Re: [PATCH] relocation truncated to fit: citus build failure on s390x

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Andres Freund <andres@anarazel.de>, Christoph Berg <myon@debian.org>, Jason Petersen <jason@citusdata.com>, pgsql-pkg-debian@postgresql.org, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2017-05-30T20:56:22Z
Lists: pgsql-hackers
On Tue, May 30, 2017 at 4:38 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> It'd be interesting if people could gather similar numbers on other
> platforms of more real-world relevance, such as ppc64.  But based on
> this small sample, I wouldn't object to just going to -fPIC across
> the board.

That seems pretty sensible to me.  I think we should aim for a
configuration that works by default.  If we use -fPIC where -fpic
would have been better, the result is that on some platforms there
might be a speed penalty, probably small.  In the reverse situation,
the build fails.  I believe that the average PostgreSQL extension
developer will prefer the first situation.  If somebody has got an
extension which is small enough to use -fpic and that person cares
about minimizing the performance penalty on s390 or other platforms
where they have this problem, then they can arrange an override in the
opposite direction.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company


Commits

  1. Always use -fPIC, not -fpic, when building shared libraries with gcc.