Re: changes for approval, part three (solaris)

[Phil Edwards <pedwards at disaster dot jaj dot com>]

On Tue, Oct 24, 2000 at 06:13:20PM -0400, David Edelsohn wrote:
> 
> 4) libio/_G_config.h now uses the GNU/Linux libc version.
[...]
> What is the purpose / intent of the current use of _G_config.h?  At least
> part of Phil's patch is working around a definition in _G_config.h which
> is not appropriate for Solaris.  Not surprisingly, AIX has problems with
> the assumptions in _G_config.h as well.

True, a lot of workarounds in a lot of places could go away if _G_config.h
were generated from, say, _G_config.h.in and included a bunch of autoconf
HAVE_ tests.  I would be more than willing to do the work.

But v3's libio is supposed to be {derived from, very close to, the exact
same as} the libio subset of glibc.  I don't know which of those is true.
Ideally, it's "the exact same as," which means that changes to v3's
libio would need to be folded back into glibc.  (Which makes me nervous.)

And I can understand if the glibc maintainers aren't interested in a lot
of autoconf tests -- and there would be a lot -- since their library is
targeted for a single OS.  (Clearly that's why those #defines are the
way they are; there're even comments like "we know this is always true,"
but you already saw those.)  It's a hard issue and I can see both sides.



-- 
pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.