Re: pgsql: Fix incidental warnings from cpluspluscheck.

Thomas Munro <thomas(dot)munro(at)gmail(dot)com> writes:
> On Mon, Aug 19, 2019 at 11:01 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Remove use of "register" keyword in hashfn.c. It's obsolescent
>> according to recent C++ compilers, and no modern C compiler pays
>> much attention to it either.

> We have some more of those:

> src/backend/regex/regexec.c
> src/common/md5.c
> src/include/port/atomics/arch-x86.h
> src/include/storage/s_lock.h
> src/pl/plperl/ppport.h

I was just getting rid of the ones that cpluspluscheck complained
about, which were in function prototypes. I doubt we want to touch
any of the register local variables in s_lock.h or arch-x86.h,
because they're all associated with asm sections that require them
to be registers anyway. And ppport.h is not ours to change.
I have no objection to getting rid of the ones in regexec.c or
md5.c, though.

regards, tom lane