[DGD] state[previous_object()] in kernel??

[Shentino]

I'll be one to admit that this scheme had me a mite bewildered as well when
I first saw it.

On Fri, Jul 30, 2010 at 5:10 PM, Krister Svanlund <
krister.svanlund at gmail.com> wrote:

> On Sat, Jul 31, 2010 at 1:55 AM, Felix A. Croes <felix at dworkin.nl> wrote:
> > Krister Svanlund <krister.svanlund at gmail.com> wrote:
> >
> >> Actually no, I've read through the file several time and just can't
> >> understand why it should be a mapping where, according to me, it never
> >> should contain more then one element ([the_conn_object:
> >> the_current_state]). This is the under the assumption that the
> >> connection object never changes for a single user.
> >
> > But you previously wrote that it might have to do with logging in
> > twice as the same user; what made you change your mind?
> >
> > Your first assumption was correct.  Each connection to the server
> > will get a new connection object, so logging in twice as admin will
> > link two connection objects to the same user object, for the duration
> > of the login+password sequence.
> >
> > In user::login(), connection(previous_object()) will formally establish
> > the link with the connection object, as well as break the link with the
> > previous connection object, if there is any.  The function is defined
> > in kernel/lib/user.c.
> >
> > Regards,
> > Felix Croes
> > ___________________________________________
> > https://mail.dworkin.nl/mailman/listinfo/dgd
> >
>
> Ah, the constant problem I'm having with LPC is figuring out what
> program calls what in what order. But I think I've figured out where
> I'm being lost at the moment. I can't seem to figure out what program
> calls user::login() so I can't figure out what previous_object() is
> supposed to return. But it is starting to make sense now, thanks a lot
> for the help.
> ___________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
>