On Mon, Jun 17, 2013 at 09:58:35AM -0400, Christopher Faylor wrote: >On Mon, Jun 17, 2013 at 06:08:27AM -0400, Andrew Schulman wrote: >>> > Do we have any idea whether the problem is with mintty or with screen? >>> >>> I observed the same behaviour as the OP, but also reproduced it in >>> rxvt and xterm, so not a mintty issue. Since it works in a console, >>> the Cygwin pty layer has to be the main suspect, unless screen has >>> specfic code for Cygwin consoles or ptys? >> >>I don't think so, but there's one big, crufty old patch that I had to use to >>get >>screen to run a few years ago. >> >>The version of screen that we're using now in Cygwin, 4.0.3, is about 8 years >>old. There haven't been any official new releases since then, but development >>has continued and there have been a lot of improvements. I'm in the process >>of >>building an updated package from the latest git, and that version appears not >>to >>need any of the patches I used to have to use to get screen to compile and run >>in Cygwin. >> >>So, I suggest that we wait until the new screen test release comes out, >>within a >>few days, and see if that fixes the problem. > >I looked at this briefly over the weekend and saw a problem in Cygwin's >pty handling code. I don't know exactly what the problem is yet but there >is something not quite right there. Something is closing a windows handle >and that's causing a SIGHUP to be sent to the pty owner. That's not right. > >I'll fix this problem eventually but it will be interesting to see if a >newer screen behaves any better.
This problem should be fixed in the upcoming snapshot. It was a longstanding bug in the pty code. Thanks for providing the succinct steps required to duplicate the problem. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
