From owner-freebsd-questions Fri Nov 17 09:46:58 1995 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id JAA18069 for questions-outgoing; Fri, 17 Nov 1995 09:46:58 -0800 Received: from austin.polstra.com (austin.polstra.com [198.211.214.10]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id JAA18064 ; Fri, 17 Nov 1995 09:46:47 -0800 Received: (from jdp@localhost) by austin.polstra.com (8.6.12/8.6.12) id JAA21432; Fri, 17 Nov 1995 09:46:19 -0800 Date: Fri, 17 Nov 1995 09:46:19 -0800 From: John Polstra Message-Id: <199511171746.JAA21432@austin.polstra.com> To: chuckr@glue.umd.edu Subject: Re: Shared Libraries...almost there... In-Reply-To: Organization: Polstra & Co., Seattle, WA Cc: freebsd-questions@FreeBSD.org, freebsd-hackers@FreeBSD.org Sender: owner-questions@FreeBSD.org Precedence: bulk In article chuckr@glue.umd.edu writes: > You know there's two version of that pic flag stuff, which stands for > Position Independent Code (I think). Yes, that's right. > The -fPIC is supposed to be > stronger in force than the -fpic, but most code does not need the strong > -fPIC. I don't know how to tell the difference, except to try it and > find out. Actually, for the i386, there is no difference between the two. They're exactly the same. On some machines (Motrola 68K architecture, for example), "-fpic" generates certain address offsets with 16-bit fields, whereas "-fPIC" uses 32-bit fields. The first form is OK for most programs, but really large programs overflow the 16-bit fields and require "-fPIC". -- John Polstra jdp@polstra.com John D. Polstra & Co., Inc. Seattle, Washington USA "Self-knowledge is always bad news." -- John Barth