For example,

.nr a 0 1 .nr xx 0 5 .nr foo 0 -2 \n+a, \n+a, \n+a, \n+a, \n+a .br \n-(xx, \n-(xx, \n-(xx, \n-(xx, \n-(xx .br \n+[foo], \n+[foo], \n+[foo], \n+[foo], \n+[foo]

produces

1, 2, 3, 4, 5 -5, -10, -15, -20, -25 -2, -4, -6, -8, -10

To change the increment value without changing the value of a register (a in the example), the following can be used:

.nr a \na 10

5.7.4 Assigning Formats

When a register is used in the text of an input file (as opposed to part of an expression), it is textually replaced (or interpolated) with a representation of that number. This output format can be changed to a variety of formats (numbers, Roman numerals, etc.). This is done using the af request.

Request: .af ident format

Change the output format of a number register. The first argument ident is the name of the number register to be changed, and the second argument format is the output format. The following output formats are available:

1

Decimal arabic numbers. This is the default format: 0, 1, 2, 3, '....

0...0

Decimal numbers with as many digits as specified. So, `00' would result in printing numbers as 01, 02, 03, ....

In fact, any digit instead of zero will do; gtroff only counts how many digits are specified. As a consequence, af's default format `1' could be specified as `0' also (and exactly this is returned by the \g escape, see below).

I

Upper-case Roman numerals: 0, I, II, III, IV, ....

i

Lower-case Roman numerals: 0, i, ii, iii, iv, ....

A

Upper-case letters: 0, A, B, C, ..., Z, AA, AB, ....

a

Lower-case letters: 0, a, b, c, ..., z, aa, ab, ....

Omitting the number register format causes a warning of type `missing'. See section 5.34 Debugging, for more details. Specifying a nonexistent format causes an error.

The following example produces `10, X, j, 010':

.nr a 10 .af a 1 \" the default format \na, .af a I \na, .af a a \na, .af a 001 \na

The largest number representable for the `i' and `I' formats is 39999 (or -39999); UNIX troff uses `z' and `w' to represent 10000 and 5000 in Roman numerals, and so does gtroff. Currently, the correct glyphs of Roman numeral five thousand and Roman numeral ten thousand (Unicode code points U+2182 and U+2181, respectively) are not available.

If ident doesn't exist, it is created.

Changing the output format of a read-only register causes an error. It is necessary to first copy the register's value to a writeable register, then apply the af request to this other register.

Escape: \gi\delimII\

Escape: \g(\arg\\delimII\

Escape: \g[\arg\\delimII\

Return the current format of the specified register ident (one-character name i, two-character name id). For example, `\ga' after the previous example would produce the string `000'. If the register hasn't been defined yet, nothing is returned.

.F

This string-valued register returns the current input file name.

.H

Horizontal resolution in basic units.

.V

Vertical resolution in basic units.

seconds

The number of seconds after the minute, normally in the range 0 to 59, but can be up to 61 to allow for leap seconds. Initialized at start-up of gtroff.

minutes

The number of minutes after the hour, in the range 0 to 59. Initialized at start-up of gtroff.

hours

The number of hours past midnight, in the range 0 to 23. Initialized at start-up of gtroff.

dw

Day of the week (1-7).

dy

Day of the month (1-31).

mo

Current month (1-12).

year

The current year.

yr

The current year minus 1900. Unfortunately, the documentation of UNIX Version 7's troff had a year 2000 bug: It incorrectly claimed that yr contains the last two digits of the year. That claim has never been true of either AT&T troff or GNU troff. Old troff input that looks like this:

'\" The following line stopped working after 1999 This document was formatted in 19\n(yr.

can be corrected as follows:

This document was formatted in \n[year].

or, to be portable to older troff versions, as follows:

.nr y4 1900+\n(yr This document was formatted in \n(y4.

.c

c.

The current input line number. Register `.c' is read-only, whereas `c.' (a gtroff extension) is writable also, affecting both `.c' and `c.'.

ln

The current output line number after a call to the nm request to activate line numbering.

See section 5.32 Miscellaneous, for more information about line numbering.

.x

The major version number. For example, if the version number is 1.03 then .x contains `1'.

.y

The minor version number. For example, if the version number is 1.03 then .y contains `03'.

.Y

The revision number of groff.

$$

The process ID of gtroff.

.g

Always 1. Macros should use this to determine whether they are running under GNU troff.

.A

If the command line option `-a' is used to produce an ASCII approximation of the output, this is set to 1, zero otherwise. See section 2.1 Options.

.P

This register is set to 1 (and to 0 otherwise) if the current page is actually being printed, i.e., if the `-o' option is being used to only print selected pages. See section 2.1 Options, for more information.

.T

If gtroff is called with the `-T' command line option, the number register .T is set to 1, and zero otherwise. See section 2.1 Options.

Additionally, gtroff predefines a single read-write string register .T which contains the current output device (for example, `latin1' or `ps').

5.8 Manipulating Filling and Adjusting

Various ways of causing breaks were given in 5.1.5 Implicit Line Breaks. The br request likewise causes a break. Several other requests also cause breaks, but implicitly. These are bp, ce, cf, fi, fl, in, nf, rj, sp, ti, and trf.

Request: .br \arg\

Break the current line, i.e., the input collected so far is emitted without adjustment.

If the no-break control character is used, gtroff suppresses the break:

a 'br b => a b

Initially, gtroff fills and adjusts text to both margins. Filling can be disabled via the nf request and re-enabled with the fi request.

Request: .fi \arg\

Register: \n[.u]

Activate fill mode (which is the default). This request implicitly enables adjusting; it also inserts a break in the text currently being filled. The read-only number register .u is set to 1.

The fill mode status is associated with the current environment (see section 5.27 Environments).

See 5.15 Line Control, for interaction with the \c escape.

Request: .nf \arg\

Activate no-fill mode. Input lines are output as-is, retaining line breaks and ignoring the current line length. This command implicitly disables adjusting; it also causes a break. The number register .u is set to 0.

The fill mode status is associated with the current environment (see section 5.27 Environments).

See 5.15 Line Control, for interaction with the \c escape.

Request: .ad [{mode

Register: \n[.j]

Set adjusting mode.

Activation and deactivation of adjusting is done implicitly with calls to the fi or nf requests.

mode can have one of the following values:

l

Adjust text to the left margin. This produces what is traditionally called ragged-right text.

r

Adjust text to the right margin, producing ragged-left text.

c

Center filled text. This is different to the ce request which only centers text without filling.

b

n

Justify to both margins. This is the default used by gtroff.

With no argument, gtroff adjusts lines in the same way it did before adjusting was deactivated (with a call to na, for example).

text .ad r text .ad c text .na text .ad \" back to centering text

The current adjustment mode is available in the read-only number register .j; it can be stored and subsequently used to set adjustment.

The adjustment mode status is associated with the current environment (see section 5.27 Environments).

Request: .na \arg\

Disable adjusting. This request won't change the current adjustment mode: A subsequent call to ad uses the previous adjustment setting.

The adjustment mode status is associated with the current environment (see section 5.27 Environments).

Request: .brp \arg\

Escape: \p\delimI\\arg\\delimII\

Adjust the current line and cause a break.

In most cases this produces very ugly results since gtroff doesn't have a sophisticated paragraph building algorithm (as TeX have, for example); instead, gtroff fills and adjusts a paragraph line by line:

This is an uninteresting sentence. This is an uninteresting sentence.\p This is an uninteresting sentence.

is formatted as

This is an uninteresting sentence. This is an uninteresting sentence. This is an uninteresting sentence.

Request: .ss word_space_size [{sentence_space_size

Register: \n[.ss]

Register: \n[.sss]

Change the minimum size of a space between filled words. It takes its units as one twelfth of the space width parameter for the current font. Initially both the word_space_size and sentence_space_size are 12.

If two arguments are given to the ss request, the second argument sets the sentence space size. If the second argument is not given, sentence space size is set to word_space_size. The sentence space size is used in two circumstances: If the end of a sentence occurs at the end of a line in fill mode, then both an inter-word space and a sentence space are added; if two spaces follow the end of a sentence in the middle of a line, then the second space is a sentence space. If a second argument is never given to the ss request, the behaviour of UNIX troff is the same as that exhibited by GNU troff. In GNU troff, as in UNIX troff, a sentence should always be followed by either a newline or two spaces.

The read-only number registers .ss and .sss hold the values of the parameters set by the first and second arguments of the ss request.

The word space and sentence space values are associated with the current environment (see section 5.27 Environments).

Contrary to AT&T troff, this request is not ignored if a TTY output device is used; the given values are then rounded down to a multiple of 12 (see section 5.35 Implementation Differences).

The request is ignored if there is no parameter.

Request: .ce [{nnn

Register: \n[.ce]

Center text. While the `.ad c' request also centers text, it fills the text as well. ce does not fill the text it affects. This request causes a break. The number of lines still to be centered is associated with the current environment (see section 5.27 Environments).

The following example demonstrates the differences. Here the input:

.ll 4i .ce 1000 This is a small text fragment which shows the differences between the `.ce' and the `.ad c' request. .ce 0 .ad c This is a small text fragment which shows the differences between the `.ce' and the `.ad c' request.

And here the result:

This is a small text fragment which shows the differences between the `.ce' and the `.ad c' request. This is a small text fragment which shows the differences between the `.ce' and the `.ad c' request.

With no arguments, ce centers the next line of text. nnn specifies the number of lines to be centered. If the argument is zero or negative, centering is disabled.

The basic length for centering text is the line length (as set with the ll request) minus the indentation (as set with the in request). Temporary indentation is ignored.

As can be seen in the previous example, it is a common idiom to turn on centering for a large number of lines, and to turn off centering after text to be centered. This is useful for any request which takes a number of lines as an argument.

The .ce read-only number register contains the number of lines remaining to be centered, as set by the ce request.

Request: .rj [{nnn

Register: \n[.rj]

Justify unfilled text to the right margin. Arguments are identical to the ce request. The .rj read-only number register is the number of lines to be right-justified as set by the rj request. This request causes a break. The number of lines still to be right-justified is associated with the current environment (see section 5.27 Environments).

5.9 Manipulating Hyphenation

As discussed in 5.1.2 Hyphenation, gtroff hyphenates words. There are a number of ways to influence hyphenation.

Request: .hy [{mode

Register: \n[.hy]

Enable hyphenation. The request has an optional numeric argument, mode, to restrict hyphenation if necessary:

1

The default argument if mode is omitted. Hyphenate without restrictions. This is also the start-up value of gtroff.

2

Do not hyphenate the last word on a page or column.

4

Do not hyphenate the last two characters of a word.

8

Do not hyphenate the first two characters of a word.

Values in the previous table are additive. For example, the value 12 causes gtroff to neither hyphenate the last two nor the first two characters of a word.

The current hyphenation restrictions can be found in the read-only number register `.hy'.

The hyphenation mode is associated with the current environment (see section 5.27 Environments).

Request: .nh \arg\

Disable hyphenation (i.e., set the hyphenation mode to zero). Note that the hyphenation mode of the last call to hy is not remembered.

The hyphenation mode is associated with the current environment (see section 5.27 Environments).

Request: .hlm [{nnn

Register: \n[.hlm]

Register: \n[.hlc]

Set the maximum number of consecutive hyphenated lines to nnn. If this number is negative, there is no maximum. The default value is -1 if nnn is omitted. This value is associated with the current environment (see section 5.27 Environments). Only lines output from a given environment count towards the maximum associated with that environment. Hyphens resulting from \% are counted; explicit hyphens are not.

The current setting of hlm is available in the .hlm read-only number register. Also the number of immediately preceding consecutive hyphenated lines are available in the read-only number register `.hlc'.

Request: .hw word1 word2 @dots{

Define how word1, word2, etc. are to be hyphenated. The words must be given with hyphens at the hyphenation points. For example:

.hw in-sa-lub-rious

Besides the space character, any character whose hyphenation code value is zero can be used to separate the arguments of hw (see the documentation for the hcode request below for more information). In addition, this request can be used more than once.

Hyphenation exceptions specified with the hw request are associated with the current hyphenation language; it causes an error if there is no current hyphenation language.

This request is ignored if there is no parameter.

In old versions of troff there was a limited amount of space to store such information; fortunately, with gtroff, this is no longer a restriction.

Escape: \%\delimI\\arg\\delimII\

Escape: \:

To tell gtroff how to hyphenate words on the fly, use the \% escape, also known as the hyphenation character. Preceding a word with this character prevents it from being hyphenated; putting it inside a word indicates to gtroff that the word may be hyphenated at that point. Note that this mechanism only affects that one occurrence of the word; to change the hyphenation of a word for the entire document, use the hw request.

The \: escape inserts a zero-width break point (that is, the word breaks but without adding a hyphen).

... check the /var/log/\:httpd/\:access_log file ...

Note that \X and \Y start a word, that is, the \% escape in (say) ` \X'...'\%foobar' and ` \Y'...'\%foobar' no longer prevents hyphenation but inserts a hyphenation point at the beginning of `foobar'; most likely this isn't what you want to do.

Request: .hc [{char

Change the hyphenation character to char. This character then works the same as the \% escape, and thus, no longer appears in the output. Without an argument, hc resets the hyphenation character to be \% (the default) only.

The hyphenation character is associated with the current environment (see section 5.27 Environments).

Request: .hpf pattern_file

Request: .hpfa pattern_file

Request: .hpfcode a b [c d @dots{

Read in a file of hyphenation patterns. This file is searched for in the same way as `name.tmac' (or `tmac.name') is searched for if the `-mname' option is specified.

It should have the same format as (simple) TeX patterns files. More specifically, the following scanning rules are implemented.

• A percent sign starts a comment (up to the end of the line) even if preceded by a backslash.

• No support for `digraphs' like \$.

• ^^xx (x is 0-9 or a-f) and ^^x (character code of x in the range 0-127) are recognized; other use of ^ causes an error.

• No macro expansion.

• hpf checks for the expression \patterns{...} (possibly with whitespace before and after the braces). Everything between the braces is taken as hyphenation patterns. Consequently, { and } are not allowed in patterns.

• Similarly, \hyphenation{...} gives a list of hyphenation exceptions.

• \endinput is recognized also.

• For backwards compatibility, if \patterns is missing, the whole file is treated as a list of hyphenation patterns (only recognizing the % character as the start of a comment).

If no hpf request is specified (either in the document or in a macro package), gtroff won't hyphenate at all.

The hpfa request appends a file of patterns to the current list.

The hpfcode request defines mapping values for character codes in hyphenation patterns. hpf or hpfa then apply the mapping (after reading the patterns) before replacing or appending them to the current list of patterns. Its arguments are pairs of character codes -- integers from 0 to 255. The request maps character code a to code b, code c to code d, and so on. You can use character codes which would be invalid otherwise.

The set of hyphenation patterns is associated with the current language set by the hla request. The hpf request is usually invoked by the `troffrc' or `troffrc-end' file; by default, `troffrc' loads hyphenation patterns for American English (in file `hyphen.us').

A second call to hpf (for the same language) will replace the hyphenation patterns with the new ones.

Invoking hpf causes an error if there is no current hyphenation language.

Request: .hcode c1 code1 c2 code2 @dots{

Set the hyphenation code of character c1 to code1, that of c2 to code2, etc. A hyphenation code must be a single input character (not a special character) other than a digit or a space. Initially each lower-case letter (`a'-`z') has its hyphenation code set to itself, and each upper-case letter (`A'-`Z') has a hyphenation code which is the lower-case version of itself.

This request is ignored if it has no parameter.

Request: .hym [{length

Register: \n[.hym]

Set the (right) hyphenation margin to length. If the current adjustment mode is not `b' or `n', the line is not hyphenated if it is shorter than length. Without an argument, the hyphenation margin is reset to its default value, which is 0. The default scaling indicator for this request is `m'. The hyphenation margin is associated with the current environment (see section 5.27 Environments).

A negative argument resets the hyphenation margin to zero, emitting a warning of type `range'.

The current hyphenation margin is available in the .hym read-only number register.

Request: .hys [{hyphenation_space

Register: \n[.hys]

Set the hyphenation space to hyphenation_space. If the current adjustment mode is `b' or `n', don't hyphenate the line if it can be justified by adding no more than hyphenation_space extra space to each word space. Without argument, the hyphenation space is set to its default value, which is 0. The default scaling indicator for this request is `m'. The hyphenation space is associated with the current environment (see section 5.27 Environments).

A negative argument resets the hyphenation space to zero, emitting a warning of type `range'.

The current hyphenation space is available in the .hys read-only number register.

Request: .shc [{glyph

Set the soft hyphen character to glyph.@footnotegroff_99.html{@dfn{Soft hyphen character is a misnomer since it is an output glyph. If the argument is omitted, the soft hyphen character is set to the default glyph \(hy (this is the start-up value of gtroff also). The soft hyphen character is the glyph that is inserted when a word is hyphenated at a line break. If the soft hyphen character does not exist in the font of the character immediately preceding a potential break point, then the line is not broken at that point. Neither definitions (specified with the char request) nor translations (specified with the tr request) are considered when finding the soft hyphen character.

Request: .hla language

Register: \n[.hla]

Set the current hyphenation language to the string language. Hyphenation exceptions specified with the hw request and hyphenation patterns specified with the hpf and hpfa requests are both associated with the current hyphenation language. The hla request is usually invoked by the `troffrc' or the `troffrc-end' files; `troffrc' sets the default language to `us'.

The current hyphenation language is available as a string in the read-only number register `.hla'.

.ds curr_language \n[.hla] \*[curr_language] => us

5.10 Manipulating Spacing

Request: .sp [{distance

Space downwards distance. With no argument it advances 1@w{ line. A negative argument causes gtroff to move up the page the specified distance. If the argument is preceded by a `|' then gtroff moves that distance from the top of the page. This request causes a line break. The default scaling indicator is `v'.

Request: .ls [{nnn

Register: \n[.L]

Output nnn-1 blank lines after each line of text. With no argument, gtroff uses the previous value before the last ls call.

.ls 2 \" This causes double-spaced output .ls 3 \" This causes triple-spaced output .ls \" Again double-spaced

The line spacing is associated with the current environment (see section 5.27 Environments).

The read-only number register .L contains the current line spacing setting.

See section 5.19.1 Changing Type Sizes, for the requests vs and pvs as alternatives to ls.

Escape: \x'spacing'

Register: \n[.a]

Sometimes, extra vertical spacing is only needed occasionally, e.g. to allow space for a tall construct (like an equation). The \x escape does this. The escape is given a numerical argument, usually enclosed in quotes (like `\x'3p''); the default scaling indicator is `v'. If this number is positive extra vertical space is inserted below the current line. A negative number adds space above. If this escape is used multiple times on the same line, the maximum of the values is used.

See section 5.6.3 Escapes, for details on parameter delimiting characters.

The .a read-only number register contains the most recent (nonnegative) extra vertical line space.

Using \x can be necessary in combination with the \b escape, as the following example shows.

This is a test with the \[rs]b escape. .br This is a test with the \[rs]b escape. .br This is a test with \b'xyz'\x'-1m'\x'1m'. .br This is a test with the \[rs]b escape. .br This is a test with the \[rs]b escape.

produces

This is a test with the \b escape. This is a test with the \b escape. x This is a test with y. z This is a test with the \b escape. This is a test with the \b escape.

Request: .ns \arg\

Request: .rs \arg\

Register: \n[.ns]

Enable no-space mode. In this mode, spacing (either via sp or via blank lines) is disabled. The bp request to advance to the next page is also disabled, except if it is accompanied by a page number (see 5.17 Page Control, for more information). This mode ends when actual text is output or the rs request is encountered which ends no-space mode. The read-only number register .ns is set to 1 as long as no-space mode is active.

This request is useful for macros that conditionally insert vertical space before the text starts (for example, a paragraph macro could insert some space except when it is the first paragraph after a section header).

5.11 Tabs and Fields

A tab character (ASCII char 9, EBCDIC char@w{ 5) causes a horizontal movement to the next tab stop (much like it did on a typewriter).

Escape: \t\delimI\\arg\\delimII\

This escape is a non-interpreted tab character. In copy mode (see section 5.22.1 Copy-in Mode), \t is the same as a real tab character.

Request: .ta [{n1

Register: \n[.tabs]

Change tab stop positions. This request takes a series of tab specifiers as arguments (optionally divided into two groups with the letter `T') which indicate where each tab stop is to be (overriding any previous settings).

Tab stops can be specified absolutely, i.e., as the distance from the left margin. For example, the following sets 6 tab stops every one inch.

.ta 1i 2i 3i 4i 5i 6i

Tab stops can also be specified using a leading `+' which means that the specified tab stop is set relative to the previous tab stop. For example, the following is equivalent to the previous example.

.ta 1i +1i +1i +1i +1i +1i

gtroff supports an extended syntax to specify repeat values after the `T' mark (these values are always taken as relative) -- this is the usual way to specify tabs set at equal intervals. The following is, yet again, the same as the previous examples. It does even more since it defines an infinite number of tab stops separated by one inch.

.ta T 1i

Now we are ready to interpret the full syntax given at the beginning: Set tabs at positions n1, n2, ..., nn and then set tabs at nn+r1, nn+r2, ..., nn+rn and then at nn+rn+r1, nn+rn+r2, ..., nn+rn+rn, and so on.

Example: `4c +6c T 3c 5c 2c' is equivalent to @samp{4c 10c 13c 18c 20c 23c 28c 30c ....

The material in each tab column (i.e., the column between two tab stops) may be justified to the right or left or centered in the column. This is specified by appending `R', `L', or `C' to the tab specifier. The default justification is `L'. Example:

.ta 1i 2iC 3iR

Some notes:

• The default unit of the ta request is `m'.

• A tab stop is converted into a non-breakable horizontal movement which can be neither stretched nor squeezed. For example,

  .ds foo a\tb\tc .ta T 5i \*[foo]

  creates a single line which is a bit longer than 10 inches (a string is used to show exactly where the tab characters are). Now consider the following:

  .ds bar a\tb b\tc .ta T 5i \*[bar]

  gtroff first converts the tab stops of the line into unbreakable horizontal movements, then splits the line after the second `b' (assuming a sufficiently short line length). Usually, this isn't what the user wants.

• Superfluous tabs (i.e., tab characters which do not correspond to a tab stop) are ignored except the first one which delimits the characters belonging to the last tab stop for right-justifying or centering. Consider the following example

  .ds Z foo\tbar\tfoo .ds ZZ foo\tbar\tfoobar .ds ZZZ foo\tbar\tfoo\tbar .ta 2i 4iR \*[Z] .br \*[ZZ] .br \*[ZZZ] .br

  which produces the following output:

  foo bar foo foo bar foobar foo bar foobar

  The first line right-justifies the second `foo' relative to the tab stop. The second line right-justifies `foobar'. The third line finally right-justifies only `foo' because of the additional tab character which marks the end of the string belonging to the last defined tab stop.

• Tab stops are associated with the current environment (see section 5.27 Environments).

• Calling ta without an argument removes all tab stops.

• The start-up value of gtroff is `T 0.5i' in troff mode and `T 0.8i' in nroff mode (the latter is done with an explicit call to the ta request in the file `tty.tmac'.

The read-only number register .tabs contains a string representation of the current tab settings suitable for use as an argument to the ta request.

.ds tab-string \n[.tabs] \*[tab-string] => T120u

The troff version of the Plan 9 operating system uses register .S for the same purpose.

Request: .tc [{fill-glyph

Normally gtroff fills the space to the next tab stop with whitespace. This can be changed with the tc request. With no argument gtroff reverts to using whitespace, which is the default. The value of this tab repetition character is associated with the current environment (see section 5.27 Environments).@footnotegroff_101.html{Tab repetition character is a misnomer since it is an output glyph.

Request: .linetabs n

Register: \n[.linetabs]

If n is missing or not zero, enable line-tabs mode, or disable it otherwise (the default). In line-tabs mode, gtroff computes tab distances relative to the (current) output line instead of the input line.

For example, the following code:

.ds x a\t\c .ds y b\t\c .ds z c .ta 1i 3i \*x \*y \*z

in normal mode, results in the output

a b c

in line-tabs mode, the same code outputs

a b c

Line-tabs mode is associated with the current environment. The read-only register .linetabs is set to 1 if in line-tabs mode, and 0 in normal mode.

5.11.1 Leaders
5.11.2 Fields

5.11.1 Leaders

Sometimes it may may be desirable to use the tc request to fill a particular tab stop with a given glyph (for example dots in a table of contents), but also normal tab stops on the rest of the line. For this gtroff provides an alternate tab mechanism, called leaders which does just that.

A leader character (character code 1) behaves similarly to a tab character: It moves to the next tab stop. The only difference is that for this movement, the fill glyph defaults to a period character and not to space.

Escape: \a\delimI\\arg\\delimII\

This escape is a non-interpreted leader character. In copy mode (see section 5.22.1 Copy-in Mode), \a is the same as a real leader character.

Request: .lc [{fill-glyph

Declare the leader repetition character.@footnotegroff_102.html{@dfn{Leader repetition character is a misnomer since it is an output glyph. Without an argument, leaders act the same as tabs (i.e., using whitespace for filling). gtroff's start-up value is a dot (`.'). The value of the leader repetition character is associated with the current environment (see section 5.27 Environments).

For a table of contents, to name an example, tab stops may be defined so that the section number is one tab stop, the title is the second with the remaining space being filled with a line of dots, and then the page number slightly separated from the dots.

.ds entry 1.1\tFoo\a\t12 .lc . .ta 1i 5i +.25i \*[entry]

This produces

1.1 Foo.......................................... 12

5.11.2 Fields

Fields are a more general way of laying out tabular data. A field is defined as the data between a pair of delimiting characters. It contains substrings which are separated by padding characters. The width of a field is the distance on the input line from the position where the field starts to the next tab stop. A padding character inserts stretchable space similar to TeX's \hss command (thus it can even be negative) to make the sum of all substring lengths plus the stretchable space equal to the field width. If more than one padding character is inserted, the available space is evenly distributed among them.

Request: .fc [{delim-char

Define a delimiting and a padding character for fields. If the latter is missing, the padding character defaults to a space character. If there is no argument at all, the field mechanism is disabled (which is the default). Note that contrary to e.g. the tab repetition character, delimiting and padding characters are not associated to the current environment (see section 5.27 Environments).

Example:

.fc # ^ .ta T 3i #foo^bar^smurf# .br #foo^^bar^smurf#

and here the result:

foo bar smurf foo bar smurf

5.12 Character Translations

The control character (`.') and the no-break control character (`'') can be changed with the cc and c2 requests, respectively.

Request: .cc [{c

Set the control character to c. With no argument the default control character `.' is restored. The value of the control character is associated with the current environment (see section 5.27 Environments).

Request: .c2 [{c

Set the no-break control character to c. With no argument the default control character `'' is restored. The value of the no-break control character is associated with the current environment (see section 5.27 Environments).

Request: .eo \arg\

Disable the escape mechanism completely. After executing this request, the backslash character `\' no longer starts an escape sequence.

This request can be very helpful in writing macros since it is not necessary then to double the escape character. Here an example:

.\" This is a simplified version of the .\" .BR request from the man macro package .eo .de BR . ds result \& . while (\n[.$] >= 2) \{\ . as result \fB\$1\fR\$2 . shift 2 . \} . if \n[.$] .as result \fB\$1 \*[result] . ft R .. .ec

Request: .ec [{c

Set the escape character to c. With no argument the default escape character `\' is restored. It can be also used to re-enable the escape mechanism after an eo request.

Note that changing the escape character globally will likely break macro packages since gtroff has no mechanism to `intern' macros, i.e., to convert a macro definition into an internal form which is independent of its representation (TeX has this mechanism). If a macro is called, it is executed literally.

Request: .ecs \arg\

Request: .ecr \arg\

The ecs request saves the current escape character in an internal register. Use this request in combination with the ec request to temporarily change the escape character.

The ecr request restores the escape character saved with ecs. Without a previous call to ecs, this request sets the escape character to \.

Escape: \, \delimI\\arg\\delimII\

Escape: \e\delimI\\arg\\delimII\

Escape: \E\delimI\\arg\\delimII\

Print the current escape character (which is the backslash character `\' by default).

\\ is a `delayed' backslash; more precisely, it is the default escape character followed by a backslash, which no longer has special meaning due to the leading escape character. It is not an escape sequence in the usual sense! In any unknown escape sequence \X the escape character is ignored and X is printed. But if X is equal to the current escape character, no warning is emitted.

As a consequence, only at top-level or in a diversion a backslash glyph is printed; in copy-in mode, it expands to a single backslash which then combines with the following character to an escape sequence.

The \E escape differs from \e by printing an escape character that is not interpreted in copy mode. Use this to define strings with escapes that work when used in copy mode (for example, as a macro argument). The following example defines strings to begin and end a superscript:

.ds { \v'-.3m'\s'\Es[.s]*60/100' .ds } \s0\v'.3m'

Another example to demonstrate the differences between the various escape sequences, using a strange escape character, `-'.

.ec - .de xxx --A'123' .. .xxx => -A'foo'

The result is surprising for most users, expecting `1' since `foo' is a valid identifier. What has happened? As mentioned above, the leading escape character makes the following character ordinary. Written with the default escape character the sequence `--' becomes `\-' -- this is the minus sign.

If the escape character followed by itself is a valid escape sequence, only \E yields the expected result:

.ec - .de xxx -EA'123' .. .xxx => 1

Escape: \.\delimI\\arg\\delimII\

Similar to \\, the sequence \. isn't a real escape sequence. As before, a warning message is suppressed if the escape character is followed by a dot, and the dot itself is printed.

.de foo . nop foo . . de bar . nop bar \\.. . .. .foo .bar => foo bar

The first backslash is consumed while the macro is read, and the second is swallowed while exexuting macro foo.

A translation is a mapping of an input character to an output glyph. The mapping occurs at output time, i.e., the input character gets assigned the metric information of the mapped output character right before input tokens are converted to nodes (see section 5.33 gtroff Internals, for more on this process).

Request: .tr {a

Request: .trin {a

Translate character a to glyph b, character c to glyph d, etc. If there is an odd number of arguments, the last one is translated to an unstretchable space (`\ ').

The trin request is identical to tr, but when you unformat a diversion with asciify it ignores the translation. See section 5.26 Diversions, for details about the asciify request.

Some notes:

• Special characters (\(xx, \[xxx], \C'xxx', \', \`, \-, \_), glyphs defined with the char request, and numbered glyphs (\N'xxx') can be translated also.

• The \e escape can be translated also.

• Characters can be mapped onto the \% and \~ escapes (but \% and \~ can't be mapped onto another glyph).

• The following characters can't be translated: space (with one exception, see below), backspace, newline, leader (and \a), tab (and \t).

• Translations are not considered for finding the soft hyphen character set with the shc request.

• The pair `c\&' (this is an arbitrary character@w{ c followed by the zero width space character) maps this character to nothing.

  .tr a\& foo bar => foo br

  It is even possible to map the space character to nothing:

  .tr aa \& foo bar => foobar

  As shown in the example, the space character can't be the first character/glyph pair as an argument of tr. Additionally, it is not possible to map the space character to any other glyph; requests like `.tr aa x' undo `.tr aa \&' instead.

  If justification is active, lines are justified in spite of the `empty' space character (but there is no minimal distance, i.e. the space character, between words).

• After an output glyph has been constructed (this happens at the moment immediately before the glyph is appended to an output glyph list, either by direct output, in a macro, diversion, or string), it is no longer affected by tr.

• Translating character to glyphs where one of them or both are undefined is possible also; tr does not check whether the entities in its argument do exist.

  See section 5.33 gtroff Internals.

• troff no longer has a hard-coded dependency on Latin-1; all charXXX entities have been removed from the font description files. This has a notable consequence which shows up in warnings like can't find character with input code XXX if the tr request isn't handled properly.

  Consider the following translation:

  .tr éÉ

  This maps input character é onto glyph É, which is identical to glyph char201. But this glyph intentionally doesn't exist! Instead, \[char201] is treated as an input character entity and is by default mapped onto \['E], and gtroff doesn't handle translations of translations.

  The right way to write the above translation is

  .tr é\['E]

  With other words, the first argument of tr should be an input character or entity, and the second one a glyph entity.

• Without an argument, the tr request is ignored.

Request: .trnt {a

trnt is the same as the tr request except that the translations do not apply to text that is transparently throughput into a diversion with \!. See section 5.26 Diversions, for more information.

For example,

.tr ab .di x \!.tm a .di .x

prints `b' to the standard error stream; if trnt is used instead of tr it prints `a'.

5.13 Troff and Nroff Mode

Originally, nroff and troff were two separate programs, the former for TTY output, the latter for everything else. With GNU troff, both programs are merged into one executable, sending its output to a device driver (grotty for TTY devices, grops for POSTSCRIPT, etc.) which interprets the intermediate output of gtroff. For UNIX troff it makes sense to talk about Nroff mode and Troff mode since the differences are hardcoded. For GNU troff, this distinction is not appropriate because gtroff simply takes the information given in the font files for a particular device without handling requests specially if a TTY output device is used.

Usually, a macro package can be used with all output devices. Nevertheless, it is sometimes necessary to make a distinction between TTY and non-TTY devices: gtroff provides two built-in conditions `n' and `t' for the if, ie, and while requests to decide whether gtroff shall behave like nroff or like troff.

Request: .troff \arg\

Make the `t' built-in condition true (and the `n' built-in condition false) for if, ie, and while conditional requests. This is the default if gtroff (not groff) is started with the `-R' switch to avoid loading of the start-up files `troffrc' and `troffrc-end'. Without `-R', gtroff stays in troff mode if the output device is not a TTY (e.g. `ps').

Request: .nroff \arg\

Make the `n' built-in condition true (and the `t' built-in condition false) for if, ie, and while conditional requests. This is the default if gtroff uses a TTY output device; the code for switching to nroff mode is in the file `tty.tmac' which is loaded by the start-up file troffrc.

See section 5.21 Conditionals and Loops, for more details on built-in conditions.

5.14 Line Layout

The following drawing shows the dimensions which gtroff uses for placing a line of output onto the page. They are labeled with the request which manipulates each dimension.

-->| in |<-- |<-----------ll------------>| +----+----+----------------------+----+ | : : : | +----+----+----------------------+----+ -->| po |<-- |<--------paper width---------------->|

These dimensions are:

po

Page offset -- this is the leftmost position of text on the final output, defining the left margin.

in

Indentation -- this is the distance from the left margin where text is printed.

ll

Line length -- this is the distance from the left margin to right margin.

A simple demonstration:

.ll 3i This is text without indentation. The line length has been set to 3\~inch. .in +.5i .ll -.5i Now the left and right margins are both increased. .in .ll Calling .in and .ll without parameters restore the previous values.

Result:

This is text without indenta- tion. The line length has been set to 3 inch. Now the left and right margins are both increased. Calling .in and .ll without parameters restore the previ- ous values.

Request: .po [{offset

Request: .po @t{+

Request: .po @t{-

Register: \n[.o]

Set horizontal page offset to offset (or increment or decrement the current value by offset). Note that this request does not cause a break, so changing the page offset in the middle of text being filled may not yield the expected result. The initial value is 1i. For TTY output devices, it is set to 0 in the startup file `troffrc'; the default scaling indicator is `m' (and not `v' as incorrectly documented in the original UNIX troff manual).

The current page offset can be found in the read-only number register `.o'.

If po is called without an argument, the page offset is reset to the previous value before the last call to po.

.po 3i \n[.o] => 720 .po -1i \n[.o] => 480 .po \n[.o] => 720

Request: .in [{indent

Request: .in @t{+

Request: .in @t{-

Register: \n[.i]

Set indentation to indent (or increment or decrement the current value by indent). This request causes a break. Initially, there is no indentation.

If in is called without an argument, the indentation is reset to the previous value before the last call to in. The default scaling indicator is `m'.

The indentation is associated with the current environment (see section 5.27 Environments).

If a negative indentation value is specified (which is not allowed), gtroff emits a warning of type `range' and sets the indentation to zero.

The effect of in is delayed until a partially collected line (if it exists) is output. A temporary indent value is reset to zero also.

The current indentation (as set by in) can be found in the read-only number register `.i'.

Request: .ti offset

Request: .ti @t{+

Request: .ti @t{-

Register: \n[.in]

Temporarily indent the next output line by offset. If an increment or decrement value is specified, adjust the temporary indentation relative to the value set by the in request.

This request causes a break; its value is associated with the current environment (see section 5.27 Environments). The default scaling indicator is `m'. A call of ti without an argument is ignored.

If the total indentation value is negative (which is not allowed), gtroff emits a warning of type `range' and sets the temporary indentation to zero. `Total indentation' is either offset if specified as an absolute value, or the temporary plus normal indentation, if offset is given as a relative value.

The effect of ti is delayed until a partially collected line (if it exists) is output.

The read-only number register .in is the indentation that applies to the current output line.

The difference between .i and .in is that the latter takes into account whether a partially collected line still uses the old indentation value or a temporary indentation value is active.

Request: .ll [{length

Request: .ll @t{+

Request: .ll @t{-

Register: \n[.l]

Register: \n[.ll]

Set the line length to length (or increment or decrement the current value by length). Initially, the line length is set to 6.5i. The effect of ll is delayed until a partially collected line (if it exists) is output. The default scaling indicator is `m'.

If ll is called without an argument, the line length is reset to the previous value before the last call to ll. If a negative line length is specified (which is not allowed), gtroff emits a warning of type `range' and sets the line length to zero.

The line length is associated with the current environment (see section 5.27 Environments).

The current line length (as set by ll) can be found in the read-only number register `.l'. The read-only number register .ll is the line length that applies to the current output line.

Similar to .i and .in, the difference between .l and .ll is that the latter takes into account whether a partially collected line still uses the old line length value.

5.15 Line Control

It is important to understand how gtroff handles input and output lines.

Many escapes use positioning relative to the input line. For example, this

This is a \h'|1.2i'test. This is a \h'|1.2i'test.

produces

This is a test. This is a test.

The main usage of this feature is to define macros which act exactly at the place where called.

.\" A simple macro to underline a word .de underline . nop \\$1\l'|0\[ul]' ..

In the above example, `|0' specifies a negative distance from the current position (at the end of the just emitted argument \$1) back to the beginning of the input line. Thus, the `\l' escape draws a line from right to left.

gtroff makes a difference between input and output line continuation; the latter is also called interrupting a line.

Escape: @t{\RET\delimI\\arg\\delimII\

Escape: \c\delimI\\arg\\delimII\

Register: \n[.int]

Continue a line. \RET (this is a backslash at the end of a line immediately followed by a newline) works on the input level, suppressing the effects of the following newline in the input.

This is a \ .test => This is a .test

The `|' operator is also affected.

\c works on the output level. Anything after this escape on the same line is ignored, except \R which works as usual. Anything before \c on the same line will be appended to the current partial output line. The next non-command line after an interrupted line counts as a new input line.

The visual results depend on whether no-fill mode is active.

• If no-fill mode is active (using the nf request), the next input text line after \c will be handled as a continuation of the same input text line.

  .nf This is a \c test. => This is a test.

• If fill mode is active (using the fi request), a word interrupted with \c will be continued with the text on the next input text line, without an intervening space.

  This is a te\c st. => This is a test.

Note that an intervening control line which causes a break is stronger than \c, flushing out the current partial line in the usual way.

The .int register contains a positive value if the last output line was interrupted with \c; this is associated with the current environment (see section 5.27 Environments).

5.16 Page Layout

gtroff provides some very primitive operations for controlling page layout.

Request: .pl [{length

Request: .pl @t{+

Request: .pl @t{-

Register: \n[.p]

Set the page length to length (or increment or decrement the current value by length). This is the length of the physical output page. The default scaling indicator is `v'.

The current setting can be found in the read-only number register `.p'.

Note that this only specifies the size of the page, not the top and bottom margins. Those are not set by gtroff directly. See section 5.25 Traps, for further information on how to do this.

Negative pl values are possible also, but not very useful: No trap is sprung, and each line is output on a single page (thus suppressing all vertical spacing).

If no argument or an invalid argument is given, pl sets the page length to 11i.

gtroff provides several operations which help in setting up top and bottom titles (or headers and footers).

Request: .tl @t{'

Print a title line. It consists of three parts: a left justified portion, a centered portion, and a right justified portion. The argument separator `'' can be replaced with any character not occurring in the title line. The `%' character is replaced with the current page number. This character can be changed with the pc request (see below).

Without argument, tl is ignored.

Some notes:

• A title line is not restricted to the top or bottom of a page.

• tl prints the title line immediately, ignoring a partially filled line (which stays untouched).

• It is not an error to omit closing delimiters. For example, `.tl /foo' is equivalent to `.tl /foo///': It prints a title line with the left justified word `foo'; the centered and right justfied parts are empty.

• tl accepts the same parameter delimiting characters as the \A escape; see 5.6.3 Escapes.

Request: .lt [{length

Request: .lt @t{+

Request: .lt @t{-

Register: \n[.lt]

The title line is printed using its own line length, which is specified (or incremented or decremented) with the lt request. Initially, the title line length is set to 6.5i. If a negative line length is specified (which is not allowed), gtroff emits a warning of type `range' and sets the title line length to zero. The default scaling indicator is `m'. If lt is called without an argument, the title length is reset to the previous value before the last call to lt.

The current setting of this is available in the .lt read-only number register; it is associated with the current environment (see section 5.27 Environments).

Request: .pn page

Request: .pn @t{+

Request: .pn @t{-

Register: \n[.pn]

Change (increase or decrease) the page number of the next page. The only argument is the page number; the request is ignored without a parameter.

The read-only number register .pn contains the number of the next page: either the value set by a pn request, or the number of the current page plus 1.

Register: \n[%]

A read-write register holding the current page number.

Request: .pc [{char

Change the page number character (used by the tl request) to a different character. With no argument, this mechanism is disabled. Note that this doesn't affect the number register %.

See section 5.25 Traps.

5.17 Page Control

Request: .bp [{page

Request: .bp @t{+

Request: .bp @t{-

Stop processing the current page and move to the next page. This request causes a break. It can also take an argument to set (increase, decrease) the page number of the next page. The only difference between bp and pn is that pn does not cause a break or actually eject a page.

.de newpage \" define macro 'bp \" begin page 'sp .5i \" vertical space .tl 'left top'center top'right top' \" title 'sp .3i \" vertical space .. \" end macro

bp has no effect if not called within the top-level diversion (see section 5.26 Diversions).

Request: .ne [{space

It is often necessary to force a certain amount of space before a new page occurs. This is most useful to make sure that there is not a single orphan line left at the bottom of a page. The ne request ensures that there is a certain distance, specified by the first argument, before the next page is triggered (see 5.25 Traps, for further information). The default scaling indicator for ne is `v'; the default value of space is 1v if no argument is given.

For example, to make sure that no fewer than 2 lines get orphaned, do the following before each paragraph:

.ne 2 text text text

ne will then automatically cause a page break if there is space for one line only.

Request: .sv [{space

Request: .os \arg\

sv is similar to the ne request; it reserves the specified amount of vertical space. If the desired amount of space exists before the next trap (or the bottom page boundary if no trap is set), the space is output immediately (ignoring a partially filled line which stays untouched). If there is not enough space, it is stored for later output via the os request. The default value is 1v if no argument is given; the default scaling indicator is `v'.

Both sv and os ignore no-space mode. While the sv request allows negative values for space, os will ignore them.

Register: \n[nl]

This register contains the current vertical position. If the vertical position is zero and the top of page transition hasn't happened yet, nl is set to negative value. gtroff itself does this at the very beginning of a document before anything has been printed, but the main usage is to plant a header trap on a page if this page has already started.

Consider the following:

.de xxx . sp . tl ''Header'' . sp .. . First page. .bp .wh 0 xxx .nr nl (-1) Second page.

Result:

First page. ... Header Second page. ...

Without resetting nl to a negative value, the just planted trap would be active beginning with the next page, not the current one.

See section 5.26 Diversions, for a comparison with the .h and .d registers.

5.18 Fonts

gtroff can switch fonts at any point in the text.

The basic set of fonts is `R', `I', `B', and `BI'. These are Times Roman, Italic, Bold, and Bold Italic. For non-TTY devices, there is also at least one symbol font which contains various special symbols (Greek, mathematics).

5.18.1 Changing Fonts
5.18.2 Font Families
5.18.3 Font Positions
5.18.4 Using Symbols
5.18.5 Special Fonts
5.18.6 Artificial Fonts
5.18.7 Ligatures and Kerning

5.18.1 Changing Fonts

Request: .ft [{font

Escape: \ff\delimII\

Escape: @t{\f(\arg\\delimII\

Escape: @t{\f[\arg\\delimII\

The ft request and the \f escape change the current font to font (one-character name f, two-character name fn).

If font is a style name (as set with the sty request or with the styles command in the `DESC' file), use it within the current font family (as set with the fam request, \F escape, or with the family command in the `DESC' file).

With no argument or using `P' as an argument, .ft switches to the previous font. Use \f[] to do this with the escape. The old syntax forms \fP or \f[P] are also supported.

Fonts are generally specified as upper-case strings, which are usually 1 to 4 characters representing an abbreviation or acronym of the font name. This is no limitation, just a convention.

The example below produces two identical lines.

eggs, bacon, .ft B spam .ft and sausage. eggs, bacon, \fBspam\fP and sausage.

Note that \f doesn't produce an input token in gtroff. As a consequence, it can be used in requests like mc (which expects a single character as an argument) to change the font on the fly:

.mc \f[I]x\f[]

See section 5.18.3 Font Positions, for an alternative syntax.

Request: .ftr f [{g

Translate font f to font g. Whenever a font named@w{ f is referred to in a \f escape sequence, or in the ft, ul, bd, cs, tkf, special, fspecial, fp, or sty requests, font g is used. If g is missing or equal to f the translation is undone.

5.18.2 Font Families

Due to the variety of fonts available, gtroff has added the concept of font families and font styles. The fonts are specified as the concatenation of the font family and style. Specifying a font without the family part causes gtroff to use that style of the current family.

Currently, fonts for the devices `-Tps', `-Tdvi', and `-Tlbp' are set up to this mechanism. By default, gtroff uses the Times family with the four styles `R', `I', `B', and `BI'.

This way, it is possible to use the basic four fonts and to select a different font family on the command line (see section 2.1 Options).

Request: .fam [{family

Register: \n[.fam]

Escape: \Ff\delimII\

Escape: @t{\F(\arg\\delimII\

Escape: @t{\F[\arg\\delimII\

Register: \n[.fn]

Switch font family to family (one-character name f, two-character name fm). If no argument is given, switch back to the previous font family. Use \F[] to do this with the escape. Note that \FP doesn't work; it selects font family `P' instead.

The value at start-up is `T'. The current font family is available in the read-only number register `.fam' (this is a string-valued register); it is associated with the current environment.

spam, .fam H \" helvetica family spam, \" used font is family H + style R = HR .ft B \" family H + style B = font HB spam, .fam T \" times family spam, \" used font is family T + style B = TB .ft AR \" font AR (not a style) baked beans, .ft R \" family T + style R = font TR and spam.

Note that \F doesn't produce an input token in gtroff. As a consequence, it can be used in requests like mc (which expects a single character as an argument) to change the font family on the fly:

.mc \F[P]x\F[]

The `.fn' register contains the current real font name of the current font. This is a string-valued register. If the current font is a style, the value of \n[.fn] is the proper concatenation of family and style name.

Request: .sty n style

Associate style with font position n. A font position can be associated either with a font or with a style. The current font is the index of a font position and so is also either a font or a style. If it is a style, the font that is actually used is the font which name is the concatenation of the name of the current family and the name of the current style. For example, if the current font is 1 and font position 1 is associated with style `R' and the current font family is `T', then font `TR' will be used. If the current font is not a style, then the current family is ignored. If the requests cs, bd, tkf, uf, or fspecial are applied to a style, they will instead be applied to the member of the current family corresponding to that style.

n must be a non-negative integer value.

The default family can be set with the `-f' option (see section 2.1 Options). The styles command in the `DESC' file controls which font positions (if any) are initially associated with styles rather than fonts. For example, the default setting for POSTSCRIPT fonts

styles R I B BI

is equivalent to

.sty 1 R .sty 2 I .sty 3 B .sty 4 BI

fam and \F always check whether the current font position is valid; this can give surprising results if the current font position is associated with a style.

In the following example, we want to access the POSTSCRIPT font FooBar from the font family Foo:

.sty \n[.fp] Bar .fam Foo => warning: can't find font `FooR'

The default font position at start-up is 1; for the POSTSCRIPT device, this is associated with style `R', so gtroff tries to open FooR.

A solution to this problem is to use a dummy font like the following:

.fp 0 dummy TR \" set up dummy font at position 0 .sty \n[.fp] Bar \" register style `Bar' .ft 0 \" switch to font at position 0 .fam Foo \" activate family `Foo' .ft Bar \" switch to font `FooBar'

See section 5.18.3 Font Positions.

5.18.3 Font Positions

For the sake of old phototypesetters and compatibility with old versions of troff, gtroff has the concept of font positions, on which various fonts are mounted.

Request: .fp pos font [{external-name

Register: \n[.f]

Register: \n[.fp]

Mount font font at position pos (which must be a non-negative integer). This numeric position can then be referred to with font changing commands. When gtroff starts it is using font position 1 (which must exist; position 0 is unused usually at start-up).

The current font in use, as a font position, is available in the read-only number register `.f'. This can be useful to remember the current font for later recall. It is associated with the current environment (see section 5.27 Environments).

.nr save-font \n[.f] .ft B ... text text text ... .ft \n[save-font]

The number of the next free font position is available in the read-only number register `.fp'. This is useful when mounting a new font, like so:

.fp \n[.fp] NEATOFONT

Fonts not listed in the `DESC' file are automatically mounted on the next available font position when they are referenced. If a font is to be mounted explicitly with the fp request on an unused font position, it should be mounted on the first unused font position, which can be found in the .fp register. Although gtroff does not enforce this strictly, it is not allowed to mount a font at a position whose number is much greater (approx. 1000 positions) than that of any currently used position.

The fp request has an optional third argument. This argument gives the external name of the font, which is used for finding the font description file. The second argument gives the internal name of the font which is used to refer to the font in gtroff after it has been mounted. If there is no third argument then the internal name is used as the external name. This feature makes it possible to use fonts with long names in compatibility mode.

Both the ft request and the \f escape have alternative syntax forms to access font positions.

Request: .ft nnn

Escape: \fn\delimII\

Escape: @t{\f(\arg\\delimII\

Escape: @t{\f[\arg\\delimII\

Change the current font position to nnn (one-digit position@w{ n, two-digit position nn), which must be a non-negative integer.

If nnn is associated with a style (as set with the sty request or with the styles command in the `DESC' file), use it within the current font family (as set with the fam request, the \F escape, or with the family command in the `DESC' file).

this is font 1 .ft 2 this is font 2 .ft \" switch back to font 1 .ft 3 this is font 3 .ft this is font 1 again

See section 5.18.1 Changing Fonts, for the standard syntax form.

5.18.4 Using Symbols

A glyph is a graphical representation of a character. While a character is an abstract entity containing semantic information, a glyph is something which can be actually seen on screen or paper. It is possible that a character has multiple glyph representation forms (for example, the character `A' can be either written in a roman or an italic font, yielding two different glyphs); sometimes more than one character maps to a single glyph (this is a ligature -- the most common is `fi').

A symbol is simply a named glyph. Within gtroff, all glyph names of a particular font are defined in its font file. If the user requests a glyph not available in this font, gtroff looks up an ordered list of special fonts. By default, the POSTSCRIPT output device supports the two special fonts `SS' (slanted symbols) and `S' (symbols) (the former is looked up before the latter). Other output devices use different names for special fonts. Fonts mounted with the fonts keyword in the `DESC' file are globally available. To install additional special fonts locally (i.e. for a particular font), use the fspecial request.

In summary, gtroff tries the following to find a given symbol:

• If the symbol has been defined with the char request, use it. This hides a symbol with the same name in the current font.

• Check the current font.

• If the symbol has been defined with the fchar request, use it.

• Check all fonts given with the fspecial request, in the order of appearance in fspecial calls.

• Check all fonts given with the special request, in the order of appearance in special calls (inclusively the special fonts defined in the `DESC' file, which come first).

• As a last resort, consult all fonts loaded up to now (in the order they have been called the first time) for special fonts and check them.

See section 8.2 Font Files, and 5.18.5 Special Fonts, for more details.

Escape: @t{, ({\delimI\\arg\\delimII\

Escape: @t{, [{\delimI\\arg\\delimII\

Insert a symbol name (two-character name nm). There is no special syntax for one-character names -- the natural form `\n' would collide with escapes.@footnotegroff_114.html{Note that a one-character symbol is not the same as an input character, i.e., the character a is not the same as \[a]. By default, groff defines only a single one-character symbol, \[-]; it is usually accessed as \-. On the other hand, gtroff has the special feature that \[charXXX] is the same as the input character with character code XXX. For example, \[char97] is identical to the letter a if ASCII encoding is active.

If name is undefined, a warning of type `char' is generated, and the escape is ignored. See section 5.34 Debugging, for information about warnings.

The list of available symbols is device dependent; see the groff_char(7) man page for a complete list for the given output device. For example, say

man -Tdvi groff_char > groff_char.dvi

for a list using the default DVI fonts (not all versions of the man program support the `-T' option). If you want to use an additional macro package to change the used fonts, groff must be called directly:

groff -Tdvi -mec -man groff_char.7 > groff_char.dvi

Escape: \C'xxx'

Typeset the glyph named xxx.@footnotegroff_114.html{\C is actually a misnomer since it accesses an output glyph. Normally it is more convenient to use \[xxx], but \C has the advantage that it is compatible with newer versions of AT&T troff and is available in compatibility mode.

Escape: \N'n'

Typeset the glyph with code n in the current font (n is not the input character code). The number n can be any non-negative decimal integer. Most devices only have glyphs with codes between 0 and 255; the Unicode output device uses codes in the range 0--65535. If the current font does not contain a glyph with that code, special fonts are not searched. The \N escape sequence can be conveniently used in conjunction with the char request:

.char \[phone] \f[ZD]\N'37'

The code of each glyph is given in the fourth column in the font description file after the charset command. It is possible to include unnamed glyphs in the font description file by using a name of `---'; the \N escape sequence is the only way to use these.

Some escape sequences directly map onto special glyphs.

Escape: \'\delimI\\arg\\delimII\

This is a backslash followed by the apostrophe character, ASCII character 0x27 (EBCDIC character 0x7D). The same as \[aa], the acute accent.

Escape: \`\delimI\\arg\\delimII\

This is a backslash followed by ASCII character 0x60 (EBCDIC character 0x79 usually). The same as \[ga], the grave accent.

Escape: \-\delimI\\arg\\delimII\

This is the same as \[-], the minus sign in the current font.

Request: .cflags n c1 c2 @dots{

Input characters and symbols have certain properties associated with it.@footnotegroff_114.html{Note that the output glyphs themselves don't have such properties. For gtroff, a glyph is a numbered box with a given width, depth, and height, nothing else. All manipulations with the cflags request work on the input level. These properties can be modified with the cflags request. The first argument is the sum of the desired flags and the remaining arguments are the characters or symbols to have those properties. It is possible to omit the spaces between the characters or symbols.

1

The character ends sentences (initially characters `.?!' have this property).

2

Lines can be broken before the character (initially no characters have this property).

4

Lines can be broken after the character (initially the character `-' and the symbols `\(hy' and `\(em' have this property).

8

The character overlaps horizontally (initially the symbols `\(ul\(rn\(ru' have this property).

16

The character overlaps vertically (initially symbol `\(br' has this property).

32

An end-of-sentence character followed by any number of characters with this property is treated as the end of a sentence if followed by a newline or two spaces; in other words the character is transparent for the purposes of end-of-sentence recognition -- this is the same as having a zero space factor in TeX (initially characters `"')]*' and the symbols `\(dg\(rq' have this property).

Request: .char g [{string

Request: .fchar g [{string

Define a new glyph g to be string (which can be empty).@footnotegroff_114.html{char is a misnomer since an output glyph is defined. Every time glyph g needs to be printed, string is processed in a temporary environment and the result is wrapped up into a single object. Compatibility mode is turned off and the escape character is set to `\' while string is being processed. Any emboldening, constant spacing or track kerning is applied to this object rather than to individual characters in string.

A glyph defined by this request can be used just like a normal glyph provided by the output device. In particular, other characters can be translated to it with the tr or trin requests; it can be made the leader character by the lc request; repeated patterns can be drawn with the glyph using the \l and \L escape sequences; words containing the glyph can be hyphenated correctly if the hcode request is used to give the glyph's symbol a hyphenation code.

There is a special anti-recursion feature: Use of g within the glyph's definition is handled like normal characters and symbols not defined with char.

Note that the tr and trin requests take precedence if char accesses the same symbol.

.tr XY X => Y .char X Z X => Y .tr XX X => Z

The fchar request defines a fallback glyph: gtroff only checks for glyphs defined with fchar if it cannot find the glyph in the current font. gtroff carries out this test before checking special fonts.

Request: .rchar c1 c2 @dots{

Remove the definitions of glyphs c1, c2,@w{ .... This undoes the effect of a char or fchar request.

It is possible to omit the whitespace between arguments.

See section 7.1 Special Characters.

5.18.5 Special Fonts

Special fonts are those that gtroff searches when it cannot find the requested glyph in the current font. The Symbol font is usually a special font.

gtroff provides the following two requests to add more special fonts. See section 5.18.4 Using Symbols, for a detailed description of the glyph searching mechanism in gtroff.

Usually, only non-TTY devices have special fonts.

Request: .special s1 s2 @dots{

Request: .fspecial f s1 s2 @dots{

Use the special request to define special fonts. They are appended to the list of global special fonts in the given order. The first entries in this list are the fonts defined with the fonts command in the `DESC' file which are marked as special in the corresponding font description files.

Use the fspecial request to designate special fonts only when font f font is active. They are appended to the list of special fonts for f in the given order. Initially, this list is empty.

5.18.6 Artificial Fonts

There are a number of requests and escapes for artificially creating fonts. These are largely vestiges of the days when output devices did not have a wide variety of fonts, and when nroff and troff were separate programs. Most of them are no longer necessary in GNU troff. Nevertheless, they are supported.

Escape: \H'height'

Escape: \H'+\delimII\

Escape: \H'-\delimII\

Change (increment, decrement) the height of the current font, but not the width. If height is zero, restore the original height. Default scaling indicator is `z'.

Currently, only the `-Tps' device supports this feature.

Note that \H doesn't produce an input token in gtroff. As a consequence, it can be used in requests like mc (which expects a single character as an argument) to change the font on the fly:

.mc \H'+5z'x\H'0'

In compatibility mode, gtroff behaves differently: If an increment or decrement is used, it is always taken relative to the current point size and not relative to the previously selected font height. Thus,

.cp 1 \H'+5'test \H'+5'test

prints the word `test' twice with the same font height (five points larger than the current font size).

Escape: \S'slant'

Slant the current font by slant degrees. Positive values slant to the right.

Currently, only the `-Tps' device supports this feature.

Note that \S doesn't produce an input token in gtroff. As a consequence, it can be used in requests like mc (which expects a single character as an argument) to change the font on the fly:

.mc \S'20'x\S'0'

This request is incorrectly documented in the original UNIX troff manual; the slant is always set to an absolute value.

Request: .ul [{lines

The ul request normally underlines subsequent lines if a TTY output device is used. Otherwise, the lines are printed in italics (only the term `underlined' is used in the following). The single argument is the number of input lines to be underlined; with no argument, the next line is underlined. If lines is zero or negative, stop the effects of ul (if it was active). Requests and empty lines do not count for computing the number of underlined input lines, even if they produce some output like tl. Lines inserted by macros (e.g. invoked by a trap) do count.

At the beginning of ul, the current font is stored and the underline font is activated. Within the span of a ul request, it is possible to change fonts, but after the last line affected by ul the saved font is restored.

This number of lines still to be underlined is associated with the current environment (see section 5.27 Environments). The underline font can be changed with the uf request.

The ul request does not underline spaces.

Request: .cu [{lines

The cu request is similar to ul but underlines spaces as well (if a TTY output device is used).

Request: .uf font

Set the underline font (globally) used by ul and cu. By default, this is the font at position 2. font can be either a non-negative font position or the name of a font.

Request: .bd font [{offset

Request: .bd font1 font2 [{offset

Register: \n[.b]

Artificially create a bold font by printing each glyph twice, slightly offset.

Two syntax forms are available.

• Imitate a bold font unconditionally. The first argument specifies the font to embolden, and the second is the number of basic units, minus one, by which the two glyphs are offset. If the second argument is missing, emboldening is turned off.

  font can be either a non-negative font position or the name of a font.

  offset is available in the .b read-only register if a special font is active; in the bd request, its default unit is `u'.

• Imitate a bold form conditionally. Embolden font1 by offset only if font font2 is the current font. This command can be issued repeatedly to set up different emboldening values for different current fonts. If the second argument is missing, emboldening is turned off for this particular current font.

  This affects special fonts only (either set up with the special command in font files or with the fspecial request).

Request: .cs font [{width

Switch to and from constant glyph space mode. If activated, the width of every glyph is width/36 ems. The em size is given absolutely by em-size; if this argument is missing, the em value is taken from the current font size (as set with the ps request) when the font is effectively in use. Without second and third argument, constant glyph space mode is deactivated.

Default scaling indicator for em-size is `z'; width is an integer.

5.18.7 Ligatures and Kerning

Ligatures are groups of characters that are run together, i.e, producing a single glyph. For example, the letters `f' and `i' can form a ligature `fi' as in the word `file'. This produces a cleaner look (albeit subtle) to the printed output. Usually, ligatures are not available in fonts for TTY output devices.

Most POSTSCRIPT fonts support the fi and fl ligatures. The C/A/T typesetter that was the target of AT&T troff also supported `ff', `ffi', and `ffl' ligatures. Advanced typesetters or `expert' fonts may include ligatures for `ft' and `ct', although GNU troff does not support these (yet).

Request: .lg [{flag

Register: \n[.lg]

Switch the ligature mechanism on or off; if the parameter is non-zero or missing, ligatures are enabled, otherwise disabled. Default is on. The current ligature mode can be found in the read-only number register .lg (set to 1 or 2 if ligatures are enabled, 0 otherwise).

Setting the ligature mode to 2 enables the two-character ligatures (fi, fl, and ff) and disables the three-character ligatures (ffi and ffl).

Pairwise kerning is another subtle typesetting mechanism that modifies the distance between a glyph pair to improve readability. In most cases (but not always) the distance is decreased. For example, compare the combination of the letters `V' and `A'. With kerning, `VA' is printed. Without kerning it appears as `VA'. Typewriter-like fonts and fonts for terminals where all glyphs have the same width don't use kerning.

Request: .kern [{flag

Register: \n[.kern]

Switch kerning on or off. If the parameter is non-zero or missing, enable pairwise kerning, otherwise disable it. The read-only number register .kern is set to 1 if pairwise kerning is enabled, 0 otherwise.

If the font description file contains pairwise kerning information, glyphs from that font are kerned. Kerning between two glyphs can be inhibited by placing \& between them: `V\&A'.

See section 8.2.2 Font File Format.

Track kerning expands or reduces the space between glyphs. This can be handy, for example, if you need to squeeze a long word onto a single line or spread some text to fill a narrow column. It must be used with great care since it is usually considered bad typography if the reader notices the effect.

Request: .tkf f s1 n1 s2 n2

Enable track kerning for font f. If the current font is f the width of every glyph is increased by an amount between n1 and n2 (n1, n2 can be negative); if the current point size is less than or equal to s1 the width is increased by n1; if it is greater than or equal to s2 the width is increased by n2; if the point size is greater than or equal to s1 and less than or equal to s2 the increase in width is a linear function of the point size.

The default scaling indicator is `z' for s1 and s2, `p' for n1 and n2.

Note that the track kerning amount is added even to the rightmost glyph in a line; for large values it is thus recommended to increase the line length by the same amount to compensate it.

Sometimes, when typesetting letters of different fonts, more or less space at such boundaries are needed. There are two escapes to help with this.

Escape: \/\delimI\\arg\\delimII\

Increase the width of the preceding glyph so that the spacing between that glyph and the following glyph is correct if the following glyph is a roman glyph. For example, if an italic f is immediately followed by a roman right parenthesis, then in many fonts the top right portion of the f overlaps the top left of the right parenthesis. Use this escape sequence whenever an italic glyph is immediately followed by a roman glyph without any intervening space. This small amount of space is also called italic correction.

Escape: \,\delimI\\arg\\delimII\

Modify the spacing of the following glyph so that the spacing between that glyph and the preceding glyph is correct if the preceding glyph is a roman glyph. Use this escape sequence whenever a roman glyph is immediately followed by an italic glyph without any intervening space. In analogy to above, this space could be called left italic correction, but this term isn't used widely.

Escape: \&\delimI\\arg\\delimII\

Insert a zero-width character, which is invisible. Its intended use is to stop interaction of a character with its surrounding.

• It prevents the insertion of extra space after an end-of-sentence character.

  Test. Test. => Test. Test. Test.\& Test. => Test. Test.

• It prevents interpretation of a control character at the beginning of an input line.

  .Test => warning: `Test' not defined \&.Test => .Test

• It prevents kerning between two glyphs.

  VA => VA V\&A => VA

• It is needed to map an arbitrary character to nothing in the tr request (see section 5.12 Character Translations).

Escape: \)\delimI\\arg\\delimII\

This escape is similar to \& except that it behaves like a character declared with the cflags request to be transparent for the purposes of an end-of-sentence character.

Its main usage is in macro definitions to protect against arguments starting with a control character.

.de xxx \)\\$1 .. .de yyy \&\\$1 .. This is a test.\c .xxx ' This is a test. =>This is a test.' This is a test. This is a test.\c .yyy ' This is a test. =>This is a test.' This is a test.

5.19 Sizes

gtroff uses two dimensions with each line of text, type size and vertical spacing. The type size is approximately the height of the tallest glyph.@footnotegroff_118.html{This is usually the parenthesis. Note that in most cases the real dimensions of the glyphs in a font are not related to its type size! For example, the standard POSTSCRIPT font families `Times Roman', `Helvetica', and `Courier' can't be used together at 10pt; to get acceptable output, the size of `Helvetica' has to be reduced by one point, and the size of `Courier' must be increased by one point. Vertical spacing is the amount of space gtroff allows for a line of text; normally, this is about 20% larger than the current type size. Ratios smaller than this can result in hard-to-read text; larger than this, it spreads the text out more vertically (useful for term papers). By default, gtroff uses 10 point type on 12 point spacing.

The difference between type size and vertical spacing is known, by typesetters, as leading (this is pronounced `ledding').

5.19.1 Changing Type Sizes
5.19.2 Fractional Type Sizes

Request: .ps [{size

Request: .ps +

Request: .ps -

Escape: \ssize\delimII\

Register: \n[.s]

Use the ps request or the \s escape to change (increase, decrease) the type size (in points). Specify size as either an absolute point size, or as a relative change from the current size. The size 0, or no argument, goes back to the previous size.

Default scaling indicator of size is `z'. If size is zero or negative, it is set to 1u.

The read-only number register .s returns the point size in points as a decimal fraction. This is a string. To get the point size in scaled points, use the .ps register instead.

.s is associated with the current environment (see section 5.27 Environments).

snap, snap, .ps +2 grin, grin, .ps +2 wink, wink, \s+2nudge, nudge,\s+8 say no more! .ps 10

The \s escape may be called in a variety of ways. Much like other escapes there must be a way to determine where the argument ends and the text begins. Any of the following forms are valid:

\sn

Set the point size to n points. n must be either 0 or in the range 4 to 39.

\s+n

\s-n

Increase or decrease the point size by n points. n must be exactly one digit.

\s(nn

Set the point size to nn points. nn must be exactly two digits.

\s+(nn

\s-(nn

\s(+nn

\s(-nn

Increase or decrease the point size by nn points. nn must be exactly two digits.

Note that \s doesn't produce an input token in gtroff. As a consequence, it can be used in requests like mc (which expects a single character as an argument) to change the font on the fly:

.mc \s[20]x\s[0]

See section 5.19.2 Fractional Type Sizes, for yet another syntactical form of using the \s escape.

Request: .ds name [{string

Request: .ds1 name [{string

Escape: \*n\delimII\

Escape: @t{\*(\arg\\delimII\

Escape: @t{\*[\arg\\delimII\

Define and access a string variable name (one-character name@w{ n, two-character name nm). If name already exists, ds overwrites the previous definition. Only the syntax form using brackets can take arguments which are handled identically to macro arguments; the single exception is that a closing bracket as an argument must be enclosed in double quotes. See section 5.6.1.1 Request Arguments, and 5.22.2 Parameters.

Example:

.ds foo a \\$1 test . This is \*[foo nice]. => This is a nice test.

The \* escape interpolates (expands in-place) a previously-defined string variable. To be more precise, the stored string is pushed onto the input stack which is then parsed by gtroff. Similar to number registers, it is possible to nest strings, i.e. string variables can be called within string variables.

If the string named by the \* escape does not exist, it is defined as empty, and a warning of type `mac' is emitted (see 5.34 Debugging, for more details).

Caution: Unlike other requests, the second argument to the ds request takes up the entire line including trailing spaces. This means that comments on a line with such a request can introduce unwanted space into a string.

.ds UX \s-1UNIX\s0\u\s-3tm\s0\d \" UNIX trademark

Instead the comment should be put on another line or have the comment escape adjacent with the end of the string.

.ds UX \s-1UNIX\s0\u\s-3tm\s0\d\" UNIX trademark

To produce leading space the string can be started with a double quote. No trailing quote is needed; in fact, any trailing quote is included in your string.

.ds sign " Yours in a white wine sauce,

Strings are not limited to a single line of text. A string can span several lines by escaping the newlines with a backslash. The resulting string is stored without the newlines.

.ds foo lots and lots \ of text are on these \ next several lines

It is not possible to have real newlines in a string. To put a single double quote character into a string, use two consecutive double quote characters.

The ds1 request turns off compatibility mode while interpreting a string. To be more precise, a @dfn{compatibility save input token is inserted at the beginning of the string, and a compatibility restore input token at the end.

.nr xxx 12345 .ds aa The value of xxx is \\n[xxx]. .ds1 bb The value of xxx ix \\n[xxx]. . .cp 1 . \*(aa => warning: number register `[' not defined => The value of xxx is 0xxx]. \*(bb => The value of xxx ix 12345.

Strings, macros, and diversions (and boxes) share the same name space. Internally, even the same mechanism is used to store them. This has some interesting consequences. For example, it is possible to call a macro with string syntax and vice versa.

.de xxx a funny test. .. This is \*[xxx] => This is a funny test. .ds yyy a funny test This is .yyy => This is a funny test.

Diversions and boxes can be also called with string syntax.

Another consequence is that you can copy one-line diversions or boxes to a string.

.di xxx a \fItest\fR .br .di .ds yyy This is \*[xxx]\c \*[yyy]. => This is a test.

As the previous example shows, it is possible to store formatted output in strings. The \c escape prevents the insertion of an additional blank line in the output.

Copying diversions longer than a single output line produces unexpected results.

.di xxx a funny .br test .br .di .ds yyy This is \*[xxx]\c \*[yyy]. => test This is a funny.

Usually, it is not predictable whether a diversion contains one or more output lines, so this mechanism should be avoided. With UNIX troff, this was the only solution to strip off a final newline from a diversion. Another disadvantage is that the spaces in the copied string are already formatted, making them unstretchable. This can cause ugly results.

A clean solution to this problem is available in GNU troff, using the requests chop to remove the final newline of a diversion, and unformat to make the horizontal spaces stretchable again.

.box xxx a funny .br test .br .box .chop xxx .unformat xxx This is \*[xxx]. => This is a funny test.

See section 5.33 gtroff Internals, for more information.

Request: .as name [{string

Request: .as1 name [{string

The as request is similar to ds but appends string to the string stored as name instead of redefining it. If name doesn't exist yet, it is created.

.as sign " with shallots, onions and garlic,

The as1 request is similar to as, but compatibility mode is switched off while the appended string is interpreted. To be more precise, a compatibility save input token is inserted at the beginning of the appended string, and a compatibility restore input token at the end.

Rudimentary string manipulation routines are given with the next two requests.

Request: .substring str n1 [{n2

Replace the string named str with the substring defined by the indices n1 and n2. The first character in the string has index 0. If n2 is omitted, it is taken to be equal to the string's length. If the index value n1 or n2 is negative, it is counted from the end of the string, going backwards: The last character has index -1, the character before the last character has index -2, etc.

.ds xxx abcdefgh .substring xxx 1 -4 \*[xxx] => bcde

Request: .length reg str

Compute the number of characters of str and return it in the number register reg. If reg doesn't exist, it is created. str is read in copy mode.

.ds xxx abcd\h'3i'efgh .length yyy \n[xxx] \n[yyy] => 14

Request: .rn xx yy

Rename the request, macro, diversion, or string xx to yy.

Request: .rm xx

Remove the request, macro, diversion, or string xx. gtroff treats subsequent invocations as if the object had never been defined.

Request: .als new old

Create an alias named new for the request, string, macro, or diversion object named old. The new name and the old name are exactly equivalent (it is similar to a hard rather than a soft link). If old is undefined, gtroff generates a warning of type `mac' and ignores the request.

Request: .chop xx

Remove (chop) the last character from the macro, string, or diversion named xx. This is useful for removing the newline from the end of diversions that are to be interpolated as strings. This command can be used repeatedly; see 5.33 gtroff Internals, for details on nodes inserted additionally by gtroff.

See section 5.5 Identifiers, and 5.6.3.1 Comments.

5.21 Conditionals and Loops

5.21.1 Operators in Conditionals
5.21.2 if-else
5.21.3 while

5.21.1 Operators in Conditionals

In if and while requests, there are several more operators available:

e

o

True if the current page is even or odd numbered (respectively).

n

True if the document is being processed in nroff mode (i.e., the .nroff command has been issued).

t

True if the document is being processed in troff mode (i.e., the .troff command has been issued).

v

Always false. This condition is for compatibility with other troff versions only.

'xxx'yyy'

True if the string xxx is equal to the string yyy. Other characters can be used in place of the single quotes; the same set of delimiters as for the \D escape is used (see section 5.6.3 Escapes). gtroff formats the strings before being compared:

.ie "|"\fR|\fP" \ true .el \ false => true

The resulting motions, glyph sizes, and fonts have to match,@footnotegroff_123.html{The created output nodes must be identical. See section 5.33 gtroff Internals. and not the individual motion, size, and font requests. In the previous example, `|' and `\fR|\fP' both result in a roman `|' glyph with the same point size and at the same location on the page, so the strings are equal. If `.ft I' had been added before the `.ie', the result would be "false" because (the first) `|' produces an italic `|' rather than a roman one.

r xxx

True if there is a number register named xxx.

d xxx

True if there is a string, macro, diversion, or request named xxx.

m xxx

True if there is a color named xxx.

c g

True if there is a glyph g available@footnotegroff_123.html{The name of this conditional operator is a misnomer since it tests names of output glyphs.

(7)

Margin character is a misnomer since it is an output glyph. The first argument is the glyph to be printed. The second argument is the distance away from the right margin. If missing, the previously set value is used; default is 10pt). For text lines that are too long (that is, longer than the text length plus dist), the margin character is directly appended to the lines.

With no arguments the margin character is turned off. If this occurs before a break, no margin character is printed.

For empty lines and lines produced by the tl request no margin character is emitted.

The margin character is associated with the current environment (see section 5.27 Environments).

This is quite useful for indicating text that has changed, and, in fact, there are programs available for doing this (they are called nrchbar and changebar and can be found in any `comp.sources.unix' archive.

.ll 3i .mc | This paragraph is highlighted with a margin character. .sp Note that vertical space isn't marked. .br \& .br But we can fake it with `\&'.

Result:

This paragraph is highlighted | with a margin character. | Note that vertical space isn't | marked. | | But we can fake it with `\&'. |

Request: .psbb filename

Register: \n[llx]

Register: \n[lly]

Register: \n[urx]

Register: \n[ury]

Retrieve the bounding box of the PostScript image found in filename. The file must conform to Adobe's Document Structuring Conventions (DSC); the command searches for a %%BoundingBox comment and extracts the bounding box values into the number registers llx, lly, urx, and ury. If an error occurs (for example, psbb cannot find the %%BoundingBox comment), it sets the four number registers to zero.

5.33 gtroff Internals

gtroff processes input in three steps. One or more input characters are converted to an input token.@footnotegroff_144.html{Except the escapes \f, \F, \H, \m, \M, \R, \s, and \S which are processed immediately if not in copy-in mode. Then, one or more input tokens are converted to an output node. Finally, output nodes are converted to the intermediate output language understood by all output devices.

Actually, before step one happens, gtroff converts certain escape sequences into reserved input characters (not accessible by the user); such reserved characters are used for other internal processing also -- this is the very reason why not all characters are valid input. See section 5.5 Identifiers, for more on this topic.

For example, the input string `fi\[:u]' is converted into a character token `f', a character token `i', and a special token `:u' (representing u umlaut). Later on, the character tokens `f' and `i' are merged to a single output node representing the ligature glyph `fi' (provided the current font has a glyph for this ligature); the same happens with `:u'. All output glyph nodes are `processed' which means that they are invariably associated with a given font, font size, advance width, etc. During the formatting process, gtroff itself adds various nodes to control the data flow.

Macros, diversions, and strings collect elements in two chained lists: a list of input tokens which have been passed unprocessed, and a list of output nodes. Consider the following the diversion.

.di xxx a \!b c .br .di

It contains these elements.

node list	token list	element number
line start node	--	1
glyph node a	--	2
word space node	--	3
--	b	4
--	\n	5
glyph node c	--	6
vertical size node	--	7
vertical size node	--	8
--	\n	9

Elements 1, 7, and 8 are inserted by gtroff; the latter two (which are always present) specify the vertical extent of the last line, possibly modified by \x. The br request finishes the current partial line, inserting a newline input token which is subsequently converted to a space when the diversion is reread. Note that the word space node has a fixed width which isn't stretchable anymore. To convert horizontal space nodes back to input tokens, use the unformat request.

Macros only contain elements in the token list (and the node list is empty); diversions and strings can contain elements in both lists.

Note that the chop request simply reduces the number of elements in a macro, string, or diversion by one. Exceptions are compatibility save and compatibility ignore input tokens which are ignored. The substring request also ignores those input tokens.

Some requests like tr or cflags work on glyph identifiers only; this means that the associated glyph can be changed without destroying this association. This can be very helpful for substituting glyphs. In the following example, we assume that glyph `foo' isn't available by default, so we provide a substitution using the fchar request and map it to input character `x'.

.fchar \[foo] foo .tr x \[foo]

Now let us assume that we install an additional special font `bar' which has glyph `foo'.

.special bar .rchar \[foo]

Since glyphs defined with fchar are searched before glyphs in special fonts, we must call rchar to remove the definition of the fallback glyph. Anyway, the translation is still active; `x' now maps to the real glyph `foo'.

5.34 Debugging

gtroff is not easy to debug, but there are some useful features and strategies for debugging.

Request: .lf line filename

Change the line number and the file name gtroff shall use for error and warning messages. line is the input line number of the next line.

Without argument, the request is ignored.

This is a debugging aid for documents which are split into many files, then put together with soelim and other preprocessors. Usually, it isn't invoked manually.

Request: .tm string

Request: .tm1 string

Request: .tmc string

Send string to the standard error output; this is very useful for printing debugging messages among other things.

string is read in copy mode.

The tm request ignores leading spaces of string; tm1 handles its argument similar to the ds request: a leading double quote in string is stripped to allow initial blanks.

The tmc request is similar to tm1 but does not append a newline (as is done in tm and tm1).

Request: .ab [{string

Similar to the tm request, except that it causes gtroff to stop processing. With no argument it prints `User Abort.' to standard error.

Request: .ex \arg\

The ex request also causes gtroff to stop processing; see also 5.30 I/O.

When doing something involved it is useful to leave the debugging statements in the code and have them turned on by a command line flag.

.if \n(DB .tm debugging output

To activate these statements say

groff -rDB=1 file

If it is known in advance that there will be many errors and no useful output, gtroff can be forced to suppress formatted output with the `-z' flag.

Request: .pm \arg\

Print the entire symbol table on stderr. Names of all defined macros, strings, and diversions are print together with their size in bytes. Since gtroff sometimes adds nodes by itself, the returned size can be larger than expected.

This request differs from UNIX troff: gtroff reports the sizes of diversions, ignores an additional argument to print only the total of the sizes, and the size isn't returned in blocks of 128 characters.