avcodec/htmlsubtitles: Replace very slow redundant sscanf() calls by cleaner and...
[ffmpeg.git] / libavcodec / htmlsubtitles.c
1 /*
2  * Copyright (c) 2010  Aurelien Jacobs <aurel@gnuage.org>
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20
21 #include "libavutil/avstring.h"
22 #include "libavutil/common.h"
23 #include "libavutil/parseutils.h"
24 #include "htmlsubtitles.h"
25
26 static int html_color_parse(void *log_ctx, const char *str)
27 {
28     uint8_t rgba[4];
29     if (av_parse_color(rgba, str, strcspn(str, "\" >"), log_ctx) < 0)
30         return -1;
31     return rgba[0] | rgba[1] << 8 | rgba[2] << 16;
32 }
33
34 enum {
35     PARAM_UNKNOWN = -1,
36     PARAM_SIZE,
37     PARAM_COLOR,
38     PARAM_FACE,
39     PARAM_NUMBER
40 };
41
42 typedef struct SrtStack {
43     char tag[128];
44     char param[PARAM_NUMBER][128];
45 } SrtStack;
46
47 static void rstrip_spaces_buf(AVBPrint *buf)
48 {
49     if (av_bprint_is_complete(buf))
50         while (buf->len > 0 && buf->str[buf->len - 1] == ' ')
51             buf->str[--buf->len] = 0;
52 }
53
54 int ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
55 {
56     char *param, buffer[128], tmp[128];
57     int len, tag_close, sptr = 1, line_start = 1, an = 0, end = 0;
58     SrtStack stack[16];
59     int closing_brace_missing = 0;
60
61     stack[0].tag[0] = 0;
62     strcpy(stack[0].param[PARAM_SIZE],  "{\\fs}");
63     strcpy(stack[0].param[PARAM_COLOR], "{\\c}");
64     strcpy(stack[0].param[PARAM_FACE],  "{\\fn}");
65
66     for (; !end && *in; in++) {
67         switch (*in) {
68         case '\r':
69             break;
70         case '\n':
71             if (line_start) {
72                 end = 1;
73                 break;
74             }
75             rstrip_spaces_buf(dst);
76             av_bprintf(dst, "\\N");
77             line_start = 1;
78             break;
79         case ' ':
80             if (!line_start)
81                 av_bprint_chars(dst, *in, 1);
82             break;
83         case '{':    /* skip all {\xxx} substrings except for {\an%d}
84                         and all microdvd like styles such as {Y:xxx} */
85             len = 0;
86             an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
87
88             if (!closing_brace_missing) {
89                 if (   (an != 1 && in[1] == '\\')
90                     || (in[1] && strchr("CcFfoPSsYy", in[1]) && in[2] == ':')) {
91                     char *bracep = strchr(in+2, '}');
92                     if (bracep) {
93                         in = bracep;
94                         break;
95                     } else
96                         closing_brace_missing = 1;
97                 }
98             }
99
100             av_bprint_chars(dst, *in, 1);
101             break;
102         case '<':
103             tag_close = in[1] == '/';
104             len = 0;
105             if (sscanf(in+tag_close+1, "%127[^>]>%n", buffer, &len) >= 1 && len > 0) {
106                 const char *tagname = buffer;
107                 while (*tagname == ' ')
108                     tagname++;
109                 if ((param = strchr(tagname, ' ')))
110                     *param++ = 0;
111                 if ((!tag_close && sptr < FF_ARRAY_ELEMS(stack)) ||
112                     ( tag_close && sptr > 0 && !strcmp(stack[sptr-1].tag, tagname))) {
113                     int i, j, unknown = 0;
114                     in += len + tag_close;
115                     if (!tag_close)
116                         memset(stack+sptr, 0, sizeof(*stack));
117                     if (!strcmp(tagname, "font")) {
118                         if (tag_close) {
119                             for (i=PARAM_NUMBER-1; i>=0; i--)
120                                 if (stack[sptr-1].param[i][0])
121                                     for (j=sptr-2; j>=0; j--)
122                                         if (stack[j].param[i][0]) {
123                                             av_bprintf(dst, "%s", stack[j].param[i]);
124                                             break;
125                                         }
126                         } else {
127                             while (param) {
128                                 if (!strncmp(param, "size=", 5)) {
129                                     unsigned font_size;
130                                     param += 5 + (param[5] == '"');
131                                     if (sscanf(param, "%u", &font_size) == 1) {
132                                         snprintf(stack[sptr].param[PARAM_SIZE],
133                                              sizeof(stack[0].param[PARAM_SIZE]),
134                                              "{\\fs%u}", font_size);
135                                     }
136                                 } else if (!strncmp(param, "color=", 6)) {
137                                     param += 6 + (param[6] == '"');
138                                     snprintf(stack[sptr].param[PARAM_COLOR],
139                                          sizeof(stack[0].param[PARAM_COLOR]),
140                                          "{\\c&H%X&}",
141                                          html_color_parse(log_ctx, param));
142                                 } else if (!strncmp(param, "face=", 5)) {
143                                     param += 5 + (param[5] == '"');
144                                     len = strcspn(param,
145                                                   param[-1] == '"' ? "\"" :" ");
146                                     av_strlcpy(tmp, param,
147                                                FFMIN(sizeof(tmp), len+1));
148                                     param += len;
149                                     snprintf(stack[sptr].param[PARAM_FACE],
150                                              sizeof(stack[0].param[PARAM_FACE]),
151                                              "{\\fn%s}", tmp);
152                                 }
153                                 if ((param = strchr(param, ' ')))
154                                     param++;
155                             }
156                             for (i=0; i<PARAM_NUMBER; i++)
157                                 if (stack[sptr].param[i][0])
158                                     av_bprintf(dst, "%s", stack[sptr].param[i]);
159                         }
160                     } else if (tagname[0] && !tagname[1] && strspn(tagname, "bisu") == 1) {
161                         av_bprintf(dst, "{\\%c%d}", tagname[0], !tag_close);
162                     } else {
163                         unknown = 1;
164                         snprintf(tmp, sizeof(tmp), "</%s>", tagname);
165                     }
166                     if (tag_close) {
167                         sptr--;
168                     } else if (unknown && !strstr(in, tmp)) {
169                         in -= len + tag_close;
170                         av_bprint_chars(dst, *in, 1);
171                     } else
172                         av_strlcpy(stack[sptr++].tag, tagname,
173                                    sizeof(stack[0].tag));
174                     break;
175                 }
176             }
177         default:
178             av_bprint_chars(dst, *in, 1);
179             break;
180         }
181         if (*in != ' ' && *in != '\r' && *in != '\n')
182             line_start = 0;
183     }
184
185     if (!av_bprint_is_complete(dst))
186         return AVERROR(ENOMEM);
187
188     while (dst->len >= 2 && !strncmp(&dst->str[dst->len - 2], "\\N", 2))
189         dst->len -= 2;
190     dst->str[dst->len] = 0;
191     rstrip_spaces_buf(dst);
192
193     return 0;
194 }