Sync our DTrace infrastructure with c.h's definition of type bool.
authorTom Lane
Wed, 6 Nov 2019 16:11:40 +0000 (11:11 -0500)
committerTom Lane
Wed, 6 Nov 2019 16:11:45 +0000 (11:11 -0500)
commitd4d0cd6ee23e2a111a24cbb7cc0be2b8453c8d4c
tree193243a5bfb31f2559e4514b549281b76c53e288
parent39a6210f9e0fb3f285934f275f6dbffbc7c3fae9
Sync our DTrace infrastructure with c.h's definition of type bool.

Since commit d26a810eb, we've defined bool as being either _Bool from
, or "unsigned char"; but that commit overlooked the fact
that probes.d has "#define bool char".  For consistency, make it say
"unsigned char" instead.  This should be strictly a cosmetic change,
but it seems best to be in sync.

Formally, in the now-normal case where we're using , it'd
be better to write "#define bool _Bool".  However, then we'd need
some build infrastructure to inject that configuration choice into
probes.d, and it doesn't seem worth the trouble.  We only use
 if sizeof(_Bool) is 1, so having DTrace think that
bool parameters are "unsigned char" should be close enough.

Back-patch to v12 where d26a810eb came in.

Discussion: https://postgr.es/m/CAA4eK1LmaKO7Du9M9Lo=kxGU8sB6aL8fa3sF6z6d5yYYVe3BuQ@mail.gmail.com
doc/src/sgml/monitoring.sgml
src/backend/utils/probes.d