[9fans] sources.cs.bell-labs.com

[9fans] sources.cs.bell-labs.com

[Roman V. Shaposhnik]

Hi guys!

Amazingly enough I've never had a chance to create an
account for myself @sources.cs.bell-labs.com. And now
that I finally decided to do so, the URL:
   http://plan9.bell-labs.com/plan9/newaccount.html
seems to be gone.

So, the question is -- how do I set things up so that
   $ 9fs sources.cs.bell-labs.com
works?

Thanks,
Roman.

Re: [9fans] sources.cs.bell-labs.com

[gdiaz]

hello

Access to sources is anonymous since a couple of years (try 9fs sources),  you only need an account if you're going to write in sources (contrib/). I think if you want an account to write there, you should contact someone at the labs (Geoff i think) or 9trouble.

slds.

gabi




> Hi guys!
>
> Amazingly enough I've never had a chance to create an
> account for myself @sources.cs.bell-labs.com. And now
> that I finally decided to do so, the URL:
>    http://plan9.bell-labs.com/plan9/newaccount.html
> seems to be gone.
>
> So, the question is -- how do I set things up so that
>    $ 9fs sources.cs.bell-labs.com
> works?
>
> Thanks,
> Roman.

Re: [9fans] sources.cs.bell-labs.com

[Roman V. Shaposhnik]

Hi

On Wed, 2008-07-30 at 12:51 +0200, gdiaz@9grid.es wrote:
> hello
>
> Access to sources is anonymous since a couple of years (try 9fs sources),

Thanks!!! mount -n solved my issue 100%.

> you only need an account if you're going to write in sources (contrib/). I think if you want an account to write there, you should contact someone at the labs (Geoff i think) or 9trouble.

Not yet ;-)

Thanks,
Roman.

Re: [9fans] sources.cs.bell-labs.com

[Roman V. Shaposhnik]

On Wed, 2008-07-30 at 12:51 +0200, gdiaz@9grid.es wrote:
> hello
>
> Access to sources is anonymous since a couple of years (try 9fs sources)

Although, on a second though, if sources.cs.bell-labs.com doesn't
require authentication it probably should reply with Rerror
to a Tauth message. And it doesn't:
   term% aux/9pcon -n tcp!sources.cs.bell-labs.com
   Tversion 8192 9P2000
        -> Tversion tag 65535 msize 8192 version '9P2000'
        <- Rversion tag 65535 msize 8192 version '9P2000'
   Tauth 1 'roman' ''
        -> Tauth tag 1 afid 1 uname roman aname
        <- Rauth tag 1 qid (0000000000000001 0 A)

Thanks,
Roman.

Re: [9fans] sources.cs.bell-labs.com

[Russ Cox]

> Although, on a second though, if sources.cs.bell-labs.com doesn't
> require authentication it probably should reply with Rerror
> to a Tauth message. And it doesn't:
>    term% aux/9pcon -n tcp!sources.cs.bell-labs.com
>    Tversion 8192 9P2000
>         -> Tversion tag 65535 msize 8192 version '9P2000'
>         <- Rversion tag 65535 msize 8192 version '9P2000'
>    Tauth 1 'roman' ''
>         -> Tauth tag 1 afid 1 uname roman aname
>         <- Rauth tag 1 qid (0000000000000001 0 A)

You're right for most servers: the convention
is that if the server doesn't want authentication,
it sends back an Rerror to Tauth.

Sources is a bit different: it doesn't require
authentication, but it will accept it.
This is how it can tell who should be writing
to each contrib directory, for example.

The result of sources following your lead
is that you need to use -n if you really intend
not to authenticate.

Russ

Re: [9fans] sources.cs.bell-labs.com

[a]

// Sources is a bit different: it doesn't require
// authentication, but it will accept it.

Is this just 'listen -N' in fossilcons(8)?

Given that listen and users are fossil-wide, rather
than fsys-wide, it seems like duplicating what
sources is doing at a "normal" Plan 9 installation
would involve running a second fossil. If -N were
on open (which seems reasonably close to open
having -A), that wouldn't be the case. Correct?
Anthony

Re: [9fans] sources.cs.bell-labs.com

[Russ Cox]

> Is this just 'listen -N' in fossilcons(8)?
> Given that listen and users are fossil-wide, rather
> than fsys-wide, it seems like duplicating what
> sources is doing at a "normal" Plan 9 installation
> would involve running a second fossil. If -N were
> on open (which seems reasonably close to open
> having -A), that wouldn't be the case. Correct?

That depends on what you mean by "duplicating what
sources is doing."  Sources runs two fossils (one it boots
from and one it serves to the world) in two different
authentication domains.  You could pull that off with
a single fossil instance (i.e., 9P and its clients could deal
with seeing different authentication domains for different
file systems), but it would be a fair amount of work to get
everything lined up right.  Much easier just to run
two different fossils.  Putting -N on open would be just
one of many changes.

If you had just one authentication domain, then
having an open -N would let you make one of your
file systems not require authentication, like sources,
but still use a single 9P port.  I'm not sure how useful
that really is.

Russ

Re: [9fans] sources.cs.bell-labs.com

[ron minnich]

On Thu, Jul 17, 2008 at 4:57 AM, erik quanstrom <quanstro@quanstro.net> wrote:
> appears not to be in dns.  did i miss a change?

nope, hang in there. It's not you.

ron

[9fans] sources.cs.bell-labs.com

[erik quanstrom]

appears not to be in dns.  did i miss a change?

- erik