Re: pl/perl extension fails on Windows

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Christoph Berg <myon(at)debian(dot)org>
Cc: Sandeep Thakkar <sandeep(dot)thakkar(at)enterprisedb(dot)com>, Ashutosh Sharma <ashu(dot)coek88(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: pl/perl extension fails on Windows
Date: 2017-08-14 17:16:29
Message-ID: 11955.1502730989@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Christoph Berg <myon(at)debian(dot)org> writes:
> HEAD as of 5a5c2feca still has the same problem on kfreebsd. Is there
> anything I could dump so we understand the problem better?

Yeah, I did not expect that 5a5c2feca would change anything on
non-Windows.

What we need to do is verify that PL/Perl's idea of
sizeof(PerlInterpreter) is different from Perl's own idea, and then
find out why --- ie, just which fields have different size/alignment
in the two compiles.

You mentioned upthread that configure shows this:

> checking for CFLAGS recommended by Perl... -D_REENTRANT -D_GNU_SOURCE
> -DDEBIAN -fwrapv -fno-strict-aliasing -pipe -I/usr/local/include
> -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64
> checking for CFLAGS to compile embedded Perl... -DDEBIAN

If the source of the problem is the same mechanism as it was for the
other platforms, then presumably the issue is that we need one or more
of
-D_REENTRANT
-D_GNU_SOURCE
-D_LARGEFILE_SOURCE
-D_FILE_OFFSET_BITS=64
to be defined while building PL/Perl. Now, it couldn't be -D_GNU_SOURCE
that's at issue, because we turn that on in src/template/linux:

# Force _GNU_SOURCE on; plperl is broken with Perl 5.8.0 otherwise
CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"

(That ancient comment is pretty interesting in this connection, isn't it.)

And I'd have thought that _LARGEFILE_SOURCE=1 and _FILE_OFFSET_BITS=64
were the default behavior on any modern platform anyway, but maybe
kfreebsd is weird about that. Anyway, you could try sticking combinations
of these symbols into perl_embed_ccflags in src/Makefile.global and
rebuilding PL/Perl to see if the problem goes away; if that works it would
give us a leg up on where the problem is.

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Mengxing Liu 2017-08-14 17:21:38 Re: [GSOC][weekly report 9] Eliminate O(N^2) scaling from rw-conflict tracking in serializable transactions
Previous Message Alvaro Herrera 2017-08-14 17:12:19 Re: shared memory based stat collector (was: Sharing record typmods between backends)