This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
autodoc.pl: Special case macros whose name begins with Perl_
[perl5.git] / autodoc.pl
CommitLineData
94bdecf9 1#!/usr/bin/perl -w
6294c161
DM
2#
3# Unconditionally regenerate:
4#
5# pod/perlintern.pod
6# pod/perlapi.pod
7#
8# from information stored in
9#
10# embed.fnc
11# plus all the .c and .h files listed in MANIFEST
12#
13# Has an optional arg, which is the directory to chdir to before reading
14# MANIFEST and *.[ch].
15#
52a9d53b 16# This script is invoked as part of 'make all'
151c3fe5 17#
f554dfc5
MH
18# '=head1' are the only headings looked for. If the first non-blank line after
19# the heading begins with a word character, it is considered to be the first
20# line of documentation that applies to the heading itself. That is, it is
21# output immediately after the heading, before the first function, and not
22# indented. The next input line that is a pod directive terminates this
23# heading-level documentation.
94bdecf9 24
1fcde0e9
KW
25# The meanings of the flags fields in embed.fnc and the source code is
26# documented at the top of embed.fnc.
27
56a0c332 28use strict;
a64c954a 29
7882b24a
NC
30if (@ARGV) {
31 my $workdir = shift;
32 chdir $workdir
33 or die "Couldn't chdir to '$workdir': $!";
34}
3d7c117d
MB
35require './regen/regen_lib.pl';
36require './regen/embed_lib.pl';
7882b24a 37
7b1f0a98
KW
38my @specialized_docs = sort qw( perlguts
39 perlxs
40 perlxstut
b87d9527 41 perlclib
a1a5a9c8 42 warnings
678f21a2 43 perlapio
31b83e34 44 perlcall
0f292d69 45 perlfilter
6a6f8717 46 perlmroapi
1a2a04a3 47 config.h
7b1f0a98 48 );
1a2a04a3
KW
49sub name_in_pod($) {
50 my $name = shift;
51 return "F<$name>" if $name =~ /\./;
52 return "L<$name>";
53}
54my $other_places_api = join " ", map { name_in_pod($_) } sort @specialized_docs, 'perlintern';
55my $other_places_intern = join " ", map { name_in_pod($_) } sort @specialized_docs, 'perlapi';
7b1f0a98 56
1a2a04a3 57@specialized_docs = map { name_in_pod($_) } sort @specialized_docs;
b87d9527
KW
58$specialized_docs[-1] =~ s/^/and /;
59my $specialized_docs = join ", ", @specialized_docs;
60
94bdecf9 61#
346f75ff 62# See database of global and static function prototypes in embed.fnc
94bdecf9
JH
63# This is used to generate prototype headers under various configurations,
64# export symbols lists for different platforms, and macros to provide an
65# implicit interpreter context argument.
66#
67
6a235718 68my %docs;
5ce57792 69my %funcflags;
5ce57792 70my %missing;
94bdecf9
JH
71
72my $curheader = "Unknown section";
73
74sub autodoc ($$) { # parse a file and extract documentation info
75 my($fh,$file) = @_;
151c3fe5 76 my($in, $doc, $line, $header_doc);
f554dfc5
MH
77
78 # Count lines easier
79 my $get_next_line = sub { $line++; return <$fh> };
80
94bdecf9 81FUNC:
f554dfc5 82 while (defined($in = $get_next_line->())) {
94bdecf9
JH
83 if ($in=~ /^=head1 (.*)/) {
84 $curheader = $1;
151c3fe5 85
f554dfc5 86 # If the next non-space line begins with a word char, then it is
1c82f4a4 87 # the start of heading-level documentation.
20046047 88 if (defined($doc = $get_next_line->())) {
f554dfc5
MH
89 # Skip over empty lines
90 while ($doc =~ /^\s+$/) {
91 if (! defined($doc = $get_next_line->())) {
92 next FUNC;
93 }
94 }
95
151c3fe5
KW
96 if ($doc !~ /^\w/) {
97 $in = $doc;
98 redo FUNC;
99 }
100 $header_doc = $doc;
151c3fe5
KW
101
102 # Continue getting the heading-level documentation until read
103 # in any pod directive (or as a fail-safe, find a closing
104 # comment to this pod in a C language file
105HDR_DOC:
f554dfc5 106 while (defined($doc = $get_next_line->())) {
151c3fe5
KW
107 if ($doc =~ /^=\w/) {
108 $in = $doc;
109 redo FUNC;
110 }
151c3fe5
KW
111
112 if ($doc =~ m:^\s*\*/$:) {
113 warn "=cut missing? $file:$line:$doc";;
114 last HDR_DOC;
115 }
116 $header_doc .= $doc;
117 }
118 }
94bdecf9
JH
119 next FUNC;
120 }
20046047
KE
121 if ($in =~ /^=for\s+apidoc\s+(.*?)\s*\n/) {
122 my $proto_in_file = $1;
123 my $proto = $proto_in_file;
124 $proto = "||$proto" unless $proto =~ /\|/;
125 my($flags, $ret, $name, @args) = split /\s*\|\s*/, $proto;
256dda50
TC
126 $name or die <<EOS;
127Bad apidoc at $file line $.:
128 $in
129Expected:
130 =for apidoc flags|returntype|name|arg|arg|...
131 =for apidoc flags|returntype|name
132 =for apidoc name
133EOS
3dbfa774
KW
134 die "flag $1 is not legal (for function $name (from $file))"
135 if $flags =~ / ( [^AabCDdEefhiMmNnTOoPpRrSsUuWXx] ) /x;
6523e108
KW
136 next FUNC if $flags =~ /h/;
137
3dbfa774 138 die "'u' flag must also have 'm' flag' for $name" if $flags =~ /u/ && $flags !~ /m/;
0a60f600
KW
139 warn ("'$name' not \\w+ in '$proto_in_file' in $file")
140 if $flags !~ /N/ && $name !~ / ^ [_[:alpha:]] \w* $ /x;
20046047 141 my $docs = "";
94bdecf9 142DOC:
20046047 143 while (defined($doc = $get_next_line->())) {
72d4186d
KW
144
145 # Other pod commands are considered part of the current
146 # function's docs, so can have lists, etc.
147 last DOC if $doc =~ /^=(cut|for\s+apidoc|head)/;
20046047
KE
148 if ($doc =~ m:^\*/$:) {
149 warn "=cut missing? $file:$line:$doc";;
150 last DOC;
72d4186d 151 }
20046047
KE
152 $docs .= $doc;
153 }
154 $docs = "\n$docs" if $docs and $docs !~ /^\n/;
5ce57792 155
20046047 156 # If the entry is also in embed.fnc, it should be defined
8902d554 157 # completely there, but not here
20046047
KE
158 my $embed_docref = delete $funcflags{$name};
159 if ($embed_docref and %$embed_docref) {
8902d554
KW
160 warn "embed.fnc entry overrides redundant information in"
161 . " '$proto_in_file' in $file" if $flags || $ret || @args;
162 $flags = $embed_docref->{'flags'};
5514c4f1
KW
163 warn "embed.fnc entry '$name' missing 'd' flag"
164 unless $flags =~ /d/;
6523e108 165 next FUNC if $flags =~ /h/;
8902d554 166 $ret = $embed_docref->{'retval'};
20046047 167 @args = @{$embed_docref->{args}};
5514c4f1
KW
168 } elsif ($flags !~ /m/) { # Not in embed.fnc, is missing if not a
169 # macro
20046047
KE
170 $missing{$name} = $file;
171 }
5ce57792 172
8902d554 173 my $inline_where = $flags =~ /A/ ? 'api' : 'guts';
5ce57792 174
20046047 175 if (exists $docs{$inline_where}{$curheader}{$name}) {
7a6610ca
DM
176 warn "$0: duplicate API entry for '$name' in $inline_where/$curheader\n";
177 next;
178 }
20046047
KE
179 $docs{$inline_where}{$curheader}{$name}
180 = [$flags, $docs, $ret, $file, @args];
5ce57792 181
151c3fe5
KW
182 # Create a special entry with an empty-string name for the
183 # heading-level documentation.
20046047 184 if (defined $header_doc) {
151c3fe5
KW
185 $docs{$inline_where}{$curheader}{""} = $header_doc;
186 undef $header_doc;
187 }
188
20046047
KE
189 if (defined $doc) {
190 if ($doc =~ /^=(?:for|head)/) {
191 $in = $doc;
192 redo FUNC;
193 }
194 } else {
195 warn "$file:$line:$in";
196 }
197 }
94bdecf9
JH
198 }
199}
200
201sub docout ($$$) { # output the docs for one function
202 my($fh, $name, $docref) = @_;
203 my($flags, $docs, $ret, $file, @args) = @$docref;
d8c40edc 204 $name =~ s/\s*$//;
94bdecf9 205
d4e99c76
KW
206 if ($flags =~ /D/) {
207 $docs = "\n\nDEPRECATED! It is planned to remove this function from a
208future release of Perl. Do not use it for new code; remove it from
209existing code.\n\n$docs";
210 }
211 else {
58a428bb
KW
212 $docs = "\n\nNOTE: this function is experimental and may change or be
213removed without notice.\n\n$docs" if $flags =~ /x/;
d4e99c76 214 }
54c193ae
KW
215
216 # Is Perl_, but no #define foo # Perl_foo
217 my $p = $flags =~ /p/ && $flags =~ /o/ && $flags !~ /M/;
218
94bdecf9 219 $docs .= "NOTE: the perl_ form of this function is deprecated.\n\n"
20046047 220 if $flags =~ /O/;
54c193ae
KW
221 if ($p) {
222 $docs .= "NOTE: this function must be explicitly called as Perl_$name";
d7cc3209 223 $docs .= " with an aTHX_ parameter" if $flags !~ /T/;
54c193ae
KW
224 $docs .= ".\n\n"
225 }
94bdecf9 226
d8c40edc 227 print $fh "=item $name\nX<$name>\n$docs";
94bdecf9
JH
228
229 if ($flags =~ /U/) { # no usage
8b5ff177 230 warn("U and s flags are incompatible") if $flags =~ /s/;
20046047 231 # nothing
05ca4832 232 } else {
8b5ff177 233 if ($flags =~ /n/) { # no args
1fcde0e9
KW
234 warn("n flag without m") unless $flags =~ /m/;
235 warn("n flag but apparently has args") if @args;
1ded1f42
KW
236 print $fh "\t$ret\t$name";
237 } else { # full usage
1ded1f42
KW
238 my $n = "Perl_"x$p . $name;
239 my $large_ret = length $ret > 7;
240 my $indent_size = 7+8 # nroff: 7 under =head + 8 under =item
241 +8+($large_ret ? 1 + length $ret : 8)
242 +length($n) + 1;
243 my $indent;
244 print $fh "\t$ret" . ($large_ret ? ' ' : "\t") . "$n(";
245 my $long_args;
246 for (@args) {
247 if ($indent_size + 2 + length > 79) {
248 $long_args=1;
249 $indent_size -= length($n) - 3;
250 last;
251 }
252 }
253 my $args = '';
2f4e6339 254 if ($flags !~ /T/ && ($p || ($flags =~ /m/ && $name =~ /^Perl_/))) {
1ded1f42
KW
255 $args = @args ? "pTHX_ " : "pTHX";
256 if ($long_args) { print $fh $args; $args = '' }
257 }
258 $long_args and print $fh "\n";
259 my $first = !$long_args;
260 while () {
261 if (!@args or
262 length $args
263 && $indent_size + 3 + length($args[0]) + length $args > 79
264 ) {
265 print $fh
266 $first ? '' : (
267 $indent //=
268 "\t".($large_ret ? " " x (1+length $ret) : "\t")
269 ." "x($long_args ? 4 : 1 + length $n)
270 ),
271 $args, (","x($args ne 'pTHX_ ') . "\n")x!!@args;
272 $args = $first = '';
273 }
274 @args or last;
275 $args .= ", "x!!(length $args && $args ne 'pTHX_ ')
276 . shift @args;
277 }
278 if ($long_args) { print $fh "\n", substr $indent, 0, -4 }
279 print $fh ")";
280 }
8b5ff177 281 print $fh ";" if $flags =~ /s/; # semicolon "dTHR;"
1ded1f42 282 print $fh "\n\n";
94bdecf9
JH
283 }
284 print $fh "=for hackers\nFound in file $file\n\n";
285}
286
f83c6033
KW
287sub sort_helper {
288 # Do a case-insensitive dictionary sort, with only alphabetics
289 # significant, falling back to using everything for determinancy
1354d57e 290 return (uc($a =~ s/[[:^alpha:]]//r) cmp uc($b =~ s/[[:^alpha:]]//r))
f83c6033
KW
291 || uc($a) cmp uc($b)
292 || $a cmp $b;
293}
294
7b73ff98 295sub output {
5a0155e6 296 my ($podname, $header, $dochash, $missing, $footer) = @_;
6a4c4cd4
DM
297 #
298 # strip leading '|' from each line which had been used to hide
299 # pod from pod checkers.
300 s/^\|//gm for $header, $footer;
301
7882b24a 302 my $fh = open_new("pod/$podname.pod", undef,
20046047 303 {by => "$0 extracting documentation",
f1f44974 304 from => 'the C source files'}, 1);
e0492643 305
7882b24a 306 print $fh $header;
e0492643 307
7b73ff98 308 my $key;
f83c6033 309 for $key (sort sort_helper keys %$dochash) {
20046047
KE
310 my $section = $dochash->{$key};
311 print $fh "\n=head1 $key\n\n";
151c3fe5
KW
312
313 # Output any heading-level documentation and delete so won't get in
314 # the way later
315 if (exists $section->{""}) {
316 print $fh $section->{""} . "\n";
317 delete $section->{""};
318 }
20046047 319 print $fh "=over 8\n\n";
151c3fe5 320
20046047
KE
321 for my $key (sort sort_helper keys %$section) {
322 docout($fh, $key, $section->{$key});
323 }
324 print $fh "\n=back\n";
7b73ff98
NC
325 }
326
5a0155e6 327 if (@$missing) {
a23e6e20 328 print $fh "\n=head1 Undocumented functions\n\n";
2616800a 329 print $fh $podname eq 'perlapi' ? <<'_EOB_' : <<'_EOB_';
474d0ac8 330The following functions have been flagged as part of the public API,
72d33970 331but are currently undocumented. Use them at your own risk, as the
ba4591a5
KW
332interfaces are subject to change. Functions that are not listed in this
333document are not intended for public use, and should NOT be used under any
334circumstances.
335
5a4fed09
KW
336If you feel you need to use one of these functions, first send email to
337L<perl5-porters@perl.org|mailto:perl5-porters@perl.org>. It may be
338that there is a good reason for the function not being documented, and it
339should be removed from this list; or it may just be that no one has gotten
340around to documenting it. In the latter case, you will be asked to submit a
341patch to document the function. Once your patch is accepted, it will indicate
342that the interface is stable (unless it is explicitly marked otherwise) and
343usable by you.
cf5f2f8f 344_EOB_
2616800a
FC
345The following functions are currently undocumented. If you use one of
346them, you may wish to consider creating and submitting documentation for
347it.
2616800a 348_EOB_
6a4c4cd4
DM
349 print $fh "\n=over\n\n";
350
cf5f2f8f
KW
351 for my $missing (sort @$missing) {
352 print $fh "=item $missing\nX<$missing>\n\n";
5a0155e6 353 }
cf5f2f8f
KW
354 print $fh "=back\n\n";
355}
7882b24a 356 print $fh $footer, "=cut\n";
5a0155e6 357
7882b24a 358 read_only_bottom_close_and_rename($fh);
cd093254
MM
359}
360
e8e591c9
NC
361foreach (@{(setup_embed())[0]}) {
362 next if @$_ < 2;
363 my ($flags, $retval, $func, @args) = @$_;
364 s/\b(?:NN|NULLOK)\b\s+//g for @args;
bc350081 365
5ce57792 366 $funcflags{$func} = {
20046047
KE
367 flags => $flags,
368 retval => $retval,
369 args => \@args,
370 };
5ce57792
NC
371}
372
5ce57792
NC
373# glob() picks up docs from extra .c or .h files that may be in unclean
374# development trees.
741c0772
NC
375open my $fh, '<', 'MANIFEST'
376 or die "Can't open MANIFEST: $!";
377while (my $line = <$fh>) {
b87d9527 378 next unless my ($file) = $line =~ /^(\S+\.(?:[ch]|pod))\t/;
5ce57792 379
1ae6ead9 380 open F, '<', $file or die "Cannot open $file for docs: $!\n";
5ce57792
NC
381 $curheader = "Functions in file $file\n";
382 autodoc(\*F,$file);
383 close F or die "Error closing $file: $!\n";
384}
741c0772 385close $fh or die "Error whilst reading MANIFEST: $!";
5ce57792
NC
386
387for (sort keys %funcflags) {
388 next unless $funcflags{$_}{flags} =~ /d/;
6523e108 389 next if $funcflags{$_}{flags} =~ /h/;
5ce57792 390 warn "no docs for $_\n"
bc350081 391}
94bdecf9 392
5ce57792 393foreach (sort keys %missing) {
5ce57792 394 warn "Function '$_', documented in $missing{$_}, not listed in embed.fnc";
94bdecf9
JH
395}
396
5ce57792
NC
397# walk table providing an array of components in each line to
398# subroutine, printing the result
399
ff5af78d
KW
400# List of funcs in the public API that aren't also marked as core-only,
401# experimental nor deprecated.
b87d9527
KW
402my @missing_api = grep $funcflags{$_}{flags} =~ /A/
403 && $funcflags{$_}{flags} !~ /[xD]/
404 && !$docs{api}{$_}, keys %funcflags;
405output('perlapi', <<"_EOB_", $docs{api}, \@missing_api, <<"_EOE_");
6a4c4cd4
DM
406|=encoding UTF-8
407|
408|=head1 NAME
409|
410|perlapi - autogenerated documentation for the perl public API
411|
412|=head1 DESCRIPTION
413|X<Perl API> X<API> X<api>
414|
b87d9527
KW
415|This file contains most of the documentation of the perl public API, as
416|generated by F<embed.pl>. Specifically, it is a listing of functions,
417|macros, flags, and variables that may be used by extension writers. Some
418|specialized items are instead documented in $specialized_docs.
419|
420|L<At the end|/Undocumented functions> is a list of functions which have yet
421|to be documented. Patches welcome! The interfaces of these are subject to
422|change without notice.
423|
424|Anything not listed here is not part of the public API, and should not be
425|used by extension writers at all. For these reasons, blindly using functions
426|listed in proto.h is to be avoided when writing extensions.
6a4c4cd4
DM
427|
428|In Perl, unlike C, a string of characters may generally contain embedded
429|C<NUL> characters. Sometimes in the documentation a Perl string is referred
430|to as a "buffer" to distinguish it from a C string, but sometimes they are
431|both just referred to as strings.
432|
433|Note that all Perl API global variables must be referenced with the C<PL_>
434|prefix. Again, those not listed here are not to be used by extension writers,
435|and can be changed or removed without notice; same with macros.
436|Some macros are provided for compatibility with the older,
437|unadorned names, but this support may be disabled in a future release.
438|
439|Perl was originally written to handle US-ASCII only (that is characters
440|whose ordinal numbers are in the range 0 - 127).
441|And documentation and comments may still use the term ASCII, when
442|sometimes in fact the entire range from 0 - 255 is meant.
443|
444|The non-ASCII characters below 256 can have various meanings, depending on
445|various things. (See, most notably, L<perllocale>.) But usually the whole
446|range can be referred to as ISO-8859-1. Often, the term "Latin-1" (or
447|"Latin1") is used as an equivalent for ISO-8859-1. But some people treat
448|"Latin1" as referring just to the characters in the range 128 through 255, or
449|somethimes from 160 through 255.
450|This documentation uses "Latin1" and "Latin-1" to refer to all 256 characters.
451|
452|Note that Perl can be compiled and run under either ASCII or EBCDIC (See
453|L<perlebcdic>). Most of the documentation (and even comments in the code)
454|ignore the EBCDIC possibility.
455|For almost all purposes the differences are transparent.
456|As an example, under EBCDIC,
457|instead of UTF-8, UTF-EBCDIC is used to encode Unicode strings, and so
458|whenever this documentation refers to C<utf8>
459|(and variants of that name, including in function names),
460|it also (essentially transparently) means C<UTF-EBCDIC>.
461|But the ordinals of characters differ between ASCII, EBCDIC, and
462|the UTF- encodings, and a string encoded in UTF-EBCDIC may occupy a different
463|number of bytes than in UTF-8.
464|
465|The listing below is alphabetical, case insensitive.
466|
94bdecf9 467_EOB_
6a4c4cd4
DM
468|
469|=head1 AUTHORS
470|
471|Until May 1997, this document was maintained by Jeff Okamoto
7b1f0a98 472|<okamoto\@corp.hp.com>. It is now maintained as part of Perl itself.
6a4c4cd4
DM
473|
474|With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
475|Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
476|Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
477|Stephen McCamant, and Gurusamy Sarathy.
478|
7b1f0a98 479|API Listing originally by Dean Roehrich <roehrich\@cray.com>.
6a4c4cd4
DM
480|
481|Updated to be autogenerated from comments in the source by Benjamin Stuhl.
482|
483|=head1 SEE ALSO
484|
7b1f0a98 485$other_places_api
94bdecf9
JH
486_EOE_
487
79fc8511
FC
488# List of non-static internal functions
489my @missing_guts =
9f589e47 490 grep $funcflags{$_}{flags} !~ /[AS]/ && !$docs{guts}{$_}, keys %funcflags;
5a0155e6 491
7b1f0a98 492output('perlintern', <<'_EOB_', $docs{guts}, \@missing_guts, <<"_EOE_");
6a4c4cd4
DM
493|=head1 NAME
494|
495|perlintern - autogenerated documentation of purely B<internal>
20046047 496|Perl functions
6a4c4cd4
DM
497|
498|=head1 DESCRIPTION
499|X<internal Perl functions> X<interpreter functions>
500|
501|This file is the autogenerated documentation of functions in the
502|Perl interpreter that are documented using Perl's internal documentation
503|format but are not marked as part of the Perl API. In other words,
504|B<they are not for use in extensions>!
505|
7b1f0a98 506_EOB_
6a4c4cd4
DM
507|
508|=head1 AUTHORS
509|
510|The autodocumentation system was originally added to the Perl core by
511|Benjamin Stuhl. Documentation is by whoever was kind enough to
512|document their functions.
513|
514|=head1 SEE ALSO
515|
7b1f0a98
KW
516$other_places_intern
517_EOE_