freetds AT lists.ibiblio.org
Subject: FreeTDS Development Group
List archive
RE: [freetds] Further complications in compiling latest snapshot
- From: "ZIGLIO, Frediano, VF-IT" <Frediano.Ziglio AT vodafone.com>
- To: <brian AT bruns.com>, "FreeTDS Development Group" <freetds AT lists.ibiblio.org>
- Cc:
- Subject: RE: [freetds] Further complications in compiling latest snapshot
- Date: Wed, 15 Dec 2004 16:11:49 +0100
signal way should work, MSG_NOSIGNAL it's better if available (less
syscall and surely threadsafe). Some platforms (like Windows) do not
send SIGPIPE to applications.
See src/tds/net.c for an example.
freddy77
>
> Yikes,
>
> Anybody know the portable way to do this? I was getting SIGPIPE on
> dead sockets and using signal(SIGPIPE, SIG_IGN) to ignore it wasn't
> working on my Linux box.
>
> I suppose we could use autoconf to detect it and fall back to the old
> (broken) method.
>
> Brian
>
> On Wed, 15 Dec 2004 10:54:14 -0000, Thompson, Bill D (London)
> <bill_d_thompson AT ml.com> wrote:
> > Brian, Guys,
> >
> > recent changes to src/pool/member.c are incompatible with my
> > implementation of send()
> >
> > line 259:
> >
> > ret = send(puser->tds->s, buf, tds->in_len, MSG_NOSIGNAL);
> >
> > This flag to send (MSG_NOSIGNAL) is not recognised in Solaris 2.6...
> >
> > also similar code in src/pool/user.c
> >
> > Bill
-
[freetds] Further complications in compiling latest snapshot,
Thompson, Bill D (London), 12/15/2004
- Re: [freetds] Further complications in compiling latest snapshot, Brian Bruns, 12/15/2004
- <Possible follow-up(s)>
- RE: [freetds] Further complications in compiling latest snapshot, ZIGLIO, Frediano, VF-IT, 12/15/2004
-
RE: [freetds] Further complications in compiling latest snapshot,
Lowden, James K, 12/15/2004
- Re: [freetds] Further complications in compiling latest snapshot, Brian Bruns, 12/15/2004
Archive powered by MHonArc 2.6.24.