"C. Bergstr?m" writes:
> James Carlson wrote:
> > For some, that's true; for others, not so true.  I don't disagree with
> > what you're doing, but it's worth pointing out that others are headed
> > in other directions, and that licensing issues are not necessarily
> > absolutes.
> >   
> If there's no promise from [insert-name-of-company] to open source it 
> and the source isn't available *today*.. it doesn't exist.. I realize 
> that's a hard line and I may/will duplicate work, but c'est la vie..  (A 
> good example is the sun cc compilers (w/o the optimization libs) I hear 
> rumors it can be done and it's planned.. it would be *really* great, but 
> hasn't happened)

As I said, I don't disagree with what you're doing.  I understand what
you're saying, but that's not the only possible political position on
this issue.

For instance, the line the OpenSolaris distribution folks drew was
between binaries that are freely redistributable (regardless of source
restrictions) and those that are not.  It's also a valid choice, but
quite different in implications.

> > I'm not sure that's compatible at the binary level like that.
> >
> > Are you also using only open-source C/C++ compilers?
> >
> >   
> For my first pass I'm trying to reduce the number of variables to my 
> changes or minimize at least. (since I'm changing other things besides 
> removing the closed bins..)

The reason I ask is that I don't see a great deal of philosophical
difference between relying on an OCO (object code only) tool chain and
relying on a few tool-chain-related OCO components, such as libCrun.
Both are "impure" from the open-source-only standpoint I think you're
trying to take.

> 2nd pass I'm aiming at fully open toolchain as well.. With open64 I'll 
> have to make other changes to how ON is compiled.  stab debugging format 
> is explicitly being used which isn't supported by open64.. (seemingly 
> small change, but I've tested and had unexpected issues)

Do you mean the www.open64.net "open research compiler?"  If so, then
I don't doubt that there'll be issues.

I don't know what the other consolidations are doing, but for ON, we
intentionally committed to supporting only two C/C++ compilers: Sun's
WorkShop and gcc.  Even at that, we support only very specific
versions of each, and changing versions is a big deal, requiring
extensive testing, and often some code rework.  Supporting a brand new
compiler in the chain would be quite hard.

Real support here means not just getting it to compile once.  Nor is
it sufficient to integrate fixes and hope for the best in the future.
To get real support, all of the contributors need to build and test
with all of the supported compilers (so that they know they're not
contributing things that will break others), and the gatekeepers as
well have to build regularly with all of the supported build
environments.  It's a change of process as well as a change of code.

It sounds like an interesting project, but unless there's some buy-in
from the others who'd be affected, I think you might be talking about
a fork of the code.

-- 
James Carlson, Solaris Networking              <james.d.carlson at sun.com>
Sun Microsystems / 35 Network Drive        71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677

Reply via email to