pidgin: 6b89b377: 2008.08.10 - csyfek <csyfek(at)gmail.com...

[Elliott Sales de Andrade]

2008/8/10 Richard Laager <rlaager at wiktel.com>:
>
> You should have done an "mtn rename crypt.c qq_crypt.c" (and the same
> for crypt.h) to preserve the history. Unfortunately, at this point,
> there is no way to fix this.
>

Well, disapproving does seem to undo that.

> Also, you should really be splitting these commits up into logical
> changesets instead of committing huge diffs of unrelated changes.
>

Yes, that's true. Because now we might need to disapprove the whole
thing for breaking string freeze:

+	msg = g_strdup_printf(_(
+		"Reply %s(0x%02X )\n"
+		"Sent %s(0x%02X )\n"
+		"Room id %d, reply [0x%02X]: \n"
+		"%s"),
+		qq_get_room_cmd_desc(reply_cmd), reply_cmd,
+		qq_get_room_cmd_desc(room_cmd), room_cmd,
+		room_id, reply, msg_utf8);
+
+	purple_notify_error(gc, NULL, _("Failed room reply"), msg);

And:

+			if (data_len >= 0) {
+				purple_debug(PURPLE_DEBUG_WARNING, "QQ",
+					"Decrypt login reply packet with password_twice_md5, %d
bytes\n", data_len);
+			} else {
+				purple_connection_error_reason(gc, PURPLE_CONNECTION_ERROR_NETWORK_ERROR,
+					_("Can not decrypt login reply"));
+				return;
+			}