incompatible screen update

Martin Dengler martin at martindengler.com
Tue Feb 8 15:10:22 UTC 2011


O
n Tue, Feb 08, 2011 at 03:33:08PM +0100, Tomas Mraz wrote:
> On Tue, 2011-02-08 at 14:28 +0000, Martin Dengler wrote: 
> > On Tue, Feb 08, 2011 at 03:05:03PM +0100, Tomas Mraz wrote:
> > > [Let's] not to try to enforce rules such as no user process left
> > > after logout blindly on all systems.
> > 
> > Not that I disagree with your statement (I tend to agree) in general,
> > but how is "screen is running but detached" morally equivalent to
> > "logout"?
> 
> That's just a "philosophical" question.

If we're going to talk about "logout" a definition of it is useful,
"philosopical" or not :).

> For example if you detach screen you are dropped to the original
> session and you really logout out of it.  And you are not really
> connected to the detached screen - it is detached because noone is
> attached to it isn't it? :)

Agreed...

> So it is definitely not a live login session because noone can
> interact with it unless he opens a real session and runs a screen to
> reattach it.

I agree except to all intents and purposes one is meant to be
*connected* in a significant way when a screen session is active
(attached or not).

With its goal being a "window manager"[1] with "virtual terminal"s[1]
that persist even when "the whole screen session is detached"[1], it
sounds to me like it should create a "session"[2] or "live login
session" in the sense of it not being killed when there are no users
connected to its terminal.

This is what I understood when you said we should deem background
processes worthy-of-not-being-killed and why I said I tended to agree
with your statement[3] - and Lennart's[2]: they are saying the same
thing, except that "session" seems well defined (but not
"philosophically") and "logout" does not.

Martin

1. http://www.gnu.org/software/screen/

2. http://lists.fedoraproject.org/pipermail/devel/2011-February/148459.html

> I actually do believe that screen should invoke the PAM
> session setup code 
...that is, "create a session" == "login" (for purposes of this
discussion)

3. "[I]t might be very reasonable and desirable on some multiuser
system to allow users having background processes that can keep
running even after the user logs out" - although I think "multiuser"
is unnecessarily restrictive.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
Url : http://lists.fedoraproject.org/pipermail/devel/attachments/20110208/87303024/attachment.bin 


More information about the devel mailing list