This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: ia64 libjava java-signal.h build failure


>>>>> "Alexandre" == Alexandre Oliva <aoliva@redhat.com> writes:

    Alexandre> This is not fair.  You had asked to have it disabled
    Alexandre> temporarily, so that people could fix some of the C++

No, that's not quite right.  

In particular, libgcj was not included in GCC 2.95.2.

I would be happy to turn on the Java front-end, but leaving libgcj off
by default, exactly as in GCC 2.95.2.  It is libgcj that contains all
the new configury goo.

    Alexandre> Maybe you don't care for Java.

Now you are being unfair.  

My feelings towards Java happen to be predominantly positive, but my
feelings -- positive or negative -- aren't relevant to whether or not
this functionality is enabled by default, and it's not polite to
suggest that they are. :-)

It's my job to try to get the release done, using very limited
resources.  That means that we simply cannot do everything we would
have liked to have done.  I'd feel differently if it were three months
ago, and we had knocked off all the high-priority bugs.  

I'm afraid that even if this "just works" for you on multiple boxes
that it won't for other people, especially on less popular systems.
If, however, it really does work "out of the box" on all of the
primary evaluation platforms, and as many secondary ones as we can
try, then I'll change my stand.  

So, how about we make a deal?  If you can get two independent reports
(different people, different boxes, different administrative domains)
for each primary platform, and one report for three other platforms,
preferably not just different versions but actually different
operating systems, go ahead and throw the switch.  Post the testers
and the platforms with the check-in.

It's OK if Java isn't built on some of these systems; that's part of
what we want to test, too.  We want to test that turning on the flag
will either cause things to build correctly, or not build at all.
Also, please run `make check' and make sure it doesn't behave oddly.
The goal is that the usual `checkout; configure; make bootstrap; make
-k check' sequence still work smoothly.

How does that sound?

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]