groff_font - GNU roff device and font description files


The groff font and output device description formats are slight extensions of those used by AT&T device-independent troff. In distinction to the AT&T implementation, groff lacks a binary format; all files are text files. (Plan 9 troff has also abandoned the binary format.) The device and font description files for a device name are stored in a devname directory. The device description file is called DESC, and, for each font supported by the device, a font description file is called f, where is usually an abbreviation of a font’s name and/or style. For example, the ps (PostScript) device has groff font description files for Times roman (TR) and Zapf Chancery Medium italic (ZCMI), among many others, while the utf8 device (for terminal emulators) has only font descriptions for the roman, italic, bold, and bold-italic styles (R, I, B, and BI, respectively).

Device and font description files are read by the formatter, troff, and by output drivers. The programs typically delegate these files’ processing to an internal library, libgroff, ensuring their consistent interpretation.

DESC file format

The DESC file contains a series of directives; each begins a line. Their order is not important, with two exceptions: (1) the res directive must precede any papersize directive; and (2) the charset directive must come last (if at all). If a directive name is repeated, later entries in the file override previous ones (except that the paper dimensions are computed based on the res directive last seen when papersize is encountered). Spaces and/or tabs separate words and are ignored at line boundaries. Comments start with the “#” character and extend to the end of a line. Empty lines are ignored.

The default font family is fam.

fonts n F1 ... Fn

Fonts F1, ..., Fn are mounted at font positions m+1, ..., m+n where m is the number of styles (see below). This directive may extend over more than one line. A font name of 0 causes no font to be mounted at the corresponding position.

hor n

The horizontal motion quantum is basic units. Horizontal quantities are rounded to multiples of n.

image_generator program

Use program to generate PNG images from PostScript input. Under GNU/Linux, this is usually gs(1), but under other systems (notably Cygwin) it might be set to another name. The grohtml(1) driver uses this directive.

paperlength n

The vertical dimension of the output medium is basic units (deprecated: use papersize instead).

papersize format-or-dimension-pair-or-file-name ...

The dimensions of the output medium are as according to the argument, which is either a standard paper format, a pair of dimensions, or the name of a plain text file containing either of the foregoing. Recognized paper formats are the ISO and DIN formats A0A7, B0B7, C0C7, and D0D7; the U.S. formats letter, legal, tabloid, ledger, statement, and executive; and the envelope formats com10, monarch, and DL. Matching is performed without regard for lettercase.

Alternatively, the argument can be a custom paper format length,width (with no spaces before or after the comma). Both length and width must have a unit appended; valid units are “i” for inches, “c” for centimeters, “p” for points, and “P” for picas. Example: “12c,235p”. An argument that starts with a digit is always treated as a custom paper format.

Finally, the argument can be a file name (e.g., /etc/papersize); if the file can be opened, the first line is read and a match attempted against each other form. No comment syntax is supported.

More than one argument can be specified; each is scanned in turn and the first valid paper specification used.

paperwidth n

The horizontal dimension of the output medium is basic units (deprecated: use papersize instead).


Direct troff to emit the name of the source file being processed. This is achieved with the intermediate output command “x F”, which grohtml interprets.

postpro program

Use program as the postprocessor.

prepro program

Use program as a preprocessor. The html and xhtml output devices use this directive.

print program

Use program as the print spooler. If omitted, groff’s -l and -L options are ignored.

res n

The device resolution is n basic units per inch.

sizes s1 ... sn 0

The device has fonts at s1, ..., sn scaled points (see below). The list of sizes must be terminated by a 0. Each si can also be a range of sizes mn. The list can extend over more than one line.

sizescale n

A typographical point is subdivided into scaled points. The default is 1.

styles S1 ... Sm

The first m font mounting positions are associated with styles S1, ..., Sm.


The postprocessor can handle the t and u intermediate output commands.


The output device supports the complete Unicode repertoire. This directive is useful only for devices which produce character entities instead of glyphs.

If unicode is present, no charset section is required in the font description files since the Unicode handling built into groff is used. However, if there are entries in a font description file’s charset section, they either override the default mappings for those particular characters or add new mappings (normally for composite characters).

The utf8, html, and xhtml output devices use this directive.

unitwidth n

Quantities in the font description files are in basic units for fonts whose type size is scaled points.


Make the font handling module always return unscaled glyph widths. The grohtml driver uses this directive.


troff should encode named glyphs inside device control commands. The grohtml driver uses this directive.

vert n

The vertical motion quantum is basic units. Vertical quantities are rounded to multiples of n.


This directive and the rest of the file are ignored. It is recognized for compatibility with other troff implementations. In GNU troff, character set repertoire is described on a per-font basis.

troff recognizes but ignores the directives spare1, spare2, and biggestfont.

The res, unitwidth, fonts, and sizes lines are mandatory. Directives not listed above are ignored by troff but may be used by postprocessors to obtain further information about the device.

Font description file format

On typesetting output devices, each font is typically available at multiple sizes. While paper measurements in the device description file are in absolute units, measurements applicable to fonts must be proportional to the type size. groff achieves this using the precedent set by AT&T device-independent troff: one font size is chosen as a norm, and all others are scaled linearly relative to that basis. The “unit width” is the number of basic units per point when the font is rendered at this nominal size.

For instance, groff’s lbp device uses a unitwidth of 800. Its Times roman font (“TR”) has a spacewidth of 833; this is also the width of its comma, period, centered period, and mathematical asterisk, while its “M” is 2,963 basic units. Thus, an “M” on the lbp device is 2,963 basic units wide at a notional type size of 800 points. (800-point type is not practical for most purposes, but using it enables the quantities in the font description files to be expressed as integers.)

A font description file has two sections. The first is a sequence of directives, and is parsed similarly to the DESC file described above. Except for the directive names that begin the second section, their ordering is immaterial. Later directives of the same name override earlier ones, spaces and tabs are handled in the same way, and the same comment syntax is supported. Empty lines are ignored throughout.

name F

The name of the font is F. “DESC” is an invalid font name. Simple integers are valid, but their use is discouraged. (groff requests and escape sequences interpret non-negative font names as mounting positions instead. Further, a font named “0” cannot be automatically mounted by the fonts directive of a DESC file.)

spacewidth n

The width of an unadjusted inter-word space is basic units.

The directives above must appear in the first section; those below are optional.

The font’s glyphs have a slant of degrees; a positive n slants in the direction of text flow.

ligatures lig1 ... lign [0]

Glyphs lig1, ..., lign are ligatures; possible ligatures are ff, fi, fl, ffi, and ffl. For compatibility with other troff implementations, the list of ligatures may be terminated with a 0. The list of ligatures must not extend over more than one line.


The font is special: when a glyph is requested that is not present in the current font, it is sought in any mounted fonts that bear this property.

Other directives in this section are ignored by troff, but may be used by postprocessors to obtain further information about the font.

The second section contains one or two subsections. These can appear in either order; the first one encountered commences the second section. Each starts with a directive on a line by itself. A charset subsection is mandatory unless the associated DESC file contains the unicode directive. Another subsection, kernpairs, is optional.

The directive charset starts the character set subsection. (For typesetter devices, this directive is misnamed since it starts a list of glyphs, not characters.) It precedes a series of glyph descriptions, one per line. Each such glyph description comprises a set of fields separated by spaces or tabs and organized as follows.

name metrics type code [entity-name] [-- comment]

name identifies the glyph: if name is a printable character c, it corresponds to the troff ordinary character c. If name is a multi-character sequence not beginning with \, it corresponds to the GNU troff special character escape sequence “\[name]”. A name consisting of three minus signs, “---”, indicates that the glyph is unnamed: such glyphs can be accessed only by the \N escape sequence in troff. A special character named “---” can still be defined using .char and similar requests. The name\-” defines the minus sign glyph. Finally, name can be the horizontal motion escape sequences, \| and \^ (“thin” and “hair” spaces, respectively), in which case only the width metric described below is applied; a font can thus customize the widths of these spaces.

The form of the metrics field is as follows (on one line; it may be broken here for readability).

width[,[height[,[depth[,[italic-correction[,[ left-italic-correction[,[subscript-correction]]]]]]]]]]

There must not be any spaces, tabs, or newlines between these subfields, which are in basic units expressed as decimal integers. Unspecified subfields default to 0. Since there is no associated binary format, these values are not required to fit into the C language data type char as they are in AT&T device-independent troff.

The width subfield gives the width of the glyph. The height subfield gives the height of the glyph (upwards is positive); if a glyph does not extend above the baseline, it should be given a zero height, rather than a negative height. The depth subfield gives the depth of the glyph, that is, the distance below the baseline to which the glyph extends (downwards is positive); if a glyph does not extend below the baseline, it should be given a zero depth, rather than a negative depth. Italic corrections are relevant to glyphs in italic or oblique styles. The italic-correction is the amount of space that should be added after an oblique glyph to be followed immediately by an upright glyph. The left-italic-correction is the amount of space that should be added before an oblique glyph to be preceded immediately by an upright glyph. The subscript-correction is the amount of space that should be added after an oblique glyph to be followed by a subscript; it should be less than the italic correction.

For fonts used with typesetting devices, the type field gives a featural description of the glyph: it is a bit mask recording whether the glyph is an ascender, descender, both, or neither. When a \w escape sequence is interpolated, these values are bitwise or-ed together for each glyph and stored in the ct register. In font descriptions for terminal devices, all glyphs might have a type of zero, regardless of their appearance.


means the glyph lies entirely between the baseline and a horizontal line at the “x-height” of the font, as with “a”, “c”, and “x”;


means the glyph descends below the baseline, like “p”;


means the glyph ascends above the font’s x-height, like “A” or “b”); and


means the glyph is both an ascender and a descender—this is true of parentheses in some fonts.

The code field gives a numeric identifier that the postprocessor uses to render the glyph. The glyph can be specified to troff using this code by means of the \N escape sequence. The code can be any integer (that is, any integer parsable by the C standard library’s strtol(3) function).

The entity-name field defines an identifier for the glyph that the postprocessor uses to print the troff glyph name. This field is optional; it was introduced so that the grohtml output driver could encode its character set. For example, the glyph \[Po] is represented by “£” in HTML 4.0. For efficiency, these data are now compiled directly into grohtml. grops uses the field to build sub-encoding arrays for PostScript fonts containing more than 256 glyphs. Anything on the line after the entity-name field or “--” is ignored.

A line in the charset section can also have the form

name "

identifying name as another name for the glyph mentioned in the preceding line. Such aliases can be chained.

The directive kernpairs starts a list of kerning adjustments to be made to adjacent glyph pairs from this font. It contains a sequence of lines formatted as follows.

g1 g2 n

The foregoing means that when glyph g1 is typeset immediately before g2, the space between them should be increased by n. Most kerning pairs should have a negative value for n.



describes the output device name.


describes the font known as F on device name.

See also

Groff: The GNU Implementation of troff, by Trent A. Fisher and Werner Lemberg, is the primary groff manual. You can browse it interactively with “info groff”.

“Troff User’s Manual” by Joseph F. Ossanna, 1976 (revised by Brian W. Kernighan, 1992), AT&T Bell Laboratories Computing Science Technical Report No. 54, widely called simply “CSTR #54”, documents the language, device and font description file formats, and device-independent output format referred to collectively in groff documentation as “AT&T troff”.

“A Typesetter-independent TROFF” by Brian W. Kernighan, 1982, AT&T Bell Laboratories Computing Science Technical Report No. 97, provides additional insights into the device and font description file formats and device-independent output format.

groff(1), subsection “Utilities”, lists programs available for describing fonts in a variety of formats such that groff output drivers can use them.

troff(1) documents the default device and font description file search path.

groff_out(5), addftinfo(1)