Re: Fix shadow warnings in logical replication code

On Wed, Aug 30, 2023 at 09:16:38AM +1000, Peter Smith wrote:
> logicalfuncs.c:184:13: warning: declaration of ‘name’ shadows a
> previous local [-Wshadow]
> char *name = TextDatumGetCString(datum_opts[i]);
> ^
> logicalfuncs.c:105:8: warning: shadowed declaration is here [-Wshadow]
> Name name;

A bit confusing here, particularly as the name is reused with
ReplicationSlotAcquire() at the end of
pg_logical_slot_get_changes_guts() once again.

> reorderbuffer.c:4843:10: warning: declaration of ‘isnull’ shadows a
> previous local [-Wshadow]
> bool isnull;
> ^
> reorderbuffer.c:4734:11: warning: shadowed declaration is here [-Wshadow]
> bool *isnull;
> ^

Agreed as well about this one.

> walsender.c:3543:14: warning: declaration of ‘sentPtr’ shadows a
> global declaration [-Wshadow]
> XLogRecPtr sentPtr;
> ^
> walsender.c:155:19: warning: shadowed declaration is here [-Wshadow]
> static XLogRecPtr sentPtr = InvalidXLogRecPtr;
> ^

This one looks pretty serious to me, particularly as the static
sentPtr is used quite a bit. It is fortunate that the impact is
limited to the WAL sender stat function.

Fixing all these seems like a good thing in the long term, so OK for
me. Like all the fixes similar to this one, I don't see a need for a
backpatch based on their locality, even if sentPtr makes me a bit
nervous to keep even in stable branches.

There is much more going on with -Wshadow, but let's do things
incrementally, case by case.
--
Michael