We have been experiencing some very odd network printing problems using
third party products - specifically ESPUL and its twin brother NetPrint.

We have several 'environment' files of PCL code that we attach to spoolfiles
destined for our various Laserjets.  These environment files control fonts,
shading and such (i.e. give simulated green bar).

Our problem is that these environment files stop working for a period of
time almost every day.  By 'stop working', I mean that the resultant file
being transmitted to the printer does not have the PCL commands in it at
all, resulting in the report being printed in whatever the printer's default
is. This was verified with a Lanalyzer trace.

We have worked extensively with Rich Corn at RAC, who has scoured code for
any possible bug.  He sent us an advanced trace version that seemed to show
a problem - the program would receive an EOF on its first read of the
environment file (after a successful open).  This is on the same environment
file that may have worked with the prior spoolfile just fine.

What's odd is that it seems to happen around the same time every day -
during our main early morning production run.  But not always (which seems
to eliminate some odd thing involving the production jobs).  A further odd
thing - this began happening immediately after we switched to a frame relay
connection between all of our plants (which simply involved adding a gateway
in NMMGR on the HP3000).  Prior to that, we had absolutely no problems.

A few weeks ago, my boss contacted Minisoft to try out Netprint, which is
almost a clone of ESPUL (made conversion easy!).  It working like a charm -
for five days, then began to exhibit the same problems.  Often it will 'fix'
itself sometime in the morning - or we can shutdown the job and restream it
- and it works again.

It has never botched up at anytime other than the morning.  It will often
run for weeks with minimal problems (2-3 reports without env files), then
'go bad' for several weeks where almost every morning report is bad.

This falls under the category of 'real mystery', leaving all of us involved
scratching our heads.  Rich has thought possible MPE bug.  I've thought it
possible to be a compiler bug.  But why does it come and go?  Hmmm.

Anyway, as we fill the paper recycle bins again, I thought I'd post and see
what thoughts, ideas and suggestions may come from the list.  Any thoughts
where to look next?

Steve Miller
Portion Pac, Inc
[log in to unmask]