1 # BEGIN BPS TAGGED BLOCK {{{
5 # This software is Copyright (c) 1996-2014 Best Practical Solutions, LLC
6 # <sales@bestpractical.com>
8 # (Except where explicitly superseded by other copyright notices)
13 # This work is made available to you under the terms of Version 2 of
14 # the GNU General Public License. A copy of that license should have
15 # been provided with this software, but in any event can be snarfed
18 # This work is distributed in the hope that it will be useful, but
19 # WITHOUT ANY WARRANTY; without even the implied warranty of
20 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
21 # General Public License for more details.
23 # You should have received a copy of the GNU General Public License
24 # along with this program; if not, write to the Free Software
25 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
26 # 02110-1301 or visit their web page on the internet at
27 # http://www.gnu.org/licenses/old-licenses/gpl-2.0.html.
30 # CONTRIBUTION SUBMISSION POLICY:
32 # (The following paragraph is not intended to limit the rights granted
33 # to you to modify and distribute this software under the terms of
34 # the GNU General Public License and is only of importance to you if
35 # you choose to contribute your changes and enhancements to the
36 # community by submitting them to Best Practical Solutions, LLC.)
38 # By intentionally submitting any modifications, corrections or
39 # derivatives to this work, or any other work intended for use with
40 # Request Tracker, to Best Practical Solutions, LLC, you confirm that
41 # you are the copyright holder for those contributions and you grant
42 # Best Practical Solutions, LLC a nonexclusive, worldwide, irrevocable,
43 # royalty-free, perpetual, license to use, copy, create derivative
44 # works based on those contributions, and sublicense and distribute
45 # those contributions and any derivatives thereof.
47 # END BPS TAGGED BLOCK }}}
51 RT::I18N - a base class for localization of RT
61 use Locale::Maketext 1.04;
62 use Locale::Maketext::Lexicon 0.25;
63 use base 'Locale::Maketext::Fuzzy';
70 # I decree that this project's first language is English.
73 'TEST_STRING' => 'Concrete Mixer',
75 '__Content-Type' => 'text/plain; charset=utf-8',
78 # That means that lookup failures can't happen -- if we get as far
79 # as looking for something in this lexicon, and we don't find it,
80 # then automagically set $Lexicon{$key} = $key, before possibly
83 # The exception is keys that start with "_" -- they aren't auto-makeable.
90 Initializes the lexicons used for localization.
97 my @lang = RT->Config->Get('LexiconLanguages');
98 @lang = ('*') unless @lang;
100 # load default functions
101 require substr(__FILE__, 0, -3) . '/i_default.pm';
103 # Load language-specific functions
104 foreach my $file ( File::Glob::bsd_glob(substr(__FILE__, 0, -3) . "/*.pm") ) {
105 unless ( $file =~ /^([-\w\s\.\/\\~:]+)$/ ) {
106 warn("$file is tainted. not loading");
111 my ($lang) = ($file =~ /([^\\\/]+?)\.pm$/);
112 next unless grep $_ eq '*' || $_ eq $lang, @lang;
117 foreach my $l ( @lang ) {
119 Gettext => $RT::LexiconPath."/$l.po",
121 push @{ $import{$l} }, map {(Gettext => "$_/$l.po")} RT->PluginDirs('po');
122 push @{ $import{$l} }, (Gettext => $RT::LocalLexiconPath."/*/$l.po",
123 Gettext => $RT::LocalLexiconPath."/$l.po");
126 # Acquire all .po files and iterate them into lexicons
127 Locale::Maketext::Lexicon->import({ _decode => 1, %import });
135 foreach my $k (keys %{RT::I18N::} ) {
136 next if $k eq 'main::';
137 next unless index($k, '::', -2) >= 0;
138 next unless exists ${ 'RT::I18N::'. $k }{'Lexicon'};
140 my $lex = *{ ${'RT::I18N::'. $k }{'Lexicon'} }{HASH};
141 # run fetch to force load
142 my $tmp = $lex->{'foo'};
143 # XXX: untie may fail with "untie attempted
144 # while 1 inner references still exist"
145 # TODO: untie that has to lower fetch impact
146 # untie %$lex if tied %$lex;
152 Returns the encoding of the current lexicon, as yanked out of __ContentType's "charset" field.
153 If it can't find anything, it returns 'ISO-8859-1'
160 sub encoding { 'utf-8' }
163 =head2 SetMIMEEntityToUTF8 $entity
165 An utility function which will try to convert entity body into utf8.
166 It's now a wrap-up of SetMIMEEntityToEncoding($entity, 'utf-8').
170 sub SetMIMEEntityToUTF8 {
171 RT::I18N::SetMIMEEntityToEncoding(shift, 'utf-8');
176 =head2 IsTextualContentType $type
178 An utility function that determines whether $type is I<textual>, meaning
179 that it can sensibly be converted to Unicode text.
181 Currently, it returns true iff $type matches this regular expression
182 (case-insensitively):
184 ^(?:text/(?:plain|html)|message/rfc822)\b
189 sub IsTextualContentType {
191 ($type =~ m{^(?:text/(?:plain|html)|message/rfc822)\b}i) ? 1 : 0;
195 =head2 SetMIMEEntityToEncoding $entity, $encoding
197 An utility function which will try to convert entity body into specified
198 charset encoding (encoded as octets, *not* unicode-strings). It will
199 iterate all the entities in $entity, and try to convert each one into
200 specified charset if whose Content-Type is 'text/plain'.
202 This function doesn't return anything meaningful.
206 sub SetMIMEEntityToEncoding {
207 my ( $entity, $enc, $preserve_words ) = ( shift, shift, shift );
209 # do the same for parts first of all
210 SetMIMEEntityToEncoding( $_, $enc, $preserve_words ) foreach $entity->parts;
212 my $head = $entity->head;
214 my $charset = _FindOrGuessCharset($entity);
216 unless( Encode::find_encoding($charset) ) {
217 $RT::Logger->warning("Encoding '$charset' is not supported");
221 unless ( $charset ) {
222 $head->replace( "X-RT-Original-Content-Type" => $head->mime_attr('Content-Type') );
223 $head->mime_attr('Content-Type' => 'application/octet-stream');
227 SetMIMEHeadToEncoding(
229 _FindOrGuessCharset($entity, 1) => $enc,
233 # If this is a textual entity, we'd need to preserve its original encoding
234 $head->replace( "X-RT-Original-Encoding" => $charset )
235 if $head->mime_attr('content-type.charset') or IsTextualContentType($head->mime_type);
237 return unless IsTextualContentType($head->mime_type);
239 my $body = $entity->bodyhandle;
241 if ( $body && ($enc ne $charset || $enc =~ /^utf-?8(?:-strict)?$/i) ) {
242 my $string = $body->as_string or return;
244 $RT::Logger->debug( "Converting '$charset' to '$enc' for "
245 . $head->mime_type . " - "
246 . ( $head->get('subject') || 'Subjectless message' ) );
248 # NOTE:: see the comments at the end of the sub.
249 Encode::_utf8_off($string);
250 Encode::from_to( $string, $charset => $enc );
252 my $new_body = MIME::Body::InCore->new($string);
254 # set up the new entity
255 $head->mime_attr( "content-type" => 'text/plain' )
256 unless ( $head->mime_attr("content-type") );
257 $head->mime_attr( "content-type.charset" => $enc );
258 $entity->bodyhandle($new_body);
262 # NOTES: Why Encode::_utf8_off before Encode::from_to
264 # All the strings in RT are utf-8 now. Quotes from Encode POD:
266 # [$length =] from_to($octets, FROM_ENC, TO_ENC [, CHECK])
267 # ... The data in $octets must be encoded as octets and not as
268 # characters in Perl's internal format. ...
270 # Not turning off the UTF-8 flag in the string will prevent the string
275 =head2 DecodeMIMEWordsToUTF8 $raw
277 An utility method which mimics MIME::Words::decode_mimewords, but only
278 limited functionality. This function returns an utf-8 string.
280 It returns the decoded string, or the original string if it's not
281 encoded. Since the subroutine converts specified string into utf-8
282 charset, it should not alter a subject written in English.
284 Why not use MIME::Words directly? Because it fails in RT when I
285 tried. Maybe it's ok now.
289 sub DecodeMIMEWordsToUTF8 {
291 return DecodeMIMEWordsToEncoding($str, 'utf-8', @_);
294 sub DecodeMIMEWordsToEncoding {
296 my $to_charset = _CanonicalizeCharset(shift);
297 my $field = shift || '';
299 # handle filename*=ISO-8859-1''%74%E9%73%74%2E%74%78%74, parameter value
300 # continuations, and similar syntax from RFC 2231
301 if ($field =~ /^Content-(Type|Disposition)/i) {
302 # This concatenates continued parameters and normalizes encoded params
303 # to QB encoded-words which we handle below
304 $str = MIME::Field::ParamVal->parse($str)->stringify;
307 # Pre-parse by removing all whitespace between encoded words
308 my $encoded_word = qr/
311 (?:\*[^?]+)? # optional '*language'
315 ([^?]+) # encoded string
318 $str =~ s/($encoded_word)\s+(?=$encoded_word)/$1/g;
320 # Also merge quoted-printable sections together, in case multiple
321 # octets of a single encoded character were split between chunks.
322 # Though not valid according to RFC 2047, this has been seen in the
324 1 while $str =~ s/(=\?[^?]+\?[Qq]\?)([^?]+)\?=\1([^?]+)\?=/$1$2$3?=/i;
326 # XXX TODO: use decode('MIME-Header', ...) and Encode::Alias to replace our
327 # custom MIME word decoding and charset canonicalization. We can't do this
328 # until we parse before decode, instead of the other way around.
329 my @list = $str =~ m/(.*?) # prefix
335 # add everything that hasn't matched to the end of the latest
336 # string in array this happen when we have 'key="=?encoded?="; key="plain"'
337 $list[-1] .= substr($str, pos $str);
341 my ($prefix, $charset, $encoding, $enc_str, $trailing) =
343 $charset = _CanonicalizeCharset($charset);
344 $encoding = lc $encoding;
346 $trailing =~ s/\s?\t?$//; # Observed from Outlook Express
348 if ( $encoding eq 'q' ) {
349 use MIME::QuotedPrint;
350 $enc_str =~ tr/_/ /; # Observed from Outlook Express
351 $enc_str = decode_qp($enc_str);
352 } elsif ( $encoding eq 'b' ) {
354 $enc_str = decode_base64($enc_str);
356 $RT::Logger->warning("Incorrect encoding '$encoding' in '$str', "
357 ."only Q(uoted-printable) and B(ase64) are supported");
360 # now we have got a decoded subject, try to convert into the encoding
361 if ( $charset ne $to_charset || $charset =~ /^utf-?8(?:-strict)?$/i ) {
362 if ( Encode::find_encoding($charset) ) {
363 Encode::from_to( $enc_str, $charset, $to_charset );
365 $RT::Logger->warning("Charset '$charset' is not supported");
366 $enc_str =~ s/[^[:print:]]/\357\277\275/g;
367 Encode::from_to( $enc_str, 'UTF-8', $to_charset )
368 unless $to_charset eq 'utf-8';
372 # XXX TODO: RT doesn't currently do the right thing with mime-encoded headers
373 # We _should_ be preserving them encoded until after parsing is completed and
374 # THEN undo the mime-encoding.
376 # This routine should be translating the existing mimeencoding to utf8 but leaving
379 # It's legal for headers to contain mime-encoded commas and semicolons which
380 # should not be treated as address separators. (Encoding == quoting here)
382 # until this is fixed, we must escape any string containing a comma or semicolon
383 # this is only a bandaid
385 # Some _other_ MUAs encode quotes _already_, and double quotes
386 # confuse us a lot, so only quote it if it isn't quoted
388 $enc_str = qq{"$enc_str"}
389 if $enc_str =~ /[,;]/
390 and $enc_str !~ /^".*"$/
391 and $prefix !~ /"$/ and $trailing !~ /^"/
392 and (!$field || $field =~ /^(?:To$|From$|B?Cc$|Content-)/i);
394 $str .= $prefix . $enc_str . $trailing;
398 # We might have \n without trailing whitespace, which will result in
407 =head2 _FindOrGuessCharset MIME::Entity, $head_only
409 When handed a MIME::Entity will first attempt to read what charset the message is encoded in. Failing that, will use Encode::Guess to try to figure it out
411 If $head_only is true, only guesses charset for head parts. This is because header's encoding (e.g. filename="...") may be different from that of body's.
415 sub _FindOrGuessCharset {
417 my $head_only = shift;
418 my $head = $entity->head;
420 if ( my $charset = $head->mime_attr("content-type.charset") ) {
421 return _CanonicalizeCharset($charset);
424 if ( !$head_only and $head->mime_type =~ m{^text/} ) {
425 my $body = $entity->bodyhandle or return;
426 return _GuessCharset( $body->as_string );
430 # potentially binary data -- don't guess the body
431 return _GuessCharset( $head->as_string );
437 =head2 _GuessCharset STRING
439 use Encode::Guess to try to figure it out the string's encoding.
443 use constant HAS_ENCODE_GUESS => do { local $@; eval { require Encode::Guess; 1 } };
444 use constant HAS_ENCODE_DETECT => do { local $@; eval { require Encode::Detect::Detector; 1 } };
447 my $fallback = _CanonicalizeCharset('iso-8859-1');
449 # if $_[0] is null/empty, we don't guess its encoding
451 unless defined $_[0] && length $_[0];
453 my @encodings = RT->Config->Get('EmailInputEncodings');
454 unless ( @encodings ) {
455 $RT::Logger->warning("No EmailInputEncodings set, fallback to $fallback");
459 if ( $encodings[0] eq '*' ) {
461 if ( HAS_ENCODE_DETECT ) {
462 my $charset = Encode::Detect::Detector::detect( $_[0] );
464 $RT::Logger->debug("Encode::Detect::Detector guessed encoding: $charset");
465 return _CanonicalizeCharset( Encode::resolve_alias( $charset ) );
468 $RT::Logger->debug("Encode::Detect::Detector failed to guess encoding");
473 "You requested to guess encoding, but we couldn't"
474 ." load Encode::Detect::Detector module"
479 unless ( @encodings ) {
480 $RT::Logger->warning("No EmailInputEncodings set except '*', fallback to $fallback");
484 unless ( HAS_ENCODE_GUESS ) {
485 $RT::Logger->error("We couldn't load Encode::Guess module, fallback to $fallback");
489 Encode::Guess->set_suspects( @encodings );
490 my $decoder = Encode::Guess->guess( $_[0] );
491 unless ( defined $decoder ) {
492 $RT::Logger->warning("Encode::Guess failed: decoder is undefined; fallback to $fallback");
496 if ( ref $decoder ) {
497 my $charset = $decoder->name;
498 $RT::Logger->debug("Encode::Guess guessed encoding: $charset");
499 return _CanonicalizeCharset( $charset );
501 elsif ($decoder =~ /(\S+ or .+)/) {
502 my %matched = map { $_ => 1 } split(/ or /, $1);
503 return 'utf-8' if $matched{'utf8'}; # one and only normalization
505 foreach my $suspect (RT->Config->Get('EmailInputEncodings')) {
506 next unless $matched{$suspect};
507 $RT::Logger->debug("Encode::Guess ambiguous ($decoder); using $suspect");
508 return _CanonicalizeCharset( $suspect );
512 $RT::Logger->warning("Encode::Guess failed: $decoder; fallback to $fallback");
518 =head2 _CanonicalizeCharset NAME
520 canonicalize charset, return lowercase version.
521 special cases are: gb2312 => gbk, utf8 => utf-8
525 sub _CanonicalizeCharset {
526 my $charset = lc shift;
527 return $charset unless $charset;
529 # Canonicalize aliases if they're known
530 if (my $canonical = Encode::resolve_alias($charset)) {
531 $charset = $canonical;
534 if ( $charset eq 'utf8' || $charset eq 'utf-8-strict' ) {
537 elsif ( $charset eq 'euc-cn' ) {
538 # gbk is superset of gb2312/euc-cn so it's safe
540 # XXX TODO: gb18030 is an even larger, more permissive superset of gbk,
541 # but needs Encode::HanExtra installed
549 =head2 SetMIMEHeadToEncoding HEAD OLD_CHARSET NEW_CHARSET
551 Converts a MIME Head from one encoding to another. This totally violates the RFC.
552 We should never need this. But, Surprise!, MUAs are badly broken and do this kind of stuff
558 sub SetMIMEHeadToEncoding {
559 my ( $head, $charset, $enc, $preserve_words ) = ( shift, shift, shift, shift );
561 $charset = _CanonicalizeCharset($charset);
562 $enc = _CanonicalizeCharset($enc);
564 return if $charset eq $enc and $preserve_words;
566 foreach my $tag ( $head->tags ) {
567 next unless $tag; # seen in wild: headers with no name
568 my @values = $head->get_all($tag);
570 foreach my $value (@values) {
571 if ( $charset ne $enc || $enc =~ /^utf-?8(?:-strict)?$/i ) {
572 Encode::_utf8_off($value);
573 Encode::from_to( $value, $charset => $enc );
575 $value = DecodeMIMEWordsToEncoding( $value, $enc, $tag )
576 unless $preserve_words;
578 # We intentionally add a leading space when re-adding the
579 # header; Mail::Header strips it before storing, but it
580 # serves to prevent it from "helpfully" canonicalizing
581 # $head->add("Subject", "Subject: foo") into the same as
582 # $head->add("Subject", "foo");
583 $head->add( $tag, " " . $value );
589 RT::Base->_ImportOverlays();