avcodec/wmaprodec: cleanup extradata dumping
[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     while (buf->len > 0 && buf->str[buf->len - 1] == ' ')
50         buf->str[--buf->len] = 0;
51 }
52
53 void ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
54 {
55     char *param, buffer[128], tmp[128];
56     int len, tag_close, sptr = 1, line_start = 1, an = 0, end = 0;
57     SrtStack stack[16];
58
59     stack[0].tag[0] = 0;
60     strcpy(stack[0].param[PARAM_SIZE],  "{\\fs}");
61     strcpy(stack[0].param[PARAM_COLOR], "{\\c}");
62     strcpy(stack[0].param[PARAM_FACE],  "{\\fn}");
63
64     for (; !end && *in; in++) {
65         switch (*in) {
66         case '\r':
67             break;
68         case '\n':
69             if (line_start) {
70                 end = 1;
71                 break;
72             }
73             rstrip_spaces_buf(dst);
74             av_bprintf(dst, "\\N");
75             line_start = 1;
76             break;
77         case ' ':
78             if (!line_start)
79                 av_bprint_chars(dst, *in, 1);
80             break;
81         case '{':    /* skip all {\xxx} substrings except for {\an%d}
82                         and all microdvd like styles such as {Y:xxx} */
83             len = 0;
84             an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
85             if ((an != 1 && (len = 0, sscanf(in, "{\\%*[^}]}%n", &len) >= 0 && len > 0)) ||
86                 (len = 0, sscanf(in, "{%*1[CcFfoPSsYy]:%*[^}]}%n", &len) >= 0 && len > 0)) {
87                 in += len - 1;
88             } else
89                 av_bprint_chars(dst, *in, 1);
90             break;
91         case '<':
92             tag_close = in[1] == '/';
93             len = 0;
94             if (sscanf(in+tag_close+1, "%127[^>]>%n", buffer, &len) >= 1 && len > 0) {
95                 const char *tagname = buffer;
96                 while (*tagname == ' ')
97                     tagname++;
98                 if ((param = strchr(tagname, ' ')))
99                     *param++ = 0;
100                 if ((!tag_close && sptr < FF_ARRAY_ELEMS(stack)) ||
101                     ( tag_close && sptr > 0 && !strcmp(stack[sptr-1].tag, tagname))) {
102                     int i, j, unknown = 0;
103                     in += len + tag_close;
104                     if (!tag_close)
105                         memset(stack+sptr, 0, sizeof(*stack));
106                     if (!strcmp(tagname, "font")) {
107                         if (tag_close) {
108                             for (i=PARAM_NUMBER-1; i>=0; i--)
109                                 if (stack[sptr-1].param[i][0])
110                                     for (j=sptr-2; j>=0; j--)
111                                         if (stack[j].param[i][0]) {
112                                             av_bprintf(dst, "%s", stack[j].param[i]);
113                                             break;
114                                         }
115                         } else {
116                             while (param) {
117                                 if (!strncmp(param, "size=", 5)) {
118                                     unsigned font_size;
119                                     param += 5 + (param[5] == '"');
120                                     if (sscanf(param, "%u", &font_size) == 1) {
121                                         snprintf(stack[sptr].param[PARAM_SIZE],
122                                              sizeof(stack[0].param[PARAM_SIZE]),
123                                              "{\\fs%u}", font_size);
124                                     }
125                                 } else if (!strncmp(param, "color=", 6)) {
126                                     param += 6 + (param[6] == '"');
127                                     snprintf(stack[sptr].param[PARAM_COLOR],
128                                          sizeof(stack[0].param[PARAM_COLOR]),
129                                          "{\\c&H%X&}",
130                                          html_color_parse(log_ctx, param));
131                                 } else if (!strncmp(param, "face=", 5)) {
132                                     param += 5 + (param[5] == '"');
133                                     len = strcspn(param,
134                                                   param[-1] == '"' ? "\"" :" ");
135                                     av_strlcpy(tmp, param,
136                                                FFMIN(sizeof(tmp), len+1));
137                                     param += len;
138                                     snprintf(stack[sptr].param[PARAM_FACE],
139                                              sizeof(stack[0].param[PARAM_FACE]),
140                                              "{\\fn%s}", tmp);
141                                 }
142                                 if ((param = strchr(param, ' ')))
143                                     param++;
144                             }
145                             for (i=0; i<PARAM_NUMBER; i++)
146                                 if (stack[sptr].param[i][0])
147                                     av_bprintf(dst, "%s", stack[sptr].param[i]);
148                         }
149                     } else if (tagname[0] && !tagname[1] && strspn(tagname, "bisu") == 1) {
150                         av_bprintf(dst, "{\\%c%d}", tagname[0], !tag_close);
151                     } else {
152                         unknown = 1;
153                         snprintf(tmp, sizeof(tmp), "</%s>", tagname);
154                     }
155                     if (tag_close) {
156                         sptr--;
157                     } else if (unknown && !strstr(in, tmp)) {
158                         in -= len + tag_close;
159                         av_bprint_chars(dst, *in, 1);
160                     } else
161                         av_strlcpy(stack[sptr++].tag, tagname,
162                                    sizeof(stack[0].tag));
163                     break;
164                 }
165             }
166         default:
167             av_bprint_chars(dst, *in, 1);
168             break;
169         }
170         if (*in != ' ' && *in != '\r' && *in != '\n')
171             line_start = 0;
172     }
173
174     while (dst->len >= 2 && !strncmp(&dst->str[dst->len - 2], "\\N", 2))
175         dst->len -= 2;
176     dst->str[dst->len] = 0;
177     rstrip_spaces_buf(dst);
178 }