Disable incorrect character pushdown in PostgreSQL

[findepi]

Extracted from #6874

[hashhar]

LGTM

[findepi]

Turns out postgresql supports varchars and chars longer than Trino, so brought back the logic for this.

[findepi]

Had to rebase onto master, to rebase #6874, to resolve conflict, to have test run.
Sorry.

[hashhar]

Does this still preserve the char semantics of padding values?

There should be a test.

[findepi]

there is a test

[findepi]

good catch. it doesn't correctly pad on write path and during predicate pushdown (this may or may not be relevant, depending on remote's coercions, need checking)

it's pre-existing behavior though and applies to other connectors as well, see io.trino.plugin.jdbc.StandardColumnMappings#defaultCharColumnMapping

do you want to file an issue for this?

[hashhar]

#7152 - the title could be better though.

[findepi]

CI #5758 and some legit stuff

[jirassimok]

I'd put these two conditions on one line, with the comment inside the if block.

Maybe also add a bit more to the comment explaining how isDiscreteSet is related to case sensitivity. The comment doesn't really clarify much right now. (Ideally, I think CASE_INSENSITIVE_CHARACTER_PUSHDOWN would have a more detailed explanation of how that works, and then we could reference it here.)

[findepi]

I put the comment in between so it unambiguous what it refers to

point take on the issue how discrete sets translate to actual predicates, and indeed it applies to CASE_INSENSITIVE_CHARACTER_PUSHDOWN