On Sun, 2006-06-25 at 18:25 +0200, Erwin Rol wrote:
On Sun, 2006-06-25 at 17:31 +0200, Ralf Corsepius wrote:
> On Sun, 2006-06-25 at 15:53 +0200, Erwin Rol wrote:
> > On Sun, 2006-06-25 at 15:42 +0200, Ralf Corsepius wrote:
> > > On Sun, 2006-06-25 at 11:08 +0200, Erwin Rol wrote:
>
> > > > of course when using
> > > > the binaries there is no circular dependency :-)
> > > Exactly. It also has another advantage: Nothing can be "more
original"
> > > than the "original" - I.e. why rebuilding target libs when you
can use
> > > the original files?
> >
> > Because those "target libs" are open source (wrapper) libs and have
to
> > be build by someone, this is very different from a closed source UNIX
> > you mentioned where you don't have any other option but to use the
> > closed source libraries.
> To me, rebuilding something just because it can be done, is
> overengineering, because
>
> 1. There is no guarantee these cross built files will identical to the
> "original" files.
> 2. You are building cross, to run these files on a target system, i.e.
> these files must be provided by somebody for the "target system".
> 3. Target files from a Linux perspective are just "meaningless binary
> data files", just like any other binary files (jpegs, sound files).
Over-engineering or not, binaries have nothing to do in source rpms,
They do. They
are just "data".
Or would you install a raytracer, just because a jpeg being used as
image somewhere is being generated by a raytracer?
especially not for an open source distribution like Fedora. And
since
they can be build from the source, there is no need for binaries in the
first place. So to me using the binaries just sounds like
lazy-engineering to me ;-)
Absolutely not:
You are building a tool chain, you are not building nor supplying the
OS. The situation is not any different from building a native GCC under
Linux: You use the library binaries being provided elsewhere.
Ralf