[DGD] Where did all the players go?

[Blain]

How hard would it be to make DGD a firewall for the whole computer? :)

On Dec 13, 2017 4:05 AM, "Raymond Jennings" <shentino at gmail.com> wrote:

> Right now, it turns out that Kotaka's priorities are shifting into
> security and ban management code.
>
> I've started noticing some alien IPs sending HTTP headers to my telnet
> interface port...
>
> ...yes, that is a bit of a "wtf" moment.
>
> As it is I only recently made "no contact with sitebanned IPs" a hard
> policy for the mudlib, even going so far as to have BAND ask
> SYSTEM_USERD to check all outstanding connections against the siteban
> list, and if a match is found, kill that connection on the spot.
>
> This is when I noticed in my logs that HTTP headers were getting sent
> to my telnet port.
>
> Besides allowign some protocol specific handling (delegated to the
> manager of the port resonsible for the connection), I've been
> considering upgrading my siteban manager.
>
> I delegate to my connection manager most things...that way banners,
> timeouts, and error messages can be handled by the code most familiar
> with the protocol being used by connections to the port in question.
> For example, it allows HTTPD to send error messages back as a web page
>
> a) allow code itself to issue automatic sitebans on a fail2ban sorta basis
> b) as a base for a, having bans automatically expire
> c) having bans logged, at a minimum by who issued them
> d) allowing LPC code itself to issue bans (basically using the klib's
> "username" of the module in question)
>
> Noticing my test box getting nailed by alien IPs obviously misusing my
> telnet port has upped this priority
>
> On Wed, Dec 13, 2017 at 1:50 AM, Blain <blain20 at gmail.com> wrote:
> > I totally agree that DGD is way way better.  I just said that game world
> > persistence can be done with save filles when Bart said it'd need lots of
> > hacks.  This topic has been confused for a few emails now.  :)  You don't
> > need private variables, objects, and callouts in the type of presidency I
> > was originally talking about with regards to keeping track of a forest's
> > growth, for example.
> >
> > There are multiple meanings of persistence in this chain.
> >
> >
> >
> >
> > Anyway, so what MUD problems need conquering these days?  Are there
> people
> > out there still starting up MUDs?
> >
> > On Dec 13, 2017 03:44, "Raymond Jennings" <shentino at gmail.com> wrote:
> >
> >> DGD does this a lot easier.
> >>
> >> Also...as you can probably already see for yourself, doing it the
> >> save/load way at least requires cooperation from LPC not to hide
> >> anything important.
> >>
> >> Which touches on a related issue I wanted to bring up:
> >>
> >> Trust.
> >>
> >> For good security reasons, you may well want certain code and
> >> variables to be isolated from others.  This is one of the principles
> >> enforced by the kernel library's design.
> >>
> >> DGD's snapshot mechanism keeps everything preserved, but WITHOUT
> >> requring two pieces of code to get in bed together when normally they
> >> may want to stay at arms length from each other for security reasons.
> >>
> >> I'm not just talking btw about players who want to cheat, or an I3
> >> daemon that you obviously want to keep sandboxed because it interacts
> >> with alien data.
> >>
> >> I've found kernel's separation to be a very handy debugging feature
> >> and forcing internal API requests to go through the proper channels
> >> can also isolate stuff so that if something goes wrong it doesn't blow
> >> up something else.
> >>
> >> DGD's snapshot mechanism is simple and clean and gets the job done.
> >>
> >> On Wed, Dec 13, 2017 at 1:37 AM, Blain <blain20 at gmail.com> wrote:
> >> > I just don't agree that it requires a lot of hacks to convert private
> >> > variables, objects, and call_outs to a nested mapping of strings.  The
> >> lima
> >> > mudlib implanted something along these lines, though I never played
> the
> >> lib.
> >> >
> >> > I'm not knocking out downplaying persistent mechanics.  DGD saves a
> lot
> >> of
> >> > trouble with its way of implementing it.
> >> >
> >> > Still, my main gripe was about the idea of keeping the whole world in
> >> > RAM/swap and using OLC to author world items and rooms.  It's just
> not a
> >> > safe proposition in my opinion.  At least hanging a set of files which
> >> can
> >> > create a world with no snapshot should likely be the center of most
> game
> >> > designs, if not all. I fall to think of any game designs which don't
> need
> >> > to use the disk to give a game something to to back to in the event
> the
> >> > snapshots are lost or were wrong all along for some reason.
> >> >
> >> > On Dec 13, 2017 03:27, "Felix A. Croes" <felix at dworkin.nl> wrote:
> >> >
> >> >> Blain <blain20 at gmail.com> wrote:
> >> >>
> >> >> > Full persistence can be achieved using save files.
> >> >>
> >> >> Save files don't save private variables, objects, or callouts.
> >> >>
> >> >> Full persistence: you are in a fight and the next hit will determine
> the
> >> >> outcome.  If the state of the game can be saved and restored at this
> >> >> point without the player noticing, then you have full persistence.
> >> >>
> >> >> Regards,
> >> >> Felix Croes
> >> >> ____________________________________________
> >> >> https://mail.dworkin.nl/mailman/listinfo/dgd
> >> > ____________________________________________
> >> > https://mail.dworkin.nl/mailman/listinfo/dgd
> >> ____________________________________________
> >> https://mail.dworkin.nl/mailman/listinfo/dgd
> > ____________________________________________
> > https://mail.dworkin.nl/mailman/listinfo/dgd
> ____________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd