This is the mail archive of the
mailing list for the Cygwin project.
Re: screen, try #1
- From: Christopher Faylor <cgf-no-personal-reply-please at cygwin dot com>
- To: cygwin-apps at cygwin dot com
- Date: Fri, 22 Apr 2005 08:15:23 -0400
- Subject: Re: screen, try #1
- References: <firstname.lastname@example.org>
- Reply-to: cygwin-apps at cygwin dot com
On Fri, Apr 22, 2005 at 07:57:23AM -0400, Andrew Schulman wrote:
>Using Christopher's patches and disabling FIFOs as suggested, I've built a
>test version of screen:
>This version runs, but it has several problems:
>* Reattaching detached sessions succeeds with Cygwin 1.5.12, but fails with
>1.5.15. On host #1 with WinXP + Cygwin 1.5.12-1, reattaching succeeds and I
>$ ll /tmp/uscreens/S-andrex
>srw------- 1 andrex None 51 Apr 22 06:05 1944.console.CARBON
>On host #2 with WinXP + Cygwin 1.5.15-1, reattaching fails ("There is no
>screen to be resumed"), and I see
>$ ll /tmp/uscreens/S-aschulma
>-rw------- 1 ASchulma None 53 Apr 22 06:00 4056.console.D77E1BASCHULMA1
>This isn't a socket, which would explain the trouble.
>* Video colors become confused sometimes, especially when creating new windows
>within screen. For example, twice my whole console window has changed to
>reverse video, both outside of screen and in all of the windows within
>screen. In the properties of the parent console window, the colors were
>swapped, but changing them back didn't fix the problem; I finally had to
>close the whole console. Another time the text color changed to cyan, which
>happens to be my prompt color.
>* Detached sessions are still tied to their parent consoles. If I detach a
>session and then exit the parent shell, the shell doesn't exit and the zombie
>console window stays around. If I force the console window to close, screen
>There are two different patch sets that people have used to get screen
>(http://cygwin.com/ml/cygwin-apps/2005-04/msg00163.html), and the older patch
>set that Lapo found (http://www.j10n.org/files/screen-4.0.2-cygwin.patch). I
>used Christopher's patches to get these results. I also tried the older one,
>which is much larger and includes Christopher's. I found the same
>reattachment problems, but haven't had time to test the other problems yet.
Some of the patches above look reasonable, especially for getting screen
working on windows 9x. Adding a "dllimport" to a declaration for a system variable
in a .c file seems like something that should be addressed by using the appropriate
I got attach/detach working repeatedly on 1.5.15 (without fifos), so I don't know
why you'd be seeing a problem.
>Should this version be uploaded so people can test it? I've labeled it 'test'
It sounds like it is still in a pretty iffy state to me if attach/detach aren't
working. That's one of the major selling points of screen.