This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Proposal: Require that GAS be built before GCC in a unified tree


On Sat, Mar 01, 2003 at 07:38:23AM -0600, Michael S. Zick wrote:
> On Friday 28 February 2003 11:11 pm, Daniel Jacobowitz wrote:
> > On Sat, Mar 01, 2003 at 12:09:32AM -0500, Nathanael Nerode wrote:
> > > The point of this proposal is that GCC's configure.in has a ludicrous
> > > number of tests for assembler features which split into two: one if
> > > there's an actual assembler to test, one if it hasn't been built yet
> > > (which does hairy version-number examination).
> > >
> > > If we required that GAS build before GCC in a unified tree, we could
> > > eliminate that half of the code, and just use the standard tests on the
> > > newly built gas.  It would be a great savings in clarity.
> > >
> > > What do people think?
> >
> > Don't you need the version number grubbing anyway for a host-x-host
> > build, if you can't run the newly built GAS?
> >
> Build both the current-machine and the cross-machine binutils first.
> Mike

Eh, that's not reasonable.  For one thing there's no reason gas needs
to support its host.  For another it's lots of extra building.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]