Path: utzoo!attcan!uunet!amdahl!apple!bloom-beacon!tut.cis.ohio-state.edu!osu-cis!att!mtunb!jcm From: jcm@mtunb.ATT.COM (was-John McMillan) Newsgroups: unix-pc.general Subject: Re: UNIX-PX Printer Setup Keywords: 3B1 windows Message-ID: <1441@mtunb.ATT.COM> Date: 20 Mar 89 15:54:12 GMT Organization: AT&T ISL Middletown NJ USA Lines: 47 In article <721@kosman.UUCP> kevin@kosman.UUCP (Root) writes: ... >Now, the followups are confusing me a bit. If lots of printers are supported, >I expect the format conversions to be in a program somewhere, but I cannot >find it. > >Continuing <1435@mtunb.ATT.COM> jcm@mtunb.UUCP (was-John McMillan) writes: >> >>The KERNEL only sends the bits out (ref: 'wd' ioctl): there is a >>screendump program that is responible for translating it to the printer >>-- and therein's your problem -- at least the software one! Quoth the User's Manual: ioctl( wd, WIOCREAD, &pixmap) This "ioctl" causes the pixel image of the entire display to be "dumped" into the memory at "pixmap". "Pixmap" should be 15660 unsigned shorts arranged as 348 rows each containing 45 unsigned shorts. The least significant bit of the first short in the array contains the upper-left-hand display pixel. In other words, bits across the scan-line are taken sequentially from shorts in the respective row (of the 348 rows), with sequential bits- in-a-short taken from the least significant end. >Are you sure? There's no program by that name (or any other unknown progams >with names ending in 'dump') on my system. Also, I can't figure out how >to look up a 'wd' ioctl. I do see ioctl(wd,WIOCREAD,&pixmap) but that >doesn't tell me much. Since it seems to tell ME everything about how to GET the data, I'm failing to understand YOUR delemma, and leave it to others to help. >>You can generate image FILES -- since I do NOT do this, I'll defer to >>others to explain how -- and drive the printer from your own >>bitmap-to-printer converter. > >I can see how to do this if I'm writing the package, but the data flow >for the Shift-Print screen dump still looks like a kernel thing to me. ? "Data flow"? You request the data with the IOCTL. Or you use the existent routines to build a FILE of that data. "A kernel thing"? >Does anyone have better information? Translator, please !-) jc mcmillan --att!mtunb!jcm