On 17 Feb 2016, at 14:07, Chris Bennett wrote: > On Wed, Feb 17, 2016 at 07:51:28PM +0100, Tobias Ulmer wrote: >> >> The only thing wrong with lpd is nobody tedu'ed it yet. >> >> No really, it is outdated beyond rescue. If you want to write a new >> print job queueing system, sure, have fun. Maybe you can come up with >> a 'cups' that doesn't suck? >> > > Oh, I agree, it's seriously ready for display in the computer museum. > But it does work in a limited fashion. > I want something in base that handles printing. > I need a project that isn't at the skill level of pouring over code > and finding subtle errors. Way over my head to do that. > Wow all of you developers are good at that! > > I understand that this is going to take a long time to accomplish. > It will need to fill a large group of users needs or it won't be > kept in base.
In the "For what it's worth" department: Here at RPI we use the BSD-based lpr/lpd to run our print servers. Our print servers run a few hundred printers for campus, including five pretty modern plotters which see heavy use at the end of each semester. We go through a mile or two of plotter paper every semester. I'm the guy who implemented most of our printing system. To touch upon the original subject of this thread, we have not used any line printers at RPI in at least 20 years. When we did have them, they were connected to IBM mainframes, not unix print servers. I use what is basically FreeBSD's lpr/lpd running on a version of solaris from the days back when Sun was still in business. I've done enough work on lpr/lpd that I worked my way into being the maintainer of lpr/lpd for FreeBSD (although I have not done much work on it lately). I made many improvements to FreeBSD's lpr/lpd from 2000 to about 2006. In some cases I pulled in changes from OpenBSD's lpr/lpd. And the versions of lpr/lpd that I use on our print servers at RPI have some additional improvements that I keep meaning to move over to FreeBSD, however the source for RPI's lpr/lpd is also quite a mess. I need to clean it up and make sense of the changes. I do not know how many of my FreeBSD changes have made it over to OpenBSD's lpr/lpd. Over the years I have tried out both CUPS and LPRng, and both times I ran into enough pain that I've stuck with plain BSD-based lpr/lpd. There is much room for improvement in BSD's lpr/lpd, but there's also a lot of good knowledge which is not obvious and not documented. My main example is that almost every non-BSD alternative will send a job's control-file before sending the data file(s). If you are sending to a busy print server it is actually important to send the data-files first, but the reason *why* you want to do that is not documented anywhere. It's only when you run a busy print server that you realize why some things were done the way they were done. I've been keeping an eye on some of your recent changes to OpenBSD, but I haven't had the time to look them over as much as I would like to. That's been the story of my life for the last 4-5 years... -- Garance Alistair Drosehn = dro...@rpi.edu Senior Systems Programmer or g...@freebsd.org Rensselaer Polytechnic Institute; Troy, NY; USA
