4 use File::Basename qw(&basename &dirname);
7 # List explicitly here the variables you want Configure to
8 # generate. Metaconfig only looks for shell variables, so you
9 # have to mention them as if they were shell variables, not
10 # %Config entries. Thus you write
12 # to ensure Configure will look for $Config{startperl}.
14 # This forces PL files to create target in same directory as PL file.
15 # This is so that make depend always knows where to find PL derivatives.
18 $file = basename($0, '.PL');
19 $file .= '.com' if $^O eq 'VMS';
21 open OUT,">$file" or die "Can't create $file: $!";
23 print "Extracting $file (with variable substitutions)\n";
25 # In this section, perl variables will be expanded during extraction.
26 # You can use $Config{...} to use Configure variables.
28 print OUT <<"!GROK!THIS!";
30 eval 'exec $Config{perlpath} -S \$0 \${1+"\$@"}'
31 if \$running_under_some_shell;
34 # In the following, perl variables are not expanded during extraction.
36 print OUT <<'!NO!SUBS!';
38 # pod2man -- Convert POD data to formatted *roff input.
40 # Copyright 1999, 2000, 2001, 2004, 2006, 2008 Russ Allbery <rra@stanford.edu>
42 # This program is free software; you may redistribute it and/or modify it
43 # under the same terms as Perl itself.
47 use Getopt::Long qw(GetOptions);
49 use Pod::Usage qw(pod2usage);
53 # Silence -w warnings.
54 use vars qw($running_under_some_shell);
56 # Insert -- into @ARGV before any single dash argument to hide it from
57 # Getopt::Long; we want to interpret it as meaning stdin.
59 @ARGV = map { $_ eq '-' && !$stdin++ ? ('--', $_) : $_ } @ARGV;
61 # Parse our options, trying to retain backward compatibility with pod2man but
62 # allowing short forms as well. --lax is currently ignored.
64 $options{errors} = 'pod';
65 Getopt::Long::config ('bundling_override');
66 GetOptions (\%options, 'center|c=s', 'date|d=s', 'fixed=s', 'fixedbold=s',
67 'fixeditalic=s', 'fixedbolditalic=s', 'help|h', 'lax|l',
68 'name|n=s', 'official|o', 'quotes|q=s', 'release|r:s',
69 'section|s=s', 'stderr', 'verbose|v', 'utf8|u') or exit 1;
70 pod2usage (0) if $options{help};
72 # Official sets --center, but don't override things explicitly set.
73 if ($options{official} && !defined $options{center}) {
74 $options{center} = 'Perl Programmers Reference Guide';
77 # Verbose is only our flag, not a Pod::Man flag.
78 my $verbose = $options{verbose};
79 delete $options{verbose};
81 # This isn't a valid Pod::Man option and is only accepted for backward
85 # Initialize and run the formatter, pulling a pair of input and output off at
87 my $parser = Pod::Man->new (%options);
90 @files = splice (@ARGV, 0, 2);
91 print " $files[1]\n" if $verbose;
92 $parser->parse_from_file (@files);
99 pod2man - Convert POD data to formatted *roff input
102 en em --stderr stderr --utf8 UTF-8 overdo markup MT-LEVEL Allbery Solaris
103 URL troff troff-specific formatters uppercased Christiansen
107 pod2man [B<--center>=I<string>] [B<--date>=I<string>]
108 [B<--fixed>=I<font>] [B<--fixedbold>=I<font>] [B<--fixeditalic>=I<font>]
109 [B<--fixedbolditalic>=I<font>] [B<--name>=I<name>] [B<--official>]
110 [B<--quotes>=I<quotes>] [B<--release>[=I<version>]]
111 [B<--section>=I<manext>] [B<--stderr>] [B<--utf8>] [B<--verbose>]
112 [I<input> [I<output>] ...]
118 B<pod2man> is a front-end for Pod::Man, using it to generate *roff input
119 from POD source. The resulting *roff code is suitable for display on a
120 terminal using nroff(1), normally via man(1), or printing using troff(1).
122 I<input> is the file to read for POD source (the POD can be embedded in
123 code). If I<input> isn't given, it defaults to C<STDIN>. I<output>, if
124 given, is the file to which to write the formatted output. If I<output>
125 isn't given, the formatted output is written to C<STDOUT>. Several POD
126 files can be processed in the same B<pod2man> invocation (saving module
127 load and compile times) by providing multiple pairs of I<input> and
128 I<output> files on the command line.
130 B<--section>, B<--release>, B<--center>, B<--date>, and B<--official> can
131 be used to set the headers and footers to use; if not given, Pod::Man will
132 assume various defaults. See below or L<Pod::Man> for details.
134 B<pod2man> assumes that your *roff formatters have a fixed-width font
135 named C<CW>. If yours is called something else (like C<CR>), use
136 B<--fixed> to specify it. This generally only matters for troff output
137 for printing. Similarly, you can set the fonts used for bold, italic, and
138 bold italic fixed-width output.
140 Besides the obvious pod conversions, Pod::Man, and therefore pod2man also
141 takes care of formatting func(), func(n), and simple variable references
142 like $foo or @bar so you don't have to use code escapes for them; complex
143 expressions like C<$fred{'stuff'}> will still need to be escaped, though.
144 It also translates dashes that aren't used as hyphens into en dashes, makes
145 long dashes--like this--into proper em dashes, fixes "paired quotes," and
146 takes care of several other troff-specific tweaks. See L<Pod::Man> for
147 complete information.
153 =item B<-c> I<string>, B<--center>=I<string>
155 Sets the centered page header to I<string>. The default is "User
156 Contributed Perl Documentation", but also see B<--official> below.
158 =item B<-d> I<string>, B<--date>=I<string>
160 Set the left-hand footer string to this value. By default, the modification
161 date of the input file will be used, or the current date if input comes from
164 =item B<--fixed>=I<font>
166 The fixed-width font to use for verbatim text and code. Defaults to
167 C<CW>. Some systems may want C<CR> instead. Only matters for troff(1)
170 =item B<--fixedbold>=I<font>
172 Bold version of the fixed-width font. Defaults to C<CB>. Only matters
175 =item B<--fixeditalic>=I<font>
177 Italic version of the fixed-width font (actually, something of a misnomer,
178 since most fixed-width fonts only have an oblique version, not an italic
179 version). Defaults to C<CI>. Only matters for troff(1) output.
181 =item B<--fixedbolditalic>=I<font>
183 Bold italic (probably actually oblique) version of the fixed-width font.
184 Pod::Man doesn't assume you have this, and defaults to C<CB>. Some
185 systems (such as Solaris) have this font available as C<CX>. Only matters
188 =item B<-h>, B<--help>
190 Print out usage information.
192 =item B<-l>, B<--lax>
194 No longer used. B<pod2man> used to check its input for validity as a
195 manual page, but this should now be done by L<podchecker(1)> instead.
196 Accepted for backward compatibility; this option no longer does anything.
198 =item B<-n> I<name>, B<--name>=I<name>
200 Set the name of the manual page to I<name>. Without this option, the manual
201 name is set to the uppercased base name of the file being converted unless
202 the manual section is 3, in which case the path is parsed to see if it is a
203 Perl module path. If it is, a path like C<.../lib/Pod/Man.pm> is converted
204 into a name like C<Pod::Man>. This option, if given, overrides any
205 automatic determination of the name.
207 Note that this option is probably not useful when converting multiple POD
208 files at once. The convention for Unix man pages for commands is for the
209 man page title to be in all-uppercase even if the command isn't.
211 =item B<-o>, B<--official>
213 Set the default header to indicate that this page is part of the standard
214 Perl release, if B<--center> is not also given.
216 =item B<-q> I<quotes>, B<--quotes>=I<quotes>
218 Sets the quote marks used to surround CE<lt>> text to I<quotes>. If
219 I<quotes> is a single character, it is used as both the left and right
220 quote; if I<quotes> is two characters, the first character is used as the
221 left quote and the second as the right quoted; and if I<quotes> is four
222 characters, the first two are used as the left quote and the second two as
225 I<quotes> may also be set to the special value C<none>, in which case no
226 quote marks are added around CE<lt>> text (but the font is still changed for
229 =item B<-r>, B<--release>
231 Set the centered footer. By default, this is the version of Perl you run
232 B<pod2man> under. Note that some system an macro sets assume that the
233 centered footer will be a modification date and will prepend something like
234 "Last modified: "; if this is the case, you may want to set B<--release> to
235 the last modified date and B<--date> to the version number.
237 =item B<-s>, B<--section>
239 Set the section for the C<.TH> macro. The standard section numbering
240 convention is to use 1 for user commands, 2 for system calls, 3 for
241 functions, 4 for devices, 5 for file formats, 6 for games, 7 for
242 miscellaneous information, and 8 for administrator commands. There is a lot
243 of variation here, however; some systems (like Solaris) use 4 for file
244 formats, 5 for miscellaneous information, and 7 for devices. Still others
245 use 1m instead of 8, or some mix of both. About the only section numbers
246 that are reliably consistent are 1, 2, and 3.
248 By default, section 1 will be used unless the file ends in C<.pm>, in
249 which case section 3 will be selected.
253 By default, B<pod2man> puts any errors detected in the POD input in a POD
254 ERRORS section in the output manual page. If B<--stderr> is given, errors
255 are sent to standard error instead and the POD ERRORS section is
258 =item B<-u>, B<--utf8>
260 By default, B<pod2man> produces the most conservative possible *roff
261 output to try to ensure that it will work with as many different *roff
262 implementations as possible. Many *roff implementations cannot handle
263 non-ASCII characters, so this means all non-ASCII characters are converted
264 either to a *roff escape sequence that tries to create a properly accented
265 character (at least for troff output) or to C<X>.
267 This option says to instead output literal UTF-8 characters. If your
268 *roff implementation can handle it, this is the best output format to use
269 and avoids corruption of documents containing non-ASCII characters.
270 However, be warned that *roff source with literal UTF-8 characters is not
271 supported by many implementations and may even result in segfaults and
274 Be aware that, when using this option, the input encoding of your POD
275 source must be properly declared unless it is US-ASCII or Latin-1. POD
276 input without an C<=encoding> command will be assumed to be in Latin-1,
277 and if it's actually in UTF-8, the output will be double-encoded. See
278 L<perlpod(1)> for more information on the C<=encoding> command.
280 =item B<-v>, B<--verbose>
282 Print out the name of each output file as it is being generated.
288 If B<pod2man> fails with errors, see L<Pod::Man> and L<Pod::Simple> for
289 information about what those errors might mean.
293 pod2man program > program.1
294 pod2man SomeModule.pm /usr/perl/man/man3/SomeModule.3
295 pod2man --section=7 note.pod > note.7
297 If you would like to print out a lot of man page continuously, you probably
298 want to set the C and D registers to set contiguous page numbering and
299 even/odd paging, at least on some versions of man(7).
301 troff -man -rC1 -rD1 perl.1 perldata.1 perlsyn.1 ...
303 To get index entries on C<STDERR>, turn on the F register, as in:
305 troff -man -rF1 perl.1
307 The indexing merely outputs messages via C<.tm> for each major page,
308 section, subsection, item, and any C<XE<lt>E<gt>> directives. See
309 L<Pod::Man> for more details.
313 Lots of this documentation is duplicated from L<Pod::Man>.
317 For those not sure of the proper layout of a man page, here are some notes
318 on writing a proper man page.
320 The name of the program being documented is conventionally written in bold
321 (using BE<lt>E<gt>) wherever it occurs, as are all program options.
322 Arguments should be written in italics (IE<lt>E<gt>). Functions are
323 traditionally written in italics; if you write a function as function(),
324 Pod::Man will take care of this for you. Literal code or commands should
325 be in CE<lt>E<gt>. References to other man pages should be in the form
326 C<manpage(section)>, and Pod::Man will automatically format those
327 appropriately. As an exception, it's traditional not to use this form when
328 referring to module documentation; use C<LE<lt>Module::NameE<gt>> instead.
330 References to other programs or functions are normally in the form of man
331 page references so that cross-referencing tools can provide the user with
332 links and the like. It's possible to overdo this, though, so be careful not
333 to clutter your documentation with too much markup.
335 The major headers should be set out using a C<=head1> directive, and are
336 historically written in the rather startling ALL UPPER CASE format, although
337 this is not mandatory. Minor headers may be included using C<=head2>, and
338 are typically in mixed case.
340 The standard sections of a manual page are:
346 Mandatory section; should be a comma-separated list of programs or functions
347 documented by this POD page, such as:
349 foo, bar - programs to do something
351 Manual page indexers are often extremely picky about the format of this
352 section, so don't put anything in it except this line. A single dash, and
353 only a single dash, should separate the list of programs or functions from
354 the description. Do not use any markup such as CE<lt>E<gt> or
355 BE<lt>E<gt>. Functions should not be qualified with C<()> or the like.
356 The description should ideally fit on a single line, even if a man program
357 replaces the dash with a few tabs.
361 A short usage summary for programs and functions. This section is mandatory
366 Extended description and discussion of the program or functions, or the body
367 of the documentation for man pages that document something else. If
368 particularly long, it's a good idea to break this up into subsections
369 C<=head2> directives like:
373 =head2 Advanced Features
375 =head2 Writing Configuration Files
377 or whatever is appropriate for your documentation.
381 Detailed description of each of the command-line options taken by the
382 program. This should be separate from the description for the use of things
383 like L<Pod::Usage|Pod::Usage>. This is normally presented as a list, with
384 each option as a separate C<=item>. The specific option string should be
385 enclosed in BE<lt>E<gt>. Any values that the option takes should be
386 enclosed in IE<lt>E<gt>. For example, the section for the option
387 B<--section>=I<manext> would be introduced with:
389 =item B<--section>=I<manext>
391 Synonymous options (like both the short and long forms) are separated by a
392 comma and a space on the same C<=item> line, or optionally listed as their
393 own item with a reference to the canonical name. For example, since
394 B<--section> can also be written as B<-s>, the above would be:
396 =item B<-s> I<manext>, B<--section>=I<manext>
398 (Writing the short option first is arguably easier to read, since the long
399 option is long enough to draw the eye to it anyway and the short option can
400 otherwise get lost in visual noise.)
404 What the program or function returns, if successful. This section can be
405 omitted for programs whose precise exit codes aren't important, provided
406 they return 0 on success as is standard. It should always be present for
411 Exceptions, error return codes, exit statuses, and errno settings.
412 Typically used for function documentation; program documentation uses
413 DIAGNOSTICS instead. The general rule of thumb is that errors printed to
414 C<STDOUT> or C<STDERR> and intended for the end user are documented in
415 DIAGNOSTICS while errors passed internal to the calling program and
416 intended for other programmers are documented in ERRORS. When documenting
417 a function that sets errno, a full list of the possible errno values
418 should be given here.
422 All possible messages the program can print out--and what they mean. You
423 may wish to follow the same documentation style as the Perl documentation;
424 see perldiag(1) for more details (and look at the POD source as well).
426 If applicable, please include details on what the user should do to correct
427 the error; documenting an error as indicating "the input buffer is too
428 small" without telling the user how to increase the size of the input buffer
429 (or at least telling them that it isn't possible) aren't very useful.
433 Give some example uses of the program or function. Don't skimp; users often
434 find this the most useful part of the documentation. The examples are
435 generally given as verbatim paragraphs.
437 Don't just present an example without explaining what it does. Adding a
438 short paragraph saying what the example will do can increase the value of
439 the example immensely.
443 Environment variables that the program cares about, normally presented as a
444 list using C<=over>, C<=item>, and C<=back>. For example:
450 Used to determine the user's home directory. F<.foorc> in this
451 directory is read for configuration details, if it exists.
455 Since environment variables are normally in all uppercase, no additional
456 special formatting is generally needed; they're glaring enough as it is.
460 All files used by the program or function, normally presented as a list, and
461 what it uses them for. File names should be enclosed in FE<lt>E<gt>. It's
462 particularly important to document files that will be potentially modified.
466 Things to take special care with, sometimes called WARNINGS.
470 Things that are broken or just don't work quite right.
474 Bugs you don't plan to fix. :-)
478 Miscellaneous commentary.
482 Who wrote it (use AUTHORS for multiple people). Including your current
483 e-mail address (or some e-mail address to which bug reports should be sent)
484 so that users have a way of contacting you is a good idea. Remember that
485 program documentation tends to roam the wild for far longer than you expect
486 and pick an e-mail address that's likely to last if possible.
490 Programs derived from other sources sometimes have this, or you might keep
491 a modification log here. If the log gets overly long or detailed,
492 consider maintaining it in a separate file, though.
494 =item COPYRIGHT AND LICENSE
498 Copyright YEAR(s) by YOUR NAME(s)
500 (No, (C) is not needed. No, "all rights reserved" is not needed.)
502 For licensing the easiest way is to use the same licensing as Perl itself:
504 This library is free software; you may redistribute it and/or modify
505 it under the same terms as Perl itself.
507 This makes it easy for people to use your module with Perl. Note that
508 this licensing is neither an endorsement or a requirement, you are of
509 course free to choose any licensing.
513 Other man pages to check out, like man(1), man(7), makewhatis(8), or
514 catman(8). Normally a simple list of man pages separated by commas, or a
515 paragraph giving the name of a reference work. Man page references, if they
516 use the standard C<name(section)> form, don't have to be enclosed in
517 LE<lt>E<gt> (although it's recommended), but other things in this section
518 probably should be when appropriate.
520 If the package has a mailing list, include a URL or subscription
523 If the package has a web site, include a URL here.
527 In addition, some systems use CONFORMING TO to note conformance to relevant
528 standards and MT-LEVEL to note safeness for use in threaded programs or
529 signal handlers. These headings are primarily useful when documenting parts
530 of a C library. Documentation of object-oriented libraries or modules may
531 use CONSTRUCTORS and METHODS sections for detailed documentation of the
532 parts of the library and save the DESCRIPTION section for an overview; other
533 large modules may use FUNCTIONS for similar reasons. Some people use
534 OVERVIEW to summarize the description if it's quite long.
536 Section ordering varies, although NAME should I<always> be the first section
537 (you'll break some man page systems otherwise), and NAME, SYNOPSIS,
538 DESCRIPTION, and OPTIONS generally always occur first and in that order if
539 present. In general, SEE ALSO, AUTHOR, and similar material should be left
540 for last. Some systems also move WARNINGS and NOTES to last. The order
541 given above should be reasonable for most purposes.
543 Finally, as a general note, try not to use an excessive amount of markup.
544 As documented here and in L<Pod::Man>, you can safely leave Perl variables,
545 function names, man page references, and the like unadorned by markup and
546 the POD translators will figure it out for you. This makes it much easier
547 to later edit the documentation. Note that many existing translators
548 (including this one currently) will do the wrong thing with e-mail addresses
549 when wrapped in LE<lt>E<gt>, so don't do that.
551 For additional information that may be more accurate for your specific
552 system, see either L<man(5)> or L<man(7)> depending on your system manual
553 section numbering conventions.
557 L<Pod::Man>, L<Pod::Simple>, L<man(1)>, L<nroff(1)>, L<perlpod(1)>,
558 L<podchecker(1)>, L<troff(1)>, L<man(7)>
560 The man page documenting the an macro set may be L<man(5)> instead of
561 L<man(7)> on your system.
563 The current version of this script is always available from its web site at
564 L<http://www.eyrie.org/~eagle/software/podlators/>. It is also part of the
565 Perl core distribution as of 5.6.0.
569 Russ Allbery <rra@stanford.edu>, based I<very> heavily on the original
570 B<pod2man> by Larry Wall and Tom Christiansen. Large portions of this
571 documentation, particularly the sections on the anatomy of a proper man
572 page, are taken from the B<pod2man> documentation by Tom.
574 =head1 COPYRIGHT AND LICENSE
576 Copyright 1999, 2000, 2001, 2004, 2006, 2008 Russ Allbery
579 This program is free software; you may redistribute it and/or modify it
580 under the same terms as Perl itself.
586 close OUT or die "Can't close $file: $!";
587 chmod 0755, $file or die "Can't reset permissions for $file: $!\n";
588 exec("$Config{'eunicefix'} $file") if $Config{'eunicefix'} ne ':';