1
0
mirror of https://github.com/irssi/irssi.git synced 2024-09-29 04:45:57 -04:00
irssi/src/core/recode.c

209 lines
4.9 KiB
C
Raw Normal View History

/*
recode.c : irssi
Copyright (C) 1999-2000 Timo Sirainen
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
#include "module.h"
#include "settings.h"
#include "servers.h"
#include "signals.h"
#include "lib-config/iconfig.h"
#include "misc.h"
static char *translit_charset;
static gboolean term_is_utf8;
gboolean is_utf8(void)
{
return term_is_utf8;
}
static gboolean is_translit(const char *charset)
{
char *pos;
pos = stristr(charset, "//translit");
return (pos != NULL);
}
gboolean is_valid_charset(const char *charset)
{
GIConv cd;
char *to = NULL;
if (!charset || *charset == '\0')
return FALSE;
if (settings_get_bool("recode_transliterate") && !is_translit(charset))
charset = to = g_strconcat(charset, "//TRANSLIT", NULL);
cd = g_iconv_open(charset, "UTF-8");
g_free(to);
if (cd != (GIConv)-1) {
g_iconv_close(cd);
return TRUE;
}
return FALSE;
}
static char *find_conversion(const SERVER_REC *server, const char *target)
{
char *conv = NULL;
if (server != NULL && target != NULL) {
char *tagtarget = g_strdup_printf("%s/%s", server->tag, target);
conv = iconfig_get_str("conversions", tagtarget, NULL);
g_free(tagtarget);
}
if (conv == NULL && target != NULL)
conv = iconfig_get_str("conversions", target, NULL);
if (conv == NULL && server != NULL)
conv = iconfig_get_str("conversions", server->tag, NULL);
return conv;
}
static int str_is_ascii(const char *str)
{
int i;
for (i = 0; str[i] != '\0'; i++)
if (str[i] & 0x80)
return 0;
return 1;
}
char *recode_in(const SERVER_REC *server, const char *str, const char *target)
{
const char *from = NULL;
const char *to = translit_charset;
char *recoded = NULL;
gboolean str_is_utf8, recode, autodetect;
int len;
if (!str)
return NULL;
recode = settings_get_bool("recode");
if (!recode)
return g_strdup(str);
len = strlen(str);
/* Only validate for UTF-8 if an 8-bit encoding. */
str_is_utf8 = 0;
if (!str_is_ascii(str))
str_is_utf8 = g_utf8_validate(str, len, NULL);
else if (!strchr(str, '\e'))
str_is_utf8 = 1;
autodetect = settings_get_bool("recode_autodetect_utf8");
if (autodetect && str_is_utf8)
if (term_is_utf8)
return g_strdup(str);
else
from = "UTF-8";
else
from = find_conversion(server, target);
if (from)
recoded = g_convert_with_fallback(str, len, to, from, NULL, NULL, NULL, NULL);
if (!recoded) {
if (str_is_utf8)
if (term_is_utf8)
return g_strdup(str);
else
from = "UTF-8";
else
if (term_is_utf8)
from = settings_get_str("recode_fallback");
else
from = NULL;
if (from)
recoded = g_convert_with_fallback(str, len, to, from, NULL, NULL, NULL, NULL);
if (!recoded)
recoded = g_strdup(str);
}
return recoded;
}
char *recode_out(const SERVER_REC *server, const char *str, const char *target)
{
char *recoded = NULL;
const char *from = translit_charset;
const char *to = NULL;
char *translit_to = NULL;
gboolean translit, recode;
int len;
if (!str)
return NULL;
recode = settings_get_bool("recode");
if (!recode)
return g_strdup(str);
len = strlen(str);
translit = settings_get_bool("recode_transliterate");
to = find_conversion(server, target);
if (to == NULL)
/* default outgoing charset if set */
to = settings_get_str("recode_out_default_charset");
if (to && *to != '\0') {
if (translit && !is_translit(to))
to = translit_to = g_strconcat(to ,"//TRANSLIT", NULL);
recoded = g_convert(str, len, to, from, NULL, NULL, NULL);
}
g_free(translit_to);
if (!recoded)
recoded = g_strdup(str);
return recoded;
}
void recode_update_charset(void)
{
const char *charset = settings_get_str("term_charset");
term_is_utf8 = !g_ascii_strcasecmp(charset, "UTF-8");
g_free(translit_charset);
if (settings_get_bool("recode_transliterate") && !is_translit(charset))
translit_charset = g_strconcat(charset, "//TRANSLIT", NULL);
else
translit_charset = g_strdup(charset);
}
void recode_init(void)
{
settings_add_bool("misc", "recode", TRUE);
settings_add_str("misc", "recode_fallback", "CP1252");
settings_add_str("misc", "recode_out_default_charset", "");
settings_add_bool("misc", "recode_transliterate", TRUE);
settings_add_bool("misc", "recode_autodetect_utf8", TRUE);
}
void recode_deinit(void)
{
g_free(translit_charset);
}