comparison libpurple/protocols/oscar/oscar.c @ 27860:28a7992d37b2

propagate from branch 'im.pidgin.pidgin' (head 3008dd1f70a9e9ddc2bbe30c0fae7aba334836fe) to branch 'im.pidgin.pidgin.yaz' (head 6a95d36de1e6e9696d508a822e4d08106ac42092)

23410:f37b3e57fb92

oscar_encoding_to_utf8(PurpleAccount *account, const char *encoding, const char *text, int textlen)
{
	gchar *utf8 = NULL;

	if ((encoding == NULL) || encoding[0] == '\0') {
		purple_debug_info("oscar", "Empty encoding, assuming UTF-8\n");
	} else if (!g_ascii_strcasecmp(encoding, "iso-8859-1")) {
		utf8 = g_convert(text, textlen, "UTF-8", "iso-8859-1", NULL, NULL, NULL);
	} else if (!g_ascii_strcasecmp(encoding, "ISO-8859-1-Windows-3.1-Latin-1") ||
	           !g_ascii_strcasecmp(encoding, "us-ascii"))
	{

		 * from the locale character set to UTF-16 (not from UTF-8
		 * to UTF-16!) in the away message.  This hack should find
		 * and do something (un)reasonable with that, and not
		 * mess up too much else. */
		const gchar *charset = purple_account_get_string(account, "encoding", NULL);
		if (charset) {
			gsize len;
			utf8 = g_convert(text, textlen, charset, "UTF-16BE", &len, NULL, NULL);
			if (!utf8 || len != textlen || !g_utf8_validate(utf8, -1, NULL)) {
				g_free(utf8);

				&& !g_utf8_validate(text, textlen, NULL))
			utf8 = g_strdup(_("(There was an error receiving this message.  The buddy you are speaking with is probably using a different encoding than expected.  If you know what encoding he is using, you can specify it in the advanced account options for your AIM/ICQ account.)"));
		else
			utf8 = g_strndup(text, textlen);
	}

	return utf8;
}

static gchar *
oscar_utf8_try_convert(PurpleAccount *account, const gchar *msg)

}

static gchar *
purple_plugin_oscar_convert_to_utf8(const gchar *data, gsize datalen, const char *charsetstr, gboolean fallback)
{
	gchar *ret = NULL;
	GError *err = NULL;

	if ((charsetstr == NULL) || (*charsetstr == '\0'))
		return NULL;

			ret = g_strndup(data, datalen);
		else
			purple_debug_warning("oscar", "String is not valid UTF-8.\n");
	}

	return ret;
}

/**
 * This attemps to decode an incoming IM into a UTF8 string.
 *

			charsetstr1 = "ISO-8859-1";
		charsetstr2 = "UTF-8";
	} else if (charset == AIM_CHARSET_ASCII) {
		/* Should just be "ASCII" */
		charsetstr1 = "ASCII";
		charsetstr2 = purple_account_get_string(account, "encoding", OSCAR_DEFAULT_CUSTOM_ENCODING);
	} else if (charset == 0x000d) {
		/* Mobile AIM client on a Nokia 3100 and an LG VX6000 */
		charsetstr1 = "ISO-8859-1";
		charsetstr2 = purple_account_get_string(account, "encoding", OSCAR_DEFAULT_CUSTOM_ENCODING);
	} else {
		/* Unknown, hope for valid UTF-8... */
		charsetstr1 = "UTF-8";
		charsetstr2 = purple_account_get_string(account, "encoding", OSCAR_DEFAULT_CUSTOM_ENCODING);
	}

	{
		PurpleBuddy *b;
		b = purple_find_buddy(account, destsn);
		if ((b != NULL) && (PURPLE_BUDDY_IS_ONLINE(b)))
		{
			*msg = g_convert(from, -1, "UTF-16BE", "UTF-8", NULL, &msglen, NULL);
			if (*msg != NULL)
			{
				*charset = AIM_CHARSET_UNICODE;
				*charsubset = 0x0000;
				*msglen_int = msglen;

	/*
	 * XXX - We need a way to only attempt to convert if we KNOW "from"
	 * can be converted to "charsetstr"
	 */
	*msg = g_convert(from, -1, charsetstr, "UTF-8", NULL, &msglen, NULL);
	if (*msg != NULL) {
		*charset = AIM_CHARSET_CUSTOM;
		*charsubset = 0x0000;
		*msglen_int = msglen;
		return;
	}

	/*
	 * Nothing else worked, so send as UTF-16BE.
	 */
	*msg = g_convert(from, -1, "UTF-16BE", "UTF-8", NULL, &msglen, &err);
	if (*msg != NULL) {
		*charset = AIM_CHARSET_UNICODE;
		*charsubset = 0x0000;
		*msglen_int = msglen;
		return;

	curpart = args->mpmsg.parts;
	while (curpart != NULL) {
		tmp = purple_plugin_oscar_decode_im_part(account, userinfo->sn, curpart->charset,
				curpart->charsubset, curpart->data, curpart->datalen);
		if (tmp != NULL) {
			g_string_append(message, tmp);
			g_free(tmp);
		}

		curpart = curpart->next;

	if (args->type & OSCAR_CAPABILITY_CHAT)
	{
		char *encoding, *utf8name, *tmp;
		GHashTable *components;

		if (!args->info.chat.roominfo.name || !args->info.chat.roominfo.exchange) {
			g_free(message);
			return 1;
		}
		encoding = args->encoding ? oscar_encoding_extract(args->encoding) : NULL;

		components = g_hash_table_new_full(g_str_hash, g_str_equal, g_free,
				g_free);
		g_hash_table_replace(components, g_strdup("room"), utf8name);
		g_hash_table_replace(components, g_strdup("exchange"),
				g_strdup_printf("%d", args->info.chat.roominfo.exchange));
		serv_got_chat_invite(gc,
				     utf8name,
				     userinfo->sn,
				     message,
				     components);

	 * character in whatever encoding the message was sent in.  Type 1
	 * messages are always made up of only one part, so we can easily account
	 * for this suck-ass part of the protocol by splitting the string into at
	 * most 1 baby string.
	 */
	msg1 = g_strsplit(args->msg, "\376", (args->type == 0x01 ? 1 : 0));
	for (numtoks=0; msg1[numtoks]; numtoks++);
	msg2 = (gchar **)g_malloc((numtoks+1)*sizeof(gchar *));
	for (i=0; msg1[i]; i++) {
		gchar *uin = g_strdup_printf("%u", args->uin);

static void
purple_odc_send_im(PeerConnection *conn, const char *message, PurpleMessageFlags imflags)
{
	GString *msg;
	GString *data;
	gchar *tmp;
	int tmplen;
	guint16 charset, charsubset;
	GData *attribs;
	const char *start, *end, *last;
	int oscar_id = 0;

	msg = g_string_new("<HTML><BODY>");

	if (last && *last)
		g_string_append(msg, last);

	g_string_append(msg, "</BODY></HTML>");

	/* Convert the message to a good encoding */
	purple_plugin_oscar_convert_to_best_encoding(conn->od->gc,
			conn->sn, msg->str, &tmp, &tmplen, &charset, &charsubset);
	g_string_free(msg, TRUE);
	msg = g_string_new_len(tmp, tmplen);
	g_free(tmp);

	/* Append any binary data that we may have */
	if (oscar_id) {
		msg = g_string_append_len(msg, data->str, data->len);
		msg = g_string_append(msg, "</BINARY>");

	int charset = 0;
	char *encoded = NULL;

	charset = oscar_charset_check(str);
	if (charset == AIM_CHARSET_UNICODE) {
		encoded = g_convert(str, -1, "UTF-16BE", "UTF-8", NULL, ret_len, NULL);
		*encoding = "unicode-2-0";
	} else if (charset == AIM_CHARSET_CUSTOM) {
		encoded = g_convert(str, -1, "ISO-8859-1", "UTF-8", NULL, ret_len, NULL);
		*encoding = "iso-8859-1";
	} else {

		charsetstr = "us-ascii";
	else if (charset == AIM_CHARSET_UNICODE)
		charsetstr = "unicode-2-0";
	else if (charset == AIM_CHARSET_CUSTOM)
		charsetstr = "iso-8859-1";
	aim_chat_send_im(od, c->conn, 0, buf2, len, charsetstr, "en");
	g_free(buf2);
	g_free(buf);

	return 0;
}

27860:28a7992d37b2

oscar_encoding_to_utf8(PurpleAccount *account, const char *encoding, const char *text, int textlen)
{
	gchar *utf8 = NULL;

	if ((encoding == NULL) || encoding[0] == '\0') {
		purple_debug_info("yaz oscar", "Empty encoding, validate as UTF-8\n");
		if(g_utf8_validate(text, textlen, NULL)){
			gsize newlen;
			utf8 = sanitize_utf(text, textlen, &newlen);
			goto done;
		}
		// not UTF-8
		purple_debug_info("yaz oscar", "Empty encoding, assuming UTF-16BE\n");
		sanitize_ucs((gchar *)text, textlen);
		utf8 = g_convert(text, textlen, "UTF-8", "UTF-16BE", NULL, NULL, NULL);
		if(utf8){
			if(!g_utf8_validate(utf8, strlen(utf8), NULL)){
				purple_debug_info("yaz oscar", "Invalid conversion\n");
				g_free(utf8);
				utf8 = NULL;
			}
		} else {
			purple_debug_info("yaz oscar", "Conversion failed\n");
		}
	} else if (!g_ascii_strcasecmp(encoding, "iso-8859-1")) {
		utf8 = g_convert(text, textlen, "UTF-8", "iso-8859-1", NULL, NULL, NULL);
	} else if (!g_ascii_strcasecmp(encoding, "ISO-8859-1-Windows-3.1-Latin-1") ||
	           !g_ascii_strcasecmp(encoding, "us-ascii"))
	{

		 * from the locale character set to UTF-16 (not from UTF-8
		 * to UTF-16!) in the away message.  This hack should find
		 * and do something (un)reasonable with that, and not
		 * mess up too much else. */
		const gchar *charset = purple_account_get_string(account, "encoding", NULL);
		sanitize_ucs((gchar *)text, textlen);
		if (charset) {
			gsize len;
			utf8 = g_convert(text, textlen, charset, "UTF-16BE", &len, NULL, NULL);
			if (!utf8 || len != textlen || !g_utf8_validate(utf8, -1, NULL)) {
				g_free(utf8);

				&& !g_utf8_validate(text, textlen, NULL))
			utf8 = g_strdup(_("(There was an error receiving this message.  The buddy you are speaking with is probably using a different encoding than expected.  If you know what encoding he is using, you can specify it in the advanced account options for your AIM/ICQ account.)"));
		else
			utf8 = g_strndup(text, textlen);
	}
done:
	return utf8;
}

static gchar *
oscar_utf8_try_convert(PurpleAccount *account, const gchar *msg)

}

static gchar *
purple_plugin_oscar_convert_to_utf8(const gchar *data, gsize datalen, const char *charsetstr, gboolean fallback)
{
	gchar *ret = NULL, *ret2 = NULL;
	GError *err = NULL;

	if ((charsetstr == NULL) || (*charsetstr == '\0'))
		return NULL;

			ret = g_strndup(data, datalen);
		else
			purple_debug_warning("oscar", "String is not valid UTF-8.\n");
	}

	ret2 = sanitize_utf(ret, -1, NULL);
	g_free(ret);
	return ret2;
}

/**
 * This attemps to decode an incoming IM into a UTF8 string.
 *

			charsetstr1 = "ISO-8859-1";
		charsetstr2 = "UTF-8";
	} else if (charset == AIM_CHARSET_ASCII) {
		/* Should just be "ASCII" */
		charsetstr1 = "ASCII";
		charsetstr2 = "UTF-8";
	} else if (charset == AIM_CHARSET_QUIRKUTF8) {
		/* Mobile AIM client on a Nokia 3100 and an LG VX6000 */
		charsetstr1 = "UTF-8";  //iChat use 0x000d when it sends UTF-8. --yaz
		charsetstr2 = "ISO-8859-1";
	} else {
		/* Unknown, hope for valid UTF-8... */
		charsetstr1 = "UTF-8";
		charsetstr2 = purple_account_get_string(account, "encoding", OSCAR_DEFAULT_CUSTOM_ENCODING);
	}

	{
		PurpleBuddy *b;
		b = purple_find_buddy(account, destsn);
		if ((b != NULL) && (PURPLE_BUDDY_IS_ONLINE(b)))
		{
			*msg = g_convert(from, strlen(from), "UTF-16BE", "UTF-8", NULL, &msglen, NULL);
			botch_ucs(*msg, msglen);
			if (*msg != NULL)
			{
				*charset = AIM_CHARSET_UNICODE;
				*charsubset = 0x0000;
				*msglen_int = msglen;

	/*
	 * XXX - We need a way to only attempt to convert if we KNOW "from"
	 * can be converted to "charsetstr"
	 */
#ifndef _WIN32
	/* nosuke reported that this portion caused unexpected
	 * conversion from utf-8 fullwidth tilde/numbers/alphabets to
	 * halfwidth ones on windows environment. --yaz */
	*msg = g_convert(from, -1, charsetstr, "UTF-8", NULL, &msglen, NULL);
	if (*msg != NULL) {
		*charset = AIM_CHARSET_CUSTOM;
		*charsubset = 0x0000;
		*msglen_int = msglen;
		return;
	}
#endif

	/*
	 * Nothing else worked, so send as UTF-16BE.
	 */
	*msg = g_convert(from, strlen(from), "UTF-16BE", "UTF-8", NULL, &msglen, &err);
	botch_ucs(*msg, msglen);
	if (*msg != NULL) {
		*charset = AIM_CHARSET_UNICODE;
		*charsubset = 0x0000;
		*msglen_int = msglen;
		return;

	curpart = args->mpmsg.parts;
	while (curpart != NULL) {
		tmp = purple_plugin_oscar_decode_im_part(account, userinfo->sn, curpart->charset,
				curpart->charsubset, curpart->data, curpart->datalen);
		if (tmp != NULL) {
			purple_str_strip_char(tmp, 0x0d); // yaz: strip CR
//			purple_debug_info("yaz oscar", "tmp=%s",tmp);
			g_string_append(message, tmp);
			g_free(tmp);
		}

		curpart = curpart->next;

	if (args->type & OSCAR_CAPABILITY_CHAT)
	{
		char *encoding, *utf8name, *tmp;
		GHashTable *components;

//		purple_debug_info("yaz oscar", "chat request %s\n", args->msg);
		if (!args->info.chat.roominfo.name || !args->info.chat.roominfo.exchange) {
			g_free(message);
			return 1;
		}
		encoding = args->encoding ? oscar_encoding_extract(args->encoding) : NULL;

		components = g_hash_table_new_full(g_str_hash, g_str_equal, g_free,
				g_free);
		g_hash_table_replace(components, g_strdup("room"), utf8name);
		g_hash_table_replace(components, g_strdup("exchange"),
				g_strdup_printf("%d", args->info.chat.roominfo.exchange));
		purple_debug_info("yaz oscar", "about to call serv_got_chat_invite\n");
//		purple_debug_info("yaz oscar", "name=%s message=%s\n", name ? name : args->info.chat.roominfo.name, message);
		serv_got_chat_invite(gc,
				     utf8name,
				     userinfo->sn,
				     message,
				     components);

	 * character in whatever encoding the message was sent in.  Type 1
	 * messages are always made up of only one part, so we can easily account
	 * for this suck-ass part of the protocol by splitting the string into at
	 * most 1 baby string.
	 */
	msg1 = g_strsplit(args->msg, "\376", (args->type == 0x01 ? 1 : 0)); // \376 is 0xfe
	for (numtoks=0; msg1[numtoks]; numtoks++);
	msg2 = (gchar **)g_malloc((numtoks+1)*sizeof(gchar *));
	for (i=0; msg1[i]; i++) {
		gchar *uin = g_strdup_printf("%u", args->uin);

static void
purple_odc_send_im(PeerConnection *conn, const char *message, PurpleMessageFlags imflags)
{
	GString *msg;
	GString *data;
	guint16 charset;
	GData *attribs;
	const char *start, *end, *last;
	int oscar_id = 0;

	msg = g_string_new("<HTML><BODY>");

	if (last && *last)
		g_string_append(msg, last);

	g_string_append(msg, "</BODY></HTML>");

	/* iChat and AIM6 use 0x000d to send UTF8. moreover, AIM6 persists only to UTF8! --yaz */
	charset = AIM_CHARSET_QUIRKUTF8;

	/* Append any binary data that we may have */
	if (oscar_id) {
		msg = g_string_append_len(msg, data->str, data->len);
		msg = g_string_append(msg, "</BINARY>");

	int charset = 0;
	char *encoded = NULL;

	charset = oscar_charset_check(str);
	if (charset == AIM_CHARSET_UNICODE) {
		encoded = g_convert(str, strlen(str), "UTF-16BE", "UTF-8", NULL, ret_len, NULL);
		botch_ucs(encoded, *ret_len);
		*encoding = "unicode-2-0";
	} else if (charset == AIM_CHARSET_CUSTOM) {
		encoded = g_convert(str, -1, "ISO-8859-1", "UTF-8", NULL, ret_len, NULL);
		*encoding = "iso-8859-1";
	} else {

		charsetstr = "us-ascii";
	else if (charset == AIM_CHARSET_UNICODE)
		charsetstr = "unicode-2-0";
	else if (charset == AIM_CHARSET_CUSTOM)
		charsetstr = "iso-8859-1";
	aim_chat_send_im(od, c->conn, 0, buf2, len, charsetstr, "JA");
	g_free(buf2);
	g_free(buf);

	return 0;
}