1 /*****************************************************************************
2 * subsdec.c : text subtitle decoder
3 *****************************************************************************
4 * Copyright (C) 2000-2006 VLC authors and VideoLAN
7 * Authors: Gildas Bazin <gbazin@videolan.org>
8 * Samuel Hocevar <sam@zoy.org>
9 * Derk-Jan Hartman <hartman at videolan dot org>
10 * Bernie Purcell <bitmap@videolan.org>
12 * This program is free software; you can redistribute it and/or modify it
13 * under the terms of the GNU Lesser General Public License as published by
14 * the Free Software Foundation; either version 2.1 of the License, or
15 * (at your option) any later version.
17 * This program is distributed in the hope that it will be useful,
18 * but WITHOUT ANY WARRANTY; without even the implied warranty of
19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 * GNU Lesser General Public License for more details.
22 * You should have received a copy of the GNU Lesser General Public License
23 * along with this program; if not, write to the Free Software Foundation,
24 * Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
25 *****************************************************************************/
27 /*****************************************************************************
29 *****************************************************************************/
36 #include <vlc_common.h>
37 #include <vlc_plugin.h>
38 #include <vlc_codec.h>
39 #include <vlc_charset.h>
43 /*****************************************************************************
45 *****************************************************************************/
46 static const char *const ppsz_encodings
[] = {
92 static const char *const ppsz_encoding_names
[] = {
94 The character encoding name in parenthesis corresponds to that used for
95 the GetACP translation. "Windows-1252" applies to Western European
96 languages using the Latin alphabet. */
97 N_("Default (Windows-1252)"),
99 N_("Universal (UTF-8)"),
100 N_("Universal (UTF-16)"),
101 N_("Universal (big endian UTF-16)"),
102 N_("Universal (little endian UTF-16)"),
103 N_("Universal, Chinese (GB18030)"),
105 /* ISO 8859 and the likes */
107 N_("Western European (Latin-9)"), /* mostly superset of Latin-1 */
108 N_("Western European (Windows-1252)"),
109 N_("Western European (IBM 00850)"),
111 N_("Eastern European (Latin-2)"),
112 N_("Eastern European (Windows-1250)"),
114 N_("Esperanto (Latin-3)"),
116 N_("Nordic (Latin-6)"), /* Latin 6 supersedes Latin 4 */
118 N_("Cyrillic (Windows-1251)"), /* ISO 8859-5 is not practically used */
119 N_("Russian (KOI8-R)"),
120 N_("Ukrainian (KOI8-U)"),
122 N_("Arabic (ISO 8859-6)"),
123 N_("Arabic (Windows-1256)"),
125 N_("Greek (ISO 8859-7)"),
126 N_("Greek (Windows-1253)"),
128 N_("Hebrew (ISO 8859-8)"),
129 N_("Hebrew (Windows-1255)"),
131 N_("Turkish (ISO 8859-9)"),
132 N_("Turkish (Windows-1254)"),
135 N_("Thai (TIS 620-2533/ISO 8859-11)"),
136 N_("Thai (Windows-874)"),
138 N_("Baltic (Latin-7)"),
139 N_("Baltic (Windows-1257)"),
140 /* 12 -> /dev/null */
142 N_("Celtic (Latin-8)"),
145 N_("South-Eastern European (Latin-10)"),
147 N_("Simplified Chinese (ISO-2022-CN-EXT)"),
148 N_("Simplified Chinese Unix (EUC-CN)"),
149 N_("Japanese (7-bits JIS/ISO-2022-JP-2)"),
150 N_("Japanese Unix (EUC-JP)"),
151 N_("Japanese (Shift JIS)"),
152 N_("Korean (EUC-KR/CP949)"),
153 N_("Korean (ISO-2022-KR)"),
154 N_("Traditional Chinese (Big5)"),
155 N_("Traditional Chinese Unix (EUC-TW)"),
156 N_("Hong-Kong Supplementary (HKSCS)"),
158 N_("Vietnamese (VISCII)"),
159 N_("Vietnamese (Windows-1258)"),
162 static const int pi_justification
[] = { 0, 1, 2 };
163 static const char *const ppsz_justification_text
[] = {
164 N_("Center"),N_("Left"),N_("Right")};
166 #define ENCODING_TEXT N_("Subtitle text encoding")
167 #define ENCODING_LONGTEXT N_("Set the encoding used in text subtitles")
168 #define ALIGN_TEXT N_("Subtitle justification")
169 #define ALIGN_LONGTEXT N_("Set the justification of subtitles")
170 #define AUTODETECT_UTF8_TEXT N_("UTF-8 subtitle autodetection")
171 #define AUTODETECT_UTF8_LONGTEXT N_("This enables automatic detection of " \
172 "UTF-8 encoding within subtitle files.")
173 #define FORMAT_TEXT N_("Formatted Subtitles")
174 #define FORMAT_LONGTEXT N_("Some subtitle formats allow for text formatting. " \
175 "VLC partly implements this, but you can choose to disable all formatting.")
177 static int OpenDecoder ( vlc_object_t
* );
178 static void CloseDecoder ( vlc_object_t
* );
181 set_shortname( N_("Subtitles"))
182 set_description( N_("Text subtitle decoder") )
183 set_capability( "decoder", 50 )
184 set_callbacks( OpenDecoder
, CloseDecoder
)
185 set_category( CAT_INPUT
)
186 set_subcategory( SUBCAT_INPUT_SCODEC
)
188 add_integer( "subsdec-align", 0, ALIGN_TEXT
, ALIGN_LONGTEXT
,
190 change_integer_list( pi_justification
, ppsz_justification_text
)
191 add_string( "subsdec-encoding", "",
192 ENCODING_TEXT
, ENCODING_LONGTEXT
, false )
193 change_string_list( ppsz_encodings
, ppsz_encoding_names
)
194 add_bool( "subsdec-autodetect-utf8", true,
195 AUTODETECT_UTF8_TEXT
, AUTODETECT_UTF8_LONGTEXT
, false )
196 add_bool( "subsdec-formatted", true, FORMAT_TEXT
, FORMAT_LONGTEXT
,
200 /*****************************************************************************
202 *****************************************************************************/
203 #define NO_BREAKING_SPACE " "
207 int i_align
; /* Subtitles alignment on the vout */
209 vlc_iconv_t iconv_handle
; /* handle to iconv instance */
210 bool b_autodetect_utf8
;
214 static subpicture_t
*DecodeBlock ( decoder_t
*, block_t
** );
215 static subpicture_t
*ParseText ( decoder_t
*, block_t
* );
216 static char *StripTags ( char * );
217 static char *CreateHtmlSubtitle( int *pi_align
, char * );
219 /*****************************************************************************
220 * OpenDecoder: probe the decoder and return score
221 *****************************************************************************
222 * Tries to launch a decoder and return score so that the interface is able
224 *****************************************************************************/
225 static int OpenDecoder( vlc_object_t
*p_this
)
227 decoder_t
*p_dec
= (decoder_t
*)p_this
;
228 decoder_sys_t
*p_sys
;
230 switch( p_dec
->fmt_in
.i_codec
)
233 case VLC_CODEC_ITU_T140
:
239 p_dec
->pf_decode_sub
= DecodeBlock
;
240 p_dec
->fmt_out
.i_cat
= SPU_ES
;
241 p_dec
->fmt_out
.i_codec
= 0;
243 /* Allocate the memory needed to store the decoder's structure */
244 p_dec
->p_sys
= p_sys
= calloc( 1, sizeof( *p_sys
) );
250 p_sys
->iconv_handle
= (vlc_iconv_t
)-1;
251 p_sys
->b_autodetect_utf8
= false;
253 const char *encoding
;
256 /* First try demux-specified encoding */
257 if( p_dec
->fmt_in
.i_codec
== VLC_CODEC_ITU_T140
)
258 encoding
= "UTF-8"; /* IUT T.140 is always using UTF-8 */
260 if( p_dec
->fmt_in
.subs
.psz_encoding
&& *p_dec
->fmt_in
.subs
.psz_encoding
)
262 encoding
= p_dec
->fmt_in
.subs
.psz_encoding
;
263 msg_Dbg (p_dec
, "trying demuxer-specified character encoding: %s",
268 /* Second, try configured encoding */
269 if ((var
= var_InheritString (p_dec
, "subsdec-encoding")) != NULL
)
271 msg_Dbg (p_dec
, "trying configured character encoding: %s", var
);
272 if (!strcmp (var
, "system"))
277 /* ^ iconv() treats "" as nl_langinfo(CODESET) */
283 /* Third, try "local" encoding */
286 The Windows ANSI code page most commonly used for this language.
287 VLC uses this as a guess of the subtitle files character set
288 (if UTF-8 and UTF-16 autodetection fails).
289 Western European languages normally use "CP1252", which is a
290 Microsoft-variant of ISO 8859-1. That suits the Latin alphabet.
291 Other scripts use other code pages.
293 This MUST be a valid iconv character set. If unsure, please refer
294 the VideoLAN translators mailing list. */
295 encoding
= vlc_pgettext("GetACP", "CP1252");
296 msg_Dbg (p_dec
, "trying default character encoding: %s", encoding
);
299 /* Check UTF-8 autodetection */
300 if (var_InheritBool (p_dec
, "subsdec-autodetect-utf8"))
302 msg_Dbg (p_dec
, "using automatic UTF-8 detection");
303 p_sys
->b_autodetect_utf8
= true;
307 if (strcasecmp (encoding
, "UTF-8") && strcasecmp (encoding
, "utf8"))
309 p_sys
->iconv_handle
= vlc_iconv_open ("UTF-8", encoding
);
310 if (p_sys
->iconv_handle
== (vlc_iconv_t
)(-1))
311 msg_Err (p_dec
, "cannot convert from %s: %m", encoding
);
315 p_sys
->i_align
= var_InheritInteger( p_dec
, "subsdec-align" );
320 /****************************************************************************
321 * DecodeBlock: the whole thing
322 ****************************************************************************
323 * This function must be fed with complete subtitles units.
324 ****************************************************************************/
325 static subpicture_t
*DecodeBlock( decoder_t
*p_dec
, block_t
**pp_block
)
330 if( !pp_block
|| *pp_block
== NULL
)
334 if( p_block
->i_flags
& (BLOCK_FLAG_DISCONTINUITY
|BLOCK_FLAG_CORRUPTED
) )
336 block_Release( p_block
);
340 p_spu
= ParseText( p_dec
, p_block
);
342 block_Release( p_block
);
348 /*****************************************************************************
349 * CloseDecoder: clean up the decoder
350 *****************************************************************************/
351 static void CloseDecoder( vlc_object_t
*p_this
)
353 decoder_t
*p_dec
= (decoder_t
*)p_this
;
354 decoder_sys_t
*p_sys
= p_dec
->p_sys
;
356 if( p_sys
->iconv_handle
!= (vlc_iconv_t
)-1 )
357 vlc_iconv_close( p_sys
->iconv_handle
);
362 /*****************************************************************************
363 * ParseText: parse an text subtitle packet and send it to the video output
364 *****************************************************************************/
365 static subpicture_t
*ParseText( decoder_t
*p_dec
, block_t
*p_block
)
367 decoder_sys_t
*p_sys
= p_dec
->p_sys
;
368 subpicture_t
*p_spu
= NULL
;
369 char *psz_subtitle
= NULL
;
371 /* We cannot display a subpicture with no date */
372 if( p_block
->i_pts
<= VLC_TS_INVALID
)
374 msg_Warn( p_dec
, "subtitle without a date" );
378 /* Check validity of packet data */
379 /* An "empty" line containing only \0 can be used to force
380 and ephemer picture from the screen */
381 if( p_block
->i_buffer
< 1 )
383 msg_Warn( p_dec
, "no subtitle data" );
387 /* Should be resiliant against bad subtitles */
388 psz_subtitle
= malloc( p_block
->i_buffer
+ 1 );
389 if( psz_subtitle
== NULL
)
391 memcpy( psz_subtitle
, p_block
->p_buffer
, p_block
->i_buffer
);
392 psz_subtitle
[p_block
->i_buffer
] = '\0';
394 if( p_sys
->iconv_handle
== (vlc_iconv_t
)-1 )
396 if (EnsureUTF8( psz_subtitle
) == NULL
)
398 msg_Err( p_dec
, "failed to convert subtitle encoding.\n"
399 "Try manually setting a character-encoding "
400 "before you open the file." );
406 if( p_sys
->b_autodetect_utf8
)
408 if( IsUTF8( psz_subtitle
) == NULL
)
410 msg_Dbg( p_dec
, "invalid UTF-8 sequence: "
411 "disabling UTF-8 subtitles autodetection" );
412 p_sys
->b_autodetect_utf8
= false;
416 if( !p_sys
->b_autodetect_utf8
)
418 size_t inbytes_left
= strlen( psz_subtitle
);
419 size_t outbytes_left
= 6 * inbytes_left
;
420 char *psz_new_subtitle
= xmalloc( outbytes_left
+ 1 );
421 char *psz_convert_buffer_out
= psz_new_subtitle
;
422 const char *psz_convert_buffer_in
= psz_subtitle
;
424 size_t ret
= vlc_iconv( p_sys
->iconv_handle
,
425 &psz_convert_buffer_in
, &inbytes_left
,
426 &psz_convert_buffer_out
, &outbytes_left
);
428 *psz_convert_buffer_out
++ = '\0';
429 free( psz_subtitle
);
431 if( ( ret
== (size_t)(-1) ) || inbytes_left
)
433 free( psz_new_subtitle
);
434 msg_Err( p_dec
, "failed to convert subtitle encoding.\n"
435 "Try manually setting a character-encoding "
436 "before you open the file." );
440 psz_subtitle
= realloc( psz_new_subtitle
,
441 psz_convert_buffer_out
- psz_new_subtitle
);
443 psz_subtitle
= psz_new_subtitle
;
447 /* Create the subpicture unit */
448 p_spu
= decoder_NewSubpictureText( p_dec
);
451 free( psz_subtitle
);
454 p_spu
->i_start
= p_block
->i_pts
;
455 p_spu
->i_stop
= p_block
->i_pts
+ p_block
->i_length
;
456 p_spu
->b_ephemer
= (p_block
->i_length
== 0);
457 p_spu
->b_absolute
= false;
459 subpicture_updater_sys_t
*p_spu_sys
= p_spu
->updater
.p_sys
;
461 p_spu_sys
->align
= SUBPICTURE_ALIGN_BOTTOM
| p_sys
->i_align
;
462 p_spu_sys
->text
= StripTags( psz_subtitle
);
463 if( var_InheritBool( p_dec
, "subsdec-formatted" ) )
464 p_spu_sys
->html
= CreateHtmlSubtitle( &p_spu_sys
->align
, psz_subtitle
);
466 free( psz_subtitle
);
471 /* Function now handles tags with attribute values, and tries
472 * to deal with &' commands too. It no longer modifies the string
473 * in place, so that the original text can be reused
475 static char *StripTags( char *psz_subtitle
)
477 char *psz_text_start
;
480 psz_text
= psz_text_start
= malloc( strlen( psz_subtitle
) + 1 );
481 if( !psz_text_start
)
484 while( *psz_subtitle
)
486 if( *psz_subtitle
== '<' )
488 if( strncasecmp( psz_subtitle
, "<br/>", 5 ) == 0 )
491 psz_subtitle
+= strcspn( psz_subtitle
, ">" );
493 else if( *psz_subtitle
== '&' )
495 if( !strncasecmp( psz_subtitle
, "<", 4 ))
498 psz_subtitle
+= strcspn( psz_subtitle
, ";" );
500 else if( !strncasecmp( psz_subtitle
, ">", 4 ))
503 psz_subtitle
+= strcspn( psz_subtitle
, ";" );
505 else if( !strncasecmp( psz_subtitle
, "&", 5 ))
508 psz_subtitle
+= strcspn( psz_subtitle
, ";" );
510 else if( !strncasecmp( psz_subtitle
, """, 6 ))
513 psz_subtitle
+= strcspn( psz_subtitle
, ";" );
517 /* Assume it is just a normal ampersand */
523 *psz_text
++ = *psz_subtitle
;
526 /* Security fix: Account for the case where input ends early */
527 if( *psz_subtitle
== '\0' ) break;
532 char *psz
= realloc( psz_text_start
, strlen( psz_text_start
) + 1 );
533 if( psz
) psz_text_start
= psz
;
535 return psz_text_start
;
538 /* Try to respect any style tags present in the subtitle string. The main
539 * problem here is a lack of adequate specs for the subtitle formats.
540 * SSA/ASS and USF are both detail spec'ed -- but they are handled elsewhere.
541 * SAMI has a detailed spec, but extensive rework is needed in the demux
542 * code to prevent all this style information being excised, as it presently
544 * That leaves the others - none of which were (I guess) originally intended
545 * to be carrying style information. Over time people have used them that way.
546 * In the absence of specifications from which to work, the tags supported
547 * have been restricted to the simple set permitted by the USF DTD, ie. :
548 * Basic: <br>, <i>, <b>, <u>, <s>
560 * There is also the further restriction that the subtitle be well-formed
561 * as an XML entity, ie. the HTML sentence:
562 * <b><i>Bold and Italics</b></i>
563 * doesn't qualify because the tags aren't nested one inside the other.
564 * <text> tags are automatically added to the output to ensure
566 * If the text doesn't qualify for any reason, a NULL string is
567 * returned, and the rendering engine will fall back to the
568 * plain text version of the subtitle.
570 /* TODO: highly suboptimal, offset should be cached */
571 static void HtmlNPut( char **ppsz_html
, const char *psz_text
, int i_max
)
573 char *psz_html
= *ppsz_html
;
574 if( psz_html
== NULL
)
577 const size_t i_offset
= strlen(psz_html
);
578 const size_t i_len
= strnlen(psz_text
, i_max
);
580 psz_html
= realloc( psz_html
, i_offset
+ i_len
+ 1 );
581 if( psz_html
!= NULL
)
583 memcpy( psz_html
+ i_offset
, psz_text
, i_len
);
584 psz_html
[i_offset
+ i_len
] = '\0';
588 *ppsz_html
= psz_html
;
591 static void HtmlPut( char **ppsz_html
, const char *psz_text
)
593 HtmlNPut( ppsz_html
, psz_text
, INT_MAX
);
596 static void HtmlCopy( char **ppsz_html
, char **ppsz_subtitle
, const char *psz_text
)
598 HtmlPut( ppsz_html
, psz_text
);
599 *ppsz_subtitle
+= strlen(psz_text
);
602 static char *CreateHtmlSubtitle( int *pi_align
, char *psz_subtitle
)
604 char *psz_tag
= malloc( 1 );
605 if( psz_tag
== NULL
)
608 char *psz_html
= malloc( 1 );
609 if( psz_html
== NULL
)
617 bool b_has_align
= false;
619 HtmlPut( &psz_html
, "<text>" );
622 while( *psz_subtitle
)
624 if( *psz_subtitle
== '\n' )
626 HtmlPut( &psz_html
, "<br/>" );
629 else if( *psz_subtitle
== '<' )
631 if( !strncasecmp( psz_subtitle
, "<br/>", 5 ))
633 HtmlCopy( &psz_html
, &psz_subtitle
, "<br/>" );
635 else if( !strncasecmp( psz_subtitle
, "<b>", 3 ) )
637 HtmlCopy( &psz_html
, &psz_subtitle
, "<b>" );
638 HtmlPut( &psz_tag
, "b" );
640 else if( !strncasecmp( psz_subtitle
, "<i>", 3 ) )
642 HtmlCopy( &psz_html
, &psz_subtitle
, "<i>" );
643 HtmlPut( &psz_tag
, "i" );
645 else if( !strncasecmp( psz_subtitle
, "<u>", 3 ) )
647 HtmlCopy( &psz_html
, &psz_subtitle
, "<u>" );
648 HtmlPut( &psz_tag
, "u" );
650 else if( !strncasecmp( psz_subtitle
, "<s>", 3 ) )
652 HtmlCopy( &psz_html
, &psz_subtitle
, "<s>" );
653 HtmlPut( &psz_tag
, "s" );
655 else if( !strncasecmp( psz_subtitle
, "<font ", 6 ))
657 const char *psz_attribs
[] = { "face=", "family=", "size=",
658 "color=", "outline-color=", "shadow-color=",
659 "outline-level=", "shadow-level=", "back-color=",
662 HtmlCopy( &psz_html
, &psz_subtitle
, "<font " );
663 HtmlPut( &psz_tag
, "f" );
665 while( *psz_subtitle
!= '>' )
670 while (*psz_subtitle
== ' ')
673 for( k
=0; psz_attribs
[ k
]; k
++ )
675 int i_len
= strlen( psz_attribs
[ k
] );
677 if( !strncasecmp( psz_subtitle
, psz_attribs
[k
], i_len
) )
680 HtmlPut( &psz_html
, psz_attribs
[k
] );
681 psz_subtitle
+= i_len
;
683 /* <font color= red */
684 while (*psz_subtitle
== ' ')
688 if( *psz_subtitle
== '"' )
691 i_len
= strcspn( psz_subtitle
, "\"" );
693 else if( *psz_subtitle
== '\'' )
696 i_len
= strcspn( psz_subtitle
, "'" );
700 i_len
= strcspn( psz_subtitle
, " \t>" );
702 HtmlPut( &psz_html
, "\"" );
703 HtmlNPut( &psz_html
, psz_subtitle
, i_len
);
704 HtmlPut( &psz_html
, "\"" );
706 psz_subtitle
+= i_len
;
707 if( *psz_subtitle
== '\"' || *psz_subtitle
== '\'' )
712 if( psz_attribs
[ k
] == NULL
)
714 /* Jump over unrecognised tag */
715 int i_len
= strcspn( psz_subtitle
, "\"" );
716 if( psz_subtitle
[i_len
] == '\"' )
718 i_len
+= 1 + strcspn( &psz_subtitle
[i_len
+ 1], "\"" );
719 if( psz_subtitle
[i_len
] == '\"' )
722 /* Not a tag, something else we do not understand */
726 psz_subtitle
+= i_len
;
729 HtmlNPut( &psz_html
, psz_subtitle
, strspn(psz_subtitle
, " ") );
731 HtmlPut( &psz_html
, ">" );
734 else if( !strncmp( psz_subtitle
, "</", 2 ))
736 bool b_match
= false;
737 bool b_ignore
= false;
738 int i_len
= (psz_tag
? strlen(psz_tag
) : 0) - 1;
739 char *psz_lastTag
= NULL
;
743 psz_lastTag
= psz_tag
+ i_len
;
746 switch( *psz_lastTag
)
749 b_match
= !strncasecmp( psz_subtitle
, "</b>", 4 );
753 b_match
= !strncasecmp( psz_subtitle
, "</i>", 4 );
757 b_match
= !strncasecmp( psz_subtitle
, "</u>", 4 );
761 b_match
= !strncasecmp( psz_subtitle
, "</s>", 4 );
765 b_match
= !strncasecmp( psz_subtitle
, "</font>", 7 );
769 i_len
= strcspn( psz_subtitle
, ">" );
770 b_match
= psz_subtitle
[i_len
] == '>';
779 /* Not well formed -- kill everything */
786 HtmlNPut( &psz_html
, psz_subtitle
, i_len
);
788 psz_subtitle
+= i_len
;
790 else if( ( psz_subtitle
[1] < 'a' || psz_subtitle
[1] > 'z' ) &&
791 ( psz_subtitle
[1] < 'A' || psz_subtitle
[1] > 'Z' ) )
793 /* We have a single < */
794 HtmlPut( &psz_html
, "<" );
799 /* We have an unknown tag or a single < */
801 /* Search for the next tag or end of tag or end of string */
802 char *psz_stop
= psz_subtitle
+ 1 + strcspn( &psz_subtitle
[1], "<>" );
803 char *psz_closing
= strstr( psz_subtitle
, "/>" );
805 if( psz_closing
&& psz_closing
< psz_stop
)
807 /* We have a self closed tag, remove it */
808 psz_subtitle
= &psz_closing
[2];
810 else if( *psz_stop
== '>' )
814 snprintf( psz_match
, sizeof(psz_match
), "</%s", &psz_subtitle
[1] );
815 psz_match
[strcspn( psz_match
, " \t>" )] = '\0';
817 if( strstr( psz_subtitle
, psz_match
) )
819 /* We have the closing tag, ignore it TODO */
820 psz_subtitle
= &psz_stop
[1];
821 HtmlPut( &psz_tag
, "I" );
825 int i_len
= psz_stop
+ 1 - psz_subtitle
;
827 /* Copy the whole data */
828 for( ; i_len
> 0; i_len
--, psz_subtitle
++ )
830 if( *psz_subtitle
== '<' )
831 HtmlPut( &psz_html
, "<" );
832 else if( *psz_subtitle
== '>' )
833 HtmlPut( &psz_html
, ">" );
835 HtmlNPut( &psz_html
, psz_subtitle
, 1 );
841 /* We have a single < */
842 HtmlPut( &psz_html
, "<" );
847 else if( *psz_subtitle
== '&' )
849 if( !strncasecmp( psz_subtitle
, "<", 4 ))
851 HtmlCopy( &psz_html
, &psz_subtitle
, "<" );
853 else if( !strncasecmp( psz_subtitle
, ">", 4 ))
855 HtmlCopy( &psz_html
, &psz_subtitle
, ">" );
857 else if( !strncasecmp( psz_subtitle
, "&", 5 ))
859 HtmlCopy( &psz_html
, &psz_subtitle
, "&" );
863 HtmlPut( &psz_html
, "&" );
867 else if( *psz_subtitle
== '>' )
869 HtmlPut( &psz_html
, ">" );
872 else if( psz_subtitle
[0] == '{' && psz_subtitle
[1] == '\\' &&
873 strchr( psz_subtitle
, '}' ) )
875 /* Check for forced alignment */
877 !strncmp( psz_subtitle
, "{\\an", 4 ) && psz_subtitle
[4] >= '1' && psz_subtitle
[4] <= '9' && psz_subtitle
[5] == '}' )
879 static const int pi_vertical
[3] = { SUBPICTURE_ALIGN_BOTTOM
, 0, SUBPICTURE_ALIGN_TOP
};
880 static const int pi_horizontal
[3] = { SUBPICTURE_ALIGN_LEFT
, 0, SUBPICTURE_ALIGN_RIGHT
};
881 const int i_id
= psz_subtitle
[4] - '1';
884 *pi_align
= pi_vertical
[i_id
/3] | pi_horizontal
[i_id
%3];
886 /* TODO fr -> rotation */
888 /* Hide {\stupidity} */
889 psz_subtitle
= strchr( psz_subtitle
, '}' ) + 1;
891 else if( psz_subtitle
[0] == '{' &&
892 ( psz_subtitle
[1] == 'Y' || psz_subtitle
[1] == 'y' )
893 && psz_subtitle
[2] == ':' && strchr( psz_subtitle
, '}' ) )
895 // FIXME: We don't do difference between Y and y, and we should.
896 if( psz_subtitle
[3] == 'i' )
898 HtmlPut( &psz_html
, "<i>" );
899 HtmlPut( &psz_tag
, "i" );
901 if( psz_subtitle
[3] == 'b' )
903 HtmlPut( &psz_html
, "<b>" );
904 HtmlPut( &psz_tag
, "b" );
906 if( psz_subtitle
[3] == 'u' )
908 HtmlPut( &psz_html
, "<u>" );
909 HtmlPut( &psz_tag
, "u" );
911 psz_subtitle
= strchr( psz_subtitle
, '}' ) + 1;
913 else if( psz_subtitle
[0] == '{' && psz_subtitle
[2] == ':' && strchr( psz_subtitle
, '}' ) )
915 // Hide other {x:y} atrocities, like {c:$bbggrr} or {P:x}
916 psz_subtitle
= strchr( psz_subtitle
, '}' ) + 1;
918 else if( psz_subtitle
[0] == '\\' && psz_subtitle
[1] )
920 if( psz_subtitle
[1] == 'N' || psz_subtitle
[1] == 'n' )
922 HtmlPut( &psz_html
, "<br/>" );
925 else if( psz_subtitle
[1] == 'h' )
927 /* Non breakable space */
928 HtmlPut( &psz_html
, NO_BREAKING_SPACE
);
933 HtmlPut( &psz_html
, "\\" );
939 HtmlNPut( &psz_html
, psz_subtitle
, 1 );
943 /* Check for double whitespace */
944 # error This test does not make sense.
945 if( ( *psz_html
== ' ' || *psz_html
== '\t' ) &&
946 ( *(psz_html
-1) == ' ' || *(psz_html
-1) == '\t' ) )
948 HtmlPut( &psz_html
, NO_BREAKING_SPACE
);
957 while( psz_tag
&& *psz_tag
)
960 char *psz_last
= &psz_tag
[strlen(psz_tag
)-1];
964 HtmlPut( &psz_html
, "</b>" );
967 HtmlPut( &psz_html
, "</i>" );
970 HtmlPut( &psz_html
, "</u>" );
973 HtmlPut( &psz_html
, "</s>" );
976 HtmlPut( &psz_html
, "</font>" );
983 /* Close not well formed subtitle */
984 HtmlPut( &psz_html
, "</text>" );