Re: ia64 libjava java-signal.h build failure

[Alexandre Oliva <aoliva at redhat dot com>]

On Apr 21, 2001, Mark Mitchell <mark@codesourcery.com> wrote:

> My understanding was that there are two predicates here.

> One predicate says whether or not to enable Java on those platforms
> where it is known to work.

> The other says on what platforms it is known to work.  This is the one
> that turns off libgcj on certain systems.

Nope.  There are two predicates.  One that enables or disables the
Java compiler by default, the other enables or disables the Java
libraries, but it won't enable them if the Java compiler is disabled.
I've been trying to get the first enabled by default, since it's the
one that takes precedence over the other, so that we can at least find
out whether libjava is broken on certain platforms, and disable it on
a per-platform basis, in the release.

The earlier we do this, the earlier we'll get reports about libjava
being broken, and the earlier we'll fix GCC for the release.  If we
don't take this step, we'll end up with a GCC without a Java compiler
enabled by default, which is a regression.  The libraries may still be
disabled, as they were in GCC 2.95.2, but it's the compiler I'm
talking about.

> Then, if you want to enable libgcj on more systems than it is already
> enabled on, some additional work will be required -- but only on that
> platform.

Precisely.  With the current set up, it's pretty much disabled on all
systems, and that's ok.  (Perhaps we should even have a
--enable-libjava configure switch to override this decision?).  What I
want is a Java compiler enabled by default, so that I don't risk
forgetting to enable Java in my weekly bootstraps.

Some of the involved platforms take the whole week for a full
build&test, so, if you really want this to be done for a third time,
we won't be able to do anything before the end of April, and,
eventually, someone will decide it's too late to make this change.

Let's please do it before the next snapshot, and then disable libjava
whenever we get a report that it doesn't work on a certain platform.
That's the right way to do it.  We really don't want Java disabled by
default on platforms in which it is known to work, and that's the
situation we have now.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me