Re: BUG #3983: pgxs files still missing in win32 install (8.3.1)

From: Magnus Hagander <magnus(at)hagander(dot)net>
To: Joe Conway <mail(at)joeconway(dot)com>
Cc: Dave Page <dpage(at)pgadmin(dot)org>, Alvaro Herrera <alvherre(at)commandprompt(dot)com>, Mike Leahy <mgleahy(at)alumni(dot)uwaterloo(dot)ca>, pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #3983: pgxs files still missing in win32 install (8.3.1)
Date: 2008-04-07 06:54:29
Message-ID: 20080407085429.662fcfae@mha-laptop
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Joe Conway wrote:
> Dave Page wrote:
> > On Tue, Mar 18, 2008 at 12:47 AM, Alvaro Herrera
> > <alvherre(at)commandprompt(dot)com> wrote:
> >> Mike Leahy wrote:
> >> > Is this actually a bug, or is there a specific reason that the
> >> > pgxs files omitted in the windows installer? I noticed the
> >> > 8.3.1 installer has been posted (as the updates had appeared on
> >> > my linux machines). From what I can tell, the 8.3.1 version of
> >> > the win32 installer is still missing pgxs.mk and any related
> >> > files.
> >>
> >> It is a bug in the Win32 installer. Please report it on the Win32
> >> installer project, http://pgfoundry.org/projects/pginstaller/
> >
> > It's not an installer bug - pgxs doesn't make sense to the VC++
> > PostgreSLQ build because we don't use (or configure) any of the
> > build system on which it relies.
> >
> > I wonder if we can either build whatever is needed to make pgxs work
> > (I'm guessing Makefile.global and more) when we run the perl scripts
> > that generate the project files, or come up with a VC++ alternative
> > to pgxs.
>
> I'm finally getting back to this, and finding that plr.dll built
> against a MinGW compiled postgres will not load in running cluster
> built with VC++ (i.e. the standard binary install). Is that to be
> expected? If so, can you give me pointers (either direct guidance or
> literal URLs) on how to build postgres and related extensions with
> VC++?

In general, mingw built modules should load just fine in msvc built
postgres. AFAIK, that's how PostGIS does it for 8.3 (though I know Mark
is working on getting MSVC build support for them). Debugging may be a
bit harder (since they use different kinds of debug symbols - postgres
uses Windows style and mingw uses mingw style), but it should certainly
load.

What trouble exactly are you seeing?

//Magnus

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message j6m 2008-04-07 08:49:42 BUG #4096: PG 8.3.1. confused about remaining disk space
Previous Message Tom Lane 2008-04-07 06:02:31 Re: CREATEDB and CREATEROLE privileges cannot vacuum pg_authid and others