[cups] Custom backend with fork()

[Helge Blischke]

Some time ago I had lots of experience with high end production printers
which have plenty of disk storage to buffer jobs and provide more than one
printing engine which can work in parallel. 
My experience was that – provided the data link is sufficiently fast
(e.g. LAN) – CUPS is getting rid of the jobs fast enough so that there is
no real need for a CUPS backend to process several jobs in parallel.

Helge

> Am 29.04.2015 um 15:12 schrieb Johannes Meixner <jsmeix at suse.de>:
> 
> 
> Hello,
> 
> On Apr 29 05:32 Michael Sweet wrote (excerpt):
>> support processing multiple jobs destined for the same
>> printer in parallel
> 
> As far as I understand it "processing multiple jobs in parallel"
> usually means that the filtering runs in parallel but not that
> several backends send data simultaneously to one same printer device.
> 
> E.g. see
> the "parallel printing / queues possible?" mail thread on
> http://cups.org/pipermail/cups/2012-November/thread.html
> and
> http://www.cups.org/str.php?L1382
> and
> http://www.cups.org/str.php?L4085
> 
> 
> But technically high-end network printer devices could be able
> to receive data simultaneously via several network connections
> so that it could make sense to run several backends in parallel
> for one same printer device.
> 
> Several years ago I tested once how a usual network printer behaves
> when it gets data simultaneously via several network connections...
> ( wasn't much fun ;-)
> 
> 
> Kind Regards
> Johannes Meixner
> -- 
> SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard, Jennifer Guild,
> Dilip Upmanyu, Graham Norton - HRB 21284 (AG Nuernberg)
> 
> _______________________________________________
> cups mailing list
> cups at cups.org
> https://www.cups.org/mailman/listinfo/cups