Re: Libpq problem on Windows. - Mailing list pgsql-hackers

From Magnus Hagander
Subject Re: Libpq problem on Windows.
Date
Msg-id 6BCB9D8A16AC4241919521715F4D8BCE475DF2@algol.sollentuna.se
Whole thread Raw
In response to Libpq problem on Windows.  ("Dave Page" <dpage@vale-housing.co.uk>)
Responses Re: Libpq problem on Windows.
List pgsql-hackers
Bruce, while we're pondering on how to solve this, can you put this up
on the open items list so we don't miss it? It's a pretty major issue.

//Magnus

> -----Original Message-----
> From: Dave Page [mailto:dpage@vale-housing.co.uk]
> Sent: Thursday, September 30, 2004 11:35 PM
> To: pgsql-hackers@postgresql.org
> Subject: [HACKERS] Libpq problem on Windows.
>
> I posted a message to pgsql-hackers-win32 regarding a problem
> with libpq exports on Win32 the other day, but have yet to
> receive any replies.
> Magnus & I have discussed the problem over IM and we both
> believe it is important to fix, but neither of us are fluent
> enough in make-ese to do so.
>
> Basically the issue is that the symbols exported by the mingw
> build of libpq.dll do not match those exported by the
> VC++/Borland builds. What we seem to get is:
>
> - Mingw exports *all* symbols appropiate for the given build.
> - VC++/Borland builds appear to export only the published
> API, but not the SSL related symbols (even with SSL enabled
> in the build).
>
> I can probably sort the second problem by adding a second set
> of .def files for SSL builds, however it seems to me that the
> mingw build should not export all symbols. Whether or nt that
> is changed, the 2 build types certainly need to be brought
> into sync otherwise Windows users may end up seeing
> unexplained crashes.
>
> Any comments on what should be done (and whether I should
> hack the BCC/VC++ stuff)?
>
> Regards, Dave.
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 8: explain analyze is your friend
>
>


pgsql-hackers by date:

Previous
From: "Simon Riggs"
Date:
Subject: Re: shared memory release following failed lock acquirement.
Next
From: Reini Urban
Date:
Subject: Re: open item: tablespace handing in pg_dump/pg_restore