Subject: Re: [netatalk-admins] PAP end-of-job timeouts?
From: Frank Morton (fmorton@base2inc.com)
Date: Wed Mar 03 1999 - 12:07:29 EST
>> In this case, how does pap know then the job is completed?
>> And say an error message is generated right at the end of printing,
>> will pap wait for it?
>
>It doesn't know, and the error is lost. This is why pap implements the
>protocol correctly. At some point, Apple (or Adobe, probably) changed
>the Laserwriter driver to not wait for the printer's EOF. This made
>Mac printing suddenly faster, since a job could be loaded in the
>printer's memory, long before the engine ripped it. Printers that came
>out after this change (particularly those using Abode's PAP
>implementation) would sometimes get confused and never send EOF. Very
>frustrating, as this was when just about the time we were adding
>accounting to netatalk. In any case, we changed the default in psf to
>reduce the number of "why does lpd hang" messages we get to the list.
Would there be any elegant way to act like my kludge that waits
until the printer goes %%idle? I understand that this might not
be complete, as some other source might be sending the next
job, thereby hanging up a completed job until another is complete
waiting for an idle state. But, in my case, as I suspect many others
do also, most of our printers are accessed exclusively through
a spooling system, so this conflict never exists.
I'd love to see something like this so I can trash my kludge.
It has been working for two years, but it pains me to know
how it is working.
Thanks.
This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:16:23 EST