[Oberon] PPP and overall Oberon coordination
eas-lab at absamail.co.za
eas-lab at absamail.co.za
Fri May 9 06:58:36 CEST 2003
peter_easthope wrote:
> es> 2. I (or a collaborator) adds a PPPMain.SetUser with an
> escape feature.
>
> Can we keep all the authentication in one
> command? Separation of NetSystem.SetUser
> and PPPMain.SetUser seems unOberonian.
On the contrary, I think the attempt to integrate NetSystem,
Dialer and ppp makes for complexity instead of 'modularlity'.
------
Do we agree on the following:
1. there is possibly a loginID and loginPassword to the ISP
*before* entereing the ppp level.
2. these 2 strings can contain any character sequence;
therefore NO character can serve as an 'escape' char - at this stage.
3. once ppp level is entered 'escape' chars are applicable.
----- this post from 'the authority' relates to the above -----
From: James Carlson <james.d.carlson at sun.com>
Subject: Re: pppd & sighup
...[snip]...
> Any ideas as to a logical way to work through each variable?
Not at this point. You *MUST* know your correct user name and
passphrase in order to authenticate via CHAP. If you don't, you're
stuck.
One remaining possibility is that your passphrase contains a special
character, and you don't have it properly quoted. The one to be
careful with is '\' -- it needs to be doubled (as \\) to pass through
the parser. (And, of course, it *is* case-sensitive.)
Another remaining possibility is that your peer doesn't actually want
you to use CHAP, but rather to use PAP. Just change the name of the
file from chap-secrets to pap-secrets to see.
One last possibility is that your chat script is actually causing the
problem. Most modern ISPs do *not* want you to log in via the chat
script before starting PPP. Instead, you just dial the number, wait
for CONNECT, and then launch straight into PPP -- no "login:", no
password.
Sorry I can't help more than that ...
---------------- end of expert opinion. -----------
I'm waiting for a written acceptance (or denial) that the pre-ppp
level ISP login and password can't be parsed by using ANY ascii-char
(including""") as a token delimiter.
> es> ... using an escape syntax (%, \, 0x, ... whatever you
> like). I still think the syntax with ':' and '@' isn't bad.
>
> My preference is for symmetry between the
> arguments and to quote each argument rather
> than introduce escape characters which I will
> forget. This is the syntax in EBNF (to the
> best of my small skill).
Well, the first consideration must be reliability to work with any
valid input. This consideration must over-ride symmetry and
ease of use.
Chris Glur
--
PS. How IS the UserID and password parsed in the login dialog
with the ISP ?
Apparently by the 'line terminator' - <CR> and/or <LF>
and perhaps a time pause ?
Can we use this ?
More information about the Oberon
mailing list