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 Message-ID: From: "Town, Brad" To: "'cygwin AT cygwin DOT com'" Subject: RE: O_NONBLOCK serial I/O blocks on 95, OK on NT Date: Thu, 7 Dec 2000 09:35:51 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain I now remember that *my* errors were due to passing NULL for some values that should be ignored. Under NT/2000, that worked, but under 95, they'd fail with ERROR_INVALID_PARAMETER. Even though the parameters should have been ignored, I still had to provide a valid pointer. Since Cygwin's serial I/O works at all under Win9x, then that's almost certainly not the problem. I wouldn't count on an upgrade doing much for this problem. Bruce Edge [mailto:bedge AT sattel DOT com] wrote: > "Town, Brad" wrote: > > > > Microsoft's overlapped I/O for serial ports is fully > implemented, though > > it's just different enough to drive you crazy. I've had > times when serial > > I/O code (straight Win32 API stuff) would work perfectly > under NT/2000, but > > would fail under Win9x because some parameters were > considered invalid. I > > wish I could remember the details. > > So do I :-) > > Does anyone know if anything changed WRT serial I/O on 98 or ME? > I have to support _one_ of the crappy OSs. I was wondering if > an upgrade would > be worth it. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com