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