Moinak Ghosh wrote:
Shawn Walker wrote:
On 26/06/07, Eric Boutilier <[EMAIL PROTECTED]> wrote:
On Tue, 26 Jun 2007, Peter Tribble wrote:
> By people, what's the target audience at this stage? I think my
question is
> really whether the first "release" is aimed at users, or whether
there should
> be a 0.0 "release" solely for the purpose of Indiana bootstrapping?
Good question. So that raises the issue of "non-emancipated"
(non-redistributable) files. That is, should there be a 0.0 or 0.1
release that includes them in it? (I vote no, FWIW.)
And if not, wouldn't that be the main constraint here?
Non-emancipated is not the same as non-redistributable. Remember that
there are binary "blobs" that are redistributable. There are several
files right now that can be redistributed but have not been
emancipated that are very important for the basic system. (libm.so
comes to mind... I think).
libm is no longer a good example for this because it has been
opensourced
sometime back:
http://dlc.sun.com/osol/devpro/downloads/current/
I'd say libc_i18n.a is one of the problems apart from a bunch of
drivers,
parts of the crypto framework etc. A few of the drivers can be replaced
with opensource equivalents but not all and some are essential.
These are
all redistributable closed source binaries. These are the biggest
constraints
I think.
Forgot the add the other big piece. The SUN Studio C++ runtime libs like
libCrun, libCstd etc. are closed binary redistributables. These are
necessary
for SUN Studio compiled C++ stuff to run like, Firefox, Thunderbird and
OpenOffice.
One way to overcome would be to do static linking, but once is still
technically redistributing closed bits. It is possible to compile
Firefox and
Thunderbird using gcc but OpenOffice + gcc + Solaris is a nightmare. I
had tried in the past when libCrun, libCstd were not redistributable and
gave up.
Regards,
Moinak.
_______________________________________________
indiana-discuss mailing list
[email protected]
http://opensolaris.org/mailman/listinfo/indiana-discuss
