GROFF_OUT(5) File Formats Manual GROFF_OUT(5)
NAME
groff_out - groff intermediate output format
DESCRIPTION
This manual page describes the intermediate output format of the GNU
roff(7) text processing system groff(1). This output is produced by a
run of the GNU troff(1) program. It contains already all device-spe-
cific information, but it is not yet fed into a device postprocessor
program.
As the GNU roff processor groff(1) is a wrapper program around troff
that automatically calls a postprocessor, this output does not show up
normally. This is why it is called intermediate within the groff sys-
tem. The groff program provides the option -Z to inhibit postprocess-
ing, such that the produced intermediate output is sent to standard
output just like calling troff manually.
In this document, the term troff output describes what is output by the
GNU troff program, while intermediate output refers to the language
that is accepted by the parser that prepares this output for the post-
processors. This parser is smarter on whitespace and implements obso-
lete elements for compatibility, otherwise both formats are the same.
Both formats can be viewed directly with gxditview(1).
The main purpose of the intermediate output concept is to facilitate
the development of postprocessors by providing a common programming in-
terface for all devices. It has a language of its own that is com-
pletely different from the groff(7) language. While the groff language
is a high-level programming language for text processing, the interme-
diate output language is a kind of low-level assembler language by
specifying all positions on the page for writing and drawing.
The pre-groff roff versions are denoted as classical troff. The inter-
mediate output produced by groff is fairly readable, while classical
troff output was hard to understand because of strange habits that are
still supported, but not used any longer by GNU troff.
LANGUAGE CONCEPTS
During the run of troff, the roff input is cracked down to the informa-
tion on what has to be printed at what position on the intended device.
So the language of the intermediate output format can be quite small.
Its only elements are commands with or without arguments. In this doc-
ument, the term "command" always refers to the intermediate output lan-
guage, never to the roff language used for document formatting. There
are commands for positioning and text writing, for drawing, and for de-
vice controlling.
Separation
Classical troff output had strange requirements on whitespace. The
groff output parser, however, is smart about whitespace by making it
maximally optional. The whitespace characters, i.e., the tab, space,
and newline characters, always have a syntactical meaning. They are
never printable because spacing within the output is always done by po-
sitioning commands.
Any sequence of space or tab characters is treated as a single syntac-
tical space. It separates commands and arguments, but is only required
when there would occur a clashing between the command code and the ar-
guments without the space. Most often, this happens when variable
length command names, arguments, argument lists, or command clusters
meet. Commands and arguments with a known, fixed length need not be
separated by syntactical space.
A line break is a syntactical element, too. Every command argument can
be followed by whitespace, a comment, or a newline character. Thus a
syntactical line break is defined to consist of optional syntactical
space that is optionally followed by a comment, and a newline charac-
ter.
The normal commands, those for positioning and text, consist of a sin-
gle letter taking a fixed number of arguments. For historical reasons,
the parser allows stacking of such commands on the same line, but for-
tunately, in groff intermediate output, every command with at least one
argument is followed by a line break, thus providing excellent read-
ability.
The other commands -- those for drawing and device controlling -- have
a more complicated structure; some recognize long command names, and
some take a variable number of arguments. So all D and x commands were
designed to request a syntactical line break after their last argument.
Only one command, `x X' has an argument that can stretch over several
lines, all other commands must have all of their arguments on the same
line as the command, i.e., the arguments may not be split by a line
break.
Empty lines, i.e., lines containing only space and/or a comment, can
occur everywhere. They are just ignored.
Argument Units
Some commands take integer arguments that are assumed to represent val-
ues in a measurement unit, but the letter for the corresponding scale
indicator is not written with the output command arguments; see
groff(7) and Groff: The GNU Implementation of troff, the groff Texinfo
manual, for more on this topic. Most commands assume the scale indica-
tor u, the basic unit of the device, some use z, the scaled point unit
of the device, while others, such as the color commands expect plain
integers. Note that these scale indicators are relative to the chosen
device. They are defined by the parameters specified in the device's
DESC file; see groff_font(5).
Note that single characters can have the eighth bit set, as can the
names of fonts and special characters (this is, glyphs). The names of
glyphs and fonts can be of arbitrary length. A glyph that is to be
printed will always be in the current font.
A string argument is always terminated by the next whitespace character
(space, tab, or newline); an embedded # character is regarded as part
of the argument, not as the beginning of a comment command. An integer
argument is already terminated by the next non-digit character, which
then is regarded as the first character of the next argument or com-
mand.
Document Parts
A correct intermediate output document consists of two parts, the pro-
logue and the body.
The task of the prologue is to set the general device parameters using
three exactly specified commands. The groff prologue is guaranteed to
consist of the following three lines (in that order):
x T device
x res n h v
x init
with the arguments set as outlined in subsection "Device Control Com-
mands" below. However, the parser for the intermediate output format
is able to swallow additional whitespace and comments as well.
The body is the main section for processing the document data. Syntac-
tically, it is a sequence of any commands different from the ones used
in the prologue. Processing is terminated as soon as the first x stop
command is encountered; the last line of any groff intermediate output
always contains such a command.
Semantically, the body is page oriented. A new page is started by a
p command. Positioning, writing, and drawing commands are always done
within the current page, so they cannot occur before the first p com-
mand. Absolute positioning (by the H and V commands) is done relative
to the current page, all other positioning is done relative to the cur-
rent location within this page.
COMMAND REFERENCE
This section describes all intermediate output commands, the classical
commands as well as the groff extensions.
Comment Command
#anything<end-of-line>
A comment. Ignore any characters from the # character up to the
next newline character.
This command is the only possibility for commenting in the intermediate
output. Each comment can be preceded by arbitrary syntactical space;
every command can be terminated by a comment.
Simple Commands
The commands in this subsection have a command code consisting of a
single character, taking a fixed number of arguments. Most of them are
commands for positioning and text writing. These commands are smart
about whitespace. Optionally, syntactical space can be inserted be-
fore, after, and between the command letter and its arguments. All of
these commands are stackable, i.e., they can be preceded by other sim-
ple commands or followed by arbitrary other commands on the same line.
A separating syntactical space is only necessary when two integer argu-
ments would clash or if the preceding argument ends with a string argu-
ment.
C xxx<white-space>
Print a glyph (special character) named xxx. The trailing syn-
tactical space or line break is necessary to allow glyph names
of arbitrary length. The glyph is printed at the current print
position; the glyph's size is read from the font file. The
print position is not changed.
c c Print glyph with single-letter name c at the current print posi-
tion; the glyph's size is read from the font file. The print
position is not changed.
f n Set font to font number n (a non-negative integer).
H n Move right to the absolute vertical position n (a non-negative
integer in basic units u) relative to left edge of current page.
h n Move n (a non-negative integer) basic units u horizontally to
the right. [CSTR #54] allows negative values for n also, but
groff doesn't use this.
m color-scheme [component ...]
Set the color for text (glyphs), line drawing, and the outline
of graphic objects using different color schemes; the analogous
command for the filling color of graphic objects is DF. The
color components are specified as integer arguments between 0
and 65536. The number of color components and their meaning
vary for the different color schemes. These commands are gener-
ated by the groff escape sequence \m. No position changing.
These commands are a groff extension.
mc cyan magenta yellow
Set color using the CMY color scheme, having the 3 color
components cyan, magenta, and yellow.
md Set color to the default color value (black in most
cases). No component arguments.
mg gray
Set color to the shade of gray given by the argument, an
integer between 0 (black) and 65536 (white).
mk cyan magenta yellow black
Set color using the CMYK color scheme, having the 4 color
components cyan, magenta, yellow, and black.
mr red green blue
Set color using the RGB color scheme, having the 3 color
components red, green, and blue.
N n Print glyph with index n (an integer, normally non-negative) of
the current font. The print position is not changed. If
-T html or -T xhtml is used, negative values are emitted also to
indicate an unbreakable space with given width. For example,
N -193 represents an unbreakable space which has a width of
193u. This command is a groff extension.
n b a Inform the device about a line break, but no positioning is done
by this command. In classical troff, the integer arguments b
and a informed about the space before and after the current line
to make the intermediate output more human readable without per-
forming any action. In groff, they are just ignored, but they
must be provided for compatibility reasons.
p n Begin a new page in the outprint. The page number is set to n.
This page is completely independent of pages formerly processed
even if those have the same page number. The vertical position
on the outprint is automatically set to 0. All positioning,
writing, and drawing is always done relative to a page, so a
p command must be issued before any of these commands.
s n Set point size to n scaled points (this is unit z in GNU troff).
Classical troff used the unit points (p) instead; see section
"Compatibility" below.
t xyz...<white-space>
t xyz... dummy-arg<white-space>
Print a word, i.e., a sequence of glyphs with single-letter
names x, y, z, etc., terminated by a space character or a line
break; an optional second integer argument is ignored (this al-
lows the formatter to generate an even number of arguments).
The first glyph should be printed at the current position, the
current horizontal position should then be increased by the
width of the first glyph, and so on for each glyph. The widths
of the glyph are read from the font file, scaled for the current
point size, and rounded to a multiple of the horizontal resolu-
tion. Special characters (glyphs with names longer than a sin-
gle letter) cannot be printed using this command; use the C com-
mand for those glyphs. This command is a groff extension; it is
only used for devices whose DESC file contains the tcommand key-
word; see groff_font(5).
u n xyz...<white-space>
Print word with track kerning. This is the same as the t com-
mand except that after printing each glyph, the current horizon-
tal position is increased by the sum of the width of that glyph
and n (an integer in basic units u). This command is a groff
extension; it is only used for devices whose DESC file contains
the tcommand keyword; see groff_font(5).
V n Move down to the absolute vertical position n (a non-negative
integer in basic units u) relative to upper edge of current
page.
v n Move n basic units u down (n is a non-negative integer).
[CSTR #54] allows negative values for n also, but groff doesn't
use this.
w Informs about a paddable whitespace to increase readability.
The spacing itself must be performed explicitly by a move com-
mand.
Graphics Commands
Each graphics or drawing command in the intermediate output starts with
the letter D followed by one or two characters that specify a subcom-
mand; this is followed by a fixed or variable number of integer argu-
ments that are separated by a single space character. A D command may
not be followed by another command on the same line (apart from a com-
ment), so each D command is terminated by a syntactical line break.
troff output follows the classical spacing rules (no space between com-
mand and subcommand, all arguments are preceded by a single space char-
acter), but the parser allows optional space between the command let-
ters and makes the space before the first argument optional. As usual,
each space can be any sequence of tab and space characters.
Some graphics commands can take a variable number of arguments. In
this case, they are integers representing a size measured in basic
units u. The h arguments stand for horizontal distances where positive
means right, negative left. The v arguments stand for vertical dis-
tances where positive means down, negative up. All these distances are
offsets relative to the current location.
Unless indicated otherwise, each graphics command directly corresponds
to a similar groff \D escape sequence; see groff(7).
Unknown D commands are assumed to be device-specific. Its arguments
are parsed as strings; the whole information is then sent to the post-
processor.
In the following command reference, the syntax element <line-break>
means a syntactical line break as defined in subsection "Separation"
above.
D~ h1 v1 h2 v2 ... hn vn<line-break>
Draw B-spline from current position to offset (h1, v1), then to
offset (h2, v2) if given, etc., up to (hn, vn). This command
takes a variable number of argument pairs; the current position
is moved to the terminal point of the drawn curve.
Da h1 v1 h2 v2<line-break>
Draw arc from current position to (h1, v1)+(h2, v2) with center
at (h1, v1); then move the current position to the final point
of the arc.
DC d<line-break>
DC d dummy-arg<line-break>
Draw a solid circle using the current fill color with diameter d
(integer in basic units u) with leftmost point at the current
position; then move the current position to the rightmost point
of the circle. An optional second integer argument is ignored
(this allows the formatter to generate an even number of argu-
ments). This command is a groff extension.
Dc d<line-break>
Draw circle line with diameter d (integer in basic units u) with
leftmost point at the current position; then move the current
position to the rightmost point of the circle.
DE h v<line-break>
Draw a solid ellipse in the current fill color with a horizontal
diameter of h and a vertical diameter of v (both integers in ba-
sic units u) with the leftmost point at the current position;
then move to the rightmost point of the ellipse. This command
is a groff extension.
De h v<line-break>
Draw an outlined ellipse with a horizontal diameter of h and a
vertical diameter of v (both integers in basic units u) with the
leftmost point at current position; then move to the rightmost
point of the ellipse.
DF color-scheme [component ...]<line-break>
Set fill color for solid drawing objects using different color
schemes; the analogous command for setting the color of text,
line graphics, and the outline of graphic objects is m. The
color components are specified as integer arguments between 0
and 65536. The number of color components and their meaning
vary for the different color schemes. These commands are gener-
ated by the groff escape sequences \D'F ...' and \M (with no
other corresponding graphics commands). No position changing.
This command is a groff extension.
DFc cyan magenta yellow<line-break>
Set fill color for solid drawing objects using the CMY
color scheme, having the 3 color components cyan, ma-
genta, and yellow.
DFd <line-break>
Set fill color for solid drawing objects to the default
fill color value (black in most cases). No component ar-
guments.
DFg gray<line-break>
Set fill color for solid drawing objects to the shade of
gray given by the argument, an integer between 0 (black)
and 65536 (white).
DFk cyan magenta yellow black<line-break>
Set fill color for solid drawing objects using the CMYK
color scheme, having the 4 color components cyan, ma-
genta, yellow, and black.
DFr red green blue<line-break>
Set fill color for solid drawing objects using the RGB
color scheme, having the 3 color components red, green,
and blue.
Df n<line-break>
The argument n must be an integer in the range -32767 to 32767.
0<=n<=1000
Set the color for filling solid drawing objects to a
shade of gray, where 0 corresponds to solid white, 1000
(the default) to solid black, and values in between to
intermediate shades of gray; this is obsoleted by command
DFg.
n<0 or n>1000
Set the filling color to the color that is currently be-
ing used for the text and the outline, see command m.
For example, the command sequence
mg 0 0 65536
Df -1
sets all colors to blue.
No position changing. This command is a groff extension.
Dl h v<line-break>
Draw line from current position to offset (h, v) (integers in
basic units u); then set current position to the end of the
drawn line.
Dp h1 v1 h2 v2 ... hn vn<line-break>
Draw a polygon line from current position to offset (h1, v1),
from there to offset (h2, v2), etc., up to offset (hn, vn), and
from there back to the starting position. For historical rea-
sons, the position is changed by adding the sum of all arguments
with odd index to the actual horizontal position and the even
ones to the vertical position. Although this doesn't make sense
it is kept for compatibility. This command is a groff exten-
sion.
DP h1 v1 h2 v2 ... hn vn<line-break>
The same macro as the corresponding Dp command with the same ar-
guments, but draws a solid polygon in the current fill color
rather than an outlined polygon. The position is changed in the
same way as with Dp. This command is a groff extension.
Dt n<line-break>
Set the current line thickness to n (an integer in basic
units u) if n>0; if n=0 select the smallest available line
thickness; if n<0 set the line thickness proportional to the
point size (this is the default before the first Dt command was
specified). For historical reasons, the horizontal position is
changed by adding the argument to the actual horizontal posi-
tion, while the vertical position is not changed. Although this
doesn't make sense it is kept for compatibility. This command
is a groff extension.
Device Control Commands
Each device control command starts with the letter x followed by a
space character (optional or arbitrary space/tab in groff) and a sub-
command letter or word; each argument (if any) must be preceded by a
syntactical space. All x commands are terminated by a syntactical line
break; no device control command can be followed by another command on
the same line (except a comment).
The subcommand is basically a single letter, but to increase readabil-
ity, it can be written as a word, i.e., an arbitrary sequence of char-
acters terminated by the next tab, space, or newline character. All
characters of the subcommand word but the first are simply ignored.
For example, troff outputs the initialization command x i as x init and
the resolution command x r as x res. But writings like x i_like_groff
and x roff_is_groff are accepted as well to mean the same commands.
In the following, the syntax element <line-break> means a syntactical
line break as defined in subsection "Separation" above.
xF name<line-break>
(Filename control command)
Use name as the intended name for the current file in error re-
ports. This is useful for remembering the original file name
when groff uses an internal piping mechanism. The input file is
not changed by this command. This command is a groff extension.
xf n s<line-break>
(font control command)
Mount font position n (a non-negative integer) with font named s
(a text word); see groff_font(5).
xH n<line-break>
(Height control command)
Set character height to n (a positive integer in scaled
points z). Classical troff used the unit points (p) instead;
see section "Compatibility" below.
xi <line-break>
(init control command)
Initialize device. This is the third command of the prologue.
xp <line-break>
(pause control command)
Parsed but ignored. The classical documentation reads pause de-
vice, can be restarted.
xr n h v<line-break>
(resolution control command)
Resolution is n, while h is the minimal horizontal motion, and v
the minimal vertical motion possible with this device; all argu-
ments are positive integers in basic units u per inch. This is
the second command of the prologue.
xS n<line-break>
(Slant control command)
Set slant to n degrees (an integer in basic units u).
xs <line-break>
(stop control command)
Terminates the processing of the current file; issued as the
last command of any intermediate troff output.
xt <line-break>
(trailer control command)
Generate trailer information, if any. In groff, this is actu-
ally just ignored.
xT xxx<line-break>
(Typesetter control command)
Set name of device to word xxx, a sequence of characters ended
by the next whitespace character. The possible device names co-
incide with those from the groff -T option. This is the first
command of the prologue.
xu n<line-break>
(underline control command)
Configure underlining of spaces. If n is 1, start underlining
of spaces; if n is 0, stop underlining of spaces. This is
needed for the cu request in nroff mode and is ignored other-
wise. This command is a groff extension.
xX anything<line-break>
(X-escape control command)
Send string anything uninterpreted to the device. If the line
following this command starts with a + character this line is
interpreted as a continuation line in the following sense. The
+ is ignored, but a newline character is sent instead to the de-
vice, the rest of the line is sent uninterpreted. The same ap-
plies to all following lines until the first character of a line
is not a + character. This command is generated by the groff
escape sequence \X. The line-continuing feature is a groff ex-
tension.
Obsolete Command
In classical troff output, emitting a single glyph was mostly done by a
very strange command that combined a horizontal move and the printing
of a glyph. It didn't have a command code, but is represented by a
3-character argument consisting of exactly 2 digits and a character.
ddc Move right dd (exactly two decimal digits) basic units u, then
print glyph with single-letter name c.
In groff, arbitrary syntactical space around and within this
command is allowed to be added. Only when a preceding command
on the same line ends with an argument of variable length a sep-
arating space is obligatory. In classical troff, large clusters
of these and other commands were used, mostly without spaces;
this made such output almost unreadable.
For modern high-resolution devices, this command does not make sense
because the width of the glyphs can become much larger than two decimal
digits. In groff, this is only used for the devices X75, X75-12, X100,
and X100-12. For other devices, the commands t and u provide a better
functionality.
POSTPROCESSING
The roff postprocessors are programs that have the task to translate
the intermediate output into actions that are sent to a device. A de-
vice can be some piece of hardware such as a printer, or a software
file format suitable for graphical or text processing. The groff sys-
tem provides powerful means that make the programming of such postpro-
cessors an easy task.
There is a library function that parses the intermediate output and
sends the information obtained to the device via methods of a class
with a common interface for each device. So a groff postprocessor must
only redefine the methods of this class. For details, see the refer-
ence in section "Files" below.
EXAMPLES
This section presents the intermediate output generated from the same
input for three different devices. The input is the sentence hell
world fed into groff on the command line.
•igh-resolution device ps
shell> echo "hell world" | groff -Z -T ps
x T ps
x res 72000 1 1
x init
p1
x font 5 TR
f5
s10000
V12000
H72000
thell
wh2500
tw
H96620
torld
n12000 0
x trailer
V792000
x stop
This output can be fed into the postprocessor grops(1) to get its rep-
resentation as a PostScript file, or gropdf(1) to output directly to
PDF.
•ow-resolution device latin1
This is similar to the high-resolution device except that the posi-
tioning is done at a minor scale. Some comments (lines starting with
#) were added for clarification; they were not generated by the for-
matter.
shell> "hell world" | groff -Z -T latin1
# prologue
x T latin1
x res 240 24 40
x init
# begin a new page
p1
# font setup
x font 1 R
f1
s10
# initial positioning on the page
V40
H0
# write text `hell'
thell
# inform about a space, and do it by a horizontal jump
wh24
# write text `world'
tworld
# announce line break, but do nothing because ...
n40 0
# ... the end of the document has been reached
x trailer
V2640
x stop
This output can be fed into the postprocessor grotty(1) to get a for-
matted text document.
•lassical style output
As a computer monitor has a very low resolution compared to modern
printers the intermediate output for the X devices can use the jump-
and-write command with its 2-digit displacements.
shell> "hell world" | groff -Z -T X100
x T X100
x res 100 1 1
x init
p1
x font 5 TR
f5
s10
V16
H100
# write text with old-style jump-and-write command
ch07e07l03lw06w11o07r05l03dh7
n16 0
x trailer
V1100
x stop
This output can be fed into the postprocessor xditview(1x) or
gxditview(1) for displaying in X.
Due to the obsolete jump-and-write command, the text clusters in the
classical output are almost unreadable.
COMPATIBILITY
The intermediate output language of the classical troff was first docu-
mented in [CSTR #97] . The groff intermediate output format is compat-
ible with this specification except for the following features.
•he classical quasi device independence is not yet implemented.
•he old hardware was very different from what we use today. So the
groff devices are also fundamentally different from the ones in clas-
sical troff. For example, the classical PostScript device was called
post and had a resolution of 720 units per inch, while groff's ps de-
vice has a resolution of 72000 units per inch. Maybe, by implement-
ing some rescaling mechanism similar to the classical quasi device
independence, these could be integrated into modern groff.
•he B-spline command D~ is correctly handled by the intermediate out-
put parser, but the drawing routines aren't implemented in some of
the postprocessor programs.
•he argument of the commands s and x H has the implicit unit scaled
point z in groff, while classical troff had point (p). This isn't an
incompatibility, but a compatible extension, for both units coincide
for all devices without a sizescale parameter, including all classi-
cal and the groff text devices. The few groff devices with a
sizescale parameter either did not exist, had a different name, or
seem to have had a different resolution. So conflicts with classical
devices are very unlikely.
•he position changing after the commands Dp, DP, and Dt is illogical,
but as old versions of groff used this feature it is kept for compat-
ibility reasons.
The differences between groff and classical troff are documented in
groff_diff(7).
FILES
/usr/local/share/groff/1.22.4/font/devname/DESC
Device description file for device name.
src/libs/libdriver/input.cpp
Defines the parser and postprocessor for the intermediate out-
put. It is located relative to the top directory of the groff
source tree. This parser is the definitive specification of the
groff intermediate output format.
AUTHORS
James Clark wrote an early version of this document, which described
only the differences between ditroff(7)'s output format and that of GNU
roff. The present version was completely rewritten in 2001 by Bernd
Warken <groff-bernd.warken-72@web.de>.
SEE ALSO
A reference like groff(7) refers to a manual page; here groff in sec-
tion 7 of the man page documentation system. To read the example, look
up section 7 in your desktop help system or call from the shell prompt
shell> man 7 groff
For more details, see man(1).
groff(1)
option -Z and further readings on groff.
groff(7)
for details of the groff language such as numerical units and
escape sequences.
groff_font(5)
for details on the device scaling parameters of the DESC file.
troff(1)
generates the device-independent intermediate output.
roff(7)
for historical aspects and the general structure of roff sys-
tems.
groff_diff(7)
The differences between the intermediate output in groff and
classical troff.
gxditview(1)
Viewer for the intermediate output.
grodvi(1), grohtml(1), grolbp(1), grolj4(1), grops(1), grotty(1)
the groff postprocessor programs.
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".
The classical troff output language is described in two AT&T Bell Labs
CSTR documents available on-line at Bell Labs CSTR site <http://
cm.bell-labs.com/cm/cs/cstr.html>.
[CSTR #97]
A Typesetter-independent TROFF by Brian Kernighan is the origi-
nal and most comprehensive documentation on the output language;
see CSTR #97 <http://cm.bell-labs.com/cm/cs/cstr/97.ps.gz>.
[CSTR #54]
The 1992 revision of the Nroff/Troff User's Manual by J. F. Os-
sanna and Brian Kernighan isn't as comprehensive as [CSTR #97]
regarding the output language; see CSTR #54 <http://
cm.bell-labs.com/cm/cs/cstr/54.ps.gz>.
groff 1.22.4 17 December 2018 GROFF_OUT(5)
man2web Home...