os9@nyit.UUCP (OS9 news) (02/14/86)
OS-9 Digest Number 5, 13-February-1986 Sorry about the two weeks silence on this newsgroup - we're having some communications problems. -bp Contents: 1. CoCo OS-9 2.00.00 Jim Jones 2. Reply to Jim O's posting about Termcaps Bob Larson 3. Another reply to Jim O's posting utfyzx!rph 4. OS-9 for the QL? Jim Jones 5. The format of this newsgroup Bruce Perens 6. Pipes and get/setstat Jim Jones --------------------------------------------------------------------- From: uokvax!emjej (Jim Jones) To: philabs!nyit!os9 Subject: CoCo OS-9 2.00.00 I received the upgrade to 2.00.00 the other day; the following are items and changes that caught my eye. [moderator's note: Radio Shack part number 700-2331, OS-9 Version 02.00.00, $25.00, bring your old OS-9 kit or receipt. -bp] 1. The upgrade includes two diskettes: one with the new OS, and one with the old chestnut BOOT program (for early disk controllers) and the CONFIG program. Interesting item: judging by the docs, the BOOT/ CONFIG disk is a (CoCo) OS-9 disk, because they mention a certain directory that lives on it. This directory, MODULES, has the various device drivers and descriptors that are known, so it looks like you can indeed set up your own non-RS stuff to be seen by CONFIG (which, BTW, is a menu-driven program for generating your own bootable OS-9 disk). 2. There's a (primitive, to be sure, but better than nothing and suitably small) help program. Haven't seen the innards of the help files, so one might be able to put in more extensive stuff--but I don't know for sure. 3. 2.00.00 has inherited the OS-9/68000 "iniz" command for attaching devices. (deiniz? I didn't see it.) 4. tmode/xmode and get/setstat have new knobs to twist. 5. Some utilities look at those knobs to decide how to format their output (huzzah). 6. The Section Two toc entry mentions the addition of networking, but the real Sect. 2 is mum. What's up, Tandy? 7. VIRQ, as has been mentioned before. 8. Some of the SCF device drivers know about hangups, and DTR (huzzah!). Speaking of those added knobs for SCF devices--some of those "reserved for future expansion" look kinda neat, for example, V.KANJI (kanji are Chinese ideographs, as referred to by Japanese). What's V.KBUF and V.MODAPR? Your guess is at least as good as mine. James "inquiring minds want to know" Jones To: nyit!os9 Subject: Re: mod.os.os9 digest #1 (Jim O's article) From: sdcrdcf!blarson (Bob Larson) >From: lsuc!jimomura Tue Jan 14 14:23:58 1986 > 1. Termcaps -- Is there any established practice as to where "termcaps" > files go in OS-9? I'm personally leaning towards sticking it in > the "SYS" directory. Os9/68000 put the TERMSET file in /dd/SYS. This file has some termial desciption, but has many assumptions that arn't always true and only uses a subset of the capabilities described in termcap. I don't know of any available routines to parse the TERMSET file. I think as a user community we should support a file termcap file in SYS. > > 3. While discussing Modula-2 on BIX, I came up with the idea of > creating a common-block handling system for OS-9. The modules > would be a 'cbman' (common block manager) and 'cb' device > descriptor. This would be an optional set of modules used > for languages which could support common blocks for data > structure sharing between free-standing modules. Has anyone > else considered this? There seem to be two ways of handling shared memory now. 1. Linking to a module and using it as shared memory. Does this work on level 2 systems? 2. Ramdisk. I don't see that your proposal has much of an advantage over it, but I am willing to listen. > 4. LANS -- Has anyone done any work regarding LANS yet? Yes, a japanese company developed a good networking package with microware's support. This should be available from microware eventually. It was demonstrated at the os9 seminar. Using this you can access any device on a remote system, not just disks. Different network hardware can be handled by different device drivers and descripters. (netman is a new file manager.) > Cheers! -- Jim O. Bob Larson ihnp4!sdcrdcf!blarson Blarson@Usc-Ecl.Arpa From: utfyzx!rph Subject: Termcaps on SuperOS9 Keywords: 6809, PET, SuperOS9, Termcaps Since Jim Omura raised the point of termcaps for OS-9, I'll also take this opportunity to point out that SuperOS9 has a cursory one (pun not completely intended) for use by a full screen editor. At the moment it is rather brief, as it consists of only the following: SP9000:$00$15XY$00:$02:$04:$0C:$06:$03:$12:$13: and resides in /D0/SYS/termset. The entries mean COMPUTER:DirectCursor(Col,Row):DeleteLine:DeleteChar:ClearScreen: ClearToEndOfLine:InsertLine:RvsOn:RvsOff though I don't know what the extra $00 in the first entry are. Presumably they must be flags since the terminal driver never needs padding. I would have preferred some letter codes, but in any case /D0/SYS is certainly the place for it, and termset is as good a name as any. Subject: OS-9 for the QL--Query Date: 27 Jan 86 16:17:40 CST (Mon) From: uokvax!emjej (Jim Jones) Keywords: QL, Sinclair Has anyone ported OS-9 to the Sinclair QL? (I know there's some OS-9 activity in England, because the folks who did the port to the Emerald Computers SBC are from there, according to a flyer from Emerald I read a while back.) That would be a nice low-cost beastie to have running OS-9. James Jones From: nyit!bp (Bruce Perens) Subject: format of this newsgroup I am going to try dropping the `digest' format of this newsgroup for a while. I'll relay individual messages as they come in, instead of waiting for a clump of four or five to make up a digest. That should make the group more interactive. Bruce P. Subject: pipes and get/setstat Date: 27 Jan 86 15:20:51 CST (Mon) From: uokvax!emjej (Jim Jones) Does anyone know what get/setstat options are supported for pipes? I'd like very much for there to be one that tells me whether data are present, for the following reason: I want to extend an editor I've been working on, to add some commands of the sort one sees in ex, i.e. <range>w !command and <range>r !command. I'd also like to do <range> w | command (! was taken...maybe I should change that...nah, too many people are used to !) which would write the range as stdin to a process running the command, delete the lines written, and replace them with stdout from the command. Ideally, I should do this without having to stoop to temporary files--that's what pipes are for, right? However, the fixed capacity of pipes means I can get into trouble. Consider the following: Some filters are like sort--they have to read all their input before they can write any output, so any method I use has to be able to write all of the text out before reading, or it will block forever. On the other hand, some filters put text out at about the same rate they take it in. The pipe leading from the filter might clog if one tried to write out all the lines being filtered, and after that the pipe to the filter. (What was that 10cc song..."deadlock holiday" or something like that...:-) The only way I can figure to do the job right is to check the pipe coming into the editor (not read, which could block--just check for the presence of data) after each byte written to the pipe leading out of the editor. SCF devices--at least some of them--support a getstat call that tells whether there are any bytes waiting to come in. Do pipes? If not, why not? How easy would it be to modify the driver? the file manager? to allow such a request if it's not allowed currently? James Jones -- --- 516-686-7644 (Bruce Perens or Alex Arthur) {allegra,seismo,decvax,vax135,ihnp4,mcvax}!philabs!nyit!os9 nyit!os9%suny-sb.CSnet@CSnet-Relay.ARPA OS-9 and BASIC09 are trademarks of Microware Systems Corporation and Motorola. mod.os.os9 is a personal (not an NYIT) project.