🚀 go-pugleaf

I'm answering part of my own question.

---On Wed, 16 Dec 1992 22:15:08 GMT, hamm@ssdt-bluestein.sps.mot.com (Steve Hamm) said:

Steve> I just switched NNTP servers, and now if I post something small
Steve> it works (like a one-line test, or this note) but if I post
Steve> something larger (say 100 lines) gnus hangs, saying "writing to
Steve> process: no more processes, nntpd".  Anyone see this before
Steve> and/or have a fix?  Thanks.

I got into nntp.el and found the routine nntp-send-region-to-server.  In
it is the following comment and code fragment:

    ;; `process-send-region' does not work if text to be sent is very
    ;;  large. I don't know maximum size of text sent correctly.
    (let ((last nil)
	  (size 100))			;Size of text sent at once.

Since the value of size is empirical, I figured that I could try
decreasing it and see what happens.  I got it down to 10 before things
were quite reliable.

Now, anyone care to tell me *why* this works?  (Other hints: my
machine was just upgraded to a sparc 10 and the new nntp server is
apparently much slower than my old nntp server.)  Seems like it must
be a handshake problem, but this fix seems rather inefficient --
anyone know a way of improving the efficiency, or a better workaround?

--Steve
--
Steve Hamm --- Motorola Inc. 3501 Ed Bluestein Blvd, MD-M2, Austin TX 78762
hamm@ssdt-bluestein.sps.mot.com  UUCP: ...cs.utexas.edu!oakhill!austoto!hamm