corrected init script name for apache2
[koha.git] / misc / translator / tmpl_process3.pl
bloba944647e844b3ff45676b05c8c57bb0e8a7fca4c
1 #!/usr/bin/perl
2 # This file is part of Koha
3 # Parts copyright 2003-2004 Paul Poulain
4 # Parts copyright 2003-2004 Jerome Vizcaino
5 # Parts copyright 2004 Ambrose Li
7 =head1 NAME
9 tmpl_process3.pl - Alternative version of tmpl_process.pl
10 using gettext-compatible translation files
12 =cut
14 use strict;
15 use Getopt::Long;
16 use Locale::PO;
17 use File::Temp qw( :POSIX );
18 use TmplTokenizer;
19 use VerboseWarnings qw( :warn :die );
21 ###############################################################################
23 use vars qw( @in_files $in_dir $str_file $out_dir $quiet );
24 use vars qw( @excludes $exclude_regex );
25 use vars qw( $recursive_p );
26 use vars qw( $pedantic_p );
27 use vars qw( $href );
28 use vars qw( $type ); # file extension (DOS form without the dot) to match
29 use vars qw( $charset_in $charset_out );
31 ###############################################################################
33 sub find_translation ($) {
34 my($s) = @_;
35 my $key = $s;
36 if ($s =~ /\S/s) {
37 $key = TmplTokenizer::string_canon($key);
38 $key = TmplTokenizer::charset_convert($key, $charset_in, $charset_out);
39 $key = TmplTokenizer::quote_po($key);
41 return defined $href->{$key}
42 && !$href->{$key}->fuzzy
43 && length Locale::PO->dequote($href->{$key}->msgstr)?
44 Locale::PO->dequote($href->{$key}->msgstr): $s;
47 sub text_replace_tag ($$) {
48 my($t, $attr) = @_;
49 my $it;
50 # value [tag=input], meta
51 my $tag = lc($1) if $t =~ /^<(\S+)/s;
52 my $translated_p = 0;
53 for my $a ('alt', 'content', 'title', 'value') {
54 if ($attr->{$a}) {
55 next if $a eq 'content' && $tag ne 'meta';
56 next if $a eq 'value' && ($tag ne 'input'
57 || (ref $attr->{'type'} && $attr->{'type'}->[1] =~ /^(?:hidden|radio|text)$/)); # FIXME
58 my($key, $val, $val_orig, $order) = @{$attr->{$a}}; #FIXME
59 if ($val =~ /\S/s) {
60 my $s = find_translation($val);
61 if ($attr->{$a}->[1] ne $s) { #FIXME
62 $attr->{$a}->[1] = $s; # FIXME
63 $attr->{$a}->[2] = ($s =~ /"/s)? "'$s'": "\"$s\""; #FIXME
64 $translated_p = 1;
69 if ($translated_p) {
70 $it = "<$tag"
71 . join('', map {
72 sprintf(' %s=%s', $_, $attr->{$_}->[2]) #FIXME
73 } sort {
74 $attr->{$a}->[3] <=> $attr->{$b}->[3] #FIXME
75 } keys %$attr)
76 . '>';
77 } else {
78 $it = $t;
80 return $it;
83 sub text_replace (**) {
84 my($h, $output) = @_;
85 for (;;) {
86 my $s = TmplTokenizer::next_token $h;
87 last unless defined $s;
88 my($kind, $t, $attr) = ($s->type, $s->string, $s->attributes);
89 if ($kind eq TmplTokenType::TEXT) {
90 print $output find_translation($t);
91 } elsif ($kind eq TmplTokenType::TEXT_PARAMETRIZED) {
92 my $fmt = find_translation($s->form);
93 print $output TmplTokenizer::parametrize($fmt, 1, $s, sub {
94 $_ = $_[0];
95 my($kind, $t, $attr) = ($_->type, $_->string, $_->attributes);
96 $kind == TmplTokenType::TAG && %$attr?
97 text_replace_tag($t, $attr): $t });
98 } elsif ($kind eq TmplTokenType::TAG && %$attr) {
99 print $output text_replace_tag($t, $attr);
100 } elsif ($s->has_js_data) {
101 for my $t (@{$s->js_data}) {
102 # FIXME for this whole block
103 if ($t->[0]) {
104 printf $output "%s%s%s", $t->[2], find_translation $t->[3],
105 $t->[2];
106 } else {
107 print $output $t->[1];
110 } elsif (defined $t) {
111 print $output $t;
116 sub listfiles ($$$) {
117 my($dir, $type, $action) = @_;
118 my @it = ();
119 if (opendir(DIR, $dir)) {
120 my @dirent = readdir DIR; # because DIR is shared when recursing
121 closedir DIR;
122 for my $dirent (@dirent) {
123 my $path = "$dir/$dirent";
124 if ($dirent =~ /^\./ || $dirent eq 'CVS' || $dirent eq 'RCS'
125 || (defined $exclude_regex && $dirent =~ /^(?:$exclude_regex)$/)) {
127 } elsif (-f $path) {
128 push @it, $path if (!defined $type || $dirent =~ /\.(?:$type)$/) || $action eq 'install';
129 } elsif (-d $path && $recursive_p) {
130 push @it, listfiles($path, $type, $action);
133 } else {
134 warn_normal "$dir: $!", undef;
136 return @it;
139 ###############################################################################
141 sub mkdir_recursive ($) {
142 my($dir) = @_;
143 local($`, $&, $', $1);
144 $dir = $` if $dir ne /^\/+$/ && $dir =~ /\/+$/;
145 my ($prefix, $basename) = ($dir =~ /\/([^\/]+)$/s)? ($`, $1): ('.', $dir);
146 mkdir_recursive($prefix) if $prefix ne '.' && !-d $prefix;
147 if (!-d $dir) {
148 print STDERR "Making directory $dir..." unless $quiet;
149 # creates with rwxrwxr-x permissions
150 mkdir($dir, 0775) || warn_normal "$dir: $!", undef;
154 ###############################################################################
156 sub usage ($) {
157 my($exitcode) = @_;
158 my $h = $exitcode? *STDERR: *STDOUT;
159 print $h <<EOF;
160 Usage: $0 create [OPTION]
161 or: $0 update [OPTION]
162 or: $0 install [OPTION]
163 or: $0 --help
164 Create or update PO files from templates, or install translated templates.
166 -i, --input=SOURCE Get or update strings from SOURCE file.
167 SOURCE is a directory if -r is also specified.
168 -o, --outputdir=DIRECTORY Install translation(s) to specified DIRECTORY
169 --pedantic-warnings Issue warnings even for detected problems
170 which are likely to be harmless
171 -r, --recursive SOURCE in the -i option is a directory
172 -s, --str-file=FILE Specify FILE as the translation (po) file
173 for input (install) or output (create, update)
174 -x, --exclude=REGEXP Exclude files matching the given REGEXP
175 --help Display this help and exit
176 -q, --quiet no output to screen (except for errors)
178 The -o option is ignored for the "create" and "update" actions.
179 Try `perldoc $0 for perhaps more information.
181 exit($exitcode);
184 ###############################################################################
186 sub usage_error (;$) {
187 for my $msg (split(/\n/, $_[0])) {
188 print STDERR "$msg\n";
190 print STDERR "Try `$0 --help for more information.\n";
191 exit(-1);
194 ###############################################################################
196 GetOptions(
197 'input|i=s' => \@in_files,
198 'outputdir|o=s' => \$out_dir,
199 'recursive|r' => \$recursive_p,
200 'str-file|s=s' => \$str_file,
201 'exclude|x=s' => \@excludes,
202 'quiet|q' => \$quiet,
203 'pedantic-warnings|pedantic' => sub { $pedantic_p = 1 },
204 'help' => \&usage,
205 ) || usage_error;
207 VerboseWarnings::set_application_name $0;
208 VerboseWarnings::set_pedantic_mode $pedantic_p;
210 # keep the buggy Locale::PO quiet if it says stupid things
211 $SIG{__WARN__} = sub {
212 my($s) = @_;
213 print STDERR $s unless $s =~ /^Strange line in [^:]+: #~/s
216 my $action = shift or usage_error('You must specify an ACTION.');
217 usage_error('You must at least specify input and string list filenames.')
218 if !@in_files || !defined $str_file;
220 # Type match defaults to *.tmpl plus *.inc if not specified
221 $type = "tmpl|inc" if !defined($type);
223 # Check the inputs for being files or directories
224 for my $input (@in_files) {
225 usage_error("$input: Input must be a file or directory.\n"
226 . "(Symbolic links are not supported at the moment)")
227 unless -d $input || -f $input;;
230 # Generates the global exclude regular expression
231 $exclude_regex = '(?:'.join('|', @excludes).')' if @excludes;
233 # Generate the list of input files if a directory is specified
234 if (-d $in_files[0]) {
235 die "If you specify a directory as input, you must specify only it.\n"
236 if @in_files > 1;
238 # input is a directory, generates list of files to process
239 $in_dir = $in_files[0];
240 $in_dir =~ s/\/$//; # strips the trailing / if any
241 @in_files = listfiles($in_dir, $type, $action);
242 } else {
243 for my $input (@in_files) {
244 die "You cannot specify input files and directories at the same time.\n"
245 unless -f $input;
249 # restores the string list from file
250 $href = Locale::PO->load_file_ashash($str_file);
252 # guess the charsets. HTML::Templates defaults to iso-8859-1
253 if (defined $href) {
254 die "$str_file: PO file is corrupted, or not a PO file\n"
255 unless defined $href->{'""'};
256 $charset_out = TmplTokenizer::charset_canon $2
257 if $href->{'""'}->msgstr =~ /\bcharset=(["']?)([^;\s"'\\]+)\1/;
258 $charset_in = 'utf-8';
259 # for my $msgid (keys %$href) {
260 # if ($msgid =~ /\bcharset=(["']?)([^;\s"'\\]+)\1/) {
261 # my $candidate = TmplTokenizer::charset_canon $2;
262 # die "Conflicting charsets in msgid: $charset_in vs $candidate => $msgid\n"
263 # if defined $charset_in && $charset_in ne $candidate;
264 # $charset_in = $candidate;
268 if (!defined $charset_in) {
269 $charset_in = TmplTokenizer::charset_canon 'utf-8';
270 warn "Warning: Can't determine original templates' charset, defaulting to $charset_in\n";
273 my $xgettext = './xgettext.pl'; # actual text extractor script
274 my $st;
276 if ($action eq 'create') {
277 # updates the list. As the list is empty, every entry will be added
278 if (!-s $str_file) {
279 warn "Removing empty file $str_file\n";
280 unlink $str_file || die "$str_file: $!\n";
282 die "$str_file: Output file already exists\n" if -f $str_file;
283 my($tmph1, $tmpfile1) = tmpnam();
284 my($tmph2, $tmpfile2) = tmpnam();
285 close $tmph2; # We just want a name
286 # Generate the temporary file that acts as <MODULE>/POTFILES.in
287 for my $input (@in_files) {
288 print $tmph1 "$input\n";
290 close $tmph1;
291 # Generate the specified po file ($str_file)
292 $st = system ($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2);
293 # Run msgmerge so that the pot file looks like a real pot file
294 # We need to help msgmerge a bit by pre-creating a dummy po file that has
295 # the headers and the "" msgid & msgstr. It will fill in the rest.
296 if ($st == 0) {
297 # Merge the temporary "pot file" with the specified po file ($str_file)
298 # FIXME: msgmerge(1) is a Unix dependency
299 # FIXME: need to check the return value
300 unless (-f $str_file) {
301 local(*INPUT, *OUTPUT);
302 open(INPUT, "<$tmpfile2");
303 open(OUTPUT, ">$str_file");
304 while (<INPUT>) {
305 print OUTPUT;
306 last if /^\n/s;
308 close INPUT;
309 close OUTPUT;
311 $st = system('msgmerge', '-U', '-s', $str_file, $tmpfile2);
312 } else {
313 error_normal "Text extraction failed: $xgettext: $!\n", undef;
314 error_additional "Will not run msgmerge\n", undef;
316 # unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
317 # unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
319 } elsif ($action eq 'update') {
320 my($tmph1, $tmpfile1) = tmpnam();
321 my($tmph2, $tmpfile2) = tmpnam();
322 close $tmph2; # We just want a name
323 # Generate the temporary file that acts as <MODULE>/POTFILES.in
324 for my $input (@in_files) {
325 print $tmph1 "$input\n";
327 close $tmph1;
328 # Generate the temporary file that acts as <MODULE>/<LANG>.pot
329 $st = system($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2,
330 '--po-mode',
331 (defined $charset_in? ('-I', $charset_in): ()),
332 (defined $charset_out? ('-O', $charset_out): ()));
333 if ($st == 0) {
334 # Merge the temporary "pot file" with the specified po file ($str_file)
335 # FIXME: msgmerge(1) is a Unix dependency
336 # FIXME: need to check the return value
337 $st = system('msgmerge', '-U', '-s', $str_file, $tmpfile2);
338 } else {
339 error_normal "Text extraction failed: $xgettext: $!\n", undef;
340 error_additional "Will not run msgmerge\n", undef;
342 # unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
343 # unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
345 } elsif ($action eq 'install') {
346 if(!defined($out_dir)) {
347 usage_error("You must specify an output directory when using the install method.");
350 if ($in_dir eq $out_dir) {
351 warn "You must specify a different input and output directory.\n";
352 exit -1;
355 # Make sure the output directory exists
356 # (It will auto-create it, but for compatibility we should not)
357 -d $out_dir || die "$out_dir: The directory does not exist\n";
359 # Try to open the file, because Locale::PO doesn't check :-/
360 open(INPUT, "<$str_file") || die "$str_file: $!\n";
361 close INPUT;
363 # creates the new tmpl file using the new translation
364 for my $input (@in_files) {
365 die "Assertion failed"
366 unless substr($input, 0, length($in_dir) + 1) eq "$in_dir/";
367 # print "$input / $type\n";
368 if (!defined $type || $input =~ /\.(?:$type)$/) {
369 my $h = TmplTokenizer->new( $input );
370 $h->set_allow_cformat( 1 );
371 VerboseWarnings::set_input_file_name $input;
373 my $target = $out_dir . substr($input, length($in_dir));
374 my $targetdir = $` if $target =~ /[^\/]+$/s;
375 mkdir_recursive($targetdir) unless -d $targetdir;
376 print STDERR "Creating $target...\n" unless $quiet;
377 open( OUTPUT, ">$target" ) || die "$target: $!\n";
378 text_replace( $h, *OUTPUT );
379 close OUTPUT;
380 } else {
381 # just copying the file
382 my $target = $out_dir . substr($input, length($in_dir));
383 my $targetdir = $` if $target =~ /[^\/]+$/s;
384 mkdir_recursive($targetdir) unless -d $targetdir;
385 system("cp -f $input $target");
386 print STDERR "Copying $input...\n" unless $quiet;
390 } else {
391 usage_error('Unknown action specified.');
394 if ($st == 0) {
395 printf "The %s seems to be successful.\n", $action unless $quiet;
396 } else {
397 printf "%s FAILED.\n", "\u$action" unless $quiet;
399 exit 0;
401 ###############################################################################
403 =head1 SYNOPSIS
405 ./tmpl_process3.pl [ I<tmpl_process.pl options> ]
407 =head1 DESCRIPTION
409 This is an alternative version of the tmpl_process.pl script,
410 using standard gettext-style PO files. While there still might
411 be changes made to the way it extracts strings, at this moment
412 it should be stable enough for general use; it is already being
413 used for the Chinese and Polish translations.
415 Currently, the create, update, and install actions have all been
416 reimplemented and seem to work.
418 =head2 Features
420 =over
422 =item -
424 Translation files in standard Uniforum PO format.
425 All standard tools including all gettext tools,
426 plus PO file editors like kbabel(1) etc.
427 can be used.
429 =item -
431 Minor changes in whitespace in source templates
432 do not generally require strings to be re-translated.
434 =item -
436 Able to handle <TMPL_VAR> variables in the templates;
437 <TMPL_VAR> variables are usually extracted in proper context,
438 represented by a short %s placeholder.
440 =item -
442 Able to handle text input and radio button INPUT elements
443 in the templates; these INPUT elements are also usually
444 extracted in proper context,
445 represented by a short %S or %p placeholder.
447 =item -
449 Automatic comments in the generated PO files to provide
450 even more context (line numbers, and the names and types
451 of the variables).
453 =item -
455 The %I<n>$s (or %I<n>$p, etc.) notation can be used
456 for change the ordering of the variables,
457 if such a reordering is required for correct translation.
459 =item -
461 If a particular <TMPL_VAR> should not appear in the
462 translation, it can be suppressed with the %0.0s notation.
464 =item -
466 Using the PO format also means translators can add their
467 own comments in the translation files, if necessary.
469 =item -
471 Create, update, and install actions are all based on the
472 same scanner module. This ensures that update and install
473 have the same idea of what is a translatable string;
474 attribute names in tags, for example, will not be
475 accidentally translated.
477 =back
479 =head1 NOTES
481 Anchors are represented by an <AI<n>> notation.
482 The meaning of this non-standard notation might not be obvious.
484 The create action calls xgettext.pl to do the actual work;
485 the update action calls xgettext.pl and msgmerge(1) to do the
486 actual work.
488 =head1 BUGS
490 xgettext.pl must be present in the current directory; the
491 msgmerge(1) command must also be present in the search path.
492 The script currently does not check carefully whether these
493 dependent commands are present.
495 Locale::PO(3) has a lot of bugs. It can neither parse nor
496 generate GNU PO files properly; a couple of workarounds have
497 been written in TmplTokenizer and more is likely to be needed
498 (e.g., to get rid of the "Strange line" warning for #~).
500 This script may not work in Windows.
502 There are probably some other bugs too, since this has not been
503 tested very much.
505 =head1 SEE ALSO
507 xgettext.pl,
508 TmplTokenizer.pm,
509 msgmerge(1),
510 Locale::PO(3),
511 translator_doc.txt
513 http://www.saas.nsw.edu.au/koha_wiki/index.php?page=DifficultTerms
515 =cut