Date: Tue, 1 Jun 1999 18:31:47 -0400 (EDT) From: Chuck Robey <chuckr@picnic.mat.net> To: Nik Clayton <nik@nothing-going-on.demon.co.uk> Cc: hackers@FreeBSD.ORG Subject: Re: Kernel config script Message-ID: <Pine.BSF.4.10.9906011809150.82061-100000@picnic.mat.net> In-Reply-To: <19990601185323.B73490@catkin.nothing-going-on.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 1 Jun 1999, Nik Clayton wrote: > On Sun, May 30, 1999 at 11:21:57PM -0400, Chuck Robey wrote: > > You guys should be aware that work is going on to change, in a rather > > major way, not just the config file, not just the configuration method, > > but the entire way that devices are detected and drivers added. > > Is this documented anywhere? Not the fact that things are going to change, > but what the user visible component of that change is? I'd hate for the > Handbook et al to suddenly be seriously out of date when a new config > mechanism is upon us. I don't think so, but I wouldn't press it right now. I'm answering this privately, because there is some controversy, and some really hurt feelings, over some of it. You see, some folks in Japan went off on their own and developed a "newconfig", which has a lot fo things in common with the work that Peter's gone and done, but also has some basic differences. They worked pretty much totally in silence. so when core told Peter to go ahead, the Japanese group *finally* opened their mouth, and there have been a lot of very hurt feelings, because of the large amount of work that was expended, and now is going to be lost. What was at the bottom of it was the idea of _communications_ itself, which Peter is/was doing, and the Japanese were not at all. They were just off on their own, assuming that they had a clear field, and that anything they developed would be accepted. There was no debate over features at all, which very correctly scared core. The thing that's finally been accepted is a pretty complete rewrite of the bus code, so things like devices that suddenly show up, like for plug in cards in portable machines, can announce themselves and get loaded dynamically. As much as possible, the entire idea of needing a config file is going to go away. Some of this is because of the success of the kld kernel loadable modules. There is more work on loadable modules, for dependency checking/loading, and better links between module file names, and module code names. This is a very fluid thing, and development sometimes takes a step backwards. Things are changing very quickly, which means that right now is an incredibly bad time to decide to make automated config tools, because the problem that automated config tools are intended to solve is going to be completely elminated, not just eased. Like I said, what's really needed now is a weekly summary of development of -current and -committers. In odoing it, tho, we *don't* want to add some huge extra mail load on Peter, Doug, Daniel, Mike, and others working towards all this. Many people who hang around hackers and are completely clueless, want to force folks to do it their way, because they won't spend the time to follow the lists and find out that *their* way is now obsolete. It gets complicated, doesn't it? ----------------------------+----------------------------------------------- Chuck Robey | Interests include any kind of voice or data chuckr@picnic.mat.net | communications topic, C programming, and Unix. 213 Lakeside Drive Apt T-1 | Greenbelt, MD 20770 | I run picnic (FreeBSD-current) (301) 220-2114 | and jaunt (Solaris7). ----------------------------+----------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.10.9906011809150.82061-100000>