job holding

Bill Wyatt anonymous at easysw.com
Tue Aug 10 08:44:05 PDT 2004


I've spent the last week or so going through the Cups docs and source,
and find that there appears to be no facility for rescheduling a job
upon error (except FAX jobs).

I'm trying to write my own backend, which I did under Solaris's print
system with few problems. I can get it working fine - when it works.
However, when an error is encountered (e.g. a timeout), I want to exit
the job, and have the system reschedule some time later, presumeably
after emailing a message to an operator. Cups currently simply dequeues
the job.

Cups seems to not have an error strategy - apparently every backend
has to manage its own strategy for sleeping, retrying, etc. And, although the cupsd.conf  file has a "ServerAdmin" field for an email field,
*nowhere* is it used, other than to fill it in from the .conf file.

Am I right in these statements? Is there no rescheduling policy? use of
email notifications?

It seems to me the Cups system docs are fine when everything goes right,
but are very poor at explaining what to do when errors crop up.





More information about the cups mailing list