On Tue, Jul 25, 2017 at 11:32 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Robert Haas <robertmhaas@gmail.com> writes:
>> On Tue, Jul 25, 2017 at 11:00 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>>> Hm, I had the idea that we were already asking ExtUtils::Embed for that,
>>> but now I see we only inquire about LDFLAGS not CCFLAGS. Yes, this sounds
>>> like a promising avenue to pursue.
>>>
>>> It would be useful to see the results of
>>> perl -MExtUtils::Embed -e ccopts
>>> on one of the affected installations, and compare that to the problematic
>>> field(s).
>
>> Why ccopts rather than ccflags?
>
> I was looking at the current code which fetches ldopts, and analogizing.
> Don't know the difference between ccflags and ccopts.
Oh, here I was thinking you were 3 steps ahead of me. :-)
Per "perldoc ExtUtils::Embed", ccopts() = perl_inc() plus ccflags()
plus ccdlflags().
On my system:
[rhaas pgsql]$ perl -MExtUtils::Embed -e 'for (qw(ccopts ccflags
ccdlflags perl_inc)) { print "==$_==\n"; eval "$_()"; print "\n\n";
};'
==ccopts==-fno-common -DPERL_DARWIN -mmacosx-version-min=10.12 -pipe -Os
-fno-strict-aliasing -fstack-protector-strong -I/opt/local/include
-DPERL_USE_SAFE_PUTENV
-I/opt/local/lib/perl5/5.24/darwin-thread-multi-2level/CORE
==ccflags==-fno-common -DPERL_DARWIN -mmacosx-version-min=10.12 -pipe -Os
-fno-strict-aliasing -fstack-protector-strong -I/opt/local/include
-DPERL_USE_SAFE_PUTENV
==ccdlflags==
==perl_inc==-I/opt/local/lib/perl5/5.24/darwin-thread-multi-2level/CORE
I don't have a clear sense of whether ccopts() or ccflags() is what we
want here, but FWIW ccopts() is more inclusive.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company