Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Date: Wed, 6 Dec 2000 14:37:06 -0500 From: Christopher Faylor To: cygwin support Subject: Re: O_NONBLOCK serial I/O blocks on 95, OK on NT Message-ID: <20001206143706.A9630@redhat.com> Reply-To: cygwin AT cygwin DOT com Mail-Followup-To: cygwin support References: <3A2E8EBA DOT 5EEB6BA3 AT sattel DOT com> <4 DOT 3 DOT 1 DOT 2 DOT 20001206142217 DOT 02101860 AT pop DOT ma DOT ultranet DOT com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.11i In-Reply-To: <4.3.1.2.20001206142217.02101860@pop.ma.ultranet.com>; from lhall@rfk.com on Wed, Dec 06, 2000 at 02:23:29PM -0500 On Wed, Dec 06, 2000 at 02:23:29PM -0500, Larry Hall (RFK Partners, Inc) wrote: >At 02:08 PM 12/6/2000, Bruce Edge wrote: >>I'm opening /dev/com1 as follows: >> fd = open( "/dev/com1", O_RDWR | O_NOCTTY | O_NONBLOCK | O_BINARY ); >> >>on NT the O_NONBLOCK works fine. >>on win 95, read( fd, buf, max_len ) blocks. >> >>I don't think the termios stuff should affect this, at least if it did, I >>should be having the same problem under NT. >>Is there a fix for this? >>Or, is there any way to check for chars pending before reading? >>I'm already using select(), which shouldn't even drop out of the wait until >>there's data. > > >Perhaps this is related to this? > >http://sources.redhat.com/ml/cygwin/2000-11/msg01544.html > >If so, you may find some insight by reading this thread... It's unlikely that this has any bearing. This is a thread on pipes. They use different mechanisms from serial I/O. Serial I/O in cygwin uses MS overlapped I/O. Maybe this isn't completely implemented on w9x. cgf -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com