How do I ensure job order

[Petrus Rossouw]

BTW, I looked at the scheduler's code and can not find a way to configure the system to do this.
The problem with our environment is that for some reason the LPD backend keeps on crapping out when it sends commands to the remote print servers.
This in turns tells the scheduler to stop the job and printer.
If I set the error policy to retry, then the scheduler puts the job on hold and will merrily continue with the next job.

I don't want it to continue with the next job, and I don't want it to cancel the first job after X retries.
This is how old LPD deamons used to work, and our applications expect this behaviour.

Can we please get an enhancement in CUPS which adds an error policy which retries the job, but does not hop on to the next job?
The retry interval is still ok, but I should be able to set the retry limit to infinate (maybe -1 as a value).