Ian Collins writes:
> James, I think you either missed my point, or more likely I wasn't clear. 
> 
> I was suggesting that it might be interesting to come up with a means of
> building the barest minimum of OpenSolaris that could boot and be used
> to build the rest.  A bit like the bootstrap build at the beginning of
> gcc, which can be built with any C compiler which is then used to build
> the rest.

That suggestion still has the same inherent problem.

It's not a matter of the volume of code that's subject to this new
foreign-build rule -- which I would suspect to be rather large,
including most of the kernel -- but rather the existence of it.  If it
exists, then contributors to Open Solaris will be _required_ to
maintain it.  It represents a tax on all contributors to support a
usage model that is, at least in my opinion, marginally interesting at
best.

Let's return to the original problem.  What is the problem being
solved here?  Why is building Open Solaris on Linux an interesting
solution to that problem?  It can't be the development tools -- the
same and better tools are already available on Solaris.  So why
bother?

I think your gcc analogy rings false in this case.  With gcc, if
you're on a platform that's unsupported but with a CPU that gcc does
support, then you need that bootstrap process to make gcc supported on
the platform.  When you've gone through this process, you end up with
a compiled version of gcc that runs on that platform.

However, in this case, we're not talking about making Open Solaris
"run" on Linux.  Solaris itself *is* a platform.  The sort of
bootstrapping needed for a platform is very much different from the
sort of bootstrapping needed for a compiler, because hardware issues
are involved.

So, instead of bootstrapping, it sounds like we're talking about
cross-compilation of Solaris to produce binaries that will run on a
_separate_ target system.  If Linux were somehow perceived as a richer
host environment, that might be interesting.  Even so, it runs into
the problems described previously: maintaining the ability to do
cross-compilation is a tax on all projects.  Is the goal worth the
cost?

-- 
James Carlson, KISS Network                    <[EMAIL PROTECTED]>
Sun Microsystems / 1 Network Drive         71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to