On Tue, Nov 26, 2002 at 08:54:29PM -0500, Michael Stone wrote: > On Tue, Nov 26, 2002 at 05:07:51PM -0700, Joel Baker wrote: > >In the origional message, I merely pointed out that keeping such things > >properly encapsulated is crucial, if you EVER want to be able to run on any > >other kernel. > > Which original message? The one I saw said "Certainly it will have a > hard time working on any of the BSDs anytime soon, if it relies on devfs > more than trivially" and "Use of /proc should also, prefferably, be > limited to traditional /proc items and not the Linux view." You didn't > say anything about encapsulating them, or using them only where > appropriate, you just said to avoid them. I suppose it boils down to what you consider "relies on" to mean. If things are hidden behind a module that can easily be replaced so that it never touches devfs, then I don't consider it "rely on" devfs. I said to avoid *relying on* them, not to avoid *using* them when possible. > >That's all I'm asking for - careful API design, that tries very hard to > > That's what you're asking for now, and it doesn't seem nearly as > controversial as what you asked for the first time. (Seems pretty close > to what I said when I suggested you'd have to plug in some > kernel-specific code for certain functions.) Which came later. But I suggest that, at this point, we write it off as a failure of communication; it would appear that we both want more or less the same result, and just picked different words for it. Which is unfortunate, but it happens. Sorry. -- *************************************************************************** Joel Baker System Administrator - lightbearer.com lucifer@lightbearer.com http://users.lightbearer.com/lucifer/
Attachment:
pgpRDohKikedw.pgp
Description: PGP signature