Merge pull request #1413 from profanity-im/feature/mam-addtodb-nodup

MAM: Only add each message once to database

src/database.c

@@ -110,8 +110,8 @@ log_database_init(ProfAccount* account)
     // message is the message text
     // timestamp the timestamp like "2020/03/24 11:12:14"
     // type is there to distinguish: message (chat), MUC message (muc), muc pm (mucpm)
-    // stanza_id is the ID from XEP-0359: Unique and Stable Stanza IDs
+    // stanza_id is the ID in <message>
-    // archive_id is the ID from XEP-0313: Message Archive Management
+    // archive_id is the stanza-id from from XEP-0359: Unique and Stable Stanza IDs used for XEP-0313: Message Archive Management
     // replace_id is the ID from XEP-0308: Last Message Correction
     // encryption is to distinguish: none, omemo, otr, pgp
     // marked_read is 0/1 whether a message has been marked as read via XEP-0333: Chat Markers
@@ -335,7 +335,7 @@ _add_to_db(ProfMessage* message, char* type, const Jid* const from_jid, const Ji
 
     char* escaped_message = str_replace(message->plain, "'", "''");
 
-    if (asprintf(&query, "INSERT INTO `ChatLogs` (`from_jid`, `from_resource`, `to_jid`, `to_resource`, `message`, `timestamp`, `stanza_id`, `replace_id`, `type`, `encryption`) VALUES ('%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s')",
+    if (asprintf(&query, "INSERT INTO `ChatLogs` (`from_jid`, `from_resource`, `to_jid`, `to_resource`, `message`, `timestamp`, `stanza_id`, `archive_id`, `replace_id`, `type`, `encryption`) SELECT '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s', '%s' WHERE NOT EXISTS (SELECT 1 FROM `ChatLogs` WHERE `archive_id` = '%s')",
                  from_jid->barejid,
                  from_jid->resourcepart ? from_jid->resourcepart : "",
                  to_jid->barejid,
@@ -343,9 +343,11 @@ _add_to_db(ProfMessage* message, char* type, const Jid* const from_jid, const Ji
                  escaped_message,
                  date_fmt,
                  message->id ? message->id : "",
+                 message->stanzaid ? message->stanzaid : "",
                  message->replace_id ? message->replace_id : "",
                  type,
-                 enc)
+                 enc,
+                 message->stanzaid)
         == -1) {
         log_error("log_database_add(): SQL query. could not allocate memory");
         return;

src/xmpp/message.c

@@ -85,7 +85,7 @@ static void _handle_muc_private_message(xmpp_stanza_t* const stanza);
 static void _handle_conference(xmpp_stanza_t* const stanza);
 static void _handle_captcha(xmpp_stanza_t* const stanza);
 static void _handle_receipt_received(xmpp_stanza_t* const stanza);
-static void _handle_chat(xmpp_stanza_t* const stanza, gboolean is_mam, gboolean is_carbon);
+static void _handle_chat(xmpp_stanza_t* const stanza, gboolean is_mam, gboolean is_carbon, const char *result_id);
 static void _handle_ox_chat(xmpp_stanza_t* const stanza, ProfMessage* message, gboolean is_mam);
 static xmpp_stanza_t* _handle_carbons(xmpp_stanza_t* const stanza);
 static void _send_message_stanza(xmpp_stanza_t* const stanza);
@@ -232,7 +232,7 @@ _message_handler(xmpp_conn_t* const conn, xmpp_stanza_t* const stanza, void* con
         }
 
         if (msg_stanza) {
-            _handle_chat(msg_stanza, FALSE, is_carbon);
+            _handle_chat(msg_stanza, FALSE, is_carbon, NULL);
         }
     } else {
         // none of the allowed types
@@ -281,6 +281,7 @@ message_init(void)
     message->to_jid = NULL;
     message->id = NULL;
     message->originid = NULL;
+    message->stanzaid = NULL;
     message->replace_id = NULL;
     message->body = NULL;
     message->encrypted = NULL;
@@ -313,6 +314,10 @@ message_free(ProfMessage* message)
         xmpp_free(ctx, message->originid);
     }
 
+    if (message->stanzaid) {
+        xmpp_free(ctx, message->stanzaid);
+    }
+
     if (message->replace_id) {
         xmpp_free(ctx, message->replace_id);
     }
@@ -1217,7 +1222,7 @@ _handle_carbons(xmpp_stanza_t* const stanza)
 }
 
 static void
-_handle_chat(xmpp_stanza_t* const stanza, gboolean is_mam, gboolean is_carbon)
+_handle_chat(xmpp_stanza_t* const stanza, gboolean is_mam, gboolean is_carbon, const char *result_id)
 {
     // some clients send the mucuser namespace with private messages
     // if the namespace exists, and the stanza contains a body element, assume its a private message
@@ -1280,6 +1285,26 @@ _handle_chat(xmpp_stanza_t* const stanza, gboolean is_mam, gboolean is_carbon)
         message->id = strdup(id);
     }
 
+    if (is_mam) {
+        // MAM has XEP-0359 stanza-id as <result id="">
+        if (result_id) {
+            message->stanzaid = strdup(result_id);
+        } else {
+            log_warning("MAM received with no result id");
+        }
+    } else {
+        // live messages use XEP-0359 <stanza-id>
+        // TODO: add to muc too
+        char* stanzaid = NULL;
+        xmpp_stanza_t* stanzaidst = stanza_get_child_by_name_and_ns(stanza, STANZA_NAME_STANZA_ID, STANZA_NS_STABLE_ID);
+        if (stanzaidst) {
+            stanzaid = (char*)xmpp_stanza_get_attribute(stanzaidst, STANZA_ATTR_ID);
+            if (stanzaid) {
+                message->stanzaid = strdup(stanzaid);
+            }
+        }
+    }
+
     // replace id for XEP-0308: Last Message Correction
     xmpp_stanza_t* replace_id_stanza = xmpp_stanza_get_child_by_ns(stanza, STANZA_NS_LAST_MESSAGE_CORRECTION);
     if (replace_id_stanza) {
@@ -1379,7 +1404,7 @@ _handle_ox_chat(xmpp_stanza_t* const stanza, ProfMessage* message, gboolean is_m
 static gboolean
 _handle_mam(xmpp_stanza_t* const stanza)
 {
-    xmpp_stanza_t* result = stanza_get_child_by_name_and_ns(stanza, "result", STANZA_NS_MAM2);
+    xmpp_stanza_t* result = stanza_get_child_by_name_and_ns(stanza, STANZA_NAME_RESULT, STANZA_NS_MAM2);
     if (!result) {
         return FALSE;
     }
@@ -1390,8 +1415,12 @@ _handle_mam(xmpp_stanza_t* const stanza)
         return FALSE;
     }
 
+    // <result xmlns='urn:xmpp:mam:2' queryid='f27' id='5d398-28273-f7382'>
+    // same as <stanza-id> from XEP-0359 for live messages
+    const char* result_id = xmpp_stanza_get_id(result);
+
     xmpp_stanza_t* message_stanza = xmpp_stanza_get_child_by_ns(forwarded, "jabber:client");
-    _handle_chat(message_stanza, TRUE, FALSE);
+    _handle_chat(message_stanza, TRUE, FALSE, result_id);
 
     return TRUE;
 }

src/xmpp/stanza.h

@@ -110,6 +110,8 @@
 #define STANZA_NAME_COMMAND          "command"
 #define STANZA_NAME_CONFIGURE        "configure"
 #define STANZA_NAME_ORIGIN_ID        "origin-id"
+#define STANZA_NAME_STANZA_ID        "stanza-id"
+#define STANZA_NAME_RESULT           "result"
 #define STANZA_NAME_MINIMIZE         "minimize"
 
 // error conditions

src/xmpp/xmpp.h

@@ -155,7 +155,10 @@ typedef struct prof_message_t
     char* originid;
     /* <replace id> XEP-0308 LMC */
     char* replace_id;
-    /* for MAM we will need archive_id (stanza-id in XEP-0359) (see database.c) */
+    /* stanza-id from XEP 0359. Used for MAM. archive_id in our database (see database.c)
+     * coming in as <stanza-id> for live messages
+     * coming in as <result id=""> for MAM messages*/
+    char *stanzaid;
     /* The raw body from xmpp message, either plaintext or OTR encrypted text */
     char* body;
     /* The encrypted message as for PGP */