Re: AROS and patches to the binutils

[Fabio Alemagna <falemagn at studenti dot unina dot it>]

On 17 Jan 2003, Nick Clifton wrote:

> Hi Fabio,
>
> > I was wondering whether would it be better, both for you and me, to make
> > little, incremental patches to the codebase, or to finish what I have in
> > mind and then make a big patch.
>
> Another point to bear in mind with the small-incremental patch
> approach is that you must not break the ability to build and run the
> binutils configured with the "--enable-targets=all" option.  So if you
> add AROS support to the configuration files, you must also be able
> build an AROS targeted toolchain.  (It does not have to work, but it
> does have to build).

As I wrote in my email, everything compiles fine, and the linker produces
executables runnable by AROS, however there's nothing in it of what I have
in mind to implement now, it's basicaly the same as the linux counterpart,
and you still have to use the -r option to build "executables".

So do you say that that's not enough? Should I at least reach a state in
which executables are really executables (in the ELF sense)? If so, I
could do it easily by implementing the 1st approach I described in one of
my past emails (making the -q option always active, so that relocations
are included in the executable), but is that an approach that you like and
that would make you accept the patch?

Otherwise, should I go directly for the 3rd approach I described and only
then submit the patch? Would that be accepted?

Fabio Alemagna