// Returns true if user_cp implies that calling mp_charset_guess() on the
// input data is required to determine the real codepage. This is the case
// if user_cp is not a real iconv codepage, but a magic value that requests
// for example ENCA charset auto-detection.
bool mp_charset_requires_guess(const char *user_cp)
{
bstr res[2] = {{0}};
split_colon(user_cp, 2, res);
return bstrcasecmp0(res[0], "enca") == 0 ||
bstrcasecmp0(res[0], "guess") == 0;
}
// Returns true if user_cp implies that calling mp_charset_guess() on the
// input data is required to determine the real codepage. This is the case
// if user_cp is not a real iconv codepage, but a magic value that requests
// for example ENCA charset auto-detection.
bool mp_charset_requires_guess(const char *user_cp)
{
bstr res[2] = {{0}};
int r = split_colon(user_cp, 2, res);
// Note that "utf8" is the UTF-8 codepage, while "utf8:..." specifies UTF-8
// by default, plus a codepage that is used if the input is not UTF-8.
return bstrcasecmp0(res[0], "enca") == 0 ||
bstrcasecmp0(res[0], "guess") == 0 ||
(r > 1 && bstrcasecmp0(res[0], "utf-8") == 0) ||
(r > 1 && bstrcasecmp0(res[0], "utf8") == 0);
}
// Runs charset auto-detection on the input buffer, and returns the result.
// If auto-detection fails, NULL is returned.
// If user_cp doesn't refer to any known auto-detection (for example because
// it's a real iconv codepage), user_cp is returned without even looking at
// the buf data.
const char *mp_charset_guess(struct mp_log *log, bstr buf, const char *user_cp,
int flags)
{
if (!mp_charset_requires_guess(user_cp))
return user_cp;
// Do our own UTF-8 detection, because at least ENCA seems to get it
// wrong sometimes (suggested by divVerent).
int r = bstr_validate_utf8(buf);
if (r >= 0 || (r > -8 && (flags & MP_ICONV_ALLOW_CUTOFF)))
return "UTF-8";
bstr params[3] = {{0}};
split_colon(user_cp, 3, params);
bstr type = params[0];
char lang[100];
snprintf(lang, sizeof(lang), "%.*s", BSTR_P(params[1]));
const char *fallback = params[2].start; // last item, already 0-terminated
const char *res = NULL;
#if HAVE_ENCA
if (bstrcasecmp0(type, "enca") == 0)
res = enca_guess(log, buf, lang);
#endif
#if HAVE_LIBGUESS
if (bstrcasecmp0(type, "guess") == 0)
res = libguess_guess(log, buf, lang);
#endif
if (bstrcasecmp0(type, "utf8") == 0 || bstrcasecmp0(type, "utf-8") == 0) {
if (!fallback)
fallback = params[1].start; // must be already 0-terminated
}
if (res) {
mp_dbg(log, "%.*s detected charset: '%s'\n",
BSTR_P(type), res);
} else {
res = fallback;
mp_dbg(log, "Detection with %.*s failed: fallback to %s\n",
BSTR_P(type), res && res[0] ? res : "broken UTF-8/Latin1");
}
if (!res && !(flags & MP_STRICT_UTF8))
res = "UTF-8-BROKEN";
return res;
}
// Runs charset auto-detection on the input buffer, and returns the result.
// If auto-detection fails, NULL is returned.
// If user_cp doesn't refer to any known auto-detection (for example because
// it's a real iconv codepage), user_cp is returned without even looking at
// the buf data.
const char *mp_charset_guess(bstr buf, const char *user_cp)
{
if (!mp_charset_requires_guess(user_cp))
return user_cp;
bstr params[3] = {{0}};
split_colon(user_cp, 3, params);
bstr type = params[0];
char lang[100];
snprintf(lang, sizeof(lang), "%.*s", BSTR_P(params[1]));
const char *fallback = params[2].start; // last item, already 0-terminated
const char *res = NULL;
#ifdef CONFIG_ENCA
if (bstrcasecmp0(type, "enca") == 0)
res = enca_guess(buf, lang);
#endif
#ifdef CONFIG_LIBGUESS
if (bstrcasecmp0(type, "guess") == 0)
res = libguess_guess(buf, lang);
#endif
if (res) {
mp_msg(MSGT_SUBREADER, MSGL_DBG2, "%.*s detected charset: '%s'\n",
BSTR_P(type), res);
} else {
res = fallback;
mp_msg(MSGT_SUBREADER, MSGL_DBG2,
"Detection with %.*s failed: fallback to %s\n",
BSTR_P(type), res && res[0] ? res : "no conversion");
}
return res;
}
// Runs charset auto-detection on the input buffer, and returns the result.
// If auto-detection fails, NULL is returned.
// If user_cp doesn't refer to any known auto-detection (for example because
// it's a real iconv codepage), user_cp is returned without even looking at
// the buf data.
// The return value may (but doesn't have to) be allocated under talloc_ctx.
const char *mp_charset_guess(void *talloc_ctx, struct mp_log *log, bstr buf,
const char *user_cp, int flags)
{
if (!mp_charset_requires_guess(user_cp))
return user_cp;
bool use_auto = strcasecmp(user_cp, "auto") == 0;
if (use_auto) {
#if HAVE_UCHARDET
user_cp = "uchardet";
#elif HAVE_ENCA
user_cp = "enca";
#else
user_cp = "UTF-8:UTF-8-BROKEN";
#endif
}
bstr params[3] = {{0}};
split_colon(user_cp, 3, params);
bstr type = params[0];
char lang[100];
snprintf(lang, sizeof(lang), "%.*s", BSTR_P(params[1]));
const char *fallback = params[2].start; // last item, already 0-terminated
const char *res = NULL;
if (use_auto) {
res = ms_bom_guess(buf);
if (res)
type = bstr0("auto");
}
#if HAVE_ENCA
if (bstrcasecmp0(type, "enca") == 0)
res = enca_guess(log, buf, lang);
#endif
#if HAVE_LIBGUESS
if (bstrcasecmp0(type, "guess") == 0)
res = libguess_guess(log, buf, lang);
#endif
#if HAVE_UCHARDET
if (bstrcasecmp0(type, "uchardet") == 0)
res = mp_uchardet(talloc_ctx, log, buf);
#endif
if (bstrcasecmp0(type, "utf8") == 0 || bstrcasecmp0(type, "utf-8") == 0) {
if (!fallback)
fallback = params[1].start; // must be already 0-terminated
int r = bstr_validate_utf8(buf);
if (r >= 0 || (r > -8 && (flags & MP_ICONV_ALLOW_CUTOFF)))
res = "utf-8";
}
if (res) {
mp_dbg(log, "%.*s detected charset: '%s'\n", BSTR_P(type), res);
} else {
res = fallback;
mp_dbg(log, "Detection with %.*s failed: fallback to %s\n",
BSTR_P(type), res && res[0] ? res : "broken UTF-8/Latin1");
}
if (!res && !(flags & MP_STRICT_UTF8))
res = "UTF-8-BROKEN";
mp_verbose(log, "Using charset '%s'.\n", res);
return res;
}
