Bug 21637: Fixed upercase letter in EasyAnalyticalRecords syspref
[koha.git] / misc / translator / tmpl_process3.pl
blobe39e64c3908e02d3b5bf777f79791db7a179c30a
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 use FindBin;
8 use lib $FindBin::Bin;
10 =head1 NAME
12 tmpl_process3.pl - Alternative version of tmpl_process.pl
13 using gettext-compatible translation files
15 =cut
17 use strict;
18 #use warnings; FIXME - Bug 2505
19 use File::Basename;
20 use Getopt::Long;
21 use Locale::PO;
22 use File::Temp qw( :POSIX );
23 use TmplTokenizer;
24 use VerboseWarnings qw( :warn :die );
26 ###############################################################################
28 use vars qw( @in_dirs @filenames @match @nomatch $str_file $out_dir $quiet );
29 use vars qw( @excludes $exclude_regex );
30 use vars qw( $recursive_p );
31 use vars qw( $pedantic_p );
32 use vars qw( $href );
33 use vars qw( $type ); # file extension (DOS form without the dot) to match
34 use vars qw( $charset_in $charset_out );
36 ###############################################################################
38 sub find_translation ($) {
39 my($s) = @_;
40 my $key = $s;
41 if ($s =~ /\S/s) {
42 $key = TmplTokenizer::string_canon($key);
43 $key = TmplTokenizer::charset_convert($key, $charset_in, $charset_out);
44 $key = TmplTokenizer::quote_po($key);
46 if (defined $href->{$key} && !$href->{$key}->fuzzy && length Locale::PO->dequote($href->{$key}->msgstr)){
47 if ($s =~ /^(\s+)/){
48 return $1 . Locale::PO->dequote($href->{$key}->msgstr);
50 else {
51 return Locale::PO->dequote($href->{$key}->msgstr);
54 else {
55 return $s;
59 sub text_replace_tag ($$) {
60 my($t, $attr) = @_;
61 my $it;
63 # value [tag=input], meta
64 my $tag = lc($1) if $t =~ /^<(\S+)/s;
65 my $translated_p = 0;
66 for my $a ('alt', 'content', 'title', 'value', 'label', 'placeholder') {
67 if ($attr->{$a}) {
68 next if $a eq 'label' && $tag ne 'optgroup';
69 next if $a eq 'content' && $tag ne 'meta';
70 next if $a eq 'value' && ($tag ne 'input' || (ref $attr->{'type'} && $attr->{'type'}->[1] =~ /^(?:checkbox|hidden|radio)$/)); # FIXME
72 my($key, $val, $val_orig, $order) = @{$attr->{$a}}; #FIXME
73 if ($val =~ /\S/s) {
74 my $s = find_translation($val);
75 if ($attr->{$a}->[1] ne $s) { #FIXME
76 $attr->{$a}->[1] = $s; # FIXME
77 $attr->{$a}->[2] = ($s =~ /"/s)? "'$s'": "\"$s\""; #FIXME
78 $translated_p = 1;
83 if ($translated_p) {
84 $it = "<$tag"
85 . join('', map { if ($_ ne '/'){
86 sprintf(' %s="%s"', $_, $attr->{$_}->[1]);
88 else {
89 sprintf(' %s',$_);
92 } sort {
93 $attr->{$a}->[3] <=> $attr->{$b}->[3] #FIXME
94 } keys %$attr);
95 $it .= '>';
97 else {
98 $it = $t;
100 return $it;
103 sub text_replace (**) {
104 my($h, $output) = @_;
105 for (;;) {
106 my $s = TmplTokenizer::next_token $h;
107 last unless defined $s;
108 my($kind, $t, $attr) = ($s->type, $s->string, $s->attributes);
109 if ($kind eq C4::TmplTokenType::TEXT) {
110 print $output find_translation($t);
111 } elsif ($kind eq C4::TmplTokenType::TEXT_PARAMETRIZED) {
112 my $fmt = find_translation($s->form);
113 print $output TmplTokenizer::parametrize($fmt, 1, $s, sub {
114 $_ = $_[0];
115 my($kind, $t, $attr) = ($_->type, $_->string, $_->attributes);
116 $kind == C4::TmplTokenType::TAG && %$attr?
117 text_replace_tag($t, $attr): $t });
118 } elsif ($kind eq C4::TmplTokenType::TAG && %$attr) {
119 print $output text_replace_tag($t, $attr);
120 } elsif ($s->has_js_data) {
121 for my $t (@{$s->js_data}) {
122 # FIXME for this whole block
123 if ($t->[0]) {
124 printf $output "%s%s%s", $t->[2], find_translation $t->[3],
125 $t->[2];
126 } else {
127 print $output $t->[1];
130 } elsif (defined $t) {
131 # Quick fix to bug 4472
132 $t = "<!DOCTYPE stylesheet [" if $t =~ /DOCTYPE stylesheet/ ;
133 print $output $t;
138 sub listfiles {
139 my($dir, $type, $action) = @_;
140 my $filenames = join ('|', @filenames); # used to update strings from this file
141 my $match = join ('|', @match); # use only this files
142 my $nomatch = join ('|', @nomatch); # do no use this files
143 my @it = ();
144 if (opendir(DIR, $dir)) {
145 my @dirent = readdir DIR; # because DIR is shared when recursing
146 closedir DIR;
147 for my $dirent (@dirent) {
148 my $path = "$dir/$dirent";
149 if ($dirent =~ /^\./ || $dirent eq 'CVS' || $dirent eq 'RCS'
150 || (defined $exclude_regex && $dirent =~ /^(?:$exclude_regex)$/)) {
152 } elsif (-f $path) {
153 my $basename = fileparse( $path );
154 push @it, $path
155 if ( not @filenames or $basename =~ /($filenames)/i )
156 and ( not @match or $basename =~ /($match)/i ) # files to include
157 and ( not @nomatch or $basename !~ /($nomatch)/i ) # files not to include
158 and (!defined $type || $dirent =~ /\.(?:$type)$/) || $action eq 'install';
159 } elsif (-d $path && $recursive_p) {
160 push @it, listfiles($path, $type, $action);
163 } else {
164 warn_normal "$dir: $!", undef;
166 return @it;
169 ###############################################################################
171 sub mkdir_recursive ($) {
172 my($dir) = @_;
173 local($`, $&, $', $1);
174 $dir = $` if $dir ne /^\/+$/ && $dir =~ /\/+$/;
175 my ($prefix, $basename) = ($dir =~ /\/([^\/]+)$/s)? ($`, $1): ('.', $dir);
176 mkdir_recursive($prefix) if $prefix ne '.' && !-d $prefix;
177 if (!-d $dir) {
178 print STDERR "Making directory $dir...\n" unless $quiet;
179 # creates with rwxrwxr-x permissions
180 mkdir($dir, 0775) || warn_normal "$dir: $!", undef;
184 ###############################################################################
186 sub usage ($) {
187 my($exitcode) = @_;
188 my $h = $exitcode? *STDERR: *STDOUT;
189 print $h <<EOF;
190 Usage: $0 create [OPTION]
191 or: $0 update [OPTION]
192 or: $0 install [OPTION]
193 or: $0 --help
194 Create or update PO files from templates, or install translated templates.
196 -i, --input=SOURCE Get or update strings from SOURCE directory(s).
197 On create or update can have multiple values.
198 On install only one value.
199 -o, --outputdir=DIRECTORY Install translation(s) to specified DIRECTORY
200 --pedantic-warnings Issue warnings even for detected problems
201 which are likely to be harmless
202 -r, --recursive SOURCE in the -i option is a directory
203 -f, --filename=FILE FILE is a specific filename or part of it.
204 If given, only these files will be processed.
205 On update only relevant strings will be updated.
206 -m, --match=FILE FILE is a specific filename or part of it.
207 If given, only these files will be processed.
208 -n, --nomatch=FILE FILE is a specific filename or part of it.
209 If given, these files will not be processed.
210 -s, --str-file=FILE Specify FILE as the translation (po) file
211 for input (install) or output (create, update)
212 -x, --exclude=REGEXP Exclude dirs matching the given REGEXP
213 --help Display this help and exit
214 -q, --quiet no output to screen (except for errors)
216 The -o option is ignored for the "create" and "update" actions.
217 Try `perldoc $0` for perhaps more information.
219 exit($exitcode);
222 ###############################################################################
224 sub usage_error (;$) {
225 for my $msg (split(/\n/, $_[0])) {
226 print STDERR "$msg\n";
228 print STDERR "Try `$0 --help for more information.\n";
229 exit(-1);
232 ###############################################################################
234 GetOptions(
235 'input|i=s' => \@in_dirs,
236 'filename|f=s' => \@filenames,
237 'match|m=s' => \@match,
238 'nomatch|n=s' => \@nomatch,
239 'outputdir|o=s' => \$out_dir,
240 'recursive|r' => \$recursive_p,
241 'str-file|s=s' => \$str_file,
242 'exclude|x=s' => \@excludes,
243 'quiet|q' => \$quiet,
244 'pedantic-warnings|pedantic' => sub { $pedantic_p = 1 },
245 'help' => \&usage,
246 ) || usage_error;
248 VerboseWarnings::set_application_name $0;
249 VerboseWarnings::set_pedantic_mode $pedantic_p;
251 # keep the buggy Locale::PO quiet if it says stupid things
252 $SIG{__WARN__} = sub {
253 my($s) = @_;
254 print STDERR $s unless $s =~ /^Strange line in [^:]+: #~/s
257 my $action = shift or usage_error('You must specify an ACTION.');
258 usage_error('You must at least specify input and string list filenames.')
259 if !@in_dirs || !defined $str_file;
261 # Type match defaults to *.tt plus *.inc if not specified
262 $type = "tt|inc|xsl|xml|def" if !defined($type);
264 # Check the inputs for being directories
265 for my $in_dir ( @in_dirs ) {
266 usage_error("$in_dir: Input must be a directory.\n"
267 . "(Symbolic links are not supported at the moment)")
268 unless -d $in_dir;
271 # Generates the global exclude regular expression
272 $exclude_regex = '(?:'.join('|', @excludes).')' if @excludes;
274 my @in_files;
275 # Generate the list of input files if a directory is specified
276 # input is a directory, generates list of files to process
278 for my $fn ( @filenames ) {
279 die "You cannot specify input files and directories at the same time.\n"
280 if -d $fn;
282 for my $in_dir ( @in_dirs ) {
283 $in_dir =~ s/\/$//; # strips the trailing / if any
284 @in_files = ( @in_files, listfiles($in_dir, $type, $action));
287 # restores the string list from file
288 $href = Locale::PO->load_file_ashash($str_file);
290 # guess the charsets. HTML::Templates defaults to iso-8859-1
291 if (defined $href) {
292 die "$str_file: PO file is corrupted, or not a PO file\n" unless defined $href->{'""'};
293 $charset_out = TmplTokenizer::charset_canon $2 if $href->{'""'}->msgstr =~ /\bcharset=(["']?)([^;\s"'\\]+)\1/;
294 $charset_in = $charset_out;
295 # for my $msgid (keys %$href) {
296 # if ($msgid =~ /\bcharset=(["']?)([^;\s"'\\]+)\1/) {
297 # my $candidate = TmplTokenizer::charset_canon $2;
298 # die "Conflicting charsets in msgid: $charset_in vs $candidate => $msgid\n"
299 # if defined $charset_in && $charset_in ne $candidate;
300 # $charset_in = $candidate;
304 # BUG6464: check consistency of PO messages
305 # - count number of '%s' in msgid and msgstr
306 for my $msg ( values %$href ) {
307 my $id_count = split(/%s/, $msg->{msgid}) - 1;
308 my $str_count = split(/%s/, $msg->{msgstr}) - 1;
309 next if $id_count == $str_count ||
310 $msg->{msgstr} eq '""' ||
311 grep { /fuzzy/ } @{$msg->{_flags}};
312 warn_normal
313 "unconsistent %s count: ($id_count/$str_count):\n" .
314 " line: " . $msg->{loaded_line_number} . "\n" .
315 " msgid: " . $msg->{msgid} . "\n" .
316 " msgstr: " . $msg->{msgstr} . "\n", undef;
320 # set our charset in to UTF-8
321 if (!defined $charset_in) {
322 $charset_in = TmplTokenizer::charset_canon 'UTF-8';
323 warn "Warning: Can't determine original templates' charset, defaulting to $charset_in\n";
325 # set our charset out to UTF-8
326 if (!defined $charset_out) {
327 $charset_out = TmplTokenizer::charset_canon 'UTF-8';
328 warn "Warning: Charset Out defaulting to $charset_out\n";
330 my $xgettext = './xgettext.pl'; # actual text extractor script
331 my $st;
333 if ($action eq 'create') {
334 # updates the list. As the list is empty, every entry will be added
335 if (!-s $str_file) {
336 warn "Removing empty file $str_file\n";
337 unlink $str_file || die "$str_file: $!\n";
339 die "$str_file: Output file already exists\n" if -f $str_file;
340 my($tmph1, $tmpfile1) = tmpnam();
341 my($tmph2, $tmpfile2) = tmpnam();
342 close $tmph2; # We just want a name
343 # Generate the temporary file that acts as <MODULE>/POTFILES.in
344 for my $input (@in_files) {
345 print $tmph1 "$input\n";
347 close $tmph1;
348 warn "I $charset_in O $charset_out";
349 # Generate the specified po file ($str_file)
350 $st = system ($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2,
351 (defined $charset_in? ('-I', $charset_in): ()),
352 (defined $charset_out? ('-O', $charset_out): ())
354 # Run msgmerge so that the pot file looks like a real pot file
355 # We need to help msgmerge a bit by pre-creating a dummy po file that has
356 # the headers and the "" msgid & msgstr. It will fill in the rest.
357 if ($st == 0) {
358 # Merge the temporary "pot file" with the specified po file ($str_file)
359 # FIXME: msgmerge(1) is a Unix dependency
360 # FIXME: need to check the return value
361 unless (-f $str_file) {
362 local(*INPUT, *OUTPUT);
363 open(INPUT, "<$tmpfile2");
364 open(OUTPUT, ">$str_file");
365 while (<INPUT>) {
366 print OUTPUT;
367 last if /^\n/s;
369 close INPUT;
370 close OUTPUT;
372 $st = system("msgmerge ".($quiet?'-q':'')." -s $str_file $tmpfile2 -o - | msgattrib --no-obsolete -o $str_file");
373 } else {
374 error_normal "Text extraction failed: $xgettext: $!\n", undef;
375 error_additional "Will not run msgmerge\n", undef;
377 unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
378 unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
380 } elsif ($action eq 'update') {
381 my($tmph1, $tmpfile1) = tmpnam();
382 my($tmph2, $tmpfile2) = tmpnam();
383 close $tmph2; # We just want a name
384 # Generate the temporary file that acts as <MODULE>/POTFILES.in
385 for my $input (@in_files) {
386 print $tmph1 "$input\n";
388 close $tmph1;
389 # Generate the temporary file that acts as <MODULE>/<LANG>.pot
390 $st = system($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2,
391 '--po-mode',
392 (defined $charset_in? ('-I', $charset_in): ()),
393 (defined $charset_out? ('-O', $charset_out): ()));
394 if ($st == 0) {
395 # Merge the temporary "pot file" with the specified po file ($str_file)
396 # FIXME: msgmerge(1) is a Unix dependency
397 # FIXME: need to check the return value
398 if ( @filenames ) {
399 my ($tmph3, $tmpfile3) = tmpnam();
400 $st = system("msgcat $str_file $tmpfile2 > $tmpfile3");
401 $st = system("msgmerge ".($quiet?'-q':'')." -s $str_file $tmpfile3 -o - | msgattrib --no-obsolete -o $str_file")
402 unless $st;
403 } else {
404 $st = system("msgmerge ".($quiet?'-q':'')." -s $str_file $tmpfile2 -o - | msgattrib --no-obsolete -o $str_file");
406 } else {
407 error_normal "Text extraction failed: $xgettext: $!\n", undef;
408 error_additional "Will not run msgmerge\n", undef;
410 unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
411 unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
413 } elsif ($action eq 'install') {
414 if(!defined($out_dir)) {
415 usage_error("You must specify an output directory when using the install method.");
418 if ( scalar @in_dirs > 1 ) {
419 usage_error("You must specify only one input directory when using the install method.");
422 my $in_dir = shift @in_dirs;
424 if ($in_dir eq $out_dir) {
425 warn "You must specify a different input and output directory.\n";
426 exit -1;
429 # Make sure the output directory exists
430 # (It will auto-create it, but for compatibility we should not)
431 -d $out_dir || die "$out_dir: The directory does not exist\n";
433 # Try to open the file, because Locale::PO doesn't check :-/
434 open(INPUT, "<$str_file") || die "$str_file: $!\n";
435 close INPUT;
437 # creates the new tmpl file using the new translation
438 for my $input (@in_files) {
439 die "Assertion failed"
440 unless substr($input, 0, length($in_dir) + 1) eq "$in_dir/";
442 my $target = $out_dir . substr($input, length($in_dir));
443 my $targetdir = $` if $target =~ /[^\/]+$/s;
445 if (!defined $type || $input =~ /\.(?:$type)$/) {
446 my $h = TmplTokenizer->new( $input );
447 $h->set_allow_cformat( 1 );
448 VerboseWarnings::set_input_file_name $input;
449 mkdir_recursive($targetdir) unless -d $targetdir;
450 print STDERR "Creating $target...\n" unless $quiet;
451 open( OUTPUT, ">$target" ) || die "$target: $!\n";
452 text_replace( $h, *OUTPUT );
453 close OUTPUT;
454 } else {
455 # just copying the file
456 mkdir_recursive($targetdir) unless -d $targetdir;
457 system("cp -f $input $target");
458 print STDERR "Copying $input...\n" unless $quiet;
462 } else {
463 usage_error('Unknown action specified.');
466 if ($st == 0) {
467 printf "The %s seems to be successful.\n", $action unless $quiet;
468 } else {
469 printf "%s FAILED.\n", "\u$action" unless $quiet;
471 exit 0;
473 ###############################################################################
475 =head1 SYNOPSIS
477 ./tmpl_process3.pl [ I<tmpl_process.pl options> ]
479 =head1 DESCRIPTION
481 This is an alternative version of the tmpl_process.pl script,
482 using standard gettext-style PO files. While there still might
483 be changes made to the way it extracts strings, at this moment
484 it should be stable enough for general use; it is already being
485 used for the Chinese and Polish translations.
487 Currently, the create, update, and install actions have all been
488 reimplemented and seem to work.
490 =head2 Features
492 =over
494 =item -
496 Translation files in standard Uniforum PO format.
497 All standard tools including all gettext tools,
498 plus PO file editors like kbabel(1) etc.
499 can be used.
501 =item -
503 Minor changes in whitespace in source templates
504 do not generally require strings to be re-translated.
506 =item -
508 Able to handle <TMPL_VAR> variables in the templates;
509 <TMPL_VAR> variables are usually extracted in proper context,
510 represented by a short %s placeholder.
512 =item -
514 Able to handle text input and radio button INPUT elements
515 in the templates; these INPUT elements are also usually
516 extracted in proper context,
517 represented by a short %S or %p placeholder.
519 =item -
521 Automatic comments in the generated PO files to provide
522 even more context (line numbers, and the names and types
523 of the variables).
525 =item -
527 The %I<n>$s (or %I<n>$p, etc.) notation can be used
528 for change the ordering of the variables,
529 if such a reordering is required for correct translation.
531 =item -
533 If a particular <TMPL_VAR> should not appear in the
534 translation, it can be suppressed with the %0.0s notation.
536 =item -
538 Using the PO format also means translators can add their
539 own comments in the translation files, if necessary.
541 =item -
543 Create, update, and install actions are all based on the
544 same scanner module. This ensures that update and install
545 have the same idea of what is a translatable string;
546 attribute names in tags, for example, will not be
547 accidentally translated.
549 =back
551 =head1 NOTES
553 Anchors are represented by an <AI<n>> notation.
554 The meaning of this non-standard notation might not be obvious.
556 The create action calls xgettext.pl to do the actual work;
557 the update action calls xgettext.pl, msgmerge(1) and msgattrib(1)
558 to do the actual work.
560 =head1 BUGS
562 xgettext.pl must be present in the current directory; both
563 msgmerge(1) and msgattrib(1) must also be present in the search path.
564 The script currently does not check carefully whether these
565 dependent commands are present.
567 Locale::PO(3) has a lot of bugs. It can neither parse nor
568 generate GNU PO files properly; a couple of workarounds have
569 been written in TmplTokenizer and more is likely to be needed
570 (e.g., to get rid of the "Strange line" warning for #~).
572 This script may not work in Windows.
574 There are probably some other bugs too, since this has not been
575 tested very much.
577 =head1 SEE ALSO
579 xgettext.pl,
580 TmplTokenizer.pm,
581 msgmerge(1),
582 Locale::PO(3),
583 translator_doc.txt
585 http://www.saas.nsw.edu.au/koha_wiki/index.php?page=DifficultTerms
587 =cut