MT2116: Addons to the CSV export
[koha.git] / misc / translator / tmpl_process3.pl
blob80325d077fa3edc6c7f884bdd1728ef6633b1ea0
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 ($$) {
48 my($t, $attr) = @_;
49 my $it;
50 # value [tag=input], meta
51 my $tag = lc($1) if $t =~ /^<(\S+)/s;
52 my $translated_p = 0;
53 for my $a ('alt', 'content', 'title', 'value','label') {
54 if ($attr->{$a}) {
55 next if $a eq 'label' && $tag ne 'optgroup';
56 next if $a eq 'content' && $tag ne 'meta';
57 next if $a eq 'value' && ($tag ne 'input'
58 || (ref $attr->{'type'} && $attr->{'type'}->[1] =~ /^(?:checkbox|hidden|radio|text)$/)); # FIXME
59 my($key, $val, $val_orig, $order) = @{$attr->{$a}}; #FIXME
60 if ($val =~ /\S/s) {
61 my $s = find_translation($val);
62 if ($attr->{$a}->[1] ne $s) { #FIXME
63 $attr->{$a}->[1] = $s; # FIXME
64 $attr->{$a}->[2] = ($s =~ /"/s)? "'$s'": "\"$s\""; #FIXME
65 $translated_p = 1;
70 if ($translated_p) {
71 $it = "<$tag"
72 . join('', map {
73 sprintf(' %s=%s', $_, $attr->{$_}->[2]) #FIXME
74 } sort {
75 $attr->{$a}->[3] <=> $attr->{$b}->[3] #FIXME
76 } keys %$attr)
77 . '>';
78 } else {
79 $it = $t;
81 return $it;
84 sub text_replace (**) {
85 my($h, $output) = @_;
86 for (;;) {
87 my $s = TmplTokenizer::next_token $h;
88 last unless defined $s;
89 my($kind, $t, $attr) = ($s->type, $s->string, $s->attributes);
90 if ($kind eq TmplTokenType::TEXT) {
91 print $output find_translation($t);
92 } elsif ($kind eq TmplTokenType::TEXT_PARAMETRIZED) {
93 my $fmt = find_translation($s->form);
94 print $output TmplTokenizer::parametrize($fmt, 1, $s, sub {
95 $_ = $_[0];
96 my($kind, $t, $attr) = ($_->type, $_->string, $_->attributes);
97 $kind == TmplTokenType::TAG && %$attr?
98 text_replace_tag($t, $attr): $t });
99 } elsif ($kind eq TmplTokenType::TAG && %$attr) {
100 print $output text_replace_tag($t, $attr);
101 } elsif ($s->has_js_data) {
102 for my $t (@{$s->js_data}) {
103 # FIXME for this whole block
104 if ($t->[0]) {
105 printf $output "%s%s%s", $t->[2], find_translation $t->[3],
106 $t->[2];
107 } else {
108 print $output $t->[1];
111 } elsif (defined $t) {
112 print $output $t;
117 sub listfiles ($$$) {
118 my($dir, $type, $action) = @_;
119 my @it = ();
120 if (opendir(DIR, $dir)) {
121 my @dirent = readdir DIR; # because DIR is shared when recursing
122 closedir DIR;
123 for my $dirent (@dirent) {
124 my $path = "$dir/$dirent";
125 if ($dirent =~ /^\./ || $dirent eq 'CVS' || $dirent eq 'RCS'
126 || (defined $exclude_regex && $dirent =~ /^(?:$exclude_regex)$/)) {
128 } elsif (-f $path) {
129 push @it, $path if (!defined $type || $dirent =~ /\.(?:$type)$/) || $action eq 'install';
130 } elsif (-d $path && $recursive_p) {
131 push @it, listfiles($path, $type, $action);
134 } else {
135 warn_normal "$dir: $!", undef;
137 return @it;
140 ###############################################################################
142 sub mkdir_recursive ($) {
143 my($dir) = @_;
144 local($`, $&, $', $1);
145 $dir = $` if $dir ne /^\/+$/ && $dir =~ /\/+$/;
146 my ($prefix, $basename) = ($dir =~ /\/([^\/]+)$/s)? ($`, $1): ('.', $dir);
147 mkdir_recursive($prefix) if $prefix ne '.' && !-d $prefix;
148 if (!-d $dir) {
149 print STDERR "Making directory $dir..." unless $quiet;
150 # creates with rwxrwxr-x permissions
151 mkdir($dir, 0775) || warn_normal "$dir: $!", undef;
155 ###############################################################################
157 sub usage ($) {
158 my($exitcode) = @_;
159 my $h = $exitcode? *STDERR: *STDOUT;
160 print $h <<EOF;
161 Usage: $0 create [OPTION]
162 or: $0 update [OPTION]
163 or: $0 install [OPTION]
164 or: $0 --help
165 Create or update PO files from templates, or install translated templates.
167 -i, --input=SOURCE Get or update strings from SOURCE file.
168 SOURCE is a directory if -r is also specified.
169 -o, --outputdir=DIRECTORY Install translation(s) to specified DIRECTORY
170 --pedantic-warnings Issue warnings even for detected problems
171 which are likely to be harmless
172 -r, --recursive SOURCE in the -i option is a directory
173 -s, --str-file=FILE Specify FILE as the translation (po) file
174 for input (install) or output (create, update)
175 -x, --exclude=REGEXP Exclude files matching the given REGEXP
176 --help Display this help and exit
177 -q, --quiet no output to screen (except for errors)
179 The -o option is ignored for the "create" and "update" actions.
180 Try `perldoc $0 for perhaps more information.
182 exit($exitcode);
185 ###############################################################################
187 sub usage_error (;$) {
188 for my $msg (split(/\n/, $_[0])) {
189 print STDERR "$msg\n";
191 print STDERR "Try `$0 --help for more information.\n";
192 exit(-1);
195 ###############################################################################
197 GetOptions(
198 'input|i=s' => \@in_files,
199 'outputdir|o=s' => \$out_dir,
200 'recursive|r' => \$recursive_p,
201 'str-file|s=s' => \$str_file,
202 'exclude|x=s' => \@excludes,
203 'quiet|q' => \$quiet,
204 'pedantic-warnings|pedantic' => sub { $pedantic_p = 1 },
205 'help' => \&usage,
206 ) || usage_error;
208 VerboseWarnings::set_application_name $0;
209 VerboseWarnings::set_pedantic_mode $pedantic_p;
211 # keep the buggy Locale::PO quiet if it says stupid things
212 $SIG{__WARN__} = sub {
213 my($s) = @_;
214 print STDERR $s unless $s =~ /^Strange line in [^:]+: #~/s
217 my $action = shift or usage_error('You must specify an ACTION.');
218 usage_error('You must at least specify input and string list filenames.')
219 if !@in_files || !defined $str_file;
221 # Type match defaults to *.tmpl plus *.inc if not specified
222 $type = "tmpl|inc|xsl" if !defined($type);
224 # Check the inputs for being files or directories
225 for my $input (@in_files) {
226 usage_error("$input: Input must be a file or directory.\n"
227 . "(Symbolic links are not supported at the moment)")
228 unless -d $input || -f $input;;
231 # Generates the global exclude regular expression
232 $exclude_regex = '(?:'.join('|', @excludes).')' if @excludes;
234 # Generate the list of input files if a directory is specified
235 if (-d $in_files[0]) {
236 die "If you specify a directory as input, you must specify only it.\n"
237 if @in_files > 1;
239 # input is a directory, generates list of files to process
240 $in_dir = $in_files[0];
241 $in_dir =~ s/\/$//; # strips the trailing / if any
242 @in_files = listfiles($in_dir, $type, $action);
243 } else {
244 for my $input (@in_files) {
245 die "You cannot specify input files and directories at the same time.\n"
246 unless -f $input;
250 # restores the string list from file
251 $href = Locale::PO->load_file_ashash($str_file);
253 # guess the charsets. HTML::Templates defaults to iso-8859-1
254 if (defined $href) {
255 die "$str_file: PO file is corrupted, or not a PO file\n" unless defined $href->{'""'};
256 $charset_out = TmplTokenizer::charset_canon $2 if $href->{'""'}->msgstr =~ /\bcharset=(["']?)([^;\s"'\\]+)\1/;
257 $charset_in = $charset_out;
258 warn "Charset in/out: ".$charset_out;
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;
269 # set our charset in to UTF-8
270 if (!defined $charset_in) {
271 $charset_in = TmplTokenizer::charset_canon 'UTF-8';
272 warn "Warning: Can't determine original templates' charset, defaulting to $charset_in\n";
274 # set our charset out to UTF-8
275 if (!defined $charset_out) {
276 $charset_out = TmplTokenizer::charset_canon 'UTF-8';
277 warn "Warning: Charset Out defaulting to $charset_out\n";
279 my $xgettext = './xgettext.pl'; # actual text extractor script
280 my $st;
282 if ($action eq 'create') {
283 # updates the list. As the list is empty, every entry will be added
284 if (!-s $str_file) {
285 warn "Removing empty file $str_file\n";
286 unlink $str_file || die "$str_file: $!\n";
288 die "$str_file: Output file already exists\n" if -f $str_file;
289 my($tmph1, $tmpfile1) = tmpnam();
290 my($tmph2, $tmpfile2) = tmpnam();
291 close $tmph2; # We just want a name
292 # Generate the temporary file that acts as <MODULE>/POTFILES.in
293 for my $input (@in_files) {
294 print $tmph1 "$input\n";
296 close $tmph1;
297 warn "I $charset_in O $charset_out";
298 # Generate the specified po file ($str_file)
299 $st = system ($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2,
300 (defined $charset_in? ('-I', $charset_in): ()),
301 (defined $charset_out? ('-O', $charset_out): ())
303 # Run msgmerge so that the pot file looks like a real pot file
304 # We need to help msgmerge a bit by pre-creating a dummy po file that has
305 # the headers and the "" msgid & msgstr. It will fill in the rest.
306 if ($st == 0) {
307 # Merge the temporary "pot file" with the specified po file ($str_file)
308 # FIXME: msgmerge(1) is a Unix dependency
309 # FIXME: need to check the return value
310 unless (-f $str_file) {
311 local(*INPUT, *OUTPUT);
312 open(INPUT, "<$tmpfile2");
313 open(OUTPUT, ">$str_file");
314 while (<INPUT>) {
315 print OUTPUT;
316 last if /^\n/s;
318 close INPUT;
319 close OUTPUT;
321 $st = system('msgmerge', '-U', '-s', $str_file, $tmpfile2);
322 } else {
323 error_normal "Text extraction failed: $xgettext: $!\n", undef;
324 error_additional "Will not run msgmerge\n", undef;
326 # unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
327 # unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
329 } elsif ($action eq 'update') {
330 my($tmph1, $tmpfile1) = tmpnam();
331 my($tmph2, $tmpfile2) = tmpnam();
332 close $tmph2; # We just want a name
333 # Generate the temporary file that acts as <MODULE>/POTFILES.in
334 for my $input (@in_files) {
335 print $tmph1 "$input\n";
337 close $tmph1;
338 # Generate the temporary file that acts as <MODULE>/<LANG>.pot
339 $st = system($xgettext, '-s', '-f', $tmpfile1, '-o', $tmpfile2,
340 '--po-mode',
341 (defined $charset_in? ('-I', $charset_in): ()),
342 (defined $charset_out? ('-O', $charset_out): ()));
343 if ($st == 0) {
344 # Merge the temporary "pot file" with the specified po file ($str_file)
345 # FIXME: msgmerge(1) is a Unix dependency
346 # FIXME: need to check the return value
347 $st = system('msgmerge', '-U', '-s', $str_file, $tmpfile2);
348 } else {
349 error_normal "Text extraction failed: $xgettext: $!\n", undef;
350 error_additional "Will not run msgmerge\n", undef;
352 # unlink $tmpfile1 || warn_normal "$tmpfile1: unlink failed: $!\n", undef;
353 # unlink $tmpfile2 || warn_normal "$tmpfile2: unlink failed: $!\n", undef;
355 } elsif ($action eq 'install') {
356 if(!defined($out_dir)) {
357 usage_error("You must specify an output directory when using the install method.");
360 if ($in_dir eq $out_dir) {
361 warn "You must specify a different input and output directory.\n";
362 exit -1;
365 # Make sure the output directory exists
366 # (It will auto-create it, but for compatibility we should not)
367 -d $out_dir || die "$out_dir: The directory does not exist\n";
369 # Try to open the file, because Locale::PO doesn't check :-/
370 open(INPUT, "<$str_file") || die "$str_file: $!\n";
371 close INPUT;
373 # creates the new tmpl file using the new translation
374 for my $input (@in_files) {
375 die "Assertion failed"
376 unless substr($input, 0, length($in_dir) + 1) eq "$in_dir/";
377 # print "$input / $type\n";
378 if (!defined $type || $input =~ /\.(?:$type)$/) {
379 my $h = TmplTokenizer->new( $input );
380 $h->set_allow_cformat( 1 );
381 VerboseWarnings::set_input_file_name $input;
383 my $target = $out_dir . substr($input, length($in_dir));
384 my $targetdir = $` if $target =~ /[^\/]+$/s;
385 mkdir_recursive($targetdir) unless -d $targetdir;
386 print STDERR "Creating $target...\n" unless $quiet;
387 open( OUTPUT, ">$target" ) || die "$target: $!\n";
388 text_replace( $h, *OUTPUT );
389 close OUTPUT;
390 } else {
391 # just copying the file
392 my $target = $out_dir . substr($input, length($in_dir));
393 my $targetdir = $` if $target =~ /[^\/]+$/s;
394 mkdir_recursive($targetdir) unless -d $targetdir;
395 system("cp -f $input $target");
396 print STDERR "Copying $input...\n" unless $quiet;
400 } else {
401 usage_error('Unknown action specified.');
404 if ($st == 0) {
405 printf "The %s seems to be successful.\n", $action unless $quiet;
406 } else {
407 printf "%s FAILED.\n", "\u$action" unless $quiet;
409 exit 0;
411 ###############################################################################
413 =head1 SYNOPSIS
415 ./tmpl_process3.pl [ I<tmpl_process.pl options> ]
417 =head1 DESCRIPTION
419 This is an alternative version of the tmpl_process.pl script,
420 using standard gettext-style PO files. While there still might
421 be changes made to the way it extracts strings, at this moment
422 it should be stable enough for general use; it is already being
423 used for the Chinese and Polish translations.
425 Currently, the create, update, and install actions have all been
426 reimplemented and seem to work.
428 =head2 Features
430 =over
432 =item -
434 Translation files in standard Uniforum PO format.
435 All standard tools including all gettext tools,
436 plus PO file editors like kbabel(1) etc.
437 can be used.
439 =item -
441 Minor changes in whitespace in source templates
442 do not generally require strings to be re-translated.
444 =item -
446 Able to handle <TMPL_VAR> variables in the templates;
447 <TMPL_VAR> variables are usually extracted in proper context,
448 represented by a short %s placeholder.
450 =item -
452 Able to handle text input and radio button INPUT elements
453 in the templates; these INPUT elements are also usually
454 extracted in proper context,
455 represented by a short %S or %p placeholder.
457 =item -
459 Automatic comments in the generated PO files to provide
460 even more context (line numbers, and the names and types
461 of the variables).
463 =item -
465 The %I<n>$s (or %I<n>$p, etc.) notation can be used
466 for change the ordering of the variables,
467 if such a reordering is required for correct translation.
469 =item -
471 If a particular <TMPL_VAR> should not appear in the
472 translation, it can be suppressed with the %0.0s notation.
474 =item -
476 Using the PO format also means translators can add their
477 own comments in the translation files, if necessary.
479 =item -
481 Create, update, and install actions are all based on the
482 same scanner module. This ensures that update and install
483 have the same idea of what is a translatable string;
484 attribute names in tags, for example, will not be
485 accidentally translated.
487 =back
489 =head1 NOTES
491 Anchors are represented by an <AI<n>> notation.
492 The meaning of this non-standard notation might not be obvious.
494 The create action calls xgettext.pl to do the actual work;
495 the update action calls xgettext.pl and msgmerge(1) to do the
496 actual work.
498 =head1 BUGS
500 xgettext.pl must be present in the current directory; the
501 msgmerge(1) command must also be present in the search path.
502 The script currently does not check carefully whether these
503 dependent commands are present.
505 Locale::PO(3) has a lot of bugs. It can neither parse nor
506 generate GNU PO files properly; a couple of workarounds have
507 been written in TmplTokenizer and more is likely to be needed
508 (e.g., to get rid of the "Strange line" warning for #~).
510 This script may not work in Windows.
512 There are probably some other bugs too, since this has not been
513 tested very much.
515 =head1 SEE ALSO
517 xgettext.pl,
518 TmplTokenizer.pm,
519 msgmerge(1),
520 Locale::PO(3),
521 translator_doc.txt
523 http://www.saas.nsw.edu.au/koha_wiki/index.php?page=DifficultTerms
525 =cut