Re: pgsql: Make hstore_plperl's build even more like plperl's - Mailing list pgsql-committers

From Peter Eisentraut
Subject Re: pgsql: Make hstore_plperl's build even more like plperl's
Date
Msg-id 55450325.6090204@gmx.net
Whole thread Raw
In response to Re: pgsql: Make hstore_plperl's build even more like plperl's  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: pgsql: Make hstore_plperl's build even more like plperl's  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-committers
On 5/2/15 11:05 AM, Andrew Dunstan wrote:
> Note that the override is EXACTLY what is done unconditionally in the
> plperl GNUmakefile:
>
>    override CPPFLAGS := -I. -I$(srcdir) $(CPPFLAGS)
>    -I$(perl_archlibexp)/CORE

The override is a red herring.  It's done this way because all CPPFLAGS
settings have to be done with override, for unrelated reasons.  In a
pgxs build, however, you can set PG_CPPFLAGS, which is later put into
CPPFLAGS with override.

Also, requiring that -I$(perl_archlibexp)/CORE is last seems bizarre.
That would mean that an earlier include directory contains a file that
is also in .../CORE, and you don't want to the one in .../CORE.  In that
case, why add .../CORE at all?  We get Perl headers from .../CORE, and
that would then mean that other include directories contain some Perl
headers that you want to use in preference to the ones in .../CORE, but
at the same time you want the ones in .../CORE as a fallback.

That might possibly be the actual case, but it would be quite odd and
should be properly explained for posterity.



pgsql-committers by date:

Previous
From: Andrew Dunstan
Date:
Subject: Re: pgsql: Make hstore_plperl's build even more like plperl's
Next
From: Peter Eisentraut
Date:
Subject: Re: pgsql: Make hstore_plperl's build even more like plperl's