static const char *enca_guess(struct mp_log *log, bstr buf, const char *language)
{
// Do our own UTF-8 detection, because ENCA seems to get it wrong sometimes
// (suggested by divVerent). Explicitly allow cut-off UTF-8.
if (bstr_validate_utf8(buf) > -8)
return "UTF-8";
if (!language || !language[0])
language = "__"; // neutral language
const char *detected_cp = NULL;
EncaAnalyser analyser = enca_analyser_alloc(language);
if (analyser) {
enca_set_termination_strictness(analyser, 0);
EncaEncoding enc = enca_analyse_const(analyser, buf.start, buf.len);
const char *tmp = enca_charset_name(enc.charset, ENCA_NAME_STYLE_ICONV);
if (tmp && enc.charset != ENCA_CS_UNKNOWN)
detected_cp = tmp;
enca_analyser_free(analyser);
} else {
mp_err(log, "ENCA doesn't know language '%s'\n", language);
size_t langcnt;
const char **languages = enca_get_languages(&langcnt);
mp_err(log, "ENCA supported languages:");
for (int i = 0; i < langcnt; i++)
mp_err(log, " %s", languages[i]);
mp_err(log, "\n");
free(languages);
}
return detected_cp;
}
static const char *enca_guess(struct mp_log *log, bstr buf, const char *language)
{
if (!language || !language[0])
language = "__"; // neutral language
const char *detected_cp = NULL;
EncaAnalyser analyser = enca_analyser_alloc(language);
if (analyser) {
enca_set_termination_strictness(analyser, 0);
EncaEncoding enc = enca_analyse_const(analyser, buf.start, buf.len);
const char *tmp = enca_charset_name(enc.charset, ENCA_NAME_STYLE_ICONV);
if (tmp && enc.charset != ENCA_CS_UNKNOWN)
detected_cp = tmp;
enca_analyser_free(analyser);
} else {
mp_err(log, "ENCA doesn't know language '%s'\n", language);
size_t langcnt;
const char **languages = enca_get_languages(&langcnt);
mp_err(log, "ENCA supported languages:");
for (int i = 0; i < langcnt; i++)
mp_err(log, " %s", languages[i]);
mp_err(log, "\n");
free(languages);
}
return detected_cp;
}
/*
* DWIM
*
* Choose some suitable values of all the libenca tuning parameters.
*/
static void
dwim_libenca_options(EncaAnalyser an, const File *file)
{
const double mu = 0.005; /* derivation in 0 */
const double m = 15.0; /* value in infinity */
ssize_t size = file->buffer->pos;
size_t sgnf;
/* The number of significant characters */
if (!size)
sgnf = 1;
else
sgnf = ceil((double)size/(size/m + 1.0/mu));
enca_set_significant(an, sgnf);
/* When buffer contains whole file, require correct termination. */
if (file->size == size)
enca_set_termination_strictness(an, 1);
else
enca_set_termination_strictness(an, 0);
enca_set_filtering(an, sgnf > 2);
}
