urban@rand.org (Mike Urban) (07/06/90)
From: urban@rand.org (Mike Urban) What incipient or existing standards, if any, specify the Shell-level printer interface (lpr and its friends)? -- Mike Urban urban@rand.ORG Volume-Number: Volume 20, Number 103
domo@tsa.co.uk (Dominic Dunlop) (07/07/90)
From: Dominic Dunlop <domo@tsa.co.uk> In article <789@longway.TIC.COM> urban@rand.org (Mike Urban) writes: >What incipient or existing standards, if any, specify >the Shell-level printer interface (lpr and its friends)? Aha. That terminal ``r'' tells me you're a Berkeley-ite. Which is too bad, I'm afraid. The draft 1003.2 shell and tools standard specifies a pretty-much emasculated version of the System V.2 (or thereabouts) lp print spooling command (which some might argue was fairly impotent in the first place). As far as ``friends'' go (lpadmin, enable, lpshut and so on in the case of System V), none are specified. This is seen as 1003.7 (administration) territory. If you can get hold of a draft (somebody in Rand must have one), you'll find lots of extended description about how little a conforming application can assume from lp, and rationale about how it should be possible to implement it as a shell script, namely cat > /dev/lp The X/Open Portability Guide, edition 3, volume 1, does little better: it describes a few more options, and the cancel utility, but they're all optional -- in effect, all an XPG-conformant system has to offer is the interface described in 1003.2. All rather depressing really. -- Dominic Dunlop Volume-Number: Volume 20, Number 111
jgreely@cis.ohio-state.edu (J Greely) (07/08/90)
From: jgreely@cis.ohio-state.edu (J Greely) In article <797@longway.TIC.COM> domo@tsa.co.uk (Dominic Dunlop) writes: [discussion of POSIX specifying only minimal print standard, bare-bones SysV lp] >All rather depressing really. Actually, I find it encouraging. It means that the world won't standardize on either SysV *or* Berkeley print spooling, which means that there's room for someone to write something good. Lpr and company are a bitch for a large network, and I wouldn't even *try* to run lp on 300 machines. What's needed is a real batch system that scales to a large network (and please, Ghod, make it administerable by a mortal). -- J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely) Volume-Number: Volume 20, Number 112