Mailing-List: contact cygwin-help@sourceware.cygnus.com; run by ezmlm
Sender: cygwin-owner@sourceware.cygnus.com
From: Chris Faylor <cgf@cygnus.com>
Date: Wed, 14 Jun 2000 10:20:24 -0400
To: cygwin@sourceware.cygnus.com
Subject: Re: signal implementation bugs
Message-ID: <20000614102024.A32457@cygnus.com>
Reply-To: cygwin@sourceware.cygnus.com
Mail-Followup-To: cygwin@sourceware.cygnus.com
References: <200006141351.JAA22844@grex.cyberspace.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <200006141351.JAA22844@grex.cyberspace.org>; from tosh@cyberspace.org on Wed, Jun 14, 2000 at 09:51:30AM -0400

On Wed, Jun 14, 2000 at 09:51:30AM -0400, tosh wrote:
>i have been trying to use the latest version of Cygwin to port a simple
>program that uses TCP sockets.  the program uses blocking calls, such
>as recv(), connect(), or accept(), and uses alarm() to set timeouts on
>them.  much to my disappointment, i found that signal delivery does not
>interrupt these calls under cygwin.  not only that, but doing
>siglongjump() from the signal handler to the main program (in order to
>abort the timed out blocked call), also doesn't work.  is there any way
>to fix this?  other than using non-blocking i/o and select() which i
>assume would work.

As you have discovered, signals don't interrupt winsock functions
currently.  Cygwin also currently lacks an implementation of setsetjmp
and siglongjmp.

These are on the plate to be implemented but I don't have an ETA.

Sorry.

cgf

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com