On Sun, Sep 18, 2022 at 2:01 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> That's kind of annoying --- seems to put a serious crimp in any plans
> to check this mechanically.
I don't see why it should make a huge difference. Granted we can't
really rely on the "readability-named-parameter" check in the way we'd
hoped, but AFAICT we can rely on the
"readability-inconsistent-declaration-parameter-name" check to work
everywhere. The latter check is far more important in practice, I
think, because people don't tend to omit parameter names very often.
One further caveat here is that I seem to need to set "IgnoreMacros:
false" to get perfect results for the "inconsistent" check when the C
preprocessor is involved, as it often is (e.g., with TransactionId
params).
Even if some other limitations become apparent, we can probably afford
to allow some false negatives. I don't see any evidence of that so
far, barring this issue with unnamed parameter checking.
--
Peter Geoghegan