On Wed, Dec 30, 2020 at 4:11 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Alexander Korotkov <akorotkov@postgresql.org> writes:
> > Multirange datatypes
>
> Multiple buildfarm members are issuing similar warnings about
> multirangetypes.c:
>
> calliphoridae | 2020-12-29 23:10:11 |
/home/andres/build/buildfarm-calliphoridae/HEAD/pgsql.build/../pgsql/src/backend/utils/adt/multirangetypes.c:218:37:
warning:suggest braces around empty body in an 'else' statement [-Wempty-body]
> calliphoridae | 2020-12-29 23:10:11 |
/home/andres/build/buildfarm-calliphoridae/HEAD/pgsql.build/../pgsql/src/backend/utils/adt/multirangetypes.c:236:37:
warning:suggest braces around empty body in an 'else' statement [-Wempty-body]
>
> These are evidently unhappy with code like
>
> if ...
> ...
> }
> else
> /* include it in range_str */ ;
>
> So apparently preferred style is more like
>
> else
> {
> /* include it in range_str */
> }
>
> I don't particularly care for the wording of this comment, either:
> it looks like it's describing an action that's about to be taken,
> only there's no action there. Could it say something more like
> "we already included it in range_str", or "we'll include it in
> range_str below the switch", or whatever's appropriate?
Thank you for noticing. I've slightly refactored this function in 16d531a30a.
------
Regards,
Alexander Korotkov
