Hi! mark,

On 23 Apr 09 15:01, you wrote to me:

 ml> JAM message bases should be good for messages up to 2Gig in size... i
 ml> know that back in the heyday, when i was working on my posting tool, i
 ml> posted the complete nodelist to my JAM bases numerous times... where i
 ml> ran into problems was with the tools i used to read the messages in
 ml> the JAM bases... many of those tools were limited in their memory
 ml> allocation and thus limited in the size of a message they could
 ml> display...

Sorry to rouse you to reply to my plea.  You're right, of course.  I just
had a squiz at the latest available source for Transnet and the limits
are...

  #define MAXMSGLINE 1000
  #define MAXMSGSIZE 64000L

It's written in fugly C and I haven't been anything like a master in that
for over 11 years, so I don't know precisely what the second value is other
than it is likely to be a 64k limit with a type declaration.

 ml> i don't think that your problem is with the storage in JAM
 ml> bases unless the tools you are using to place those digest posts in
 ml> the JAM bases are similarly limited like the readers i speak of
 ml> above...

Yep, right again.  Both goldEd and SemPoint have the same trouble.  Even a
'write to disk' command from goldEd produces a truncated file - I have one
to prove it, along with the original version prior to posting by Transnet.

 ml> all that aside, i can fully understand your reasoning for wanting to
 ml> split the digests... personally, i don't do digests prefering instead
 ml> to get the invidivial messages sent as they are submitted... it save
 ml> me some work and i don't have to think about spliting the digests into
 ml> individual messages :)

As do I; I'd rather it be individual posts as well.  But I exhausted web
queries of *free* news servers and even tried a couple of servers using the
'mailer' from another tool I used to use called 'NewsGate'.  No joy, so I
had to revert to using Google Groups.

 ml> i'd suspect that like other similar *nix formats, the next From: line
 ml> would indicate the start of the next message's header... when reading
 ml> the header, it is all header until you hit the first blank line...
 ml> after that, everything is the body until the next From: line... this
 ml> provided that the From: line is the first line of each and every
 ml> header...

I'm somewhat spoiled by the Google Groups digest format.  They have a
master subject line giving the total post count for the digest, so I can
code something using a loop counter and compare posting executions with
that control value.  Plus, the digest itself has lead-in text placed before
each post that can possibly be searched for.  Same idea as yours just a
different 'target'.  :)

I have an outline approach (no postings yet) that wants to work, and should
have it all squared away Sunday arvo.

Thanks for your help and guidance again, mark.  All the best.

Cheers,
Paul.

... Tower case?  Nah, fell off the desk and landed that way.
--- GoldED+/LNX 1.1.4.7