From owner-freebsd-ports Mon Mar 20 00:38:21 1995 Return-Path: ports-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id AAA13219 for ports-outgoing; Mon, 20 Mar 1995 00:38:21 -0800 Received: from LOCALHOST (LOCALHOST [127.0.0.1]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id AAA13202; Mon, 20 Mar 1995 00:38:13 -0800 X-Authentication-Warning: freefall.cdrom.com: Host LOCALHOST didn't use HELO protocol To: Paul Traina cc: nate@sneezy.sri.com (Nate Williams), Steven Wallace , CVS-commiters@freefall.cdrom.com, freebsd-ports@FreeBSD.org Subject: Re: cvs commit: src/gnu/usr.bin/ld shlib.c In-reply-to: Your message of "Sun, 19 Mar 95 21:48:08 PST." <199503200548.VAA02959@precipice.Shockwave.COM> Date: Mon, 20 Mar 1995 00:38:12 -0800 Message-ID: <13201.795688692@freefall.cdrom.com> From: "Jordan K. Hubbard" Sender: ports-owner@FreeBSD.org Precedence: bulk If it's truly a gcc "feature" to link in /usr/local/lib then I would have to say, with 15 years experience in porting software across extremely disparate UN*X platforms, that this is something that gcc got entirely wrong. The only thing /usr/local/lib is guaranteed to be is utterly unreliable on any given machine (I have no control over it) and the minute you start pointing off into the void in hopes that something good will happen (and doing it behind the user's back, so it's not even immediately obvious when things go badly wrong!), you're setting yourself up for a fall. gcc was broken, Nate has fixed it. Whomever decided to point into an essentially randomly populated directory by default in gcc should be shot and skinned. It was a stupid decision. Period. Jordan