but used by different engines to coexist.

In particular, since the memory dump files (.fmt/.base/.mem) are now stored in subdirectories named for the engine (tex, pdftex, xetex, etc.), you must specify an engine name in order to find them. For example, cont-en.fmt typically exists for both pdftex and xetex. With the default path settings, you can use ‘--engine=/’ to look for any dump file, regardless of engine; if a dump file exists for more than one engine, it's indeterminate which one is returned. (The ‘/’ ends up specifying a normal recursive search along the path where the dumps are stored, namely ‘$TEXMF/web2c{/$engine,}’.)

Set the format for lookup to name. By default, the format is guessed from the filename, with ‘tex’ being used if nothing else fits. The recognized filename extensions (including any leading ‘.’) are also allowable names.

All formats also have a name, which is the only way to specify formats with no associated suffix. For example, for Dvips configuration files you can use ‘--format="dvips config"’. (The quotes are for the sake of the shell.)

Here's the current list of recognized names and the associated suffixes. See Supported file formats, for more information on each of these.

The strings in parentheses are abbreviations recognized only by kpsewhich (not the underlying library calls). They are provided when it would otherwise require an argument containing a space to specify the format, to simplify quoting of calls from shells.

          gf: gf
          pk: pk
          bitmap font (bitmapfont):
          tfm: .tfm
          afm: .afm
          base: .base
          bib: .bib
          bst: .bst
          cnf: .cnf
          ls-R: ls-R ls-r
          fmt: .fmt
          map: .map
          mem: .mem
          mf: .mf
          mfpool: .pool
          mft: .mft
          mp: .mp
          mppool: .pool
          MetaPost support (mpsupport):
          ocp: .ocp
          ofm: .ofm .tfm
          opl: .opl  .pl
          otp: .otp
          ovf: .ovf .vf
          ovp: .ovp  .vpl
          graphic/figure:  .eps .epsi
          tex: .tex  .sty .cls .fd .aux .bbl .def .clo .ldf
          TeX system documentation (doc):
          texpool: .pool
          TeX system sources (source):  .dtx .ins
          PostScript header:  .pro
          Troff fonts (trofffont):
          type1 fonts: .pfa .pfb
          vf: .vf
          dvips config (dvipsconfig):
          ist: .ist
          truetype fonts: .ttf .ttc .TTF .TTC .dfont
          type42 fonts: .t42 .T42
          web2c files (web2c):
          other text files (othertext):
          other binary files (otherbin):
          misc fonts (miscfont):
          web: .web  .ch
          cweb: .w .web  .ch
          enc files: .enc
          cmap files (cmap):
          subfont definition files: .sfd
          opentype fonts: .otf
          pdftex config (pdftexconfig):
          lig files: .lig
          lua: .luc .luctex .texluc .lua .luatex .texlua
          font feature files: .fea
          cid maps: .cid .cidmap
          mlbib: .mlbib .bib
          mlbst: .mlbst .bst
          clua: .dll .so
          ris: .ris
          bltxml: .bltxml

This option and ‘--path’ are mutually exclusive.

After processing the command line, read additional filenames to look up from standard input.
Turn on or off the ‘mktex’ script associated with filetype. Usual values for filetype are ‘pk’, ‘mf’, ‘tex’, and ‘tfm’. By default, all are off in Kpsewhich, even if they are enabled for TeX. This option implies setting --must-exist. See mktex scripts.
Set the mode name to string; this also only affects ‘gf’ and ‘pk’ lookups. No default: any mode will be found. See mktex script arguments.
Do everything possible to find the files, notably including searching the disk and running the ‘mktex’ scripts. By default, only the ls-R database is checked, in the interest of efficiency.
Search along the path string (colon-separated as usual), instead of guessing the search path from the filename. ‘//’ and all the usual expansions are supported (see Path expansion). This option and ‘--format’ are mutually exclusive. To output the complete directory expansion of a path, instead of doing a one-shot lookup, see ‘--expand-path’ and ‘--show-path’ in the following section.
Set the program name to name; default is ‘kpsewhich’. This can affect the search paths via the ‘.prognam’ feature in configuration files (see Config files).
Exit successfully if name is safe to open for reading or writing, respectively, else unsuccessfully. No output is written. These tests take account of the related Kpathsea configuration settings (see Calling sequence).
Report only those matches whose directory part ends with string (compared literally, except case is ignored on a case-insensitive operating system). For example, suppose there are two matches for a given name:
          kpsewhich foo.sty
          ⇒ /some/where/foo.sty

Then we can narrow the result to what we are interested in with --subdir:

          kpsewhich --subdir=where foo.sty
          ⇒ /some/where/foo.sty
          kpsewhich --subdir=place foo.sty
          ⇒ /another/place/foo.sty

The string to match must be at the end of the directory part of the match, and it is taken literally, with no pattern matching:

          kpsewhich --subdir=another foo.sty

The string to match may cross directory components:

          kpsewhich --subdir=some/where foo.sty
          ⇒ /some/where/foo.sty

--subdir implies --all; if there is more than one match, they will all be reported (in our example, both ‘where’ and ‘place’ end in ‘e’):

          kpsewhich --subdir=e
          ⇒ /some/where/foo.sty

Because of the above rules, the presence of a leading ‘/’ is important, since it “anchors” the match to a full component name:

          kpsewhich --subdir=/lace foo.sty

However, a trailing ‘/’ is immaterial (and ignored), since the match always takes place at the end of the directory part:

          kpsewhich --subdir=lace/ foo.sty
          ⇒ /another/place/foo.sty

The purpose of these rules is to make it convenient to find results only within a particular area of the tree. For instance, a given script named foo.lua might exist within both texmf-dist/scripts/pkg1/ and texmf-dist/scripts/pkg2/. By specifying, say, ‘--subdir=/pkg1’, you can be sure of getting the one you are interested in.

We only match at the end because a site might happen to install TeX in /some/coincidental/pkg1/path/, and we wouldn't want to match texmf-dist/scripts/pkg2/ that when searching for ‘/pkg1’.

Next: , Previous: Path searching options, Up: Invoking kpsewhich

3.5.2 Specially-recognized files for kpsewhich

kpsewhich recognizes a few special filenames on the command line and defaults to using the `known' file formats for them, merely to save the time and trouble of specifying the format. This is only a feature of kpsewhich; when using the Kpathsea library itself, none of these special filenames are recognized, and it's still up to the caller to specify the desired format.

Here is the list of special filenames to kpsewhich, along with their corresponding format:
dvips config

other text files

web2c files


web2c files


pdftex config’ (although pdftex.cfg is not used any more; look for the file pdftexconfig.tex instead.)


other text files

A user-specified format will override the above defaults.

Another useful configuration file in this regard is, found in texmf/texconfig/, which records various information about the above configuration files (among others).

Next: , Previous: Specially-recognized files, Up: Invoking kpsewhich

3.5.3 Auxiliary tasks

Kpsewhich provides some additional features not strictly related to path lookup:

Previous: Auxiliary tasks, Up: Invoking kpsewhich

3.5.4 Standard options

Kpsewhich accepts the standard GNU options:

Next: , Previous: Path searching, Up: Top

4 TeX support

Although the basic features in Kpathsea can be used for any type of path searching, it came about (like all libraries) with a specific application in mind: I wrote Kpathsea specifically for TeX system programs. I had been struggling with the programs I was using (Dvips, Xdvi, and TeX itself) having slightly different notions of how to specify paths; and debugging was painful, since no code was shared.

Therefore, Kpathsea provides some TeX-specific formats and features. Indeed, many of the supposedly generic path searching features were provided because they seemed useful in that conTeXt (font lookup, particularly).

Kpathsea provides a standard way to search for files of any of the supported file types; glyph fonts are a bit different than all the rest. Searches are based solely on filenames, not file contents—if a GF file is named cmr10.600pk, it will be found as a PK file.

Next: , Up: TeX support

4.1 Supported file formats

Kpathsea has support for a number of file types. Each file type has a list of environment and config file variables that are checked to define the search path, and most have a default suffix that plays a role in finding files (see the next section). Some also define additional suffixes, and/or a program to be run to create missing files on the fly.

Since environment variables containing periods, such as ‘TEXINPUTS.latex’, are not allowed on some systems, Kpathsea looks for environment variables with an underscore, e.g., ‘TEXINPUTS_latex’ (see Config files).

The following table lists the above information.

(Adobe font metrics, see Metric files) AFMFONTS; suffix ‘.afm’.
(Metafont memory dump, see Memory dumps) MFBASES, TEXMFINI; suffix ‘.base’.
(BibTeX bibliography source, see bibtex invocation) BIBINPUTS, TEXBIB; suffix ‘.bib’.
(BibTeX style, see Basic BibTeX style files) BSTINPUTS; suffix ‘.bst’.
(character map files) CMAPFONTS; suffix ‘.cmap’.
(Runtime configuration files, see Config files) TEXMFCNF; suffix ‘.cnf’.
(CWEB input files) CWEBINPUTS; suffixes ‘.w’, ‘.web’; additional suffix ‘.ch’.
dvips config
(Dvips ‘config.*’ files, such as, see Config files) TEXCONFIG.
enc files
(encoding vectors) ENCFONTS; suffix ‘.enc’.
(TeX memory dump, see Memory dumps) TEXFORMATS, TEXMFINI; suffix ‘.fmt’.
font cid map
(CJK mapping) FONTCIDMAPS suffix ‘.cid’.
font feature files
(primarily for OpenType font features) FONTFEATURES suffix ‘.fea’.
(generic font bitmap, see Glyph files) programFONTS, GFFONTS, GLYPHFONTS, TEXFONTS; suffix ‘gf’.
(Encapsulated PostScript figures, see PostScript figures) TEXPICTS, TEXINPUTS; additional suffixes: ‘.eps’, ‘.epsi’.
(makeindex style files) TEXINDEXSTYLE, INDEXSTYLE; suffix ‘.ist’.
lig files
(ligature definition files) LIGFONTS; suffix ‘.lig’.
(Filename databases, see Filename database) TEXMFDBS.
(Fontmaps, see Fontmap) TEXFONTMAPS; suffix ‘.map’.
(MetaPost memory dump, see Memory dumps) MPMEMS, TEXMFINI; suffix ‘.mem’.
MetaPost support
(MetaPost support files, used by DMP; see dmp invocation) MPSUPPORT.
(Metafont source, see mf invocation) MFINPUTS; suffix ‘.mf’; dynamic creation program: mktexmf.
(Metafont program strings, see pooltype invocation) MFPOOL, TEXMFINI; suffix ‘.pool’.
(MFT style file, see mft invocation) MFTINPUTS; suffix ‘.mft’.
misc fonts
(font-related files that don't fit the other categories) MISCFONTS
(MlBibTeX bibliography source) MLBIBINPUTS, BIBINPUTS, TEXBIB; suffixes ‘.mlbib’, ‘.mlbib’.
(MlBibTeX style) MLBSTINPUTS, BSTINPUTS; suffixes ‘.mlbst’, ‘.bst’.
(MetaPost source, see mpost invocation) MPINPUTS; suffix ‘.mp’.
(MetaPost program strings, see pooltype invocation) MPPOOL, TEXMFINI; suffix ‘.pool’.
(Omega compiled process files) OCPINPUTS;
suffix ‘.ocp’; dynamic creation program: MakeOmegaOCP.
(Omega font metrics) OFMFONTS, TEXFONTS;
suffixes ‘.ofm’, ‘.tfm’; dynamic creation program: MakeOmegaOFM.
opentype fonts
(OpenType fonts) OPENTYPEFONTS.
(Omega property lists) OPLFONTS, TEXFONTS; suffix ‘.opl’.
(Omega translation process files) OTPINPUTS; suffix ‘.otp’.
(Omega virtual fonts) OVFFONTS, TEXFONTS; suffix ‘.ovf’.
(Omega virtual property lists) OVPFONTS, TEXFONTS; suffix ‘.ovp’.
pdftex config
(PDFTeX-specific configuration files) PDFTEXCONFIG.
(packed bitmap fonts, see Glyph files) PROGRAMFONTS (program being ‘XDVI’, etc.), PKFONTS, TEXPKS, GLYPHFONTS, TEXFONTS; suffix ‘pk’; dynamic creation program: mktexpk.
PostScript header
(downloadable PostScript, see Header files) TEXPSHEADERS, PSHEADERS; additional suffix ‘.pro’.
subfont definition files
(subfont definition files) SFDFONTS suffix ‘.sfd’.
(TeX source, see tex invocation) TEXINPUTS; suffix ‘.tex’; additional suffixes: none, because such a list cannot be complete; dynamic creation program: mktextex.
TeX system documentation
(Documentation files for the TeX system) TEXDOCS.
TeX system sources
(Source files for the TeX system) TEXSOURCES.
(Architecture-independent executables distributed in the texmf trees) TEXMFSCRIPTS.
(TeX program strings, see pooltype invocation) TEXPOOL, TEXMFINI; suffix ‘.pool’.
(TeX font metrics, see Metric files) TFMFONTS, TEXFONTS; suffix ‘.tfm’; dynamic creation program: mktextfm.
Troff fonts
(Troff fonts, used by DMP; see DMP invocation) TRFONTS.
truetype fonts
(TrueType outline fonts) TTFONTS; suffixes ‘.ttf’ and ‘.TTF’, ‘.ttc’ and ‘.TTC’, ‘.dfont’.
type1 fonts
(Type 1 PostScript outline fonts, see Glyph files) T1FONTS, T1INPUTS, TEXPSHEADERS, DVIPSHEADERS; suffixes ‘.pfa’, ‘.pfb’.
type42 fonts
(Type 42 PostScript outline fonts) T42FONTS.
(virtual fonts, see Virtual fonts) VFFONTS, TEXFONTS; suffix ‘.vf’.
(WEB input files) WEBINPUTS; suffix ‘.web’; additional suffix ‘.ch’.
web2c files
(files specific to the web2c implementation) WEB2C.

There are two special cases, because the paths and environment variables always depend on the name of the program: the variable name is constructed by converting the program name to upper case, and then appending ‘INPUTS’. Assuming the program is called ‘foo’, this gives us the following table.

other text files
(text files used by ‘foo’) FOOINPUTS.
other binary files
(binary files used by ‘foo’) FOOINPUTS.

If an environment variable by these names are set, the corresponding texmf.cnf definition won't be looked at (unless, as usual, the environment variable value has an extra ‘:’). See Default expansion.

For the font variables, the intent is that:

Next: , Previous: Supported file formats, Up: TeX support

4.2 File lookup

This section describes how Kpathsea searches for most files (bitmap font searches are the exception, as described in the next section).

Here is the search strategy for a file name:

  1. If the file format defines default suffixes, and the suffix of name name is not already a known suffix for that format, try the name with each default appended, and use alternative names found in the fontmaps if necessary. Example: given ‘’, look for ‘’.
  2. Search for name, and if necessary for alternative names found in the fontmaps. Example: given ‘’, we also look for ‘’.
  3. If the file format defines a program to invoke to create missing files, run it (see mktex scripts).

The order in which we search for “suffixed” name (item 1) or the “as-is” name (item 2) is controlled by the try_std_extension_first configuration value. The default set in texmf.cnf is true, since common suffixes are already recognized: ‘babel.sty’ will only look for ‘babel.sty’, not ‘babel.sty.tex’, regardless of this setting.

When the suffix is unknown (e.g., ‘’), both names are always tried; the difference is the order in which they are tried.

try_std_extension_first only affects names being looked up which *already* have an extension. A name without an extension (e.g., ‘tex story’) will always have an extension added first.

This algorithm is implemented in the function kpathsea_find_file in the source file kpathsea/tex-file.c. You can watch it in action with the debugging options (see Debugging).

Next: , Previous: File lookup, Up: TeX support

4.3 Glyph lookup

This section describes how Kpathsea searches for a bitmap font in GF or PK format (or either) given a font name (e.g., ‘cmr10’) and a resolution (e.g., 600).

Here is an outline of the search strategy (details in the sections below) for a file name at resolution dpi. The search stops at the first successful lookup.

  1. Look for an existing file name.dpiformat in the specified format(s).
  2. If name is an alias for a file f in the fontmap file, look for f.dpi.
  3. Run an external program (typically named ‘mktexpk’) to generate the font (see mktex scripts)
  4. Look for fallback.dpi, where fallback is some last-resort font (typically ‘cmr10’).

This is implemented in kpathsea_find_glyph in kpathsea/tex-glyph.c.

Next: , Up: Glyph lookup

4.3.1 Basic glyph lookup

When Kpathsea looks for a bitmap font name at resolution dpi in a format format, it first checks each directory in the search path for a file ‘name.dpiformat’; for example, ‘cmr10.600pk’. Kpathsea looks for a PK file first, then a GF file.

If that fails, Kpathsea looks for ‘dpidpi/name.format’; for example, ‘dpi600/’. This is how fonts are typically stored on filesystems (such as DOS) that permit only three-character extensions.

If that fails, Kpathsea looks for a font with a close-enough dpi. “Close enough” is defined by the macro KPSE_BITMAP_TOLERANCE in kpathsea/tex-glyph.h to be dpi / 500 + 1. This is slightly more than the 0.2% minimum allowed by the DVI standard (CTAN:/dviware/driv-standard/level-0).

Next: , Previous: Basic glyph lookup, Up: Glyph lookup

4.3.2 Fontmap

If a bitmap font or metric file is not found with the original name (see the previous section), Kpathsea looks through any fontmap files for an alias for the original font name. These files are named and searched for along the TEXFONTMAPS environment/config file variable. All files that are found are read; earlier definitions override later ones.

This feature is intended to help in two respects:

  1. An alias name is limited in length only by available memory, not by your filesystem. Therefore, if you want to ask for ‘Times-Roman’ instead of ptmr, you can (you get ‘ptmr8r’).
  2. A few fonts have historically had multiple names: specifically, LaTeX's “circle font” has variously been known as circle10, lcircle10, and lcirc10. Aliases can make all the names equivalent, so that it no longer matters what the name of the installed file is; TeX documents will find their favorite name.

The format of fontmap files is straightforward:

If an alias has an extension, it matches only those files with that extension; otherwise, it matches anything with the same root, regardless of extension. For example, an alias ‘foo.tfm’ matches only when foo.tfm is being searched for; but an alias ‘foo’ matches foo.vf, foo.600pk, etc.

As an example, here is an excerpt from the in the Web2c distribution. It makes the circle fonts equivalent and includes automatically generated maps for most PostScript fonts available from various font suppliers.

     circle10        lcircle10
     circle10        lcirc10
     lcircle10       circle10
     lcircle10       lcirc10
     lcirc10         circle10
     lcirc10         lcircle10

Fontmaps are implemented in the file kpathsea/fontmap.c. The Fontname distribution has much more information on font naming (see Introduction).

Previous: Fontmap, Up: Glyph lookup

4.3.3 Fallback font

If a bitmap font cannot be found or created at the requested size, Kpathsea looks for the font at a set of fallback resolutions. You specify these resolutions as a colon-separated list (like search paths). Kpathsea looks first for a program-specific environment variable (e.g., DVIPSSIZES for Dvipsk), then the environment variable TEXSIZES, then a default specified at compilation time (the Make variable default_texsizes). You can set this list to be empty if you prefer to find fonts at their stated size or not at all.

Finally, if the font cannot be found even at the fallback resolutions, Kpathsea looks for a fallback font, typically cmr10. Programs must enable this feature by calling kpathsea_init_prog (see Calling sequence); the default is no fallback font.

Previous: Glyph lookup, Up: TeX support

4.4 Suppressing warnings

Kpathsea provides a way to suppress selected usually-harmless warnings; this is useful at large sites where most users are not administrators, and thus the warnings are merely a source of confusion, not a help. To do this, you set the environment variable or configuration file value TEX_HUSH to a colon-separated list of values. Here are the possibilities:

Suppress everything possible.
Suppress mismatched font checksum warnings.
Suppress warnings when a character is missing from a font that a DVI or VF file tries to typeset.
Don't suppress any warnings.
Suppress warnings about attempts to access a file whose permissions render it unreadable.
Suppresses warnings about an unimplemented or unparsable ‘\special’ command.

tex-hush.c defines the function that checks the variable value. Each driver implements its own checks where appropriate.

Next: , Previous: TeX support, Up: Top

5 Programming

This chapter is for programmers who wish to use Kpathsea. See Introduction, for the conditions under which you may do so.

Next: , Up: Programming

5.1 Programming overview

Aside from this manual, your best source of information is the source to the programs that use Kpathsea (see Introduction). Of those, Dviljk is probably the simplest, and hence a good place to start. Xdvik adds VF support and the complication of X resources. Dvipsk adds the complication of its own config files. Web2c is source code I also maintain, so it uses Kpathsea rather straightforwardly, but is of course complicated by the Web to C translation. Finally, Kpsewhich is a small utility program whose sole purpose is to exercise the main path-searching functionality.

When looking at these program sources, you should know that previous versions of the library had a different programming interface, to support re-entrancy. In that interface the library function names were prefixed with kpse_ instead of kpathsea_, and they did not need an instance variable as first argument. This change was made in 2009. Some of the programs mentioned above may still be using the previous interface.

Beyond these examples, the .h files in the Kpathsea source describe the interfaces and functionality (and of course the .c files define the actual routines, which are the ultimate documentation). pathsearch.h declares the basic searching routine. tex-file.h and tex-glyph.h define the interfaces for looking up particular kinds of files. In view of the way the headers depend on each other, it is recommended to use #include <kpathsea/kpathsea.h>, which includes every Kpathsea header.

If you want to include only specific headers, you should still consider including kpathsea/config.h before including any other Kpathsea header, as it provides symbols used in the other headers. Note that kpathsea/config.h includes kpathsea/c-auto.h, which is generated by Autoconf.

The library provides no way for an external program to register new file types: tex-file.[ch] must be modified to do this. For example, Kpathsea has support for looking up Dvips config files, even though no program other than Dvips will likely ever want to do so. I felt this was acceptable, since along with new file types should also come new defaults in texmf.cnf (and its descendant paths.h), since it's simplest for users if they can modify one configuration file for all kinds of paths.

Kpathsea does not parse any formats itself; it barely opens any files. Its primary purpose is to return filenames. The GNU font utilities does contain libraries to read TFM, GF, and PK files, as do the programs above, of course.

Next: , Previous: Programming overview, Up: Programming

5.2 Calling sequence

The typical way to use Kpathsea in your program goes something like this:

  1. Call kpathsea_new to create a new library instance. This variable must be passed as the first argument to all the following library functions. The rest of this manual will be using kpse as a placeholder for the name of this variable.
  2. Call kpathsea_set_program_name with argv[0] as the second argument; the third argument is a string or NULL. The third argument is used by Kpathsea as the program name for the .program feature of config files (see Config files). If the third argument is NULL, the value of the second argument is used. This function must be called before any other use of the Kpathsea library.

    kpathsea_set_program_name always sets the variables kpse->invocation_name and kpse->invocation_short_name. These variables are used in the error message macros defined in kpathsea/lib.h. It sets the variable kpse->program_name to the program name it uses.

    It also initializes debugging options based on the environment variable KPATHSEA_DEBUG (if that is set).

    Finally, it sets the environment variables SELFAUTOLOC, SELFAUTODIR and SELFAUTOPARENT to the location, parent and grandparent directory of the executable, removing . and .. path elements and resolving symbolic links. These are used in the default configuration file to allow people to invoke TeX from anywhere. You can use ‘kpsewhich --expand-var=\$SELFAUTOLOC’, etc., to see the values.

  3. Set debugging options. See Debugging. If your program doesn't have a debugging option already, you can define one and set kpse->debug to the number that the user supplies (as in Dviljk and Web2c), or you can just omit this altogether (people can always set KPATHSEA_DEBUG). If you do have runtime debugging already, you need to merge Kpathsea's options with yours (as in Dvipsk and Xdvik).
  4. If your program has its own configuration files that can define search paths, you should assign those paths to the client_path member in the appropriate element of the kpse->format_info array. (This array is indexed by file type; see tex-file.h.) See resident.c in Dvipsk for an example.
  5. Call kpathsea_init_prog (see proginit.c). It's useful for the DVI drivers, at least, but for other programs it may be simpler to extract the parts of it that actually apply. This does not initialize any paths, it just looks for (and sets) certain environment variables and other random information. (A search path is always initialized at the first call to find a file of that type; this eliminates much useless work, e.g., initializing the BibTeX search paths in a DVI driver.)
  6. The routine to actually find a file of type format is kpathsea_find_file. You can call kpathsea_find_file after doing only the first and second of the initialization steps above—Kpathsea automatically reads the texmf.cnf generic config files, looks for environment variables, and does expansions at the first lookup.
  7. To find PK and/or GF bitmap fonts, the routine is kpathsea_find_glyph, defined in tex-glyph.h. This returns a structure in addition to the resultant filename, because fonts can be found in so many ways. See the documentation in the source.
  8. To actually open a file, not just return a filename, call kpathsea_open_file. This function takes the name to look up and a Kpathsea file format as arguments, and returns the usual FILE *. It always assumes the file must exist, and thus will search the disk if necessary (unless the search path specified ‘!!’, etc.). In other words, if you are looking up a VF or some other file that need not exist, don't use this.
  9. TeX can write output files, via the \openout primitive; this opens a security hole vulnerable to Trojan horse attack: an unwitting user could run a TeX program that overwrites, say, ~/.rhosts. Analogous security holes exist for many other programs. To alleviate this, there is a configuration variable openout_any, which selects one of three levels of security. When it is set to ‘a’ (for “any”), no restrictions are imposed. When it is set to ‘r’ (for “restricted”), filenames beginning with ‘.’ are disallowed (except .tex because LaTeX needs it). When it is set to ‘p’ (for “paranoid”) additional restrictions are imposed: an absolute filename must refer to a file in (a subdirectory) of TEXMFOUTPUT, and any attempt to go up a directory level is forbidden (that is, paths may not contain a ‘..’ component). The paranoid setting is the default. (For backwards compatibility, ‘y’ and ‘1’ are synonyms of ‘a’, while ‘n’ and ‘0’ are synonyms for ‘r’.) The function kpathsea_out_name_ok, with a filename as second argument, returns true if that filename is acceptable to be opend for output or false otherwise.
  10. Similarly, the function kpathsea_in_name_ok, with a filename as second argument, returns true if that filename is acceptable to be opend for input or false otherwise, depending on the value of the configuration variable openin_any (with ‘a’ as default).
  11. To close the kpathsea library instance you are using, call kpathsea_finish. This function closes any open log files and frees the memory used by the instance.

Kpathsea also provides many utility routines. Some are generic: hash tables, memory allocation, string concatenation and copying, string lists, reading input lines of arbitrary length, etc. Others are filename-related: default path, tilde, and variable expansion, stat calls, etc. (Perhaps someday I'll move the former to a separate library.)

The c-*.h header files can also help your program adapt to many different systems. You will almost certainly want to use Autoconf and probably Automake for configuring and building your software if you use Kpathsea; I strongly recommend using Autoconf and Automake regardless. They are available from ftp://ftp.gnu.og/pub/gnu/.

Next: , Previous: Calling sequence, Up: Programming

5.3 Program-specific files

Many programs will need to find some configuration files. Kpathsea contains some support to make it easy to place them in their own directories. The Standard TeX directory structure (see Introduction), specifies that such files should go into a subdirectory named after the program, like ‘texmf/ttf2pk’.

Two formats, ‘kpse_program_text_format’ and ‘kpse_program_binary_format’, use .:$TEXMF/program// as their compiled-in search path. To override this default, you can use the variable PROGRAMINPUTS in the environment and/or ‘texmf.cnf’. That is to say, the name of the variable is constructed by converting the name of the program to upper case, and appending INPUTS.

The only difference between these two formats is whether kpathsea_open_file will open the files it finds in text or binary mode.

Previous: Program-specific files, Up: Programming

5.4 Programming with config files

You can (and probably should) use the same texmf.cnf configuration file that Kpathsea uses for your program. This helps installers by keeping all configuration in one place.

To retrieve a value var from config files, the best way is to call kpathsea_var_value on the string var. This will look first for an environment variable var, then a config file value. The result will be the value found or ‘NULL’. This function is declared in kpathsea/variable.h. For an example, see the shell_escape code in web2c/lib/texmfmp.c.

The routine to do variable expansion in the context of a search path (as opposed to simply retrieving a value) is kpathsea_var_expand, also declared in kpathsea/variable.h. It's generally only necessary to set the search path structure components as explained in the previous section, rather than using this yourself.

If for some reason you want to retrieve a value only from a config file, not automatically looking for a corresponding environment variable, call kpathsea_cnf_get (declared in kpathsea/cnf.h) with the string var.

No initialization calls are needed.

Previous: Programming, Up: Top