Leonard Erickson wrote in a message to Jason Fesler:

 LE> Got a question. I commented out the "SPLIT 15000" line, as
 LE> our tosser can handle anything we are going to get (if it
 LE> ever chokes, the author will rewrite it so it won't. He
 LE> believes "text unbounded" means what it says). 

 LE> Trouble is, GIGO seems to be splitting messages *anyway*.
 LE> Since we aren't feeding anybody else, we'd kinda like to not
 LE> have this happen. I *could* put in something like "SPLIT
 LE> 16711680" (max message size our messagebase format allows)
 LE> but that'd be kinda silly.

Alas, there's a default SPLIT value if you don't set one yourself.
Go ahead and set it to the max message size you wish to honor.
I'd set it to the same as your MAX_USENETBUF.

What tosser is that?  It's a work of art to handle true unbounded text,
considering the current state of fidonet  :-).  Having half
the info at the beginning of the message is cool, but having half
at the end *sucks*..

 LE> Also, as I know from sad experience, some of the software
 LE> that can't handle large messages does so based on number of
 LE> *lines*, not size. For example, one mail import program
 LE> would silently truncated inbound email at line 100, no
 LE> matter how small the message was (sent a test message
 LE> consisting of 1 to 200, each on a seperate line. It got
 LE> chopped). 

Blech!

 LE> So it might be a thought to add a "maxlines" command to
 LE> limit the max number of lines.

I'll keep it in mind, but no promises.  I've never been asked for
this one before at all, and the amount of overhead is not small.
Currently, GIGO doesnm't try and parse lines (other than headers) - it's just
one big buffer.