[Rework] Rework URL structure: user field

src/libserver/protocol.c

@@ -994,9 +994,8 @@ emails_protocol_cb (gpointer key, gpointer value, gpointer ud)
 	struct rspamd_url *url = value;
 	ucl_object_t *obj;
 
-	if (url->userlen > 0 && url->hostlen > 0 &&
-			url->host == url->user + url->userlen + 1) {
-		obj = ucl_object_fromlstring (url->user,
+	if (url->userlen > 0 && url->hostlen > 0) {
+		obj = ucl_object_fromlstring (rspamd_url_user_unsafe (url),
 				url->userlen + url->hostlen + 1);
 		ucl_array_append (cb->top, obj);
 	}

src/libserver/re_cache.c

@@ -1223,7 +1223,7 @@ rspamd_re_cache_exec_re (struct rspamd_task *task,
 					continue;
 				}
 
-				in = url->user;
+				in = rspamd_url_user_unsafe (url);
 				len = url->userlen + 1 + url->hostlen;
 				scvec[i] = (guchar *) in;
 				lenvec[i++] = len;

src/libserver/url.c

@@ -1874,7 +1874,7 @@ rspamd_url_shift (struct rspamd_url *uri, gsize nlen,
 	switch (field) {
 	case UF_SCHEMA:
 		if (uri->userlen > 0) {
-			uri->user -= shift;
+			uri->usershift -= shift;
 		}
 		if (uri->hostlen > 0) {
 			uri->host -= shift;
@@ -2038,7 +2038,7 @@ rspamd_url_parse (struct rspamd_url *uri,
 				uri->fragmentlen = complen;
 				break;
 			case UF_USERINFO:
-				uri->user = comp;
+				uri->usershift = u.field_data[i].off;
 				uri->userlen = complen;
 				break;
 			default:
@@ -3382,7 +3382,7 @@ rspamd_email_hash (gconstpointer u)
 	}
 
 	if (url->userlen > 0) {
-		rspamd_cryptobox_fast_hash_update (&st, url->user, url->userlen);
+		rspamd_cryptobox_fast_hash_update (&st, rspamd_url_user_unsafe(url), url->userlen);
 	}
 
 	return (guint)rspamd_cryptobox_fast_hash_final (&st);
@@ -3404,8 +3404,9 @@ rspamd_emails_cmp (gconstpointer a, gconstpointer b)
 				return FALSE;
 			}
 			else {
-				return rspamd_lc_cmp (u1->user, u2->user, u1->userlen) ==
-						0;
+				return (rspamd_lc_cmp (rspamd_url_user_unsafe(u1),
+						rspamd_url_user_unsafe(u2),
+						u1->userlen) == 0);
 			}
 		}
 		else {
@@ -3606,7 +3607,7 @@ static const unsigned char rspamd_url_encoding_classes[256] = {
 
 #define CHECK_URL_COMPONENT(beg, len, flags) do { \
 	for (i = 0; i < (len); i ++) { \
-		if ((rspamd_url_encoding_classes[(beg)[i]] & (flags)) == 0) { \
+		if ((rspamd_url_encoding_classes[(guchar)(beg)[i]] & (flags)) == 0) { \
 			dlen += 2; \
 		} \
 	} \
@@ -3614,10 +3615,10 @@ static const unsigned char rspamd_url_encoding_classes[256] = {
 
 #define ENCODE_URL_COMPONENT(beg, len, flags) do { \
 	for (i = 0; i < (len) && dend > d; i ++) { \
-		if ((rspamd_url_encoding_classes[(beg)[i]] & (flags)) == 0) { \
+		if ((rspamd_url_encoding_classes[(guchar)(beg)[i]] & (flags)) == 0) { \
 			*d++ = '%'; \
-			*d++ = hexdigests[((beg)[i] >> 4) & 0xf]; \
-			*d++ = hexdigests[(beg)[i] & 0xf]; \
+			*d++ = hexdigests[(guchar)((beg)[i] >> 4) & 0xf]; \
+			*d++ = hexdigests[(guchar)(beg)[i] & 0xf]; \
 		} \
 		else { \
 			*d++ = (beg)[i]; \
@@ -3638,7 +3639,7 @@ rspamd_url_encode (struct rspamd_url *url, gsize *pdlen,
 
 	CHECK_URL_COMPONENT ((guchar *)url->host, url->hostlen,
 			RSPAMD_URL_FLAGS_HOSTSAFE);
-	CHECK_URL_COMPONENT ((guchar *)url->user, url->userlen,
+	CHECK_URL_COMPONENT (rspamd_url_user_unsafe(url), url->userlen,
 			RSPAMD_URL_FLAGS_USERSAFE);
 	CHECK_URL_COMPONENT ((guchar *)url->data, url->datalen,
 			RSPAMD_URL_FLAGS_PATHSAFE);
@@ -3677,7 +3678,7 @@ rspamd_url_encode (struct rspamd_url *url, gsize *pdlen,
 	}
 
 	if (url->userlen > 0) {
-		ENCODE_URL_COMPONENT ((guchar *)url->user, url->userlen,
+		ENCODE_URL_COMPONENT (rspamd_url_user_unsafe (url), url->userlen,
 				RSPAMD_URL_FLAGS_USERSAFE);
 		*d++ = ':';
 	}

src/libserver/url.h

@@ -47,7 +47,9 @@ struct rspamd_url {
 	guint protocol;
 	guint port;
 
-	gchar *user;
+	guint usershift;
+	guint userlen;
+
 	gchar *host;
 	gchar *data;
 	gchar *query;
@@ -58,7 +60,6 @@ struct rspamd_url {
 	struct rspamd_url *phished_url;
 
 	guint protocollen;
-	guint userlen;
 	guint hostlen;
 	guint datalen;
 	guint querylen;
@@ -71,6 +72,9 @@ struct rspamd_url {
 	guint count;
 };
 
+#define rspamd_url_user(u) ((u)->userlen > 0 ? (u)->string + (u)->usershift : NULL)
+#define rspamd_url_user_unsafe(u) ((u)->string + (u)->usershift)
+
 enum uri_errno {
 	URI_ERRNO_OK = 0,           /* Parsing went well */
 	URI_ERRNO_EMPTY,        /* The URI string was empty */

src/lua/lua_url.c

@@ -198,8 +198,8 @@ lua_url_get_user (lua_State *L)
 	LUA_TRACE_POINT;
 	struct rspamd_lua_url *url = lua_check_url (L, 1);
 
-	if (url != NULL && url->url->user != NULL) {
-		lua_pushlstring (L, url->url->user, url->url->userlen);
+	if (url != NULL && rspamd_url_user (url->url) != NULL) {
+		lua_pushlstring (L, rspamd_url_user (url->url), url->url->userlen);
 	}
 	else {
 		lua_pushnil (L);
@@ -307,7 +307,10 @@ lua_url_tostring (lua_State *L)
 		if (url->url->protocol == PROTOCOL_MAILTO) {
 			gchar *tmp = g_malloc (url->url->userlen + 1 +
 								   url->url->hostlen);
-			memcpy (tmp, url->url->user, url->url->userlen);
+			if (url->url->userlen) {
+				memcpy (tmp, url->url->string + url->url->usershift, url->url->userlen);
+			}
+
 			tmp[url->url->userlen] = '@';
 			memcpy (tmp + url->url->userlen + 1, url->url->host,
 					url->url->hostlen);
@@ -675,7 +678,7 @@ lua_url_to_table (lua_State *L)
 
 		if (u->userlen > 0) {
 			lua_pushstring (L, "user");
-			lua_pushlstring (L, u->user, u->userlen);
+			lua_pushlstring (L, rspamd_url_user (u), u->userlen);
 			lua_settable (L, -3);
 		}