This is the mail archive of the
mailing list for the GCC project.
Re: Updates to --with-sysroot support
- From: Daniel Jacobowitz <drow at mvista dot com>
- To: Zack Weinberg <zack at codesourcery dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Alexandre Oliva <aoliva at redhat dot com>,Geoff Keating <geoffk at geoffk dot org>
- Date: Tue, 5 Nov 2002 09:31:05 -0500
- Subject: Re: Updates to --with-sysroot support
- References: <20021104165729.GA9266@nevyn.them.org> <20021105072939.GC10353@egil.codesourcery.com>
On Mon, Nov 04, 2002 at 11:29:39PM -0800, Zack Weinberg wrote:
> On Mon, Nov 04, 2002 at 11:57:29AM -0500, Daniel Jacobowitz wrote:
> > I spent a couple days working with this and finally found a solution for
> > (most of) my concerns. This patch:
> > - Fixes an undocumented behavior in cppinit.c which I consider to be a bug.
> > If -iprefix was used to change the compiler prefix, then first the
> > standard include directories which begin with the prefix were searched
> > (using the relocated prefix); then all directories were searched in order
> > (using the original prefix). This means that -iprefix'd versions of
> > GCC_INCLUDE_DIR (that's -iprefix's prefix followed by "include", folks)
> > would be searched before GPLUSPLUS_INCLUDE_DIR. Oops.
> > I changed this so that the order is preserved, and unrelocated versions
> > are not searched. This means that people who were using -iprefix and
> > still expecting to get GCC's <stddef.h> et al. from the normal location
> > are out of luck; they'll have to specify that path as an -isystem
> > explicitly. I think this change is correct but I'd like a second
> > opinion.
> > - No longer defines PREFIX_INCLUDE_DIR for a sysrooted compiler. This
> > one's a little more arguable but I think it's appropriate. The sysrooted
> > compiler should work like a system compiler inside of $sysroot; searching
> > $prefix/include defeats that.
> > I need this in my local setups but I won't be heartbroken if it's
> > considered wrong for the FSF tree :)
> Just for the record, I like the sound of both these changes.
> Am I correct in thinking that for a normal Unix native compiler, you
> could set sysroot to "/usr" and get the expected behavior? Or is it
> prepended to $prefix? Or what?
Well, right now I don't think it'll work correctly on a native
compiler. I haven't tested it; I think the new bits will but some of
Alex's earlier bits won't, since neither of us was really designing for
that case. The theory says you would set sysroot to "". It is
prepended to things like STANDARD_INCLUDE_DIR, LOCAL_INCLUDE_DIR,
-rpath-link /lib, etc.
MontaVista Software Debian GNU/Linux Developer