Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!pasteur!sim!scole From: scole@sim.uucp (Steven Curtis Cole) Newsgroups: comp.binaries.ibm.pc.d Subject: Re: Help with DVIEPS (Again, Desperate!) Message-ID: <21995@pasteur.Berkeley.EDU> Date: 12 Feb 90 01:59:56 GMT References: <90Feb4.120923est.4589@neat.cs.toronto.edu> Sender: news@pasteur.Berkeley.EDU Reply-To: scole@sim.UUCP (Steven Curtis Cole) Distribution: na Organization: University of California, Berkeley Lines: 31 Noticing a few messages about DVIEPS here, I'd though I'd tell you all what I've managed to accomplish with it. I got the binary in the DOSTEX distribution, and noticed memory problems as well as a few other bugs having to do with stray dots over the page. Disatisfied with its performance, I grabbed the source for the DVI drivers and recompiled with my own changes. What happens in my version: To keep memory use down, I scan each page three times, storing and printing one third of the page on each pass. The old DVIEPS I had used only two passes. Unfortunately, this increases execution time by 50%. The new version is also written for a generic IBM graphics printer, not whatever printer it was originally written for. Whatever the noise problem was disappeared in the new version. Unfortunately, the changes aren't ifdef'd very well, and I didn't feel that my changes should be submitted to the people who worry about the source for TeX. I'd be happy to contribute the source to someone who can fix it up and get it to the right people; I don't have the time now to properly take care of it. (Such is the life of a graduate student...) Note that this thing is monsterously slow. It's faster for me to upload my dvi files, and then drive to campus to use a laser printer to produce output than it is to use my poor EPSON... If you've got capability like this, spend your time getting that to work and just forget about DVIEPS... (Of course, I'm also running an old 4.77 MHz XT...) Steve Cole scole@sim.berkeley.edu