NNTP protocol via REXX
John Small
REXX program.
I have done an IPTRACE of such a posting and I think I have figured
out most of what I need. But I do have a few questions:
1) How is the Message ID determined? Is there a formal process to
determining it? It seems to be somewhat random. Is the only
requirement that it be unique?
2) Is the User Agent required? If yes, does it matter what the string
is?
3) I am literally interested only in being able to post simple text
messages. So I hope to avoid having to dig into a highly technical
reference on NNTP. Is there a resource along the lines of a "NNTP for
Dummies" which I could use to get answers to simple questions like
these?
--
John Small
Trevor Hemsley
<jsm...@os2world.net> wrote:
> I am looking into posting messages to newsgroups like this using a
> REXX program.
>
> I have done an IPTRACE of such a posting and I think I have figured
> out most of what I need. But I do have a few questions:
> 1) How is the Message ID determined? Is there a formal process to
> determining it? It seems to be somewhat random. Is the only
> requirement that it be unique?
If you don't provide one then I think many servers will do so for you.
> 2) Is the User Agent required? If yes, does it matter what the string
> is?
It isn't required.
> 3) I am literally interested only in being able to post simple text
> messages. So I hope to avoid having to dig into a highly technical
> reference on NNTP. Is there a resource along the lines of a "NNTP for
> Dummies" which I could use to get answers to simple questions like
> these?
RFC1036 is as good as any
--
Trevor Hemsley, Brighton, UK
Trevor dot Hemsley at ntlworld dot com
Steven Levine
<Trevor....@mytrousers.ntlworld.com> wrote:
Hi,
> RFC1036 is as good as any
In addition to this, John should probably look at RFC977 which
explains how to post messages to the server.
Steven
--
---------------------------------------------------------------------
Steven Levine <ste...@earthlink.bogus.net>
eCS/Warp/DIY etc. www.scoug.com www.ecomstation.com
---------------------------------------------------------------------
Jonathan de Boyne Pollard
>>
>> RFC1036 is as good as any
>>
> In addition to this, John should probably look at RFC977 which
> explains how to post messages to the server.
>
As for message IDs: There's a reason that whilst the USEFOR
documentation of the protocol, message format, and the architecture
reached RFC status, the papers on message IDs did not. The creation of
message IDs in Usenet is a non-trivial business, whose complexity is not
reduced by the fact that most of the published descriptions of how
message IDs are formed, including the USEFOR RFC drafts on the subject,
are based upon a simplistic ideal world, and don't match what widespread
real world practice (which comprises at least three major distinct ID
composition schemes) actually is. It's necessary to know what happens
in the real world in order to satisfy the uniqueness guarantee. The
idealistic descriptions are not by themselves enough for a correct
implementation. Fortunately, if one is injecting a message via NNTP,
the server is required to add a message ID header if the proto-article
does not contain one. So for a simple injection-only NUA in REXX one
can just omit the header entirely and sidestep the problem, as
mentioned. (-:
John Small
<J.deBoynePoll...@NTLWorld.COM> wrote:
> >
> >>
> >> RFC1036 is as good as any
> >>
> > In addition to this, John should probably look at RFC977 which
> > explains how to post messages to the server.
> >
> You're both out of date. RFCs 5322, 5536, and 5537 are the places to go.
Thanks for this. I'll have a look at these. The older RFC's seemed to
conflict with what I was seeing.
> .... Fortunately, if one is injecting a message via NNTP,
> the server is required to add a message ID header if the proto-article
> does not contain one. So for a simple injection-only NUA in REXX one
> can just omit the header entirely and sidestep the problem, as
> mentioned. (-:
This was already suggested by Trevor and I once I stopped trying to
generate a message ID, my code started working!
Thanks to all for your help! (I'll be back if I run into more
difficulties down the road.)
--
John Small
