/*
* Copyright (c) 2010 Aurelien Jacobs <aurel@gnuage.org>
+ * Copyright (c) 2017 Clément Bœsch <u@pkh.me>
*
* This file is part of FFmpeg.
*
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
+#include "libavutil/avassert.h"
#include "libavutil/avstring.h"
#include "libavutil/common.h"
#include "libavutil/parseutils.h"
static int html_color_parse(void *log_ctx, const char *str)
{
uint8_t rgba[4];
+ int nb_sharps = 0;
+ while (str[nb_sharps] == '#')
+ nb_sharps++;
+ str += FFMAX(0, nb_sharps - 1);
if (av_parse_color(rgba, str, strcspn(str, "\" >"), log_ctx) < 0)
return -1;
return rgba[0] | rgba[1] << 8 | rgba[2] << 16;
}
-enum {
- PARAM_UNKNOWN = -1,
- PARAM_SIZE,
- PARAM_COLOR,
- PARAM_FACE,
- PARAM_NUMBER
-};
-
-typedef struct SrtStack {
- char tag[128];
- char param[PARAM_NUMBER][128];
-} SrtStack;
-
static void rstrip_spaces_buf(AVBPrint *buf)
{
- while (buf->len > 0 && buf->str[buf->len - 1] == ' ')
- buf->str[--buf->len] = 0;
+ if (av_bprint_is_complete(buf))
+ while (buf->len > 0 && buf->str[buf->len - 1] == ' ')
+ buf->str[--buf->len] = 0;
+}
+
+/* skip all {\xxx} substrings except for {\an%d}
+ and all microdvd like styles such as {Y:xxx} */
+static void handle_open_brace(AVBPrint *dst, const char **inp, int *an, int *closing_brace_missing)
+{
+ int len = 0;
+ const char *in = *inp;
+
+ *an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
+
+ if (!*closing_brace_missing) {
+ if ( (*an != 1 && in[1] == '\\')
+ || (in[1] && strchr("CcFfoPSsYy", in[1]) && in[2] == ':')) {
+ char *bracep = strchr(in+2, '}');
+ if (bracep) {
+ *inp = bracep;
+ return;
+ } else
+ *closing_brace_missing = 1;
+ }
+ }
+
+ av_bprint_chars(dst, *in, 1);
}
-void ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
+struct font_tag {
+ char face[128];
+ int size;
+ uint32_t color;
+};
+
+/*
+ * The general politic of the convert is to mask unsupported tags or formatting
+ * errors (but still alert the user/subtitles writer with an error/warning)
+ * without dropping any actual text content for the final user.
+ */
+int ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
{
- char *param, buffer[128], tmp[128];
- int len, tag_close, sptr = 1, line_start = 1, an = 0, end = 0;
- SrtStack stack[16];
+ char *param, buffer[128];
+ int len, tag_close, sptr = 0, line_start = 1, an = 0, end = 0;
+ int closing_brace_missing = 0;
+ int i, likely_a_tag;
+
+ /*
+ * state stack is only present for fonts since they are the only tags where
+ * the state is not binary. Here is a typical use case:
+ *
+ * <font color="red" size=10>
+ * red 10
+ * <font size=50> RED AND BIG </font>
+ * red 10 again
+ * </font>
+ *
+ * On the other hand, using the state system for all the tags should be
+ * avoided because it breaks wrongly nested tags such as:
+ *
+ * <b> foo <i> bar </b> bla </i>
+ *
+ * We don't want to break here; instead, we will treat all these tags as
+ * binary state markers. Basically, "<b>" will activate bold, and "</b>"
+ * will deactivate it, whatever the current state.
+ *
+ * This will also prevents cases where we have a random closing tag
+ * remaining after the opening one was dropped. Yes, this happens and we
+ * still don't want to print a "</b>" at the end of the dialog event.
+ */
+ struct font_tag stack[16];
- stack[0].tag[0] = 0;
- strcpy(stack[0].param[PARAM_SIZE], "{\\fs}");
- strcpy(stack[0].param[PARAM_COLOR], "{\\c}");
- strcpy(stack[0].param[PARAM_FACE], "{\\fn}");
+ memset(&stack[0], 0, sizeof(stack[0]));
for (; !end && *in; in++) {
switch (*in) {
if (!line_start)
av_bprint_chars(dst, *in, 1);
break;
- case '{': /* skip all {\xxx} substrings except for {\an%d}
- and all microdvd like styles such as {Y:xxx} */
- len = 0;
- an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
- if ((an != 1 && (len = 0, sscanf(in, "{\\%*[^}]}%n", &len) >= 0 && len > 0)) ||
- (len = 0, sscanf(in, "{%*1[CcFfoPSsYy]:%*[^}]}%n", &len) >= 0 && len > 0)) {
- in += len - 1;
- } else
- av_bprint_chars(dst, *in, 1);
+ case '{':
+ handle_open_brace(dst, &in, &an, &closing_brace_missing);
break;
case '<':
+ /*
+ * "<<" are likely latin guillemets in ASCII or some kind of random
+ * style effect; see sub/badsyntax.srt in the FATE samples
+ * directory for real test cases.
+ */
+
+ likely_a_tag = 1;
+ for (i = 0; in[1] == '<'; i++) {
+ av_bprint_chars(dst, '<', 1);
+ likely_a_tag = 0;
+ in++;
+ }
+
tag_close = in[1] == '/';
+ if (tag_close)
+ likely_a_tag = 1;
+
+ av_assert0(in[0] == '<');
+
len = 0;
- if (sscanf(in+tag_close+1, "%127[^>]>%n", buffer, &len) >= 1 && len > 0) {
+
+ if (sscanf(in+tag_close+1, "%127[^<>]>%n", buffer, &len) >= 1 && len > 0) {
+ const int skip = len + tag_close;
const char *tagname = buffer;
- while (*tagname == ' ')
+ while (*tagname == ' ') {
+ likely_a_tag = 0;
tagname++;
+ }
if ((param = strchr(tagname, ' ')))
*param++ = 0;
- if ((!tag_close && sptr < FF_ARRAY_ELEMS(stack)) ||
- ( tag_close && sptr > 0 && !strcmp(stack[sptr-1].tag, tagname))) {
- int i, j, unknown = 0;
- in += len + tag_close;
- if (!tag_close)
- memset(stack+sptr, 0, sizeof(*stack));
- if (!strcmp(tagname, "font")) {
- if (tag_close) {
- for (i=PARAM_NUMBER-1; i>=0; i--)
- if (stack[sptr-1].param[i][0])
- for (j=sptr-2; j>=0; j--)
- if (stack[j].param[i][0]) {
- av_bprintf(dst, "%s", stack[j].param[i]);
- break;
- }
- } else {
- while (param) {
- if (!strncmp(param, "size=", 5)) {
- unsigned font_size;
- param += 5 + (param[5] == '"');
- if (sscanf(param, "%u", &font_size) == 1) {
- snprintf(stack[sptr].param[PARAM_SIZE],
- sizeof(stack[0].param[PARAM_SIZE]),
- "{\\fs%u}", font_size);
- }
- } else if (!strncmp(param, "color=", 6)) {
- param += 6 + (param[6] == '"');
- snprintf(stack[sptr].param[PARAM_COLOR],
- sizeof(stack[0].param[PARAM_COLOR]),
- "{\\c&H%X&}",
- html_color_parse(log_ctx, param));
- } else if (!strncmp(param, "face=", 5)) {
- param += 5 + (param[5] == '"');
- len = strcspn(param,
- param[-1] == '"' ? "\"" :" ");
- av_strlcpy(tmp, param,
- FFMIN(sizeof(tmp), len+1));
- param += len;
- snprintf(stack[sptr].param[PARAM_FACE],
- sizeof(stack[0].param[PARAM_FACE]),
- "{\\fn%s}", tmp);
+
+ /* Check if this is likely a tag */
+#define LIKELY_A_TAG_CHAR(x) (((x) >= '0' && (x) <= '9') || \
+ ((x) >= 'a' && (x) <= 'z') || \
+ ((x) >= 'A' && (x) <= 'Z') || \
+ (x) == '_' || (x) == '/')
+ for (i = 0; tagname[i]; i++) {
+ if (!LIKELY_A_TAG_CHAR(tagname[i])) {
+ likely_a_tag = 0;
+ break;
+ }
+ }
+
+ if (!av_strcasecmp(tagname, "font")) {
+ if (tag_close && sptr > 0) {
+ struct font_tag *cur_tag = &stack[sptr--];
+ struct font_tag *last_tag = &stack[sptr];
+
+ if (cur_tag->size) {
+ if (!last_tag->size)
+ av_bprintf(dst, "{\\fs}");
+ else if (last_tag->size != cur_tag->size)
+ av_bprintf(dst, "{\\fs%d}", last_tag->size);
+ }
+
+ if (cur_tag->color & 0xff000000) {
+ if (!(last_tag->color & 0xff000000))
+ av_bprintf(dst, "{\\c}");
+ else if (last_tag->color != cur_tag->color)
+ av_bprintf(dst, "{\\c&H%"PRIX32"&}", last_tag->color & 0xffffff);
+ }
+
+ if (cur_tag->face[0]) {
+ if (!last_tag->face[0])
+ av_bprintf(dst, "{\\fn}");
+ else if (strcmp(last_tag->face, cur_tag->face))
+ av_bprintf(dst, "{\\fn%s}", last_tag->face);
+ }
+ } else if (!tag_close && sptr < FF_ARRAY_ELEMS(stack) - 1) {
+ struct font_tag *new_tag = &stack[sptr + 1];
+
+ *new_tag = stack[sptr++];
+
+ while (param) {
+ if (!av_strncasecmp(param, "size=", 5)) {
+ param += 5 + (param[5] == '"');
+ if (sscanf(param, "%u", &new_tag->size) == 1)
+ av_bprintf(dst, "{\\fs%u}", new_tag->size);
+ } else if (!av_strncasecmp(param, "color=", 6)) {
+ int color;
+ param += 6 + (param[6] == '"');
+ color = html_color_parse(log_ctx, param);
+ if (color >= 0) {
+ new_tag->color = 0xff000000 | color;
+ av_bprintf(dst, "{\\c&H%"PRIX32"&}", new_tag->color & 0xffffff);
}
- if ((param = strchr(param, ' ')))
- param++;
+ } else if (!av_strncasecmp(param, "face=", 5)) {
+ param += 5 + (param[5] == '"');
+ len = strcspn(param,
+ param[-1] == '"' ? "\"" :" ");
+ av_strlcpy(new_tag->face, param,
+ FFMIN(sizeof(new_tag->face), len+1));
+ param += len;
+ av_bprintf(dst, "{\\fn%s}", new_tag->face);
}
- for (i=0; i<PARAM_NUMBER; i++)
- if (stack[sptr].param[i][0])
- av_bprintf(dst, "%s", stack[sptr].param[i]);
+ if ((param = strchr(param, ' ')))
+ param++;
}
- } else if (tagname[0] && !tagname[1] && strspn(tagname, "bisu") == 1) {
- av_bprintf(dst, "{\\%c%d}", tagname[0], !tag_close);
- } else {
- unknown = 1;
- snprintf(tmp, sizeof(tmp), "</%s>", tagname);
}
- if (tag_close) {
- sptr--;
- } else if (unknown && !strstr(in, tmp)) {
- in -= len + tag_close;
- av_bprint_chars(dst, *in, 1);
- } else
- av_strlcpy(stack[sptr++].tag, tagname,
- sizeof(stack[0].tag));
- break;
+ in += skip;
+ } else if (tagname[0] && !tagname[1] && strchr("bisu", av_tolower(tagname[0]))) {
+ av_bprintf(dst, "{\\%c%d}", (char)av_tolower(tagname[0]), !tag_close);
+ in += skip;
+ } else if (!av_strncasecmp(tagname, "br", 2) &&
+ (!tagname[2] || (tagname[2] == '/' && !tagname[3]))) {
+ av_bprintf(dst, "\\N");
+ in += skip;
+ } else if (likely_a_tag) {
+ if (!tag_close) // warn only once
+ av_log(log_ctx, AV_LOG_WARNING, "Unrecognized tag %s\n", tagname);
+ in += skip;
+ } else {
+ av_bprint_chars(dst, '<', 1);
}
+ } else {
+ av_bprint_chars(dst, *in, 1);
}
+ break;
default:
av_bprint_chars(dst, *in, 1);
break;
line_start = 0;
}
+ if (!av_bprint_is_complete(dst))
+ return AVERROR(ENOMEM);
+
while (dst->len >= 2 && !strncmp(&dst->str[dst->len - 2], "\\N", 2))
dst->len -= 2;
dst->str[dst->len] = 0;
rstrip_spaces_buf(dst);
+
+ return 0;
}