Include hash of result set metadata in prepared statement id

[beltran]

Implement the driver part of https://issues.apache.org/jira/browse/CASSANDRA-10786. This is a protocol 5 change. The problem with doing it is that it breaks protocol v5 against 3.11 so we wouldn't have a way of testing it in travis until Cassandra 4 is released.

[alourie]

@beltran we possibly could run them as a separate test with ccm running on trunk.

[Zariel]

I think fundamentally if C* are not going to port this to all V5 versions then they should reconsider how they've implemented it, I really don't want to have checks in the protocol layer like if proto == 5 && version >= 4.0.0 this just seems silly

[Zariel]

Or we just dont support v5beta anymore?

[beltran]

I believe the approach other drivers have taken is to let v5 break for 3.11 and implement it as it only had to work for 4.0. I've verified this with the python driver but I'll check the java one later.

[beltran]

It's the same for the java driver. I haven't seen it officially stated anywhere but it seems like V5 just doesn't work for 3.11. It breaks with prepared statements.

[alourie]

Maybe I could work on a patch for C* that wouldn't allow using V5 with 3.x :-)

[OleksiienkoMykyta]

Now it's not relevant, Cassandra v3 uses native protocol v4 as it's highest-supported version. The native protocol v5 was introduced in Cassandra v4.0.
About the (Include hash of result set metadata in prepared statement id), will be fixed by #1822
Should be closed.

[joao-r-reis]

This is being tracked in CASSGO-1