1#!perl
2# Generates info for Module::CoreList from this perl tree
3# run this from the root of a perl tree
4#
5# Data is on STDOUT.
6#
7# With an optional arg specifying the root of a CPAN mirror, outputs the
8# %upstream and %bug_tracker hashes too.
9
10use autodie;
11use strict;
12use warnings;
13use File::Find;
14use ExtUtils::MM_Unix;
15use version;
16use lib "Porting";
17use Maintainers qw(%Modules files_to_modules);
18use File::Spec;
19use Parse::CPAN::Meta;
20use IPC::Cmd 'can_run';
21use HTTP::Tiny;
22use IO::Uncompress::Gunzip;
23
24my $corelist_file = './dist/Module-CoreList/lib/Module/CoreList.pm';
25my $utils_file = './dist/Module-CoreList/lib/Module/CoreList/Utils.pm';
26
27my %lines;
28my %module_to_file;
29my %modlist;
30
31die "usage: $0 [ cpan-mirror/ ] [ 5.x.y] \n" unless @ARGV <= 2;
32my $cpan         = shift;
33my $raw_version  = shift || $];
34my $perl_version = version->parse("$raw_version");
35my $perl_vnum    = $perl_version->numify;
36my $perl_vstring = $perl_version->normal; # how do we get version.pm to not give us leading v?
37$perl_vstring =~ s/^v//;
38
39if ( !-f 'MANIFEST' ) {
40    die "Must be run from the root of a clean perl tree\n";
41}
42
43open( my $corelist_fh, '<', $corelist_file );
44my $corelist = join( '', <$corelist_fh> );
45close $corelist_fh;
46
47unless (
48    $corelist =~ /^%released \s* = \s* \(
49        .*?
50        $perl_vnum \s* => \s* .*?
51        \);/ismx
52    )
53{
54    warn "Adding $perl_vnum to the list of released perl versions. Please consider adding a release date.\n";
55    $corelist =~ s/^(%released \s* = \s* .*?) ( \) )
56                /$1  $perl_vnum => '????-??-??',\n  $2/ismx;
57}
58
59if ($cpan) {
60    my $modlistfile = File::Spec->catfile( $cpan, 'modules', '02packages.details.txt' );
61    my $content;
62
63    my $fh;
64    if ( -e $modlistfile ) {
65        warn "Reading the module list from $modlistfile";
66        open $fh, '<', $modlistfile;
67    } elsif ( -e $modlistfile . ".gz" ) {
68        my $zcat = can_run('gzcat') || can_run('zcat') or die "Can't find gzcat or zcat";
69        warn "Reading the module list from $modlistfile.gz";
70        open $fh, '-|', "$zcat $modlistfile.gz";
71    } else {
72        warn "About to fetch 02packages from www.cpan.org. This may take a few minutes\n";
73	my $gzipped_content = fetch_url('http://www.cpan.org/modules/02packages.details.txt.gz');
74	unless ($gzipped_content) {
75            die "Unable to read 02packages.details.txt from either your CPAN mirror or www.cpan.org";
76        }
77	IO::Uncompress::Gunzip::gunzip(\$gzipped_content, \$content, Transparent => 0)
78	    or die "Can't gunzip content: $IO::Uncompress::Gunzip::GunzipError";
79    }
80
81    if ( $fh and !$content ) {
82        local $/ = "\n";
83        $content = join( '', <$fh> );
84    }
85
86    die "Incompatible modlist format"
87        unless $content =~ /^Columns: +package name, version, path/m;
88
89    # Converting the file to a hash is about 5 times faster than a regexp flat
90    # lookup.
91    for ( split( qr/\n/, $content ) ) {
92        next unless /^([A-Za-z_:0-9]+) +[-0-9.undefHASHVERSIONvsetwhenloadingbogus]+ +(\S+)/;
93        $modlist{$1} = $2;
94    }
95}
96
97find(
98    sub {
99        if (-d) {
100          my @parts = File::Spec->splitdir($File::Find::name);
101          # be careful not to skip inc::latest
102          return $File::Find::prune = 1 if @parts == 3 and ($parts[-1] eq 'inc' or $parts[-1] eq 't');
103        }
104
105        /(\.pm|_pm\.PL)$/ or return;
106        /PPPort\.pm$/ and return;
107        my $module = $File::Find::name;
108        $module =~ /\b(demo|t|private|corpus)\b/ and return;    # demo or test modules
109        my $version = MM->parse_version($_);
110        defined $version or $version = 'undef';
111        $version =~ /\d/ and $version = "'$version'";
112
113        # some heuristics to figure out the module name from the file name
114        $module =~ s{^(lib|cpan|dist|ext|os2/OS2)/}{}
115			and $1 ne 'lib'
116            and (
117            $module =~ s{\b(\w+)/\1\b}{$1},
118            $module =~ s{^B/O}{O},
119            $module =~ s{^Devel-PPPort}{Devel},
120            $module =~ s{^libnet/}{},
121            $module =~ s{^PathTools/}{},
122            $module =~ s{REXX/DLL}{DLL},
123            $module =~ s{^Encode/encoding}{encoding},
124            $module =~ s{^IPC-SysV/}{IPC/},
125            $module =~ s{^MIME-Base64/QuotedPrint}{MIME/QuotedPrint},
126            $module =~ s{^(?:DynaLoader|Errno|Opcode|XSLoader)/}{},
127            $module =~ s{^Sys-Syslog/win32}{Sys-Syslog},
128            $module =~ s{^Time-Piece/Seconds}{Time/Seconds},
129            );
130		$module =~ s{^lib/}{}g;
131        $module =~ s{/}{::}g;
132        $module =~ s{-}{::}g;
133		$module =~ s{^.*::lib::}{}; # turns Foo/lib/Foo.pm into Foo.pm
134        $module =~ s/(\.pm|_pm\.PL)$//;
135        $lines{$module}          = $version;
136        $module_to_file{$module} = $File::Find::name;
137    },
138    'os2/OS2',
139    'lib',
140    'ext',
141	'cpan',
142	'dist'
143);
144
145-e 'configpm' and $lines{Config} = "$]";
146
147if ( open my $ucdv, "<", "lib/unicore/version" ) {
148    chomp( my $ucd = <$ucdv> );
149    $lines{Unicode} = "'$ucd'";
150    close $ucdv;
151}
152
153my $delta_data = make_corelist_delta(
154  $perl_vnum,
155  \%lines,
156  \%Module::CoreList::version
157);
158
159my $versions_in_release = "    " . $perl_vnum . " => {\n";
160$versions_in_release .= "        delta_from => $delta_data->{delta_from},\n";
161$versions_in_release .= "        changed => {\n";
162foreach my $key (sort keys $delta_data->{changed}->%*) {
163  $versions_in_release .= sprintf "            %-24s=> %s,\n", "'$key'",
164      defined $delta_data->{changed}{$key} ? "'"
165        . $delta_data->{changed}{$key} . "'" : "undef";
166}
167$versions_in_release .= "        },\n";
168$versions_in_release .= "        removed => {\n";
169for my $key (sort keys %{ $delta_data->{removed} || {} }) {
170  $versions_in_release .= sprintf "            %-24s=> %s,\n", "'$key'", 1;
171}
172$versions_in_release .= "        }\n";
173$versions_in_release .= "    },\n";
174
175$corelist =~ s/^(%delta\s*=\s*.*?)^\s*$perl_vnum\s*=>\s*{.*?},\s*(^\);)$/$1$2/ism;
176$corelist =~ s/^(%delta\s*=\s*.*?)(^\);)$/$1$versions_in_release$2/ism;
177
178exit unless %modlist;
179
180# We have to go through this two stage lookup, given how Maintainers.pl keys its
181# data by "Module", which is really a dist.
182my $file_to_M = files_to_modules( values %module_to_file );
183
184sub slurp_utf8($) {
185    open my $fh, "<:utf8", "$_[0]"
186	or die "can't open $_[0] for reading: $!";
187    return do { local $/; <$fh> };
188}
189
190sub parse_cpan_meta($) {
191    return Parse::CPAN::Meta->${
192	$_[0] =~ /\A\x7b/ ? \"load_json_string" : \"load_yaml_string"
193    }($_[0]);
194}
195
196my %module_to_upstream;
197my %module_to_dist;
198my %dist_to_meta_YAML;
199my %module_to_deprecated;
200while ( my ( $module, $file ) = each %module_to_file ) {
201    my $M = $file_to_M->{$file};
202    next unless $M;
203    next if $Modules{$M}{MAINTAINER} && $Modules{$M}{MAINTAINER} eq 'P5P';
204    $module_to_upstream{$module} = $Modules{$M}{UPSTREAM};
205    $module_to_deprecated{$module} = 1 if $Modules{$M}{DEPRECATED};
206    next
207        if defined $module_to_upstream{$module}
208            && $module_to_upstream{$module} eq 'blead';
209    my $dist = $modlist{$module};
210    unless ($dist) {
211        warn "Can't find a distribution for $module\n";
212        next;
213    }
214    $module_to_dist{$module} = $dist;
215
216    next if exists $dist_to_meta_YAML{$dist};
217
218    $dist_to_meta_YAML{$dist} = undef;
219
220    # Like it or lump it, this has to be Unix format.
221    my $meta_YAML_path = "authors/id/$dist";
222    $meta_YAML_path =~ s/(?:tar\.gz|tar\.bz2|zip|tgz)$/meta/
223	or die "ERROR: bad meta YAML path: '$meta_YAML_path'";
224    my $meta_YAML_url = 'http://www.cpan.org/' . $meta_YAML_path;
225
226    if ( -e "$cpan/$meta_YAML_path" ) {
227        $dist_to_meta_YAML{$dist} = parse_cpan_meta(slurp_utf8( $cpan . "/" . $meta_YAML_path ));
228    } elsif ( my $content = fetch_url($meta_YAML_url) ) {
229        unless ($content) {
230            warn "Failed to fetch $meta_YAML_url\n";
231            next;
232        }
233        eval { $dist_to_meta_YAML{$dist} = parse_cpan_meta($content); };
234        if ( my $err = $@ ) {
235            warn "$meta_YAML_path: ".$err;
236            next;
237        }
238    } else {
239        warn "$meta_YAML_path does not exist for $module\n";
240
241        # I tried code to open the tarballs with Archive::Tar to find and
242        # extract META.yml, but only Text-Tabs+Wrap-2006.1117.tar.gz had one,
243        # so it's not worth including.
244        next;
245    }
246}
247
248my $upstream_stanza = "%upstream = (\n";
249foreach my $module ( sort keys %module_to_upstream ) {
250    my $upstream = defined $module_to_upstream{$module} ? "'$module_to_upstream{$module}'" : 'undef';
251    $upstream_stanza .= sprintf "    %-24s=> %s,\n", "'$module'", $upstream;
252}
253$upstream_stanza .= ");";
254
255$corelist =~ s/^%upstream .*? ;$/$upstream_stanza/ismx;
256
257# Deprecation generation
258{
259  my $delta_data = make_corelist_delta(
260    $perl_vnum,
261    \%module_to_deprecated,
262    do { no warnings 'once'; \%Module::CoreList::deprecated },
263  );
264
265  my $deprecated_stanza = "    " . $perl_vnum . " => {\n";
266  $deprecated_stanza .= "        delta_from => $delta_data->{delta_from},\n";
267  $deprecated_stanza .= "        changed => {\n";
268  foreach my $key (sort keys $delta_data->{changed}->%*) {
269    $deprecated_stanza .= sprintf "            %-24s=> %s,\n", "'$key'",
270        defined $delta_data->{changed}{$key} ? "'"
271          . $delta_data->{changed}{$key} . "'" : "undef";
272  }
273  $deprecated_stanza .= "        },\n";
274  $deprecated_stanza .= "        removed => {\n";
275  for my $key (sort keys %{ $delta_data->{removed} || {} }) {
276    $deprecated_stanza .= sprintf "           %-24s=> %s,\n", "'$key'", 1;
277  }
278  $deprecated_stanza .= "        }\n";
279  $deprecated_stanza .= "    },\n";
280
281  $corelist =~ s/^(%deprecated\s*=\s*.*?)^\s*$perl_vnum\s*=>\s*{.*?},\s*(^\);)$/$1$2/ism;
282  $corelist =~ s/^(%deprecated\s*=\s*.*?)(^\);)$/$1$deprecated_stanza$2/xism;
283}
284
285my $tracker = "%bug_tracker = (\n";
286foreach my $module ( sort keys %module_to_upstream ) {
287    my $upstream = defined $module_to_upstream{$module};
288    next
289        if defined $upstream and $upstream eq 'blead';
290
291    my $bug_tracker;
292
293    my $dist = $module_to_dist{$module};
294    $bug_tracker = $dist_to_meta_YAML{$dist}->{resources}{bugtracker}
295        if $dist;
296    $bug_tracker = $bug_tracker->{web} if ref($bug_tracker) eq "HASH";
297
298    $bug_tracker = defined $bug_tracker ? quote($bug_tracker) : 'undef';
299    next if $bug_tracker eq "'https://github.com/Perl/perl5/issues'";
300	next if $bug_tracker eq "'http://rt.perl.org/perlbug/'";
301	next if $bug_tracker eq "'https://rt.perl.org/perlbug/'";
302    $tracker .= sprintf "    %-24s=> %s,\n", "'$module'", $bug_tracker;
303}
304$tracker .= ");";
305
306$corelist =~ s/^%bug_tracker .*? ;/$tracker/eismx;
307
308write_corelist($corelist,$corelist_file);
309
310open( my $utils_fh, '<', $utils_file );
311my $utils = join( '', <$utils_fh> );
312close $utils_fh;
313
314my %utils = map { ( $_ => 1 ) } parse_utils_lst();
315
316my $delta_utils = make_coreutils_delta($perl_vnum, \%utils);
317
318my $utilities_in_release = "    " . $perl_vnum . " => {\n";
319$utilities_in_release .= "        delta_from => $delta_utils->{delta_from},\n";
320$utilities_in_release .= "        changed => {\n";
321foreach my $key (sort keys $delta_utils->{changed}->%*) {
322  $utilities_in_release .= sprintf "            %-24s=> %s,\n", "'$key'",
323      defined $delta_utils->{changed}{$key} ? "'"
324        . $delta_utils->{changed}{$key} . "'" : "undef";
325}
326$utilities_in_release .= "        },\n";
327$utilities_in_release .= "        removed => {\n";
328for my $key (sort keys %{ $delta_utils->{removed} || {} }) {
329  $utilities_in_release .= sprintf "            %-24s=> %s,\n", "'$key'", 1;
330}
331$utilities_in_release .= "        }\n";
332$utilities_in_release .= "    },\n";
333
334$utils =~ s/^(my %delta\s*=\s*.*?)^\s*$perl_vnum\s*=>\s*{.*?},\s*(^\);)$/$1$2/ism;
335$utils =~ s/^(my %delta\s*=\s*.*?)(^\);)$/$1$utilities_in_release$2/ism;
336
337write_corelist($utils,$utils_file);
338
339warn "All done. Please check over the following files carefully before committing.\nThanks!\n";
340warn "$corelist_file\n$utils_file\n";
341
342sub write_corelist {
343    my $content = shift;
344    my $filename = shift;
345    open (my $clfh, ">", $filename);
346    binmode $clfh;
347    print $clfh $content;
348    close($clfh);
349}
350
351sub fetch_url {
352    my $url = shift;
353    my $http = HTTP::Tiny->new;
354    my $response = $http->get($url);
355    if ($response->{success}) {
356	return $response->{content};
357    } else {
358	warn "Error fetching $url: $response->{status} $response->{reason}\n";
359        return;
360    }
361}
362
363sub make_corelist_delta {
364  my($version, $lines, $existing) = @_;
365  # Trust core perl, if someone does use a weird version number the worst that
366  # can happen is an extra delta entry for a module.
367  my %versions = map { $_ => eval $lines->{$_} } keys %$lines;
368
369  # Ensure we have the corelist data loaded from this perl checkout, not the system one.
370  require $corelist_file;
371
372  my %deltas;
373  # Search for the release with the least amount of changes (this avoids having
374  # to ask for where this perl was branched from).
375  for my $previous (reverse sort { $a <=> $b } keys %$existing) {
376    # Shouldn't happen, but ensure we don't load weird data...
377    next if $previous > $version || $previous == $version;
378    my $delta = $deltas{$previous} = {};
379    ($delta->{changed}, $delta->{removed}) = calculate_delta(
380      $existing->{$previous}, \%versions);
381  }
382
383  my $smallest = (sort {
384      ((keys($deltas{$a}->{changed}->%*) + keys($deltas{$a}->{removed}->%*)) <=>
385       (keys($deltas{$b}->{changed}->%*) + keys($deltas{$b}->{removed}->%*))) ||
386      $b <=> $a
387    } keys %deltas)[0];
388
389  return {
390    delta_from => $smallest,
391    changed => $deltas{$smallest}{changed},
392    removed => $deltas{$smallest}{removed},
393  }
394}
395
396sub make_coreutils_delta {
397  my($version, $lines) = @_;
398  # Trust core perl, if someone does use a weird version number the worst that
399  # can happen is an extra delta entry for a module.
400  my %utilities = map { $_ => eval $lines->{$_} } keys %$lines;
401
402  # Ensure we have the corelist data loaded from this perl checkout, not the system one.
403  require $utils_file;
404
405  my %deltas;
406  # Search for the release with the least amount of changes (this avoids having
407  # to ask for where this perl was branched from).
408  for my $previous (reverse sort { $a <=> $b } keys %Module::CoreList::Utils::utilities) {
409    # Shouldn't happen, but ensure we don't load weird data...
410    next if $previous > $version || $previous == $version;
411
412    my $delta = $deltas{$previous} = {};
413    ($delta->{changed}, $delta->{removed}) = calculate_delta(
414      $Module::CoreList::Utils::utilities{$previous}, \%utilities);
415  }
416
417  my $smallest = (sort {
418      ((keys($deltas{$a}->{changed}->%*) + keys($deltas{$a}->{removed}->%*)) <=>
419       (keys($deltas{$b}->{changed}->%*) + keys($deltas{$b}->{removed}->%*))) ||
420      $b <=> $a
421    } keys %deltas)[0];
422
423  return {
424    delta_from => $smallest,
425    changed => $deltas{$smallest}{changed},
426    removed => $deltas{$smallest}{removed},
427  }
428}
429
430# Calculate (changed, removed) modules between two versions.
431sub calculate_delta {
432  my($from, $to) = @_;
433  my(%changed, %removed);
434
435  for my $package(keys %$from) {
436    if(not exists $to->{$package}) {
437      $removed{$package} = 1;
438    }
439  }
440
441  for my $package(keys %$to) {
442    if(!exists $from->{$package}
443        || (defined $from->{$package} && !defined $to->{$package})
444        || (!defined $from->{$package} && defined $to->{$package})
445        || (defined $from->{$package} && defined $to->{$package}
446            && $from->{$package} ne $to->{$package})) {
447      $changed{$package} = $to->{$package};
448    }
449  }
450
451  return \%changed, \%removed;
452}
453
454sub quote {
455    my ($str) = @_;
456    # There's gotta be something already doing this properly that we could just
457    # reuse, but I can't quite thing of where to look for it, so I'm gonna do
458    # the simplest possible thing that'll allow me to release 5.17.7.  --rafl
459    $str =~ s/'/\\'/g;
460    "'${str}'";
461}
462
463sub parse_utils_lst {
464  require File::Spec::Unix;
465  my @scripts;
466  open my $fh, '<', 'utils.lst' or die "$!\n";
467  while (<$fh>) {
468    chomp;
469    my ($file,$extra) = split m!#!;
470    $file =~ s!\s+!!g;
471    push @scripts, $file;
472    $extra =~ s!\s+!!g if $extra;
473    if ( $extra and my ($link) = $extra =~ m!^link=(.+?)$! ) {
474      push @scripts, $link;
475    }
476  }
477  return map { +( File::Spec::Unix->splitpath( $_ ) )[-1] } @scripts;
478}
479