From owner-freebsd-hackers Thu Nov 21 14:46:08 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id OAA22892 for hackers-outgoing; Thu, 21 Nov 1996 14:46:08 -0800 (PST) Received: from rocky.mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id OAA22867 for ; Thu, 21 Nov 1996 14:45:59 -0800 (PST) Received: (from nate@localhost) by rocky.mt.sri.com (8.7.5/8.7.3) id PAA12448; Thu, 21 Nov 1996 15:45:52 -0700 (MST) Date: Thu, 21 Nov 1996 15:45:52 -0700 (MST) Message-Id: <199611212245.PAA12448@rocky.mt.sri.com> From: Nate Williams To: Richard Wackerbarth Cc: Nate Williams , Terry Lambert , hackers@freebsd.org Subject: Re: Who needs Perl? We do! In-Reply-To: References: <199611212046.NAA13887@phaeton.artisoft.com> <199611211714.SAA01528@ravenock.cybercity.dk> <199611212141.OAA12035@rocky.mt.sri.com> Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Richard Wackerbarth writes: > >Richard was completely free to do what he wanted to do, but he wasn't > >going to get the 'blessing' of anyone until he had a working prototype > >that was at least as good as the current system. > > As I have pointed out a number of times, as I understand your "rules", > it is not reasonable to even attempt this. And that's where we disagree. > There are EXTENSIVE changes which must be made. If I am shooting at a > moving target, I'll never be able to catch up. The 'target' doeesn't keep moving. We have provided almost *NO* 'new' features over the 4.4 stuff other than making it work like someone expected it to, plus add the ports stuff. New stuff: 1) Working shlib support 2) Working R/O source tree (obj links and such) 3) Ports 4) Better .depend handling Since your new solution is better it shouldn't rely on the hacks and fixes we've needed to do to get things working 'the way we want' (although I disagree with some of the changes, majority rules). You need to provide (or at least show) that you can provide the same functionality with a 'better' system, or we aren't going anywhere. Heck, show me a working prototype that has 4 levels. Top level / \ Subdir1 Subdir2 / / \ P1 P2 ssubdir3 / \ P3 P4 Proof that you're way is 'better' is all that we're asking for. Can you fix the build environment so that it's easier than the current system, and that doesn't require any of the kludges the existing system needs (pre-built binaries), etc.. Make some dependencies in your tree, heck build something that requires flex, and have flex be one of the tools in the tree. Nate