| dvratil |
potential performance impact because we fetch every item for each
relation, let's hope the caching fixes that.
notes from randa:
-> at them moment the mimtype as string is used as argument and also transfered
-> till what point i should use the id, and where to transform to the name?
-> should be done
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
The mimetypes should be transmitted in FetchRelationsCommand as QByteArrays, not qint64s, then the server should resolve the mimetypes into IDs to be used as part of the SQL query (mimetypes will always hit the in-memory cache, so no performance problem there). You don't have to worry about transmitting an empty QByteArray being a waste of resources, as a matter of fact it's more effective than transmitting a null qint64 :)
Note that you also need to adjust FetchRelationsResponse to hold the mimetypes, again using QByteArrays there is OK.
| src/private/protocol.cpp | ||
|---|---|---|
| 2592 | Please adjust the debug string as well | |
This changes protocol interface, please bump the number in Protocol::version() and adjust ProtocolTest::testProtocolVersion() accordingly (you can ignore the comment in the function, we don't have a test for Relations yet :)