PG->Fabric (T-SQL): age() emitted verbatim

[NivasSA]

Title: PG->Fabric (T-SQL): age() emitted verbatim -> Msg 195 (and returns interval, no Fabric type)

age() has no Fabric built-in, yet strict() emits it verbatim. polyglot-sql 0.5.15, strict:

age(d1, timestamp '2000-01-01') -> AGE(d1, CAST('2000-01-01' AS DATETIME2(6)))

Fabric raises "Msg 195: 'AGE' is not a recognized built-in function name". age() also
returns interval (no Fabric type; cf. #301), so it must be rejected regardless.
Expected: Err(Unsupported), as polyglot already does for lpad/rpad. Sibling of #286 / #292.

[tobilg]

Thanks for the new issues @NivasSA and @VaibhaveS. To be honest I‘m unsure how I can sustain this pace of incoming issues, more than 40 only for today… I appreciate to work you put into it, but it’s like A LOT.

[tobilg]

This is fixed in the upcoming version 0.5.16 that will be available in approx. 2h

[NivasSA]

Hi @tobilg, once again thank you for your prompt fixes. I see you have already fixed the issues we reported.

The recent influx of reports is from us exercising Polyglot much more heavily than before. We genuinely appreciate the work on the project. We are evaluating whether Polyglot can support some internal projects, so as we work through the findings, we will also look for opportunities to contribute fixes and PRs where we can.

[tobilg]

Thanks for the feedback! PRs would help to speed up things I guess, still having lots of details on the issue reports as you're doing it already is great for automated resolving of the issues.

I started some internal refactoring of Polyglot on the weekend, and I think I found a way to handle most issue reports faster / better with a combination of a custom prompt and /goal from Codex.

[NivasSA]

That sounds great. We are happy to provide more detail in the issues we file if that helps streamline investigation or debugging. Let us know what information would be most useful, and we will incorporate it going forward.

For additional context, we are currently evaluating Polyglot against PostgreSQL's regression test suite. A lot of the recent reports are coming from it. Just thought it might be useful as you plan the refactoring work.

[NivasSA]

The issue is fixed, so closing it. Thanks