Validate sqlc function arguments

[ryan-berger]

Fix #1632

[kyleconroy]

Looks like there's a test failure. Sorry that it took a few days to run.

[ryan-berger]

Oops, thanks for catching that. I forgot to rerun tests. I took a look and it seems a bit of an interesting problem. The error message that the test expects comes from the Visitor which returns an error when it visits an invalid function. If I return the error in there, it will crash because the ParamStyle function is called before the Visitor is invoked If I keep it how it is, tests fail and I have to duplicate logic. I'll take a look at how to deal with this tomorrow. If you have any preferences, let me know and I'll make sure to implement it that way

…

On Sun, May 22, 2022, 1:04 AM Kyle Conroy ***@***.***> wrote: Looks like there's a test failure. Sorry that it took a few days to run. — Reply to this email directly, view it on GitHub <#1633 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADRKZHNXCPGQ7PJMVUPA7F3VLHL7NANCNFSM5WNVEYQA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[kyleconroy]

It looks like we already validate the arguments at a later spot. How about we just return early if there are no args to avoid the panic?

[ryan-berger]

@kyleconroy Fixed.

I didn't see the nil/zero check initially but once you remove that from the visitor's code, it will output the correct error message.

I've also written a test just in case. Even though it doesn't really add much to coverage, it will at least serve as a regression test for future work.

Thanks for taking your time to maintain this!