Commit | Line | Data |
9741dab0 |
1 | # Pod::Man -- Convert POD data to formatted *roff input. |
b84d8b9e |
2 | # $Id: Man.pm,v 1.30 2001/11/28 01:14:28 eagle Exp $ |
9741dab0 |
3 | # |
77003bb1 |
4 | # Copyright 1999, 2000, 2001 by Russ Allbery <rra@stanford.edu> |
9741dab0 |
5 | # |
3c014959 |
6 | # This program is free software; you may redistribute it and/or modify it |
9741dab0 |
7 | # under the same terms as Perl itself. |
8 | # |
b84d8b9e |
9 | # This module translates POD documentation into *roff markup using the man |
10 | # macro set, and is intended for converting POD documents written as Unix |
11 | # manual pages to manual pages that can be read by the man(1) command. It is |
12 | # a replacement for the pod2man command distributed with versions of Perl |
13 | # prior to 5.6. |
c9abbd5d |
14 | # |
15 | # Perl core hackers, please note that this module is also separately |
16 | # maintained outside of the Perl core as part of the podlators. Please send |
17 | # me any patches at the address above in addition to sending them to the |
18 | # standard Perl mailing lists. |
9741dab0 |
19 | |
3c014959 |
20 | ############################################################################## |
9741dab0 |
21 | # Modules and declarations |
3c014959 |
22 | ############################################################################## |
9741dab0 |
23 | |
24 | package Pod::Man; |
25 | |
b84d8b9e |
26 | require 5.005; |
9741dab0 |
27 | |
28 | use Carp qw(carp croak); |
bf202ccd |
29 | use Pod::ParseLink qw(parselink); |
9741dab0 |
30 | use Pod::Parser (); |
31 | |
32 | use strict; |
33 | use subs qw(makespace); |
34 | use vars qw(@ISA %ESCAPES $PREAMBLE $VERSION); |
35 | |
36 | @ISA = qw(Pod::Parser); |
37 | |
3c014959 |
38 | # Don't use the CVS revision as the version, since this module is also in Perl |
39 | # core and too many things could munge CVS magic revision strings. This |
40 | # number should ideally be the same as the CVS revision in podlators, however. |
b84d8b9e |
41 | $VERSION = 1.30; |
9741dab0 |
42 | |
43 | |
3c014959 |
44 | ############################################################################## |
9741dab0 |
45 | # Preamble and *roff output tables |
3c014959 |
46 | ############################################################################## |
9741dab0 |
47 | |
48 | # The following is the static preamble which starts all *roff output we |
49 | # generate. It's completely static except for the font to use as a |
ab1f1d91 |
50 | # fixed-width font, which is designed by @CFONT@, and the left and right |
3c014959 |
51 | # quotes to use for C<> text, designated by @LQOUTE@ and @RQUOTE@. $PREAMBLE |
52 | # should therefore be run through s/\@CFONT\@/<font>/g before output. |
9741dab0 |
53 | $PREAMBLE = <<'----END OF PREAMBLE----'; |
54 | .de Sh \" Subsection heading |
55 | .br |
56 | .if t .Sp |
57 | .ne 5 |
58 | .PP |
59 | \fB\\$1\fR |
60 | .PP |
61 | .. |
62 | .de Sp \" Vertical space (when we can't use .PP) |
63 | .if t .sp .5v |
64 | .if n .sp |
65 | .. |
9741dab0 |
66 | .de Vb \" Begin verbatim text |
67 | .ft @CFONT@ |
68 | .nf |
69 | .ne \\$1 |
70 | .. |
71 | .de Ve \" End verbatim text |
72 | .ft R |
73 | |
74 | .fi |
75 | .. |
76 | .\" Set up some character translations and predefined strings. \*(-- will |
77 | .\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left |
78 | .\" double quote, and \*(R" will give a right double quote. | will give a |
3c014959 |
79 | .\" real vertical bar. \*(C+ will give a nicer C++. Capital omega is used to |
80 | .\" do unbreakable dashes and therefore won't be available. \*(C` and \*(C' |
81 | .\" expand to `' in nroff, nothing in troff, for use with C<>. |
9741dab0 |
82 | .tr \(*W-|\(bv\*(Tr |
83 | .ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p' |
84 | .ie n \{\ |
85 | . ds -- \(*W- |
86 | . ds PI pi |
87 | . if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch |
88 | . if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch |
89 | . ds L" "" |
90 | . ds R" "" |
ab1f1d91 |
91 | . ds C` @LQUOTE@ |
92 | . ds C' @RQUOTE@ |
9741dab0 |
93 | 'br\} |
94 | .el\{\ |
95 | . ds -- \|\(em\| |
96 | . ds PI \(*p |
97 | . ds L" `` |
98 | . ds R" '' |
99 | 'br\} |
100 | .\" |
3c014959 |
101 | .\" If the F register is turned on, we'll generate index entries on stderr for |
102 | .\" titles (.TH), headers (.SH), subsections (.Sh), items (.Ip), and index |
103 | .\" entries marked with X<> in POD. Of course, you'll have to process the |
104 | .\" output yourself in some meaningful fashion. |
9741dab0 |
105 | .if \nF \{\ |
106 | . de IX |
107 | . tm Index:\\$1\t\\n%\t"\\$2" |
f3248e50 |
108 | .. |
9741dab0 |
109 | . nr % 0 |
110 | . rr F |
111 | .\} |
112 | .\" |
3c014959 |
113 | .\" For nroff, turn off justification. Always turn off hyphenation; it makes |
114 | .\" way too many mistakes in technical documents. |
9741dab0 |
115 | .hy 0 |
116 | .if n .na |
117 | .\" |
118 | .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2). |
119 | .\" Fear. Run. Save yourself. No user-serviceable parts. |
9741dab0 |
120 | . \" fudge factors for nroff and troff |
121 | .if n \{\ |
122 | . ds #H 0 |
123 | . ds #V .8m |
124 | . ds #F .3m |
125 | . ds #[ \f1 |
126 | . ds #] \fP |
127 | .\} |
128 | .if t \{\ |
129 | . ds #H ((1u-(\\\\n(.fu%2u))*.13m) |
130 | . ds #V .6m |
131 | . ds #F 0 |
132 | . ds #[ \& |
133 | . ds #] \& |
134 | .\} |
135 | . \" simple accents for nroff and troff |
136 | .if n \{\ |
137 | . ds ' \& |
138 | . ds ` \& |
139 | . ds ^ \& |
140 | . ds , \& |
141 | . ds ~ ~ |
142 | . ds / |
143 | .\} |
144 | .if t \{\ |
145 | . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u" |
146 | . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u' |
147 | . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u' |
148 | . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u' |
149 | . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u' |
150 | . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u' |
151 | .\} |
152 | . \" troff and (daisy-wheel) nroff accents |
153 | .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V' |
154 | .ds 8 \h'\*(#H'\(*b\h'-\*(#H' |
155 | .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#] |
156 | .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H' |
157 | .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u' |
158 | .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#] |
159 | .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#] |
160 | .ds ae a\h'-(\w'a'u*4/10)'e |
161 | .ds Ae A\h'-(\w'A'u*4/10)'E |
162 | . \" corrections for vroff |
163 | .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u' |
164 | .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u' |
165 | . \" for low resolution devices (crt and lpr) |
166 | .if \n(.H>23 .if \n(.V>19 \ |
167 | \{\ |
168 | . ds : e |
169 | . ds 8 ss |
170 | . ds o a |
171 | . ds d- d\h'-1'\(ga |
172 | . ds D- D\h'-1'\(hy |
173 | . ds th \o'bp' |
174 | . ds Th \o'LP' |
175 | . ds ae ae |
176 | . ds Ae AE |
177 | .\} |
178 | .rm #[ #] #H #V #F C |
179 | ----END OF PREAMBLE---- |
5cdeb5a2 |
180 | #`# for cperl-mode |
181 | |
9741dab0 |
182 | # This table is taken nearly verbatim from Tom Christiansen's pod2man. It |
183 | # assumes that the standard preamble has already been printed, since that's |
184 | # what defines all of the accent marks. Note that some of these are quoted |
185 | # with double quotes since they contain embedded single quotes, so use \\ |
186 | # uniformly for backslash for readability. |
187 | %ESCAPES = ( |
188 | 'amp' => '&', # ampersand |
bf202ccd |
189 | 'apos' => "'", # apostrophe |
9741dab0 |
190 | 'lt' => '<', # left chevron, less-than |
191 | 'gt' => '>', # right chevron, greater-than |
192 | 'quot' => '"', # double quote |
ab1f1d91 |
193 | 'sol' => '/', # solidus (forward slash) |
be3174d2 |
194 | 'verbar' => '|', # vertical bar |
9741dab0 |
195 | |
196 | 'Aacute' => "A\\*'", # capital A, acute accent |
197 | 'aacute' => "a\\*'", # small a, acute accent |
198 | 'Acirc' => 'A\\*^', # capital A, circumflex accent |
199 | 'acirc' => 'a\\*^', # small a, circumflex accent |
200 | 'AElig' => '\*(AE', # capital AE diphthong (ligature) |
201 | 'aelig' => '\*(ae', # small ae diphthong (ligature) |
202 | 'Agrave' => "A\\*`", # capital A, grave accent |
203 | 'agrave' => "A\\*`", # small a, grave accent |
204 | 'Aring' => 'A\\*o', # capital A, ring |
205 | 'aring' => 'a\\*o', # small a, ring |
206 | 'Atilde' => 'A\\*~', # capital A, tilde |
207 | 'atilde' => 'a\\*~', # small a, tilde |
208 | 'Auml' => 'A\\*:', # capital A, dieresis or umlaut mark |
209 | 'auml' => 'a\\*:', # small a, dieresis or umlaut mark |
210 | 'Ccedil' => 'C\\*,', # capital C, cedilla |
211 | 'ccedil' => 'c\\*,', # small c, cedilla |
212 | 'Eacute' => "E\\*'", # capital E, acute accent |
213 | 'eacute' => "e\\*'", # small e, acute accent |
214 | 'Ecirc' => 'E\\*^', # capital E, circumflex accent |
215 | 'ecirc' => 'e\\*^', # small e, circumflex accent |
216 | 'Egrave' => 'E\\*`', # capital E, grave accent |
217 | 'egrave' => 'e\\*`', # small e, grave accent |
218 | 'ETH' => '\\*(D-', # capital Eth, Icelandic |
219 | 'eth' => '\\*(d-', # small eth, Icelandic |
220 | 'Euml' => 'E\\*:', # capital E, dieresis or umlaut mark |
221 | 'euml' => 'e\\*:', # small e, dieresis or umlaut mark |
222 | 'Iacute' => "I\\*'", # capital I, acute accent |
223 | 'iacute' => "i\\*'", # small i, acute accent |
224 | 'Icirc' => 'I\\*^', # capital I, circumflex accent |
225 | 'icirc' => 'i\\*^', # small i, circumflex accent |
226 | 'Igrave' => 'I\\*`', # capital I, grave accent |
227 | 'igrave' => 'i\\*`', # small i, grave accent |
228 | 'Iuml' => 'I\\*:', # capital I, dieresis or umlaut mark |
229 | 'iuml' => 'i\\*:', # small i, dieresis or umlaut mark |
230 | 'Ntilde' => 'N\*~', # capital N, tilde |
231 | 'ntilde' => 'n\*~', # small n, tilde |
232 | 'Oacute' => "O\\*'", # capital O, acute accent |
233 | 'oacute' => "o\\*'", # small o, acute accent |
234 | 'Ocirc' => 'O\\*^', # capital O, circumflex accent |
235 | 'ocirc' => 'o\\*^', # small o, circumflex accent |
236 | 'Ograve' => 'O\\*`', # capital O, grave accent |
237 | 'ograve' => 'o\\*`', # small o, grave accent |
238 | 'Oslash' => 'O\\*/', # capital O, slash |
239 | 'oslash' => 'o\\*/', # small o, slash |
240 | 'Otilde' => 'O\\*~', # capital O, tilde |
241 | 'otilde' => 'o\\*~', # small o, tilde |
242 | 'Ouml' => 'O\\*:', # capital O, dieresis or umlaut mark |
243 | 'ouml' => 'o\\*:', # small o, dieresis or umlaut mark |
244 | 'szlig' => '\*8', # small sharp s, German (sz ligature) |
245 | 'THORN' => '\\*(Th', # capital THORN, Icelandic |
246 | 'thorn' => '\\*(th', # small thorn, Icelandic |
247 | 'Uacute' => "U\\*'", # capital U, acute accent |
248 | 'uacute' => "u\\*'", # small u, acute accent |
249 | 'Ucirc' => 'U\\*^', # capital U, circumflex accent |
250 | 'ucirc' => 'u\\*^', # small u, circumflex accent |
251 | 'Ugrave' => 'U\\*`', # capital U, grave accent |
252 | 'ugrave' => 'u\\*`', # small u, grave accent |
253 | 'Uuml' => 'U\\*:', # capital U, dieresis or umlaut mark |
254 | 'uuml' => 'u\\*:', # small u, dieresis or umlaut mark |
255 | 'Yacute' => "Y\\*'", # capital Y, acute accent |
256 | 'yacute' => "y\\*'", # small y, acute accent |
257 | 'yuml' => 'y\\*:', # small y, dieresis or umlaut mark |
bf202ccd |
258 | |
259 | 'nbsp' => '\\ ', # non-breaking space |
260 | 'shy' => '', # soft (discretionary) hyphen |
9741dab0 |
261 | ); |
262 | |
263 | |
3c014959 |
264 | ############################################################################## |
9741dab0 |
265 | # Static helper functions |
3c014959 |
266 | ############################################################################## |
9741dab0 |
267 | |
3c014959 |
268 | # Protect leading quotes and periods against interpretation as commands. Also |
269 | # protect anything starting with a backslash, since it could expand or hide |
270 | # something that *roff would interpret as a command. This is overkill, but |
271 | # it's much simpler than trying to parse *roff here. |
c9abbd5d |
272 | sub protect { |
273 | local $_ = shift; |
44464a02 |
274 | s/^([.\'\\])/\\&$1/mg; |
c9abbd5d |
275 | $_; |
276 | } |
5cdeb5a2 |
277 | |
9741dab0 |
278 | # Translate a font string into an escape. |
279 | sub toescape { (length ($_[0]) > 1 ? '\f(' : '\f') . $_[0] } |
280 | |
5cdeb5a2 |
281 | |
3c014959 |
282 | ############################################################################## |
9741dab0 |
283 | # Initialization |
3c014959 |
284 | ############################################################################## |
285 | |
286 | # Initialize the object. Here, we also process any additional options passed |
287 | # to the constructor or set up defaults if none were given. center is the |
288 | # centered title, release is the version number, and date is the date for the |
289 | # documentation. Note that we can't know what file name we're processing due |
290 | # to the architecture of Pod::Parser, so that *has* to either be passed to the |
291 | # constructor or set separately with Pod::Man::name(). |
9741dab0 |
292 | sub initialize { |
293 | my $self = shift; |
294 | |
3c014959 |
295 | # Figure out the fixed-width font. If user-supplied, make sure that they |
296 | # are the right length. |
9741dab0 |
297 | for (qw/fixed fixedbold fixeditalic fixedbolditalic/) { |
298 | if (defined $$self{$_}) { |
299 | if (length ($$self{$_}) < 1 || length ($$self{$_}) > 2) { |
ab1f1d91 |
300 | croak qq(roff font should be 1 or 2 chars,) |
301 | . qq( not "$$self{$_}"); |
9741dab0 |
302 | } |
303 | } else { |
304 | $$self{$_} = ''; |
305 | } |
306 | } |
307 | |
3c014959 |
308 | # Set the default fonts. We can't be sure what fixed bold-italic is going |
309 | # to be called, so default to just bold. |
9741dab0 |
310 | $$self{fixed} ||= 'CW'; |
311 | $$self{fixedbold} ||= 'CB'; |
312 | $$self{fixeditalic} ||= 'CI'; |
313 | $$self{fixedbolditalic} ||= 'CB'; |
314 | |
3c014959 |
315 | # Set up a table of font escapes. First number is fixed-width, second is |
316 | # bold, third is italic. |
9741dab0 |
317 | $$self{FONTS} = { '000' => '\fR', '001' => '\fI', |
318 | '010' => '\fB', '011' => '\f(BI', |
319 | '100' => toescape ($$self{fixed}), |
320 | '101' => toescape ($$self{fixeditalic}), |
321 | '110' => toescape ($$self{fixedbold}), |
322 | '111' => toescape ($$self{fixedbolditalic})}; |
323 | |
324 | # Extra stuff for page titles. |
325 | $$self{center} = 'User Contributed Perl Documentation' |
326 | unless defined $$self{center}; |
b84d8b9e |
327 | $$self{indent} = 4 unless defined $$self{indent}; |
9741dab0 |
328 | |
3c014959 |
329 | # We used to try first to get the version number from a local binary, but |
330 | # we shouldn't need that any more. Get the version from the running Perl. |
331 | # Work a little magic to handle subversions correctly under both the |
332 | # pre-5.6 and the post-5.6 version numbering schemes. |
9741dab0 |
333 | if (!defined $$self{release}) { |
c9abbd5d |
334 | my @version = ($] =~ /^(\d+)\.(\d{3})(\d{0,3})$/); |
335 | $version[2] ||= 0; |
336 | $version[2] *= 10 ** (3 - length $version[2]); |
337 | for (@version) { $_ += 0 } |
338 | $$self{release} = 'perl v' . join ('.', @version); |
9741dab0 |
339 | } |
340 | |
341 | # Double quotes in things that will be quoted. |
c9abbd5d |
342 | for (qw/center date release/) { |
343 | $$self{$_} =~ s/\"/\"\"/g if $$self{$_}; |
344 | } |
9741dab0 |
345 | |
ab1f1d91 |
346 | # Figure out what quotes we'll be using for C<> text. |
5cdeb5a2 |
347 | $$self{quotes} ||= '"'; |
ab1f1d91 |
348 | if ($$self{quotes} eq 'none') { |
349 | $$self{LQUOTE} = $$self{RQUOTE} = ''; |
350 | } elsif (length ($$self{quotes}) == 1) { |
351 | $$self{LQUOTE} = $$self{RQUOTE} = $$self{quotes}; |
352 | } elsif ($$self{quotes} =~ /^(.)(.)$/ |
353 | || $$self{quotes} =~ /^(..)(..)$/) { |
354 | $$self{LQUOTE} = $1; |
355 | $$self{RQUOTE} = $2; |
356 | } else { |
357 | croak qq(Invalid quote specification "$$self{quotes}"); |
358 | } |
359 | |
3c014959 |
360 | # Double the first quote; note that this should not be s///g as two double |
361 | # quotes is represented in *roff as three double quotes, not four. Weird, |
362 | # I know. |
ab1f1d91 |
363 | $$self{LQUOTE} =~ s/\"/\"\"/; |
364 | $$self{RQUOTE} =~ s/\"/\"\"/; |
365 | |
9741dab0 |
366 | $self->SUPER::initialize; |
367 | } |
368 | |
ab1f1d91 |
369 | # For each document we process, output the preamble first. |
9741dab0 |
370 | sub begin_pod { |
371 | my $self = shift; |
372 | |
373 | # Try to figure out the name and section from the file name. |
374 | my $section = $$self{section} || 1; |
375 | my $name = $$self{name}; |
376 | if (!defined $name) { |
377 | $name = $self->input_file; |
fe6f1558 |
378 | $section = 3 if (!$$self{section} && $name =~ /\.pm\z/i); |
379 | $name =~ s/\.p(od|[lm])\z//i; |
bf202ccd |
380 | if ($section !~ /^3/) { |
9741dab0 |
381 | require File::Basename; |
382 | $name = uc File::Basename::basename ($name); |
383 | } else { |
bf202ccd |
384 | # Assume that we're dealing with a module. We want to figure out |
385 | # the full module name from the path to the file, but we don't |
386 | # want to include too much of the path into the module name. Lose |
387 | # everything up to the first of: |
388 | # |
389 | # */lib/*perl*/ standard or site_perl module |
390 | # */*perl*/lib/ from -Dprefix=/opt/perl |
391 | # */*perl*/ random module hierarchy |
392 | # |
393 | # which works. Also strip off a leading site or site_perl |
394 | # component, any OS-specific component, and any version number |
395 | # component, and strip off an initial component of "lib" or |
396 | # "blib/lib" since that's what ExtUtils::MakeMaker creates. |
397 | # splitdir requires at least File::Spec 0.8. |
398 | require File::Spec; |
399 | my ($volume, $dirs, $file) = File::Spec->splitpath ($name); |
400 | my @dirs = File::Spec->splitdir ($dirs); |
401 | my $cut = 0; |
402 | my $i; |
403 | for ($i = 0; $i < scalar @dirs; $i++) { |
404 | if ($dirs[$i] eq 'lib' && $dirs[$i + 1] =~ /perl/) { |
405 | $cut = $i + 2; |
406 | last; |
407 | } elsif ($dirs[$i] =~ /perl/) { |
408 | $cut = $i + 1; |
409 | $cut++ if $dirs[$i + 1] eq 'lib'; |
410 | last; |
9741dab0 |
411 | } |
9741dab0 |
412 | } |
bf202ccd |
413 | if ($cut > 0) { |
414 | splice (@dirs, 0, $cut); |
415 | shift @dirs if ($dirs[0] =~ /^site(_perl)?$/); |
416 | shift @dirs if ($dirs[0] =~ /^[\d.]+$/); |
b616daaf |
417 | shift @dirs if ($dirs[0] =~ /^(.*-$^O|$^O-.*|$^O)$/); |
bf202ccd |
418 | } |
419 | shift @dirs if $dirs[0] eq 'lib'; |
420 | splice (@dirs, 0, 2) if ($dirs[0] eq 'blib' && $dirs[1] eq 'lib'); |
421 | |
422 | # Remove empty directories when building the module name; they |
423 | # occur too easily on Unix by doubling slashes. |
424 | $name = join ('::', (grep { $_ ? $_ : () } @dirs), $file); |
9741dab0 |
425 | } |
426 | } |
427 | |
3c014959 |
428 | # If $name contains spaces, quote it; this mostly comes up in the case of |
429 | # input from stdin. |
77003bb1 |
430 | $name = '"' . $name . '"' if ($name =~ /\s/); |
431 | |
9741dab0 |
432 | # Modification date header. Try to use the modification time of our |
433 | # input. |
434 | if (!defined $$self{date}) { |
435 | my $time = (stat $self->input_file)[9] || time; |
436 | my ($day, $month, $year) = (localtime $time)[3,4,5]; |
437 | $month++; |
438 | $year += 1900; |
c9abbd5d |
439 | $$self{date} = sprintf ('%4d-%02d-%02d', $year, $month, $day); |
9741dab0 |
440 | } |
441 | |
bf202ccd |
442 | # Now, print out the preamble and the title. The meaning of the arguments |
443 | # to .TH unfortunately vary by system; some systems consider the fourth |
444 | # argument to be a "source" and others use it as a version number. |
445 | # Generally it's just presented as the left-side footer, though, so it |
446 | # doesn't matter too much if a particular system gives it another |
447 | # interpretation. |
448 | # |
449 | # The order of date and release used to be reversed in older versions of |
450 | # this module, but this order is correct for both Solaris and Linux. |
ab1f1d91 |
451 | local $_ = $PREAMBLE; |
452 | s/\@CFONT\@/$$self{fixed}/; |
453 | s/\@LQUOTE\@/$$self{LQUOTE}/; |
454 | s/\@RQUOTE\@/$$self{RQUOTE}/; |
455 | chomp $_; |
bf202ccd |
456 | my $pversion = $Pod::Parser::VERSION; |
9741dab0 |
457 | print { $self->output_handle } <<"----END OF HEADER----"; |
bf202ccd |
458 | .\\" Automatically generated by Pod::Man v$VERSION, Pod::Parser v$pversion |
9741dab0 |
459 | .\\" |
460 | .\\" Standard preamble: |
3c014959 |
461 | .\\" ======================================================================== |
ab1f1d91 |
462 | $_ |
3c014959 |
463 | .\\" ======================================================================== |
9741dab0 |
464 | .\\" |
465 | .IX Title "$name $section" |
bf202ccd |
466 | .TH $name $section "$$self{date}" "$$self{release}" "$$self{center}" |
9741dab0 |
467 | .UC |
468 | ----END OF HEADER---- |
9741dab0 |
469 | |
470 | # Initialize a few per-file variables. |
b616daaf |
471 | $$self{INDENT} = 0; # Current indentation level. |
472 | $$self{INDENTS} = []; # Stack of indentations. |
473 | $$self{INDEX} = []; # Index keys waiting to be printed. |
474 | $$self{ITEMS} = 0; # The number of consecutive =items. |
475 | $$self{SHIFTWAIT} = 0; # Whether there is a shift waiting. |
476 | $$self{SHIFTS} = []; # Stack of .RS shifts. |
9741dab0 |
477 | } |
478 | |
479 | |
3c014959 |
480 | ############################################################################## |
9741dab0 |
481 | # Core overrides |
3c014959 |
482 | ############################################################################## |
9741dab0 |
483 | |
484 | # Called for each command paragraph. Gets the command, the associated |
485 | # paragraph, the line number, and a Pod::Paragraph object. Just dispatches |
486 | # the command to a method named the same as the command. =cut is handled |
487 | # internally by Pod::Parser. |
488 | sub command { |
489 | my $self = shift; |
490 | my $command = shift; |
491 | return if $command eq 'pod'; |
3c014959 |
492 | return if ($$self{EXCLUDE} && $command ne 'end'); |
ab1f1d91 |
493 | if ($self->can ('cmd_' . $command)) { |
494 | $command = 'cmd_' . $command; |
844b31e3 |
495 | $self->$command (@_); |
3c014959 |
496 | } else { |
ab1f1d91 |
497 | my ($text, $line, $paragraph) = @_; |
5cdeb5a2 |
498 | my $file; |
499 | ($file, $line) = $paragraph->file_line; |
ab1f1d91 |
500 | $text =~ s/\n+\z//; |
501 | $text = " $text" if ($text =~ /^\S/); |
502 | warn qq($file:$line: Unknown command paragraph "=$command$text"\n); |
503 | return; |
844b31e3 |
504 | } |
9741dab0 |
505 | } |
506 | |
3c014959 |
507 | # Called for a verbatim paragraph. Gets the paragraph, the line number, and a |
508 | # Pod::Paragraph object. Rofficate backslashes, untabify, put a zero-width |
509 | # character at the beginning of each line to protect against commands, and |
510 | # wrap in .Vb/.Ve. |
9741dab0 |
511 | sub verbatim { |
512 | my $self = shift; |
513 | return if $$self{EXCLUDE}; |
514 | local $_ = shift; |
515 | return if /^\s+$/; |
516 | s/\s+$/\n/; |
517 | my $lines = tr/\n/\n/; |
518 | 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me; |
519 | s/\\/\\e/g; |
520 | s/^(\s*\S)/'\&' . $1/gme; |
5cdeb5a2 |
521 | $self->makespace; |
9741dab0 |
522 | $self->output (".Vb $lines\n$_.Ve\n"); |
523 | $$self{NEEDSPACE} = 0; |
524 | } |
525 | |
3c014959 |
526 | # Called for a regular text block. Gets the paragraph, the line number, and a |
527 | # Pod::Paragraph object. Perform interpolation and output the results. |
9741dab0 |
528 | sub textblock { |
529 | my $self = shift; |
530 | return if $$self{EXCLUDE}; |
531 | $self->output ($_[0]), return if $$self{VERBATIM}; |
532 | |
bf202ccd |
533 | # Parse the tree. collapse knows about references to scalars as well as |
534 | # scalars and does the right thing with them. Tidy up any trailing |
535 | # whitespace. |
c9abbd5d |
536 | my $text = shift; |
c9abbd5d |
537 | $text = $self->parse ($text, @_); |
538 | $text =~ s/\n\s*$/\n/; |
bf202ccd |
539 | |
540 | # Output the paragraph. We also have to handle =over without =item. If |
541 | # there's an =over without =item, NEWINDENT will be set, and we need to |
542 | # handle creation of the indent here. Set WEIRDINDENT so that it will be |
543 | # cleaned up on =back. |
5cdeb5a2 |
544 | $self->makespace; |
b616daaf |
545 | if ($$self{SHIFTWAIT}) { |
bf202ccd |
546 | $self->output (".RS $$self{INDENT}\n"); |
b616daaf |
547 | push (@{ $$self{SHIFTS} }, $$self{INDENT}); |
548 | $$self{SHIFTWAIT} = 0; |
bf202ccd |
549 | } |
50a3fd2a |
550 | $self->output (protect $self->textmapfonts ($text)); |
9741dab0 |
551 | $self->outindex; |
552 | $$self{NEEDSPACE} = 1; |
553 | } |
554 | |
b84d8b9e |
555 | # Called for a formatting code. Takes a Pod::InteriorSequence object and |
9741dab0 |
556 | # returns a reference to a scalar. This scalar is the final formatted text. |
b84d8b9e |
557 | # It's returned as a reference to an array so that other formatting codes |
558 | # above us know that the text has already been processed. |
9741dab0 |
559 | sub sequence { |
560 | my ($self, $seq) = @_; |
561 | my $command = $seq->cmd_name; |
562 | |
bf202ccd |
563 | # We have to defer processing of the inside of an L<> formatting code. If |
b84d8b9e |
564 | # this code is nested inside an L<> code, return the literal raw text of |
565 | # it. |
bf202ccd |
566 | my $parent = $seq->nested; |
567 | while (defined $parent) { |
568 | return $seq->raw_text if ($parent->cmd_name eq 'L'); |
569 | $parent = $parent->nested; |
570 | } |
571 | |
9741dab0 |
572 | # Zero-width characters. |
b84d8b9e |
573 | return [ '\&' ] if ($command eq 'Z'); |
9741dab0 |
574 | |
a3e04946 |
575 | # C<>, L<>, X<>, and E<> don't apply guesswork to their contents. C<> |
576 | # needs some additional special handling. |
577 | my $literal = ($command =~ /^[CELX]$/); |
b84d8b9e |
578 | local $_ = $self->collapse ($seq->parse_tree, $literal, $command eq 'C'); |
9741dab0 |
579 | |
bf202ccd |
580 | # Handle E<> escapes. Numeric escapes that match one of the supported ISO |
581 | # 8859-1 characters don't work at present. |
9741dab0 |
582 | if ($command eq 'E') { |
2e20e14f |
583 | if (/^\d+$/) { |
b84d8b9e |
584 | return [ chr ($_) ]; |
2e20e14f |
585 | } elsif (exists $ESCAPES{$_}) { |
b84d8b9e |
586 | return [ $ESCAPES{$_} ]; |
9741dab0 |
587 | } else { |
aa212ad6 |
588 | my ($file, $line) = $seq->file_line; |
589 | warn "$file:$line: Unknown escape E<$_>\n"; |
b84d8b9e |
590 | return [ "E<$_>" ]; |
9741dab0 |
591 | } |
592 | } |
593 | |
b84d8b9e |
594 | # For all the other codes, empty content produces no output. |
9741dab0 |
595 | return '' if $_ eq ''; |
596 | |
b84d8b9e |
597 | # Handle simple formatting codes. |
9741dab0 |
598 | if ($command eq 'B') { |
b84d8b9e |
599 | return [ '\f(BS' . $_ . '\f(BE' ]; |
600 | } elsif ($command eq 'F' || $command eq 'I') { |
601 | return [ '\f(IS' . $_ . '\f(IE' ]; |
9741dab0 |
602 | } elsif ($command eq 'C') { |
b84d8b9e |
603 | return [ $self->quote_literal ($_) ]; |
9741dab0 |
604 | } |
605 | |
606 | # Handle links. |
607 | if ($command eq 'L') { |
bf202ccd |
608 | my ($text, $type) = (parselink ($_))[1,4]; |
609 | return '' unless $text; |
610 | my ($file, $line) = $seq->file_line; |
611 | $text = $self->parse ($text, $line); |
612 | $text = '<' . $text . '>' if $type eq 'url'; |
b84d8b9e |
613 | return [ $text ]; |
9741dab0 |
614 | } |
5cdeb5a2 |
615 | |
9741dab0 |
616 | # Whitespace protection replaces whitespace with "\ ". |
617 | if ($command eq 'S') { |
618 | s/\s+/\\ /g; |
b84d8b9e |
619 | return [ $_ ]; |
9741dab0 |
620 | } |
621 | |
622 | # Add an index entry to the list of ones waiting to be output. |
b84d8b9e |
623 | if ($command eq 'X') { |
624 | push (@{ $$self{INDEX} }, $_); |
625 | return ''; |
626 | } |
9741dab0 |
627 | |
628 | # Anything else is unknown. |
aa212ad6 |
629 | my ($file, $line) = $seq->file_line; |
b84d8b9e |
630 | warn "$file:$line: Unknown formatting code $command<$_>\n"; |
9741dab0 |
631 | } |
632 | |
633 | |
3c014959 |
634 | ############################################################################## |
9741dab0 |
635 | # Command paragraphs |
3c014959 |
636 | ############################################################################## |
9741dab0 |
637 | |
638 | # All command paragraphs take the paragraph and the line number. |
639 | |
640 | # First level heading. We can't output .IX in the NAME section due to a bug |
641 | # in some versions of catman, so don't output a .IX for that section. .SH |
b84d8b9e |
642 | # already uses small caps, so remove \s1 and \s-1. |
9741dab0 |
643 | sub cmd_head1 { |
644 | my $self = shift; |
645 | local $_ = $self->parse (@_); |
646 | s/\s+$//; |
647 | s/\\s-?\d//g; |
77003bb1 |
648 | s/\s*\n\s*/ /g; |
5cdeb5a2 |
649 | if ($$self{ITEMS} > 1) { |
650 | $$self{ITEMS} = 0; |
651 | $self->output (".PD\n"); |
652 | } |
50a3fd2a |
653 | $self->output ($self->switchquotes ('.SH', $self->mapfonts ($_))); |
9741dab0 |
654 | $self->outindex (($_ eq 'NAME') ? () : ('Header', $_)); |
655 | $$self{NEEDSPACE} = 0; |
656 | } |
657 | |
658 | # Second level heading. |
659 | sub cmd_head2 { |
660 | my $self = shift; |
661 | local $_ = $self->parse (@_); |
662 | s/\s+$//; |
77003bb1 |
663 | s/\s*\n\s*/ /g; |
5cdeb5a2 |
664 | if ($$self{ITEMS} > 1) { |
665 | $$self{ITEMS} = 0; |
666 | $self->output (".PD\n"); |
667 | } |
50a3fd2a |
668 | $self->output ($self->switchquotes ('.Sh', $self->mapfonts ($_))); |
9741dab0 |
669 | $self->outindex ('Subsection', $_); |
670 | $$self{NEEDSPACE} = 0; |
671 | } |
672 | |
50a3fd2a |
673 | # Third level heading. |
674 | sub cmd_head3 { |
675 | my $self = shift; |
676 | local $_ = $self->parse (@_); |
677 | s/\s+$//; |
77003bb1 |
678 | s/\s*\n\s*/ /g; |
50a3fd2a |
679 | if ($$self{ITEMS} > 1) { |
680 | $$self{ITEMS} = 0; |
681 | $self->output (".PD\n"); |
682 | } |
683 | $self->makespace; |
b616daaf |
684 | $self->output ($self->textmapfonts ('\f(IS' . $_ . '\f(IE') . "\n"); |
50a3fd2a |
685 | $self->outindex ('Subsection', $_); |
686 | $$self{NEEDSPACE} = 1; |
687 | } |
688 | |
689 | # Fourth level heading. |
690 | sub cmd_head4 { |
691 | my $self = shift; |
692 | local $_ = $self->parse (@_); |
693 | s/\s+$//; |
77003bb1 |
694 | s/\s*\n\s*/ /g; |
50a3fd2a |
695 | if ($$self{ITEMS} > 1) { |
696 | $$self{ITEMS} = 0; |
697 | $self->output (".PD\n"); |
698 | } |
699 | $self->makespace; |
700 | $self->output ($self->textmapfonts ($_) . "\n"); |
701 | $self->outindex ('Subsection', $_); |
702 | $$self{NEEDSPACE} = 1; |
703 | } |
704 | |
9741dab0 |
705 | # Start a list. For indents after the first, wrap the outside indent in .RS |
706 | # so that hanging paragraph tags will be correct. |
707 | sub cmd_over { |
708 | my $self = shift; |
709 | local $_ = shift; |
710 | unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} } |
b616daaf |
711 | if (@{ $$self{SHIFTS} } < @{ $$self{INDENTS} }) { |
9741dab0 |
712 | $self->output (".RS $$self{INDENT}\n"); |
b616daaf |
713 | push (@{ $$self{SHIFTS} }, $$self{INDENT}); |
9741dab0 |
714 | } |
715 | push (@{ $$self{INDENTS} }, $$self{INDENT}); |
716 | $$self{INDENT} = ($_ + 0); |
b616daaf |
717 | $$self{SHIFTWAIT} = 1; |
9741dab0 |
718 | } |
719 | |
720 | # End a list. If we've closed an embedded indent, we've mangled the hanging |
721 | # paragraph indent, so temporarily replace it with .RS and set WEIRDINDENT. |
722 | # We'll close that .RS at the next =back or =item. |
723 | sub cmd_back { |
724 | my $self = shift; |
725 | $$self{INDENT} = pop @{ $$self{INDENTS} }; |
726 | unless (defined $$self{INDENT}) { |
aa212ad6 |
727 | my ($file, $line, $paragraph) = @_; |
728 | ($file, $line) = $paragraph->file_line; |
729 | warn "$file:$line: Unmatched =back\n"; |
9741dab0 |
730 | $$self{INDENT} = 0; |
731 | } |
b616daaf |
732 | if (@{ $$self{SHIFTS} } > @{ $$self{INDENTS} }) { |
9741dab0 |
733 | $self->output (".RE\n"); |
b616daaf |
734 | pop @{ $$self{SHIFTS} }; |
9741dab0 |
735 | } |
736 | if (@{ $$self{INDENTS} } > 0) { |
737 | $self->output (".RE\n"); |
738 | $self->output (".RS $$self{INDENT}\n"); |
9741dab0 |
739 | } |
740 | $$self{NEEDSPACE} = 1; |
b616daaf |
741 | $$self{SHIFTWAIT} = 0; |
9741dab0 |
742 | } |
743 | |
744 | # An individual list item. Emit an index entry for anything that's |
745 | # interesting, but don't emit index entries for things like bullets and |
3c014959 |
746 | # numbers. rofficate bullets too while we're at it (so for nice output, use * |
747 | # for your lists rather than o or . or - or some other thing). Newlines in an |
748 | # item title are turned into spaces since *roff can't handle them embedded. |
9741dab0 |
749 | sub cmd_item { |
750 | my $self = shift; |
751 | local $_ = $self->parse (@_); |
752 | s/\s+$//; |
46bce7d0 |
753 | s/\s*\n\s*/ /g; |
9741dab0 |
754 | my $index; |
755 | if (/\w/ && !/^\w[.\)]\s*$/) { |
756 | $index = $_; |
fe2227f0 |
757 | $index =~ s/^\s*[-*+o.]?(?:\s+|\Z)//; |
9741dab0 |
758 | } |
bf202ccd |
759 | $_ = '*' unless $_; |
9741dab0 |
760 | s/^\*(\s|\Z)/\\\(bu$1/; |
b616daaf |
761 | if (@{ $$self{SHIFTS} } == @{ $$self{INDENTS} }) { |
9741dab0 |
762 | $self->output (".RE\n"); |
b616daaf |
763 | pop @{ $$self{SHIFTS} }; |
9741dab0 |
764 | } |
50a3fd2a |
765 | $_ = $self->textmapfonts ($_); |
5cdeb5a2 |
766 | $self->output (".PD 0\n") if ($$self{ITEMS} == 1); |
3c014959 |
767 | $self->output ($self->switchquotes ('.IP', $_, $$self{INDENT})); |
9741dab0 |
768 | $self->outindex ($index ? ('Item', $index) : ()); |
769 | $$self{NEEDSPACE} = 0; |
5cdeb5a2 |
770 | $$self{ITEMS}++; |
b616daaf |
771 | $$self{SHIFTWAIT} = 0; |
9741dab0 |
772 | } |
773 | |
774 | # Begin a block for a particular translator. Setting VERBATIM triggers |
775 | # special handling in textblock(). |
776 | sub cmd_begin { |
777 | my $self = shift; |
778 | local $_ = shift; |
779 | my ($kind) = /^(\S+)/ or return; |
780 | if ($kind eq 'man' || $kind eq 'roff') { |
781 | $$self{VERBATIM} = 1; |
782 | } else { |
783 | $$self{EXCLUDE} = 1; |
784 | } |
785 | } |
786 | |
787 | # End a block for a particular translator. We assume that all =begin/=end |
788 | # pairs are properly closed. |
789 | sub cmd_end { |
790 | my $self = shift; |
791 | $$self{EXCLUDE} = 0; |
792 | $$self{VERBATIM} = 0; |
793 | } |
794 | |
795 | # One paragraph for a particular translator. Ignore it unless it's intended |
796 | # for man or roff, in which case we output it verbatim. |
797 | sub cmd_for { |
798 | my $self = shift; |
799 | local $_ = shift; |
9741dab0 |
800 | return unless s/^(?:man|roff)\b[ \t]*\n?//; |
801 | $self->output ($_); |
802 | } |
803 | |
804 | |
3c014959 |
805 | ############################################################################## |
9741dab0 |
806 | # Escaping and fontification |
3c014959 |
807 | ############################################################################## |
9741dab0 |
808 | |
809 | # At this point, we'll have embedded font codes of the form \f(<font>[SE] |
3c014959 |
810 | # where <font> is one of B, I, or F. Turn those into the right font start or |
811 | # end codes. The old pod2man didn't get B<someI<thing> else> right; after I<> |
812 | # it switched back to normal text rather than bold. We take care of this by |
813 | # using variables as a combined pointer to our current font sequence, and set |
814 | # each to the number of current nestings of start tags for that font. Use |
815 | # them as a vector to look up what font sequence to use. |
50a3fd2a |
816 | # |
817 | # \fP changes to the previous font, but only one previous font is kept. We |
818 | # don't know what the outside level font is; normally it's R, but if we're |
3c014959 |
819 | # inside a heading it could be something else. So arrange things so that the |
820 | # outside font is always the "previous" font and end with \fP instead of \fR. |
821 | # Idea from Zack Weinberg. |
9741dab0 |
822 | sub mapfonts { |
823 | my $self = shift; |
824 | local $_ = shift; |
825 | |
826 | my ($fixed, $bold, $italic) = (0, 0, 0); |
827 | my %magic = (F => \$fixed, B => \$bold, I => \$italic); |
50a3fd2a |
828 | my $last = '\fR'; |
829 | s { \\f\((.)(.) } { |
830 | my $sequence = ''; |
831 | my $f; |
832 | if ($last ne '\fR') { $sequence = '\fP' } |
833 | ${ $magic{$1} } += ($2 eq 'S') ? 1 : -1; |
834 | $f = $$self{FONTS}{($fixed && 1) . ($bold && 1) . ($italic && 1)}; |
835 | if ($f eq $last) { |
836 | ''; |
837 | } else { |
838 | if ($f ne '\fR') { $sequence .= $f } |
839 | $last = $f; |
840 | $sequence; |
841 | } |
842 | }gxe; |
843 | $_; |
844 | } |
845 | |
846 | # Unfortunately, there is a bug in Solaris 2.6 nroff (not present in GNU |
847 | # groff) where the sequence \fB\fP\f(CW\fP leaves the font set to B rather |
3c014959 |
848 | # than R, presumably because \f(CW doesn't actually do a font change. To work |
849 | # around this, use a separate textmapfonts for text blocks where the default |
850 | # font is always R and only use the smart mapfonts for headings. |
50a3fd2a |
851 | sub textmapfonts { |
852 | my $self = shift; |
853 | local $_ = shift; |
854 | |
855 | my ($fixed, $bold, $italic) = (0, 0, 0); |
856 | my %magic = (F => \$fixed, B => \$bold, I => \$italic); |
9741dab0 |
857 | s { \\f\((.)(.) } { |
858 | ${ $magic{$1} } += ($2 eq 'S') ? 1 : -1; |
859 | $$self{FONTS}{($fixed && 1) . ($bold && 1) . ($italic && 1)}; |
860 | }gxe; |
861 | $_; |
862 | } |
863 | |
864 | |
3c014959 |
865 | ############################################################################## |
866 | # *roff-specific parsing and magic |
867 | ############################################################################## |
9741dab0 |
868 | |
869 | # Called instead of parse_text, calls parse_text with the right flags. |
870 | sub parse { |
871 | my $self = shift; |
872 | $self->parse_text ({ -expand_seq => 'sequence', |
873 | -expand_ptree => 'collapse' }, @_); |
874 | } |
5cdeb5a2 |
875 | |
b84d8b9e |
876 | # Takes a parse tree, a flag saying whether or not to treat it as literal text |
877 | # (not call guesswork on it), and a flag saying whether or not to clean some |
878 | # things up for *roff, and returns the concatenation of all of the text |
879 | # strings in that parse tree. If the literal flag isn't true, guesswork() |
880 | # will be called on all plain scalars in the parse tree. Otherwise, just |
881 | # escape backslashes in the normal case. If collapse is being called on a C<> |
882 | # code, $cleanup should be set to true and some additional cleanup will be |
883 | # done. Assumes that everything in the parse tree is either a scalar or a |
3c014959 |
884 | # reference to a scalar. |
9741dab0 |
885 | sub collapse { |
b84d8b9e |
886 | my ($self, $ptree, $literal, $cleanup) = @_; |
887 | return join ('', map { |
888 | if (ref $_) { |
889 | join ('', @$_); |
890 | } elsif ($literal) { |
891 | if ($cleanup) { |
892 | s/\\/\\e/g; |
893 | s/-/\\-/g; |
894 | s/__/_\\|_/g; |
9741dab0 |
895 | } |
b84d8b9e |
896 | $_; |
897 | } else { |
898 | $self->guesswork ($_); |
899 | } |
900 | } $ptree->children); |
9741dab0 |
901 | } |
902 | |
903 | # Takes a text block to perform guesswork on; this is guaranteed not to |
b84d8b9e |
904 | # contain any formatting codes. Returns the text block with remapping done. |
9741dab0 |
905 | sub guesswork { |
906 | my $self = shift; |
907 | local $_ = shift; |
908 | |
909 | # rofficate backslashes. |
910 | s/\\/\\e/g; |
911 | |
912 | # Ensure double underbars have a tiny space between them. |
913 | s/__/_\\|_/g; |
914 | |
bf202ccd |
915 | # Leave hyphens only if they're part of regular words and there is only |
916 | # one dash at a time. Leave a dash after the first character as a regular |
917 | # non-breaking dash, but don't let it mark the rest of the word invalid |
918 | # for hyphenation. |
919 | s/-/\\-/g; |
920 | s{ |
921 | ( (?:\G|^|\s) [a-zA-Z] ) ( \\- )? |
922 | ( (?: [a-zA-Z]+ \\-)+ ) |
923 | ( [a-zA-Z]+ ) (?=\s|\Z) |
924 | \b |
925 | } { |
926 | my ($prefix, $hyphen, $main, $suffix) = ($1, $2, $3, $4); |
927 | $hyphen ||= ''; |
928 | $main =~ s/\\-/-/g; |
929 | $prefix . $hyphen . $main . $suffix; |
930 | }egx; |
931 | |
932 | # Translate -- into a real em dash if it's used like one. |
933 | s{ (\s) \\-\\- (\s) } { $1 . '\*(--' . $2 }egx; |
934 | s{ (\b[a-zA-Z]+) \\-\\- (\s|\Z|[a-zA-Z]+\b) } { $1 . '\*(--' . $2 }egx; |
935 | |
3c014959 |
936 | # Make all caps a little smaller. Be careful here, since we don't want to |
937 | # make @ARGV into small caps, nor do we want to fix the MIME in |
9741dab0 |
938 | # MIME-Version, since it looks weird with the full-height V. |
939 | s{ |
940 | ( ^ | [\s\(\"\'\`\[\{<>] ) |
bf202ccd |
941 | ( [A-Z] [A-Z] (?: [/A-Z+:\d_\$&] | \\- )* ) |
942 | (?= [\s>\}\]\(\)\'\".?!,;] | \\*\(-- | $ ) |
c9abbd5d |
943 | } { $1 . '\s-1' . $2 . '\s0' }egx; |
9741dab0 |
944 | |
9741dab0 |
945 | # Italize functions in the form func(). |
946 | s{ |
ee89c1da |
947 | ( \b | \\s-1 ) |
9741dab0 |
948 | ( |
ee89c1da |
949 | [A-Za-z_] ([:\w]|\\s-?[01])+ \(\) |
9741dab0 |
950 | ) |
ee89c1da |
951 | } { $1 . '\f(IS' . $2 . '\f(IE' }egx; |
9741dab0 |
952 | |
953 | # func(n) is a reference to a manual page. Make it \fIfunc\fR\|(n). |
954 | s{ |
ee89c1da |
955 | ( \b | \\s-1 ) |
bf202ccd |
956 | ( [A-Za-z_] (?:[.:\w]|\\-|\\s-?[01])+ ) |
9741dab0 |
957 | ( |
ee89c1da |
958 | \( \d [a-z]* \) |
9741dab0 |
959 | ) |
ee89c1da |
960 | } { $1 . '\f(IS' . $2 . '\f(IE\|' . $3 }egx; |
9741dab0 |
961 | |
962 | # Convert simple Perl variable references to a fixed-width font. |
963 | s{ |
964 | ( \s+ ) |
965 | ( [\$\@%] [\w:]+ ) |
966 | (?! \( ) |
967 | } { $1 . '\f(FS' . $2 . '\f(FE'}egx; |
968 | |
9741dab0 |
969 | # Fix up double quotes. |
970 | s{ \" ([^\"]+) \" } { '\*(L"' . $1 . '\*(R"' }egx; |
971 | |
972 | # Make C++ into \*(C+, which is a squinched version. |
973 | s{ \b C\+\+ } {\\*\(C+}gx; |
974 | |
975 | # All done. |
976 | $_; |
977 | } |
978 | |
3c014959 |
979 | # Handles C<> text, deciding whether to put \*C` around it or not. This is a |
980 | # whole bunch of messy heuristics to try to avoid overquoting, originally from |
981 | # Barrie Slaymaker. This largely duplicates similar code in Pod::Text. |
982 | sub quote_literal { |
983 | my $self = shift; |
984 | local $_ = shift; |
985 | |
986 | # A regex that matches the portion of a variable reference that's the |
987 | # array or hash index, separated out just because we want to use it in |
988 | # several places in the following regex. |
989 | my $index = '(?: \[.*\] | \{.*\} )?'; |
990 | |
991 | # Check for things that we don't want to quote, and if we find any of |
992 | # them, return the string with just a font change and no quoting. |
993 | m{ |
994 | ^\s* |
995 | (?: |
996 | ( [\'\`\"] ) .* \1 # already quoted |
997 | | \` .* \' # `quoted' |
998 | | \$+ [\#^]? \S $index # special ($^Foo, $") |
999 | | [\$\@%&*]+ \#? [:\'\w]+ $index # plain var or func |
1000 | | [\$\@%&*]* [:\'\w]+ (?: -> )? \(\s*[^\s,]\s*\) # 0/1-arg func call |
1001 | | [+-]? [\d.]+ (?: [eE] [+-]? \d+ )? # a number |
1002 | | 0x [a-fA-F\d]+ # a hex constant |
1003 | ) |
1004 | \s*\z |
1005 | }xo && return '\f(FS' . $_ . '\f(FE'; |
1006 | |
1007 | # If we didn't return, go ahead and quote the text. |
1008 | return '\f(FS\*(C`' . $_ . "\\*(C'\\f(FE"; |
1009 | } |
1010 | |
9741dab0 |
1011 | |
3c014959 |
1012 | ############################################################################## |
9741dab0 |
1013 | # Output formatting |
3c014959 |
1014 | ############################################################################## |
9741dab0 |
1015 | |
1016 | # Make vertical whitespace. |
1017 | sub makespace { |
1018 | my $self = shift; |
5cdeb5a2 |
1019 | $self->output (".PD\n") if ($$self{ITEMS} > 1); |
1020 | $$self{ITEMS} = 0; |
1021 | $self->output ($$self{INDENT} > 0 ? ".Sp\n" : ".PP\n") |
1022 | if $$self{NEEDSPACE}; |
9741dab0 |
1023 | } |
1024 | |
3c014959 |
1025 | # Output any pending index entries, and optionally an index entry given as an |
1026 | # argument. Support multiple index entries in X<> separated by slashes, and |
1027 | # strip special escapes from index entries. |
9741dab0 |
1028 | sub outindex { |
1029 | my ($self, $section, $index) = @_; |
1030 | my @entries = map { split m%\s*/\s*% } @{ $$self{INDEX} }; |
1031 | return unless ($section || @entries); |
1032 | $$self{INDEX} = []; |
b616daaf |
1033 | my @output; |
9741dab0 |
1034 | if (@entries) { |
b616daaf |
1035 | push (@output, [ 'Xref', join (' ', @entries) ]); |
9741dab0 |
1036 | } |
1037 | if ($section) { |
9741dab0 |
1038 | $index =~ s/\\-/-/g; |
1039 | $index =~ s/\\(?:s-?\d|.\(..|.)//g; |
b616daaf |
1040 | push (@output, [ $section, $index ]); |
1041 | } |
1042 | for (@output) { |
1043 | my ($type, $entry) = @$_; |
1044 | $entry =~ s/\"/\"\"/g; |
1045 | $self->output (".IX $type " . '"' . $entry . '"' . "\n"); |
9741dab0 |
1046 | } |
9741dab0 |
1047 | } |
1048 | |
1049 | # Output text to the output device. |
1050 | sub output { print { $_[0]->output_handle } $_[1] } |
1051 | |
50a3fd2a |
1052 | # Given a command and a single argument that may or may not contain double |
1053 | # quotes, handle double-quote formatting for it. If there are no double |
1054 | # quotes, just return the command followed by the argument in double quotes. |
1055 | # If there are double quotes, use an if statement to test for nroff, and for |
1056 | # nroff output the command followed by the argument in double quotes with |
1057 | # embedded double quotes doubled. For other formatters, remap paired double |
73849855 |
1058 | # quotes to LQUOTE and RQUOTE. |
50a3fd2a |
1059 | sub switchquotes { |
1060 | my $self = shift; |
1061 | my $command = shift; |
1062 | local $_ = shift; |
1063 | my $extra = shift; |
1064 | s/\\\*\([LR]\"/\"/g; |
1065 | |
1066 | # We also have to deal with \*C` and \*C', which are used to add the |
1067 | # quotes around C<> text, since they may expand to " and if they do this |
b616daaf |
1068 | # confuses the .SH macros and the like no end. Expand them ourselves. |
1069 | # Also separate troff from nroff if there are any fixed-width fonts in use |
1070 | # to work around problems with Solaris nroff. |
50a3fd2a |
1071 | my $c_is_quote = ($$self{LQUOTE} =~ /\"/) || ($$self{RQUOTE} =~ /\"/); |
b616daaf |
1072 | my $fixedpat = join ('|', @{ $$self{FONTS} }{'100', '101', '110', '111'}); |
1073 | $fixedpat =~ s/\\/\\\\/g; |
1074 | $fixedpat =~ s/\(/\\\(/g; |
1075 | if (/\"/ || /$fixedpat/) { |
50a3fd2a |
1076 | s/\"/\"\"/g; |
3c014959 |
1077 | my $nroff = $_; |
50a3fd2a |
1078 | my $troff = $_; |
1079 | $troff =~ s/\"\"([^\"]*)\"\"/\`\`$1\'\'/g; |
3c014959 |
1080 | if ($c_is_quote && /\\\*\(C[\'\`]/) { |
1081 | $nroff =~ s/\\\*\(C\`/$$self{LQUOTE}/g; |
1082 | $nroff =~ s/\\\*\(C\'/$$self{RQUOTE}/g; |
1083 | $troff =~ s/\\\*\(C[\'\`]//g; |
1084 | } |
1085 | $nroff = qq("$nroff") . ($extra ? " $extra" : ''); |
50a3fd2a |
1086 | $troff = qq("$troff") . ($extra ? " $extra" : ''); |
3c014959 |
1087 | |
1088 | # Work around the Solaris nroff bug where \f(CW\fP leaves the font set |
1089 | # to Roman rather than the actual previous font when used in headings. |
1090 | # troff output may still be broken, but at least we can fix nroff by |
b616daaf |
1091 | # just switching the font changes to the non-fixed versions. |
1092 | $nroff =~ s/\Q$$self{FONTS}{100}\E(.*)\\f[PR]/$1/g; |
1093 | $nroff =~ s/\Q$$self{FONTS}{101}\E(.*)\\f([PR])/\\fI$1\\f$2/g; |
1094 | $nroff =~ s/\Q$$self{FONTS}{110}\E(.*)\\f([PR])/\\fB$1\\f$2/g; |
1095 | $nroff =~ s/\Q$$self{FONTS}{111}\E(.*)\\f([PR])/\\f\(BI$1\\f$2/g; |
3c014959 |
1096 | |
59548eca |
1097 | # Now finally output the command. Only bother with .ie if the nroff |
3c014959 |
1098 | # and troff output isn't the same. |
1099 | if ($nroff ne $troff) { |
59548eca |
1100 | return ".ie n $command $nroff\n.el $command $troff\n"; |
3c014959 |
1101 | } else { |
1102 | return "$command $nroff\n"; |
1103 | } |
50a3fd2a |
1104 | } else { |
1105 | $_ = qq("$_") . ($extra ? " $extra" : ''); |
1106 | return "$command $_\n"; |
1107 | } |
1108 | } |
1109 | |
9741dab0 |
1110 | __END__ |
1111 | |
3c014959 |
1112 | ############################################################################## |
9741dab0 |
1113 | # Documentation |
3c014959 |
1114 | ############################################################################## |
9741dab0 |
1115 | |
1116 | =head1 NAME |
1117 | |
1118 | Pod::Man - Convert POD data to formatted *roff input |
1119 | |
1120 | =head1 SYNOPSIS |
1121 | |
1122 | use Pod::Man; |
1123 | my $parser = Pod::Man->new (release => $VERSION, section => 8); |
1124 | |
1125 | # Read POD from STDIN and write to STDOUT. |
1126 | $parser->parse_from_filehandle; |
1127 | |
1128 | # Read POD from file.pod and write to file.1. |
1129 | $parser->parse_from_file ('file.pod', 'file.1'); |
1130 | |
1131 | =head1 DESCRIPTION |
1132 | |
1133 | Pod::Man is a module to convert documentation in the POD format (the |
1134 | preferred language for documenting Perl) into *roff input using the man |
1135 | macro set. The resulting *roff code is suitable for display on a terminal |
bf202ccd |
1136 | using L<nroff(1)>, normally via L<man(1)>, or printing using L<troff(1)>. |
1137 | It is conventionally invoked using the driver script B<pod2man>, but it can |
1138 | also be used directly. |
9741dab0 |
1139 | |
1140 | As a derived class from Pod::Parser, Pod::Man supports the same methods and |
1141 | interfaces. See L<Pod::Parser> for all the details; briefly, one creates a |
b4558dc4 |
1142 | new parser with C<< Pod::Man->new() >> and then calls either |
9741dab0 |
1143 | parse_from_filehandle() or parse_from_file(). |
1144 | |
1145 | new() can take options, in the form of key/value pairs that control the |
1146 | behavior of the parser. See below for details. |
1147 | |
1148 | If no options are given, Pod::Man uses the name of the input file with any |
1149 | trailing C<.pod>, C<.pm>, or C<.pl> stripped as the man page title, to |
1150 | section 1 unless the file ended in C<.pm> in which case it defaults to |
1151 | section 3, to a centered title of "User Contributed Perl Documentation", to |
1152 | a centered footer of the Perl version it is run with, and to a left-hand |
1153 | footer of the modification date of its input (or the current date if given |
1154 | STDIN for input). |
1155 | |
1156 | Pod::Man assumes that your *roff formatters have a fixed-width font named |
1157 | CW. If yours is called something else (like CR), use the C<fixed> option to |
1158 | specify it. This generally only matters for troff output for printing. |
1159 | Similarly, you can set the fonts used for bold, italic, and bold italic |
1160 | fixed-width output. |
1161 | |
1162 | Besides the obvious pod conversions, Pod::Man also takes care of formatting |
bf202ccd |
1163 | func(), func(3), and simple variable references like $foo or @bar so you |
9741dab0 |
1164 | don't have to use code escapes for them; complex expressions like |
1165 | C<$fred{'stuff'}> will still need to be escaped, though. It also translates |
1166 | dashes that aren't used as hyphens into en dashes, makes long dashes--like |
b4558dc4 |
1167 | this--into proper em dashes, fixes "paired quotes," makes C++ look right, |
1168 | puts a little space between double underbars, makes ALLCAPS a teeny bit |
1169 | smaller in B<troff>, and escapes stuff that *roff treats as special so that |
1170 | you don't have to. |
9741dab0 |
1171 | |
1172 | The recognized options to new() are as follows. All options take a single |
1173 | argument. |
1174 | |
1175 | =over 4 |
1176 | |
1177 | =item center |
1178 | |
1179 | Sets the centered page header to use instead of "User Contributed Perl |
1180 | Documentation". |
1181 | |
1182 | =item date |
1183 | |
1184 | Sets the left-hand footer. By default, the modification date of the input |
1185 | file will be used, or the current date if stat() can't find that file (the |
1186 | case if the input is from STDIN), and the date will be formatted as |
1187 | YYYY-MM-DD. |
1188 | |
1189 | =item fixed |
1190 | |
1191 | The fixed-width font to use for vertabim text and code. Defaults to CW. |
bf202ccd |
1192 | Some systems may want CR instead. Only matters for B<troff> output. |
9741dab0 |
1193 | |
1194 | =item fixedbold |
1195 | |
1196 | Bold version of the fixed-width font. Defaults to CB. Only matters for |
bf202ccd |
1197 | B<troff> output. |
9741dab0 |
1198 | |
1199 | =item fixeditalic |
1200 | |
1201 | Italic version of the fixed-width font (actually, something of a misnomer, |
1202 | since most fixed-width fonts only have an oblique version, not an italic |
bf202ccd |
1203 | version). Defaults to CI. Only matters for B<troff> output. |
9741dab0 |
1204 | |
1205 | =item fixedbolditalic |
1206 | |
1207 | Bold italic (probably actually oblique) version of the fixed-width font. |
1208 | Pod::Man doesn't assume you have this, and defaults to CB. Some systems |
bf202ccd |
1209 | (such as Solaris) have this font available as CX. Only matters for B<troff> |
9741dab0 |
1210 | output. |
1211 | |
bf202ccd |
1212 | =item name |
1213 | |
1214 | Set the name of the manual page. Without this option, the manual name is |
1215 | set to the uppercased base name of the file being converted unless the |
1216 | manual section is 3, in which case the path is parsed to see if it is a Perl |
1217 | module path. If it is, a path like C<.../lib/Pod/Man.pm> is converted into |
1218 | a name like C<Pod::Man>. This option, if given, overrides any automatic |
1219 | determination of the name. |
1220 | |
ab1f1d91 |
1221 | =item quotes |
1222 | |
1223 | Sets the quote marks used to surround CE<lt>> text. If the value is a |
1224 | single character, it is used as both the left and right quote; if it is two |
1225 | characters, the first character is used as the left quote and the second as |
1226 | the right quoted; and if it is four characters, the first two are used as |
1227 | the left quote and the second two as the right quote. |
1228 | |
1229 | This may also be set to the special value C<none>, in which case no quote |
1230 | marks are added around CE<lt>> text (but the font is still changed for troff |
1231 | output). |
1232 | |
9741dab0 |
1233 | =item release |
1234 | |
1235 | Set the centered footer. By default, this is the version of Perl you run |
bf202ccd |
1236 | Pod::Man under. Note that some system an macro sets assume that the |
9741dab0 |
1237 | centered footer will be a modification date and will prepend something like |
1238 | "Last modified: "; if this is the case, you may want to set C<release> to |
1239 | the last modified date and C<date> to the version number. |
1240 | |
1241 | =item section |
1242 | |
1243 | Set the section for the C<.TH> macro. The standard section numbering |
1244 | convention is to use 1 for user commands, 2 for system calls, 3 for |
1245 | functions, 4 for devices, 5 for file formats, 6 for games, 7 for |
1246 | miscellaneous information, and 8 for administrator commands. There is a lot |
1247 | of variation here, however; some systems (like Solaris) use 4 for file |
1248 | formats, 5 for miscellaneous information, and 7 for devices. Still others |
1249 | use 1m instead of 8, or some mix of both. About the only section numbers |
1250 | that are reliably consistent are 1, 2, and 3. |
1251 | |
1252 | By default, section 1 will be used unless the file ends in .pm in which case |
1253 | section 3 will be selected. |
1254 | |
1255 | =back |
1256 | |
1257 | The standard Pod::Parser method parse_from_filehandle() takes up to two |
1258 | arguments, the first being the file handle to read POD from and the second |
1259 | being the file handle to write the formatted output to. The first defaults |
1260 | to STDIN if not given, and the second defaults to STDOUT. The method |
1261 | parse_from_file() is almost identical, except that its two arguments are the |
1262 | input and output disk files instead. See L<Pod::Parser> for the specific |
1263 | details. |
1264 | |
1265 | =head1 DIAGNOSTICS |
1266 | |
1267 | =over 4 |
1268 | |
ab1f1d91 |
1269 | =item roff font should be 1 or 2 chars, not "%s" |
9741dab0 |
1270 | |
1271 | (F) You specified a *roff font (using C<fixed>, C<fixedbold>, etc.) that |
1272 | wasn't either one or two characters. Pod::Man doesn't support *roff fonts |
1273 | longer than two characters, although some *roff extensions do (the canonical |
bf202ccd |
1274 | versions of B<nroff> and B<troff> don't either). |
9741dab0 |
1275 | |
1276 | =item Invalid link %s |
1277 | |
b84d8b9e |
1278 | (W) The POD source contained a C<LE<lt>E<gt>> formatting code that |
1279 | Pod::Man was unable to parse. You should never see this error message; it |
1280 | probably indicates a bug in Pod::Man. |
9741dab0 |
1281 | |
ab1f1d91 |
1282 | =item Invalid quote specification "%s" |
1283 | |
1284 | (F) The quote specification given (the quotes option to the constructor) was |
1285 | invalid. A quote specification must be one, two, or four characters long. |
1286 | |
1287 | =item %s:%d: Unknown command paragraph "%s". |
1288 | |
1289 | (W) The POD source contained a non-standard command paragraph (something of |
1290 | the form C<=command args>) that Pod::Man didn't know about. It was ignored. |
1291 | |
aa212ad6 |
1292 | =item %s:%d: Unknown escape EE<lt>%sE<gt> |
9741dab0 |
1293 | |
1294 | (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Man didn't |
1295 | know about. C<EE<lt>%sE<gt>> was printed verbatim in the output. |
1296 | |
b84d8b9e |
1297 | =item %s:%d: Unknown formatting code %s |
9741dab0 |
1298 | |
b84d8b9e |
1299 | (W) The POD source contained a non-standard formatting code (something of |
9741dab0 |
1300 | the form C<XE<lt>E<gt>>) that Pod::Man didn't know about. It was ignored. |
1301 | |
aa212ad6 |
1302 | =item %s:%d: Unmatched =back |
9741dab0 |
1303 | |
1304 | (W) Pod::Man encountered a C<=back> command that didn't correspond to an |
1305 | C<=over> command. |
1306 | |
1307 | =back |
1308 | |
1309 | =head1 BUGS |
1310 | |
b4558dc4 |
1311 | Eight-bit input data isn't handled at all well at present. The correct |
1312 | approach would be to map EE<lt>E<gt> escapes to the appropriate UTF-8 |
1313 | characters and then do a translation pass on the output according to the |
1314 | user-specified output character set. Unfortunately, we can't send eight-bit |
1315 | data directly to the output unless the user says this is okay, since some |
1316 | vendor *roff implementations can't handle eight-bit data. If the *roff |
1317 | implementation can, however, that's far superior to the current hacked |
1318 | characters that only work under troff. |
1319 | |
1320 | There is currently no way to turn off the guesswork that tries to format |
1321 | unmarked text appropriately, and sometimes it isn't wanted (particularly |
1322 | when using POD to document something other than Perl). |
9741dab0 |
1323 | |
1324 | The NAME section should be recognized specially and index entries emitted |
1325 | for everything in that section. This would have to be deferred until the |
1326 | next section, since extraneous things in NAME tends to confuse various man |
1327 | page processors. |
1328 | |
9741dab0 |
1329 | Pod::Man doesn't handle font names longer than two characters. Neither do |
bf202ccd |
1330 | most B<troff> implementations, but GNU troff does as an extension. It would |
9741dab0 |
1331 | be nice to support as an option for those who want to use it. |
1332 | |
1333 | The preamble added to each output file is rather verbose, and most of it is |
1334 | only necessary in the presence of EE<lt>E<gt> escapes for non-ASCII |
1335 | characters. It would ideally be nice if all of those definitions were only |
1336 | output if needed, perhaps on the fly as the characters are used. |
1337 | |
9741dab0 |
1338 | Pod::Man is excessively slow. |
1339 | |
b4558dc4 |
1340 | =head1 CAVEATS |
1341 | |
1342 | The handling of hyphens and em dashes is somewhat fragile, and one may get |
1343 | the wrong one under some circumstances. This should only matter for |
1344 | B<troff> output. |
1345 | |
1346 | When and whether to use small caps is somewhat tricky, and Pod::Man doesn't |
1347 | necessarily get it right. |
1348 | |
9741dab0 |
1349 | =head1 SEE ALSO |
1350 | |
bf202ccd |
1351 | L<Pod::Parser>, L<perlpod(1)>, L<pod2man(1)>, L<nroff(1)>, L<troff(1)>, |
1352 | L<man(1)>, L<man(7)> |
9741dab0 |
1353 | |
1354 | Ossanna, Joseph F., and Brian W. Kernighan. "Troff User's Manual," |
1355 | Computing Science Technical Report No. 54, AT&T Bell Laboratories. This is |
bf202ccd |
1356 | the best documentation of standard B<nroff> and B<troff>. At the time of |
1357 | this writing, it's available at |
1358 | L<http://www.cs.bell-labs.com/cm/cs/cstr.html>. |
9741dab0 |
1359 | |
bf202ccd |
1360 | The man page documenting the man macro set may be L<man(5)> instead of |
1361 | L<man(7)> on your system. Also, please see L<pod2man(1)> for extensive |
1362 | documentation on writing manual pages if you've not done it before and |
1363 | aren't familiar with the conventions. |
9741dab0 |
1364 | |
1365 | =head1 AUTHOR |
1366 | |
bf202ccd |
1367 | Russ Allbery <rra@stanford.edu>, based I<very> heavily on the original |
1368 | B<pod2man> by Tom Christiansen <tchrist@mox.perl.com>. |
9741dab0 |
1369 | |
3c014959 |
1370 | =head1 COPYRIGHT AND LICENSE |
1371 | |
1372 | Copyright 1999, 2000, 2001 by Russ Allbery <rra@stanford.edu>. |
1373 | |
1374 | This program is free software; you may redistribute it and/or modify it |
1375 | under the same terms as Perl itself. |
1376 | |
9741dab0 |
1377 | =cut |