Re: pgindent bizarreness

Why has pgindent decided to screw up all the FD_SET calls in our code?
See for example
http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/postmaster/pgstat.c.diff?r1=1.188;r2=1.189

regards, tom lane

Tom Lane wrote:
> Why has pgindent decided to screw up all the FD_SET calls in our code?
> See for example
> http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/postmaster/pgstat.c.diff?r1=1.188;r2=1.189
>
>
>

This appears to be due to this on mingw:

/mingw/include/winsock.h:typedef struct fd_set FD_SET;
/mingw/include/winsock2.h:typedef struct fd_set FD_SET;

We have a list of excluded symbols in the typdef finding code, which
currently consists of this list:

'date','interval','timestamp','ANY'

Looks like we should add 'FD_SET' to this list.

I'm actually wondering if the list of excluded symbols should go in
pgindent rather than the typedef finding code. It would be a very simple
change to the pgindent script.

cheers

andrew

Tom Lane wrote:
> Why has pgindent decided to screw up all the FD_SET calls in our code?
> See for example
> http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/postmaster/pgstat.c.diff?r1=1.188;r2=1.189

Because the typedef list supplied by Andrew includes FD_SET as a
typedef. :-O See src/tools/pgindent/typedefs.list.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

Bruce Momjian wrote:
> Tom Lane wrote:
>
>> Why has pgindent decided to screw up all the FD_SET calls in our code?
>> See for example
>> http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/postmaster/pgstat.c.diff?r1=1.188;r2=1.189
>>
>
> Because the typedef list supplied by Andrew includes FD_SET as a
> typedef. :-O See src/tools/pgindent/typedefs.list.
>

I've committed a fix to pgindent for this. Do we want to rerun pgindent
for these files?:

src/test/examples/testlibpq2.c
src/interfaces/libpq/fe-misc.c
src/backend/postmaster/syslogger.c
src/backend/postmaster/pgstat.c
src/backend/postmaster/postmaster.c
src/backend/replication/libpqwalreceiver/libpqwalreceiver.c
src/backend/port/win32/socket.c
src/backend/libpq/auth.c
contrib/pgbench/pgbench.c

cheers

andrew

Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> I've committed a fix to pgindent for this. Do we want to rerun pgindent
> for these files?:

I think the plan is to redo pgindent near the end of beta. There's
probably no need to do it right now.

regards, tom lane

Andrew Dunstan wrote:
>
>
> Bruce Momjian wrote:
> > Tom Lane wrote:
> >
> >> Why has pgindent decided to screw up all the FD_SET calls in our code?
> >> See for example
> >> http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/postmaster/pgstat.c.diff?r1=1.188;r2=1.189
> >>
> >
> > Because the typedef list supplied by Andrew includes FD_SET as a
> > typedef. :-O See src/tools/pgindent/typedefs.list.
> >
>
> I've committed a fix to pgindent for this. Do we want to rerun pgindent
> for these files?:
>
> src/test/examples/testlibpq2.c
> src/interfaces/libpq/fe-misc.c
> src/backend/postmaster/syslogger.c
> src/backend/postmaster/pgstat.c
> src/backend/postmaster/postmaster.c
> src/backend/replication/libpqwalreceiver/libpqwalreceiver.c
> src/backend/port/win32/socket.c
> src/backend/libpq/auth.c
> contrib/pgbench/pgbench.c

Uh, I rarely run pgindent on individual files like this. If it is
FD_SET you are worried about, I suggest we find all the FD_SET symbols
and just remove the whitespace pgindent added and commit that.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

Tom Lane wrote:
> Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> > I've committed a fix to pgindent for this. Do we want to rerun pgindent
> > for these files?:
>
> I think the plan is to redo pgindent near the end of beta. There's
> probably no need to do it right now.

Sure, sounds like a plan.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com