On Wed, Apr 21, 2004 at 11:50:59AM +0200, Corinna Vinschen <[EMAIL PROTECTED]> wrote: > On Apr 21 02:36, Yitzchak Scott-Thoennes wrote: > > On Wed, Apr 21, 2004 at 10:46:59AM +0200, Corinna Vinschen <[EMAIL PROTECTED]> > > wrote: > > > Did you try it with a recent Cygwin version? I used your above test > > > application and I'm getting a EPIPE (resp. a SIGPIPE) as expected. > > > > I think I had heard that others were not seeing this failure. I'm > > attaching an strace from my test program in case that may be helpful. > > Is there anything you can think of that might make a difference? > > Would you mind to debug that by yourself? You are seeing this problem so > I guess you're the best person to debug that problem. Setting a breakpoint > to 'fhandler_socket::sendmsg(msghdr const*, int, int)' should be fine. > I can't see any reason for Cygwin to do that. There's no loop anywhere > which could run endlessly. So either WSASendTo itself hangs, which is > pretty unlikely for overlapped sockets, or WSASendTo returns SOCKET_ERROR > and WSAGetLastError returns WSA_IO_PENDING so that WSAWaitForMultipleEvents > in wsock_event::wait hangs. Any chance that a firewall or anti-virus > software is interfering?
I never had a chance to debug this, but upgrading to the 20040520 snapshot (from the 20040430) fixed this (as well as a hang using perl Net::Ping to ping 127.0.0.1 via udp that I was seeing and others apparently weren't.) Both were probably dependent on firewall (McAfee) or vpn software (Check Point VPN-1 SecureClient NG) I'm running, but seem to be better now. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
