On Wed, Apr 3, 2024 at 11:17 AM Tristan Partin <tristan@neon.tech> wrote:
> I think patch 2 makes it worse. The value in -Wswitch is that when new
> enum variants are added, the developer knows the locations to update.
> Adding a default case makes -Wswitch pointless.
>
> Patch 1 is still good. The comment change in patch 2 is good too!
It seems to me that 0001 should either remove the pg_unreachable()
call or change the break to a return, but not both. The commit message
tries to justify doing both by saying that the pg_unreachable() call
doesn't have much value, but there's not much value in omitting
pg_unreachable() from unreachable places, either, so I'm not
convinced.
I agree with Tristan's analysis of 0002.
--
Robert Haas
EDB: http://www.enterprisedb.com