1 /*****************************************************************************
2 * subsdec.c : text subtitles decoder
3 *****************************************************************************
4 * Copyright (C) 2000-2006 the VideoLAN team
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
13 * it under the terms of the GNU General Public License as published by
14 * the Free Software Foundation; either version 2 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 General Public License for more details.
22 * You should have received a copy of the GNU General Public License
23 * along with this program; if not, write to the Free Software
24 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
25 *****************************************************************************/
27 /*****************************************************************************
29 *****************************************************************************/
34 #include <vlc_common.h>
35 #include <vlc_plugin.h>
36 #include <vlc_codec.h>
37 #include <vlc_charset.h>
41 /*****************************************************************************
43 *****************************************************************************/
44 static const char *const ppsz_encodings[] = {
90 static const char *const ppsz_encoding_names[] = {
92 The character encoding name in parenthesis corresponds to that used for
93 the GetACP translation. "Windows-1252" applies to Western European
94 languages using the Latin alphabet. */
95 N_("Default (Windows-1252)"),
97 N_("Universal (UTF-8)"),
98 N_("Universal (UTF-16)"),
99 N_("Universal (big endian UTF-16)"),
100 N_("Universal (little endian UTF-16)"),
101 N_("Universal, Chinese (GB18030)"),
103 /* ISO 8859 and the likes */
105 N_("Western European (Latin-9)"), /* mostly superset of Latin-1 */
106 N_("Western European (Windows-1252)"),
107 N_("Western European (IBM 00850)"),
109 N_("Eastern European (Latin-2)"),
110 N_("Eastern European (Windows-1250)"),
112 N_("Esperanto (Latin-3)"),
114 N_("Nordic (Latin-6)"), /* Latin 6 supersedes Latin 4 */
116 N_("Cyrillic (Windows-1251)"), /* ISO 8859-5 is not practically used */
117 N_("Russian (KOI8-R)"),
118 N_("Ukrainian (KOI8-U)"),
120 N_("Arabic (ISO 8859-6)"),
121 N_("Arabic (Windows-1256)"),
123 N_("Greek (ISO 8859-7)"),
124 N_("Greek (Windows-1253)"),
126 N_("Hebrew (ISO 8859-8)"),
127 N_("Hebrew (Windows-1255)"),
129 N_("Turkish (ISO 8859-9)"),
130 N_("Turkish (Windows-1254)"),
133 N_("Thai (TIS 620-2533/ISO 8859-11)"),
134 N_("Thai (Windows-874)"),
136 N_("Baltic (Latin-7)"),
137 N_("Baltic (Windows-1257)"),
138 /* 12 -> /dev/null */
140 N_("Celtic (Latin-8)"),
143 N_("South-Eastern European (Latin-10)"),
145 N_("Simplified Chinese (ISO-2022-CN-EXT)"),
146 N_("Simplified Chinese Unix (EUC-CN)"),
147 N_("Japanese (7-bits JIS/ISO-2022-JP-2)"),
148 N_("Japanese Unix (EUC-JP)"),
149 N_("Japanese (Shift JIS)"),
150 N_("Korean (EUC-KR/CP949)"),
151 N_("Korean (ISO-2022-KR)"),
152 N_("Traditional Chinese (Big5)"),
153 N_("Traditional Chinese Unix (EUC-TW)"),
154 N_("Hong-Kong Supplementary (HKSCS)"),
156 N_("Vietnamese (VISCII)"),
157 N_("Vietnamese (Windows-1258)"),
160 static const int pi_justification[] = { 0, 1, 2 };
161 static const char *const ppsz_justification_text[] = {
162 N_("Center"),N_("Left"),N_("Right")};
164 #define ENCODING_TEXT N_("Subtitles text encoding")
165 #define ENCODING_LONGTEXT N_("Set the encoding used in text subtitles")
166 #define ALIGN_TEXT N_("Subtitles justification")
167 #define ALIGN_LONGTEXT N_("Set the justification of subtitles")
168 #define AUTODETECT_UTF8_TEXT N_("UTF-8 subtitles autodetection")
169 #define AUTODETECT_UTF8_LONGTEXT N_("This enables automatic detection of " \
170 "UTF-8 encoding within subtitles files.")
171 #define FORMAT_TEXT N_("Formatted Subtitles")
172 #define FORMAT_LONGTEXT N_("Some subtitle formats allow for text formatting. " \
173 "VLC partly implements this, but you can choose to disable all formatting.")
175 static int OpenDecoder ( vlc_object_t * );
176 static void CloseDecoder ( vlc_object_t * );
179 set_shortname( N_("Subtitles"))
180 set_description( N_("Text subtitles decoder") )
181 set_capability( "decoder", 50 )
182 set_callbacks( OpenDecoder, CloseDecoder )
183 set_category( CAT_INPUT )
184 set_subcategory( SUBCAT_INPUT_SCODEC )
186 add_integer( "subsdec-align", 0, ALIGN_TEXT, ALIGN_LONGTEXT,
188 change_integer_list( pi_justification, ppsz_justification_text )
189 add_string( "subsdec-encoding", "",
190 ENCODING_TEXT, ENCODING_LONGTEXT, false )
191 change_string_list( ppsz_encodings, ppsz_encoding_names, 0 )
192 add_bool( "subsdec-autodetect-utf8", true,
193 AUTODETECT_UTF8_TEXT, AUTODETECT_UTF8_LONGTEXT, false )
194 add_bool( "subsdec-formatted", true, FORMAT_TEXT, FORMAT_LONGTEXT,
198 /*****************************************************************************
200 *****************************************************************************/
201 #define NO_BREAKING_SPACE " "
205 int i_align; /* Subtitles alignment on the vout */
207 vlc_iconv_t iconv_handle; /* handle to iconv instance */
208 bool b_autodetect_utf8;
212 static subpicture_t *DecodeBlock ( decoder_t *, block_t ** );
213 static subpicture_t *ParseText ( decoder_t *, block_t * );
214 static char *StripTags ( char * );
215 static char *CreateHtmlSubtitle( int *pi_align, char * );
217 /*****************************************************************************
218 * OpenDecoder: probe the decoder and return score
219 *****************************************************************************
220 * Tries to launch a decoder and return score so that the interface is able
222 *****************************************************************************/
223 static int OpenDecoder( vlc_object_t *p_this )
225 decoder_t *p_dec = (decoder_t*)p_this;
226 decoder_sys_t *p_sys;
228 switch( p_dec->fmt_in.i_codec )
231 case VLC_CODEC_ITU_T140:
237 p_dec->pf_decode_sub = DecodeBlock;
238 p_dec->fmt_out.i_cat = SPU_ES;
239 p_dec->fmt_out.i_codec = 0;
241 /* Allocate the memory needed to store the decoder's structure */
242 p_dec->p_sys = p_sys = calloc( 1, sizeof( *p_sys ) );
248 p_sys->iconv_handle = (vlc_iconv_t)-1;
249 p_sys->b_autodetect_utf8 = false;
251 const char *encoding;
254 /* First try demux-specified encoding */
255 if( p_dec->fmt_in.i_codec == VLC_CODEC_ITU_T140 )
256 encoding = "UTF-8"; /* IUT T.140 is always using UTF-8 */
258 if( p_dec->fmt_in.subs.psz_encoding && *p_dec->fmt_in.subs.psz_encoding )
260 encoding = p_dec->fmt_in.subs.psz_encoding;
261 msg_Dbg (p_dec, "trying demuxer-specified character encoding: %s",
265 /* Second, try configured encoding */
266 if ((var = var_InheritString (p_dec, "subsdec-encoding")) != NULL)
268 msg_Dbg (p_dec, "trying configured character encoding: %s", var);
269 if (!strcmp (var, "system"))
274 /* ^ iconv() treats "" as nl_langinfo(CODESET) */
280 /* Third, try "local" encoding with optional UTF-8 autodetection */
283 The Windows ANSI code page most commonly used for this language.
284 VLC uses this as a guess of the subtitle files character set
285 (if UTF-8 and UTF-16 autodetection fails).
286 Western European languages normally use "CP1252", which is a
287 Microsoft-variant of ISO 8859-1. That suits the Latin alphabet.
288 Other scripts use other code pages.
290 This MUST be a valid iconv character set. If unsure, please refer
291 the VideoLAN translators mailing list. */
292 encoding = vlc_pgettext("GetACP", "CP1252");
293 msg_Dbg (p_dec, "trying default character encoding: %s", encoding);
294 if (var_InheritBool (p_dec, "subsdec-autodetect-utf8"))
296 msg_Dbg (p_dec, "using automatic UTF-8 detection");
297 p_sys->b_autodetect_utf8 = true;
301 if (strcasecmp (encoding, "UTF-8") && strcasecmp (encoding, "utf8"))
303 p_sys->iconv_handle = vlc_iconv_open ("UTF-8", encoding);
304 if (p_sys->iconv_handle == (vlc_iconv_t)(-1))
305 msg_Err (p_dec, "cannot convert from %s: %m", encoding);
309 p_sys->i_align = var_InheritInteger( p_dec, "subsdec-align" );
314 /****************************************************************************
315 * DecodeBlock: the whole thing
316 ****************************************************************************
317 * This function must be fed with complete subtitles units.
318 ****************************************************************************/
319 static subpicture_t *DecodeBlock( decoder_t *p_dec, block_t **pp_block )
324 if( !pp_block || *pp_block == NULL )
328 if( p_block->i_flags & (BLOCK_FLAG_DISCONTINUITY|BLOCK_FLAG_CORRUPTED) )
330 block_Release( p_block );
334 p_spu = ParseText( p_dec, p_block );
336 block_Release( p_block );
342 /*****************************************************************************
343 * CloseDecoder: clean up the decoder
344 *****************************************************************************/
345 static void CloseDecoder( vlc_object_t *p_this )
347 decoder_t *p_dec = (decoder_t *)p_this;
348 decoder_sys_t *p_sys = p_dec->p_sys;
350 if( p_sys->iconv_handle != (vlc_iconv_t)-1 )
351 vlc_iconv_close( p_sys->iconv_handle );
356 /*****************************************************************************
357 * ParseText: parse an text subtitle packet and send it to the video output
358 *****************************************************************************/
359 static subpicture_t *ParseText( decoder_t *p_dec, block_t *p_block )
361 decoder_sys_t *p_sys = p_dec->p_sys;
362 subpicture_t *p_spu = NULL;
363 char *psz_subtitle = NULL;
365 /* We cannot display a subpicture with no date */
366 if( p_block->i_pts <= VLC_TS_INVALID )
368 msg_Warn( p_dec, "subtitle without a date" );
372 /* Check validity of packet data */
373 /* An "empty" line containing only \0 can be used to force
374 and ephemer picture from the screen */
375 if( p_block->i_buffer < 1 )
377 msg_Warn( p_dec, "no subtitle data" );
381 /* Should be resiliant against bad subtitles */
382 psz_subtitle = malloc( p_block->i_buffer + 1 );
383 if( psz_subtitle == NULL )
385 memcpy( psz_subtitle, p_block->p_buffer, p_block->i_buffer );
386 psz_subtitle[p_block->i_buffer] = '\0';
388 if( p_sys->iconv_handle == (vlc_iconv_t)-1 )
390 if (EnsureUTF8( psz_subtitle ) == NULL)
392 msg_Err( p_dec, "failed to convert subtitle encoding.\n"
393 "Try manually setting a character-encoding "
394 "before you open the file." );
400 if( p_sys->b_autodetect_utf8 )
402 if( IsUTF8( psz_subtitle ) == NULL )
404 msg_Dbg( p_dec, "invalid UTF-8 sequence: "
405 "disabling UTF-8 subtitles autodetection" );
406 p_sys->b_autodetect_utf8 = false;
410 if( !p_sys->b_autodetect_utf8 )
412 size_t inbytes_left = strlen( psz_subtitle );
413 size_t outbytes_left = 6 * inbytes_left;
414 char *psz_new_subtitle = xmalloc( outbytes_left + 1 );
415 char *psz_convert_buffer_out = psz_new_subtitle;
416 const char *psz_convert_buffer_in = psz_subtitle;
418 size_t ret = vlc_iconv( p_sys->iconv_handle,
419 &psz_convert_buffer_in, &inbytes_left,
420 &psz_convert_buffer_out, &outbytes_left );
422 *psz_convert_buffer_out++ = '\0';
423 free( psz_subtitle );
425 if( ( ret == (size_t)(-1) ) || inbytes_left )
427 free( psz_new_subtitle );
428 msg_Err( p_dec, "failed to convert subtitle encoding.\n"
429 "Try manually setting a character-encoding "
430 "before you open the file." );
434 psz_subtitle = realloc( psz_new_subtitle,
435 psz_convert_buffer_out - psz_new_subtitle );
437 psz_subtitle = psz_new_subtitle;
441 /* Create the subpicture unit */
442 p_spu = decoder_NewSubpictureText( p_dec );
445 free( psz_subtitle );
448 p_spu->i_start = p_block->i_pts;
449 p_spu->i_stop = p_block->i_pts + p_block->i_length;
450 p_spu->b_ephemer = (p_block->i_length == 0);
451 p_spu->b_absolute = false;
453 subpicture_updater_sys_t *p_spu_sys = p_spu->updater.p_sys;
455 p_spu_sys->align = SUBPICTURE_ALIGN_BOTTOM | p_sys->i_align;
456 p_spu_sys->text = StripTags( psz_subtitle );
457 if( var_InheritBool( p_dec, "subsdec-formatted" ) )
458 p_spu_sys->html = CreateHtmlSubtitle( &p_spu_sys->align, psz_subtitle );
460 free( psz_subtitle );
465 /* Function now handles tags with attribute values, and tries
466 * to deal with &' commands too. It no longer modifies the string
467 * in place, so that the original text can be reused
469 static char *StripTags( char *psz_subtitle )
471 char *psz_text_start;
474 psz_text = psz_text_start = malloc( strlen( psz_subtitle ) + 1 );
475 if( !psz_text_start )
478 while( *psz_subtitle )
480 if( *psz_subtitle == '<' )
482 if( strncasecmp( psz_subtitle, "<br/>", 5 ) == 0 )
485 psz_subtitle += strcspn( psz_subtitle, ">" );
487 else if( *psz_subtitle == '&' )
489 if( !strncasecmp( psz_subtitle, "<", 4 ))
492 psz_subtitle += strcspn( psz_subtitle, ";" );
494 else if( !strncasecmp( psz_subtitle, ">", 4 ))
497 psz_subtitle += strcspn( psz_subtitle, ";" );
499 else if( !strncasecmp( psz_subtitle, "&", 5 ))
502 psz_subtitle += strcspn( psz_subtitle, ";" );
504 else if( !strncasecmp( psz_subtitle, """, 6 ))
507 psz_subtitle += strcspn( psz_subtitle, ";" );
511 /* Assume it is just a normal ampersand */
517 *psz_text++ = *psz_subtitle;
520 /* Security fix: Account for the case where input ends early */
521 if( *psz_subtitle == '\0' ) break;
526 char *psz = realloc( psz_text_start, strlen( psz_text_start ) + 1 );
527 if( psz ) psz_text_start = psz;
529 return psz_text_start;
532 /* Try to respect any style tags present in the subtitle string. The main
533 * problem here is a lack of adequate specs for the subtitle formats.
534 * SSA/ASS and USF are both detail spec'ed -- but they are handled elsewhere.
535 * SAMI has a detailed spec, but extensive rework is needed in the demux
536 * code to prevent all this style information being excised, as it presently
538 * That leaves the others - none of which were (I guess) originally intended
539 * to be carrying style information. Over time people have used them that way.
540 * In the absence of specifications from which to work, the tags supported
541 * have been restricted to the simple set permitted by the USF DTD, ie. :
542 * Basic: <br>, <i>, <b>, <u>, <s>
554 * There is also the further restriction that the subtitle be well-formed
555 * as an XML entity, ie. the HTML sentence:
556 * <b><i>Bold and Italics</b></i>
557 * doesn't qualify because the tags aren't nested one inside the other.
558 * <text> tags are automatically added to the output to ensure
560 * If the text doesn't qualify for any reason, a NULL string is
561 * returned, and the rendering engine will fall back to the
562 * plain text version of the subtitle.
564 static void HtmlNPut( char **ppsz_html, const char *psz_text, int i_max )
566 const int i_len = strlen(psz_text);
568 strncpy( *ppsz_html, psz_text, i_max );
569 *ppsz_html += __MIN(i_max,i_len);
572 static void HtmlPut( char **ppsz_html, const char *psz_text )
574 strcpy( *ppsz_html, psz_text );
575 *ppsz_html += strlen(psz_text);
577 static void HtmlCopy( char **ppsz_html, char **ppsz_subtitle, const char *psz_text )
579 HtmlPut( ppsz_html, psz_text );
580 *ppsz_subtitle += strlen(psz_text);
583 static char *CreateHtmlSubtitle( int *pi_align, char *psz_subtitle )
586 char *psz_tag = malloc( ( strlen( psz_subtitle ) / 3 ) + 1 );
593 size_t i_buf_size = strlen( psz_subtitle ) + 100;
594 char *psz_html_start = malloc( i_buf_size );
595 char *psz_html = psz_html_start;
596 if( psz_html_start == NULL )
603 bool b_has_align = false;
605 HtmlPut( &psz_html, "<text>" );
608 while( *psz_subtitle )
610 if( *psz_subtitle == '\n' )
612 HtmlPut( &psz_html, "<br/>" );
615 else if( *psz_subtitle == '<' )
617 if( !strncasecmp( psz_subtitle, "<br/>", 5 ))
619 HtmlCopy( &psz_html, &psz_subtitle, "<br/>" );
621 else if( !strncasecmp( psz_subtitle, "<b>", 3 ) )
623 HtmlCopy( &psz_html, &psz_subtitle, "<b>" );
624 strcat( psz_tag, "b" );
626 else if( !strncasecmp( psz_subtitle, "<i>", 3 ) )
628 HtmlCopy( &psz_html, &psz_subtitle, "<i>" );
629 strcat( psz_tag, "i" );
631 else if( !strncasecmp( psz_subtitle, "<u>", 3 ) )
633 HtmlCopy( &psz_html, &psz_subtitle, "<u>" );
634 strcat( psz_tag, "u" );
636 else if( !strncasecmp( psz_subtitle, "<s>", 3 ) )
638 HtmlCopy( &psz_html, &psz_subtitle, "<s>" );
639 strcat( psz_tag, "s" );
641 else if( !strncasecmp( psz_subtitle, "<font ", 6 ))
643 const char *psz_attribs[] = { "face=", "family=", "size=",
644 "color=", "outline-color=", "shadow-color=",
645 "outline-level=", "shadow-level=", "back-color=",
648 HtmlCopy( &psz_html, &psz_subtitle, "<font " );
649 strcat( psz_tag, "f" );
651 while( *psz_subtitle != '>' )
655 for( k=0; psz_attribs[ k ]; k++ )
657 int i_len = strlen( psz_attribs[ k ] );
659 if( !strncasecmp( psz_subtitle, psz_attribs[k], i_len ) )
662 HtmlPut( &psz_html, psz_attribs[k] );
663 psz_subtitle += i_len;
666 if( *psz_subtitle == '"' )
669 i_len = strcspn( psz_subtitle, "\"" );
671 else if( *psz_subtitle == '\'' )
674 i_len = strcspn( psz_subtitle, "'" );
678 i_len = strcspn( psz_subtitle, " \t>" );
680 HtmlPut( &psz_html, "\"" );
681 HtmlNPut( &psz_html, psz_subtitle, i_len );
682 HtmlPut( &psz_html, "\"" );
684 psz_subtitle += i_len;
685 if( *psz_subtitle == '\"' || *psz_subtitle == '\'' )
690 if( psz_attribs[ k ] == NULL )
692 /* Jump over unrecognised tag */
693 int i_len = strcspn( psz_subtitle, "\"" );
694 if( psz_subtitle[i_len] == '\"' )
696 i_len += 1 + strcspn( &psz_subtitle[i_len + 1], "\"" );
697 if( psz_subtitle[i_len] == '\"' )
700 /* Not a tag, something else we do not understand */
704 psz_subtitle += i_len;
706 while (*psz_subtitle == ' ')
707 *psz_html++ = *psz_subtitle++;
712 else if( !strncmp( psz_subtitle, "</", 2 ))
714 bool b_match = false;
715 bool b_ignore = false;
716 int i_len = strlen( psz_tag ) - 1;
717 char *psz_lastTag = NULL;
721 psz_lastTag = psz_tag + i_len;
724 switch( *psz_lastTag )
727 b_match = !strncasecmp( psz_subtitle, "</b>", 4 );
731 b_match = !strncasecmp( psz_subtitle, "</i>", 4 );
735 b_match = !strncasecmp( psz_subtitle, "</u>", 4 );
739 b_match = !strncasecmp( psz_subtitle, "</s>", 4 );
743 b_match = !strncasecmp( psz_subtitle, "</font>", 7 );
747 i_len = strcspn( psz_subtitle, ">" );
748 b_match = psz_subtitle[i_len] == '>';
757 /* Not well formed -- kill everything */
758 free( psz_html_start );
759 psz_html_start = NULL;
764 HtmlNPut( &psz_html, psz_subtitle, i_len );
766 psz_subtitle += i_len;
768 else if( ( psz_subtitle[1] < 'a' || psz_subtitle[1] > 'z' ) &&
769 ( psz_subtitle[1] < 'A' || psz_subtitle[1] > 'Z' ) )
771 /* We have a single < */
772 HtmlPut( &psz_html, "<" );
777 /* We have an unknown tag or a single < */
779 /* Search for the next tag or end of tag or end of string */
780 char *psz_stop = psz_subtitle + 1 + strcspn( &psz_subtitle[1], "<>" );
781 char *psz_closing = strstr( psz_subtitle, "/>" );
783 if( psz_closing && psz_closing < psz_stop )
785 /* We have a self closed tag, remove it */
786 psz_subtitle = &psz_closing[2];
788 else if( *psz_stop == '>' )
792 snprintf( psz_match, sizeof(psz_match), "</%s", &psz_subtitle[1] );
793 psz_match[strcspn( psz_match, " \t>" )] = '\0';
795 if( strstr( psz_subtitle, psz_match ) )
797 /* We have the closing tag, ignore it TODO */
798 psz_subtitle = &psz_stop[1];
799 strcat( psz_tag, "I" );
803 int i_len = psz_stop + 1 - psz_subtitle;
805 /* Copy the whole data */
806 for( ; i_len > 0; i_len--, psz_subtitle++ )
808 if( *psz_subtitle == '<' )
809 HtmlPut( &psz_html, "<" );
810 else if( *psz_subtitle == '>' )
811 HtmlPut( &psz_html, ">" );
813 *psz_html++ = *psz_subtitle;
819 /* We have a single < */
820 HtmlPut( &psz_html, "<" );
825 else if( *psz_subtitle == '&' )
827 if( !strncasecmp( psz_subtitle, "<", 4 ))
829 HtmlCopy( &psz_html, &psz_subtitle, "<" );
831 else if( !strncasecmp( psz_subtitle, ">", 4 ))
833 HtmlCopy( &psz_html, &psz_subtitle, ">" );
835 else if( !strncasecmp( psz_subtitle, "&", 5 ))
837 HtmlCopy( &psz_html, &psz_subtitle, "&" );
841 HtmlPut( &psz_html, "&" );
845 else if( *psz_subtitle == '>' )
847 HtmlPut( &psz_html, ">" );
850 else if( psz_subtitle[0] == '{' && psz_subtitle[1] == '\\' &&
851 strchr( psz_subtitle, '}' ) )
853 /* Check for forced alignment */
855 !strncmp( psz_subtitle, "{\\an", 4 ) && psz_subtitle[4] >= '1' && psz_subtitle[4] <= '9' && psz_subtitle[5] == '}' )
857 static const int pi_vertical[3] = { SUBPICTURE_ALIGN_BOTTOM, 0, SUBPICTURE_ALIGN_TOP };
858 static const int pi_horizontal[3] = { SUBPICTURE_ALIGN_LEFT, 0, SUBPICTURE_ALIGN_RIGHT };
859 const int i_id = psz_subtitle[4] - '1';
862 *pi_align = pi_vertical[i_id/3] | pi_horizontal[i_id%3];
864 /* TODO fr -> rotation */
866 /* Hide {\stupidity} */
867 psz_subtitle = strchr( psz_subtitle, '}' ) + 1;
869 else if( psz_subtitle[0] == '{' &&
870 ( psz_subtitle[1] == 'Y' || psz_subtitle[1] == 'y' )
871 && psz_subtitle[2] == ':' && strchr( psz_subtitle, '}' ) )
873 // FIXME: We don't do difference between Y and y, and we should.
874 if( psz_subtitle[3] == 'i' )
876 HtmlPut( &psz_html, "<i>" );
877 strcat( psz_tag, "i" );
879 if( psz_subtitle[3] == 'b' )
881 HtmlPut( &psz_html, "<b>" );
882 strcat( psz_tag, "b" );
884 if( psz_subtitle[3] == 'u' )
886 HtmlPut( &psz_html, "<u>" );
887 strcat( psz_tag, "u" );
889 psz_subtitle = strchr( psz_subtitle, '}' ) + 1;
891 else if( psz_subtitle[0] == '{' && psz_subtitle[2] == ':' && strchr( psz_subtitle, '}' ) )
893 // Hide other {x:y} atrocities, like {c:$bbggrr} or {P:x}
894 psz_subtitle = strchr( psz_subtitle, '}' ) + 1;
896 else if( psz_subtitle[0] == '\\' && psz_subtitle[1] )
898 if( psz_subtitle[1] == 'N' || psz_subtitle[1] == 'n' )
900 HtmlPut( &psz_html, "<br/>" );
903 else if( psz_subtitle[1] == 'h' )
905 /* Non breakable space */
906 HtmlPut( &psz_html, NO_BREAKING_SPACE );
911 HtmlPut( &psz_html, "\\" );
917 *psz_html = *psz_subtitle;
918 if( psz_html > psz_html_start )
920 /* Check for double whitespace */
921 if( ( *psz_html == ' ' || *psz_html == '\t' ) &&
922 ( *(psz_html-1) == ' ' || *(psz_html-1) == '\t' ) )
924 HtmlPut( &psz_html, NO_BREAKING_SPACE );
932 if( ( size_t )( psz_html - psz_html_start ) > i_buf_size - 50 )
934 const int i_len = psz_html - psz_html_start;
937 char *psz_new = realloc( psz_html_start, i_buf_size );
940 psz_html_start = psz_new;
941 psz_html = &psz_new[i_len];
946 static const char *psz_text_close = "</text>";
947 static const char *psz_tag_long = "/font>";
949 /* Realloc for closing tags and shrink memory */
950 const size_t i_length = (size_t)( psz_html - psz_html_start );
952 const size_t i_size = i_length + strlen(psz_tag_long) * strlen(psz_tag) + strlen(psz_text_close) + 1;
953 char *psz_new = realloc( psz_html_start, i_size );
956 psz_html_start = psz_new;
957 psz_html = &psz_new[i_length];
959 /* Close not well formed subtitle */
963 char *psz_last = &psz_tag[strlen(psz_tag)-1];
967 HtmlPut( &psz_html, "</b>" );
970 HtmlPut( &psz_html, "</i>" );
973 HtmlPut( &psz_html, "</u>" );
976 HtmlPut( &psz_html, "</s>" );
979 HtmlPut( &psz_html, "/font>" );
987 HtmlPut( &psz_html, psz_text_close );
992 return psz_html_start;