Mailing-List: contact cygwin-help@sourceware.cygnus.com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner@sources.redhat.com Delivered-To: mailing list cygwin@sources.redhat.com Date: Thu, 20 Sep 2001 13:35:57 +0200 From: Corinna Vinschen To: cygwin@cygwin.com Subject: Re: fcntl(F_SETFD) on a file or socket handle Message-ID: <20010920133557.G4194@cygbert.vinschen.de> Mail-Followup-To: cygwin@cygwin.com References: <077d01c1400e$2d92cdc0$d2823bd5@dmitry> <20010918101238.H22900@cygbert.vinschen.de> <008501c14183$d9eddc10$d2823bd5@dmitry> <20010920090219.A4194@cygbert.vinschen.de> <003601c141a8$8adc7760$d2823bd5@dmitry> <20010920100907.D4194@cygbert.vinschen.de> <001101c141b1$f3929010$d2823bd5@dmitry> <20010920105800.E4194@cygbert.vinschen.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20010920105800.E4194@cygbert.vinschen.de>; from cygwin@cygwin.com on Thu, Sep 20, 2001 at 10:58:00AM +0200 On Thu, Sep 20, 2001 at 10:58:00AM +0200, Corinna Vinschen wrote: > On Thu, Sep 20, 2001 at 05:53:17PM +0900, Dmitry Timoshkov wrote: > > "Corinna Vinschen" wrote: > > > > [skipped] > > > > > > I see that fhandler_socket::fcntl() calls fhandler_base::fcntl() in > > > > all cases different from F_SETFL. Perhaps my understanding of things > > > > in C++ is not enough (I'm an old hand plain C programmer). > > > > > > set_close_on_exec() is a virtual method which means, the called > > > method is dependent of the actual class of the object which calls > > > it. > > > > Thanks for clarification. > > > > > Did you get the sources of Cygwin via CVS? If so, could > > > you please check if my just checked in patch works for you? > > > > I don't get Cygwin sources from CVS, but I have looked into your patch > > after it has appeared in CVS. Using SetHandleInformation() on NT instead > > of DuplicateHandle()/CloseHandle() pair is much better IMO, and should > > not cause any side effects. > > Yeah but even if the patch helps for NT, it wouldn't solve it > for 9x. Sigh. I have checked in another change to sockets now which changes the behavior so that socket handles will not be Dup'd/Closed anymore. Instead, after an exec they will be closed when close_on_exec is set. Could you please test it as soon as possible? Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developer mailto:cygwin@cygwin.com Red Hat, Inc. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/