2 # Copyright 2009 Ian Beckwith <ianb@erislabs.net>
3 # License: GPL v2 or later.
7 $me=($0=~/(?:.*\/)?(.*)/)[0];
11 use Digest::SHA qw(sha256_hex);
15 our $CLSCANDIR="debian/clscan";
16 our $FILESCACHE="$CLSCANDIR/files.yaml";
17 our $NEWFILES="$CLSCANDIR/new.txt";
18 our $COPYRIGHTSTUB="$CLSCANDIR/copyright.in";
20 my $gpl_boilerplate=<<"EOL";
21 This program is free software: you can redistribute it and/or modify
22 it under the terms of the GNU General Public License as published by
23 the Free Software Foundation; either version 3 of the License, or
24 (at your option) any later version.
26 This program is distributed in the hope that it will be useful,
27 but WITHOUT ANY WARRANTY; without even the implied warranty of
28 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
29 GNU General Public License for more details.
31 You should have received a copy of the GNU General Public License
32 along with this program. If not, see <http://www.gnu.org/licenses/>.
35 my $lgpl2_boilerplate=<<"EOL";
36 This program is free software; you can redistribute it and/or modify it
37 under the terms of the GNU Library General Public License as published
38 by the Free Software Foundation; either version 2, or (at your option)
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
44 Library General Public License for more details.
46 You should have received a copy of the GNU Library General Public
47 License along with this program; if not, write to the Free Software
48 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
52 my $lgpl3_boilerplate=<<"EOL";
53 This program is free software: you can redistribute it and/or modify
54 it under the terms of the GNU Lesser General Public License as published by
55 the Free Software Foundation; either version 3 of the License, or
56 (at your option) any later version.
58 This program is distributed in the hope that it will be useful,
59 but WITHOUT ANY WARRANTY; without even the implied warranty of
60 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
61 GNU General Public License for more details.
63 You should have received a copy of the GNU General Public License
64 along with this program. If not, see <http://www.gnu.org/licenses/>.
67 # license overrides as specified in modules/*
68 our $module_licenses =
78 license_text => ("This file is free software; the Free Software Foundation\n" .
79 "gives unlimited permission to copy and/or distribute it,\n" .
80 "with or without modifications, as long as this notice is preserved.\n"),
85 license_text => $lgpl3_boilerplate,
90 license_text => $lgpl2_boilerplate,
95 license_text => $lgpl3_boilerplate,
100 license_text => $lgpl3_boilerplate,
102 "unmodifiable license text" =>
105 license_text => ("Everyone is permitted to copy and distribute verbatim copies\n" .
106 "of this license document, but changing it is not allowed.\n"),
108 "GPLed build tool" =>
111 license_text => $gpl_boilerplate,
116 license_text => $gpl_boilerplate,
124 our @deleted_files=();
130 my $writecopyright=0;
132 usage() unless(@ARGV);
133 usage() unless GetOptions("scan|s" => \$scan,
134 "merge|m" => \$merge,
135 "write|w" => \$writecopyright,
142 write_copyright() if ($merge || $writecopyright);
148 for my $file (@filenames)
165 unless(keys(%$files))
167 die("$me: no files known, run $0 --scan\n");
169 unless(copy($COPYRIGHTSTUB, "debian/copyright"))
171 die("$me: cannot copy $COPYRIGHTSTUB to debian/copyright: $!\n");
173 unless(open(COPYRIGHT, ">>debian/copyright"))
175 die("$me: cannot append to debian/copyright: $!\n");
178 # group files by license/license_text/copyright
180 for my $file (sort keys(%$files))
182 my $license=$files->{$file}->{license_override} || $files->{$file}->{license};
183 my $copyright=$files->{$file}->{copyright};
184 my $license_text=$files->{$file}->{license_text_override} ||
185 $files->{$file}->{license_text};
186 push(@{$licenses->{$license}->{$license_text}->{$copyright}}, $file);
190 print COPYRIGHT license_trailer(sort keys(%$licenses));
191 for my $license (sort keys(%$licenses))
193 for my $license_text (sort keys(%{$licenses->{$license}}))
195 my $licensestr=$license;
196 if(length($license_text))
199 # license_text + empty license = License: other
200 if(!length($license))
202 $licensestr = "other";
204 $licensestr .= " [REF$refnum]";
205 $refs{$licensestr}=$license_text;
209 if(!length($license)) {
210 $licensestr="unknown";
213 for my $copyright (sort keys(%{$licenses->{$license}->{$license_text}}))
215 next if(!length($license) && !length($copyright) && !length($license_text));
216 my @filelist=sort @{$licenses->{$license}->{$license_text}->{$copyright}};
217 print COPYRIGHT "Files: ", join(' ', @filelist), "\n";
218 print COPYRIGHT "Copyright: ". (length($copyright) ? $copyright : "unknown" ) . "\n";
219 print COPYRIGHT "License: $licensestr\n" if length($licensestr);
220 print COPYRIGHT "\n";
224 for my $ref (sort byref keys(%refs))
226 print COPYRIGHT "License: $ref\n";
227 my @text=split(/\n/, $refs{$ref});
228 @text=map { ($_ eq "") ? "." : $_; } @text;
229 print COPYRIGHT map { " " . $_ . "\n" } @text;
230 print COPYRIGHT "\n";
237 my $aref=($a=~/\[REF(\d+)\]/)[0];
238 my $bref=($b=~/\[REF(\d+)\]/)[0];
241 return($aref <=> $bref);
248 my @licenses_used=@_;
251 "Apache-2.0" => "Apache License Version 2.0",
252 "Artistic" => "Artistic License",
253 "GFDL-1.2" => "GNU Free Documentation License Version 1.2",
254 "GFDL-1.3" => "GNU Free Documentation License Version 1.3",
255 "GFDL" => "GNU Free Documentation License",
256 "GPL-2" => "GNU General Public License Version 2",
257 "GPL-3" => "GNU General Public License Version 3",
258 "GPL" => "GNU General Public License",
259 "LGPL-2" => "GNU Library General Public License Version 2",
260 "LGPL-2.1" => "GNU Lesser General Public License Version 2.1",
261 "LGPL-3" => "GNU Lesser General Public License Version 3",
262 "LGPL" => "GNU Lesser General Public License",
266 for my $type (reverse sort keys(%$license_data))
268 for my $license (@licenses_used)
270 if($license =~ /$type(\+|\s|$)/i)
272 $types_found{$type}=1;
277 # if just one, use standard style
278 if(keys(%types_found) == 1)
280 my ($file, $name)=each(%types_found);
281 $text .= " The complete text of the $name can be\n";
282 $text .= " found in /usr/share/common-licenses/$file\n";
286 # more than one, use table.
287 $text .= " The complete text of standard licenses referenced above\n";
288 $text .= " can be found in /usr/share/common-licenses/ as follows:\n .\n ";
289 $text .= sprintf("%-60s %s\n", "LICENSE", "FILE");
290 for my $type (sort keys(%types_found))
292 $text .= sprintf(" %-60s %s\n", $license_data->{$type}, $type);
305 if(open(LICENSECHECK, "licensecheck --copyright \"$file\"|"))
307 while(<LICENSECHECK>)
320 $copyright =~ s/^\s*Copyright\s*:\s*//;
321 $license =~ s/.*UNKNOWN.*//;
322 $license =~ s/(L?GPL) \(v([\.\d]+) or later\)/$1-$2+/i;
323 $license =~ s/(L?GPL) \(v([\.\d]+)\)/$1-$2/i;
324 $license =~ s/G?FDL \(v([\.\d]+) or later\)/GFDL-$1+/i;
325 $license =~ s/G?FDL \(v([\.\d]+)\)/GFDL-$1/i;
326 $license =~ s/Apache \(v([\.\d]+) or later\)/Apache-$1+/i;
327 $license =~ s/Apache \(v([\.\d]+)\)/Apache-$1+/i;
329 return($license, $copyright);
335 unless(open(FILE, $filename))
337 warn("$me: $filename: cannot open: $!\n");
341 for(my $i=0; $i < 15; $i++)
348 my $hash=sha256_hex($header);
349 if( (!exists($files->{$filename})) ||
350 ($files->{$filename}->{hash} ne $hash))
352 filechanged($filename, $hash, $header);
359 my($filename, $hash, $header)=@_;
360 my($license_guess, $copyright_guess)=guess_license($filename);
364 license=>$license_guess,
365 copyright=>$copyright_guess,
368 if(exists($files->{$filename}))
370 if(exists($files->{$filename}->{copyright}))
372 $new->{$filename}->{copyright}=$files->{$filename}->{copyright};
373 $new->{$filename}->{copyright_guess}=$copyright_guess;
375 if(exists($files->{$filename}->{license}))
377 $new->{$filename}->{license}=$files->{$filename}->{license};
378 $new->{$filename}->{license_guess}=$license_guess;
380 if(exists($files->{$filename}->{license_text}))
382 $new->{$filename}->{license_text}=$files->{$filename}->{license_text};
389 find(\&wanted_files, ".");
394 if(/^\.git/ || /^\.cvs/ || /^debian/ || /^modules$/ || /^\.pc/)
396 $File::Find::prune=1;
400 push(@filenames, $File::Find::name);
406 if(/^\.[^\/]/ || /^README$/ || /^COPYING$/)
408 $File::Find::prune=1;
413 unless(open(MOD, $File::Find::name))
415 warn("$me: cannot open $File::Find::name: $!\n");
426 $infiles = $inlicense = 0;
430 push(@{$overrides{$_}},@files);
452 find({ wanted => \&wanted_modules, no_chdir => 1}, "modules/");
453 for my $license (keys(%overrides))
455 if(!exists($module_licenses->{$license}))
457 die("$me: license override \"$license\" not found in \$module_licenses\n");
459 my @overridden_files=map { "./" . $_; } @{$overrides{$license}};
460 for my $file (@overridden_files)
462 my $override=$module_licenses->{$license};
463 if(length($override->{license}))
465 $files->{$file}->{license_override}=$override->{license};
467 if(length($override->{license_text}))
469 $files->{$file}->{license_text_override}=$override->{license_text};
478 unless(open(YAML,$FILESCACHE))
480 warn("$me: cannot load cache $FILESCACHE: $!\n");
495 unless(open(YAML,">$FILESCACHE"))
497 warn("$me: cannot save cache $FILESCACHE: $!\n");
500 print YAML Dump($files);
509 warn("$me: no new/changed files found\n");
511 unless(open(NEW,">$NEWFILES"))
513 die("$me: cannot write to $NEWFILES: $!\n");
515 for my $file (sort keys %$new)
517 print NEW "File: $file\n";
518 print NEW "Hash: ", $new->{$file}->{hash}, "\n";
519 print NEW "Copyright: ", $new->{$file}->{copyright}, "\n";
520 if($new->{$file}->{copyright_guess})
522 print NEW "#Copyright_guess: ", $new->{$file}->{copyright_guess}, "\n";
524 print NEW "License: ", $new->{$file}->{license}, "\n";
525 if($new->{$file}->{license_guess})
527 print NEW "#License_guess: ", $new->{$file}->{license_guess}, "\n";
529 if($new->{$file}->{license_text})
531 my @text=split(/\n/, $new->{$file}->{license_text});
532 print NEW "\t" . join("\n\t", @text), "\n";
534 print NEW "#Header: \n";
535 my @headerlines=split(/\n/, $new->{$file}->{header});
536 @headerlines=map { "#" . $_ } grep { defined $_; } @headerlines;
537 print NEW join("\n", @headerlines);
542 print NEW map { "Deleted: $_\n"; } @deleted_files;
549 unless(open(NEW, $NEWFILES))
551 die("$me: $NEWFILES: cannot open: $!\n");
558 my $in_license_text=0;
565 if($in_license_text && /^\s+(.*)/)
567 $license_text .= $1 . "\n";
573 elsif(/^File:\s*(.*)/)
576 # save previous entry
577 if(length($file) && length($hash))
581 copyright=>$copyright,
583 license_text=>$license_text };
590 $in_license_text = 0;
592 elsif(/^Hash:\s*(.*)/)
596 elsif(/^Copyright:\s*(.*)/)
600 elsif(/^License:\s*(.*)/)
606 elsif(/^Deleted:\s*(.*)/)
608 if(exists($files->{$1}))
610 delete($files->{$1});
615 warn("$me: $NEWFILES: line $line not recognized\n");
620 if(length($file) && length($hash))
622 $files->{$file}={ hash=>$hash,
623 copyright=>$copyright,
625 license_text=>$license_text };
631 my %newnames = map { $_ => 1 } @filenames;
632 for my $file (sort keys(%$files))
634 unless(exists($newnames{$file}))
636 push(@deleted_files, $file);
641 print "Removed files:\n";
642 print join("\n", @deleted_files),"\n";
659 die("usage: $me [--scan] [--merge]\n",
660 "scans for changed copyright/licenses\n",
661 " -s|-scan Scan for new files & files with changed copyright headers\n",
662 " Writes to debian/clscan/new.txt for manual review.\n",
663 " -m|--merge Merges new data from debian/clscan/new.txt\n",
664 " -w|--write Writes updated debian/copyright.\n",
665 " --merge implies --write.\n");