Path: utzoo!utgpu!watserv1!ria!csd.uwo.ca!webber From: webber@csd.uwo.ca (Robert E. Webber) Newsgroups: comp.os.minix Subject: Re: Various & Sundry questions....FLAMMABLE! Message-ID: <2573@ria.ccs.uwo.ca> Date: 27 Mar 91 05:20:27 GMT References: <1991Mar24.235828.45888@vaxb.acs.unt.edu> Sender: news@ria.ccs.uwo.ca Organization: see disclaimer Lines: 44 In article <1991Mar24.235828.45888@vaxb.acs.unt.edu> kenc@vaxb.acs.unt.edu writes: . o Is there a source to FTP manpages that work with Minix? I can get them . from the unix box at work, and (sort of) read them... If your machine at work doesn't run Version 7 unix, this may be a mixed blessing as well as a copyright problem I won't even think about touching. There is a collection of stripped-down man pages created from scratch that have been posted from time to time in this group and are supposed to be somewhere on the plains.nodak.edu archive server. Personally I think docummenting the present system is more important than modifying it, but with the book available offline and the source available online, I doubt that many people see a need for the equivalent of an online version of the book melded with traditional-style man pages. By the way, the existing man pages don't use nroff but instead come preformatted (a horrible idea that has become quite popular on System V based systems that is motivated by the slowness and historic unreliability of available nroff clones). . o A few times today I've run into mentin of job control. What is this . mysterious brew? Please correct me if I'm wrong, but doesn't it . basically mean that a 'suspended' process just doesn't get any CPU . until it's revived? Is it currently in development? Job control doesn't play a significant role in the BOOK since it is focussed on V7 ideas. Job control is essentially an approach to virtual terminals that was ported from TENEX to BSD (along with a number of other things). Thus the critical aspect of it is the ability to switch terminal control between multiple processes. Thus, it has been frequently claimed that window systems remove the necessity for traditional job control which mucked up the terminal i/o. I don't see much mention of it in my '86 POSIX draft, but the 1989 The Design and Implementation of the 4.3BSD UNIX Operating System (Leffler, McKusick, Karels, and Quarterman,, Addison- Wesley - no ISBN in my copy) indicates that POSIX job control essentially means the BSD model - so, if you are interested, it would be worthwhile tracking down both the Leffler et al book and a current version of the POSIX specs (since Minix-developers seem to currently view posixization and ansifying as progress). Given virtual consoles already in many Minixes, the ability to remove a job from the scheduling queue via nice would be useful. One of the neater things I stumbled upon in SunOS is the ability to pause a process, get a core dump of it, and then continue the process on its way. I suspect it will be less controversial if it is proposed as enhancements to the scheduler rather than called `job control.' --- BOB (webber@csd.uwo.ca)