Build progress on Mac OS X PowerPC

[Andrew John Hughes]

2009/11/16 Eric Richardson <ekrichardson at gmail.com>:
> Andrew,
>
> On Sun, Nov 15, 2009 at 7:00 AM, Andrew John Hughes
> <gnu_andrew at member.fsf.org> wrote:
>>
>> 2009/11/15 Michael Franz <mvfranz at gmail.com>:
>> > Eric,
>> >
>> > I have encountered this issue before.  I am not sure what I did to fix
>> > it.
>> > I think I used ecj instead of the Apple javac.  Check the javac perl
>> > script
>> > that IcedTea has.  The other option is to apply this patch in the
>> > required
>> > build files.
>> >
>> >
>> > http://cr.openjdk.java.net/~andrew/ecj/01/webrev.01/make/build.xml.udiff.html
>> >
>>
>> That patch is in OpenJDK:
>> http://hg.openjdk.java.net/jdk7/jdk7/langtools/rev/2aa3a1cdb094
>>
>> > Here is my initial post on this
>> >
>> > http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2009-August/006857.html
>> >
>> > If I remember correctly, there is a tool that is compiled by ant (the
>> > two in
>> > the output), but it is not compiled with Java 5 byte code.  When ant
>> > tries
>> > to run the tool you get that error.
>> >
>> > I would switch to using ecj, as you might be able to get past this
>> > issue,
>> > there is another issue that will come up and ecj acts differently than
>> > javac
>> > and you really need the ecj way of doing things.  This is my original
>> > email
>> > on this:
>> >
>> > http://mail.openjdk.java.net/pipermail/distro-pkg-dev/2009-August/006879.html
>> >
>>
>> ecj is currently the only supported method for doing a build without
>> --disable-bootstrap, but that may change in the future.
>> Note that a normal OpenJDK build requires a Sun-based implementation
>> of Java 6, due to dependence on com.sun classes and AWT classes that
>> are part of the 1.6 specification.  We do our best to get round this
>> for IcedTea bootstrapping, but that testing is based on ecj+gcj not
>> Apple's products.
>
> We appreciate this and believe the Icedtea infrastructure really is the best
> method to bring Free Java to other platforms albeit in this case a not
> purely free one. I guess with the Apple PPC we have no choice but to use
> JDK5 which is a Sun based implementation. With the bsd port changes upstream
> we wouldn't even have patching problems.
>
> BTW, what is the safest make option that will not remove the patches I have
> applied. I hand patched the portions of the patches that didn't work and
> removed them from the make file after they succeeded. I didn't have much
> luck with the make clean-patch option so I opted for this approach
> eventhough this may not be the best approach.
>
> It seems that so far I have been able to just fix problems and then run make
> again without any problems but I'm not too advanced in my understanding of
> the process and my approach so far is not very repeatable.
>
> Eric
>
>> --
>> Andrew :-)
>>
>> Free Java Software Engineer
>> Red Hat, Inc. (http://www.redhat.com)
>>
>> Support Free Java!
>> Contribute to GNU Classpath and the OpenJDK
>> http://www.gnu.org/software/classpath
>> http://openjdk.java.net
>>
>> PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
>> Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8
>
>

If there are patches that won't apply for a good reason, we can
actually knock them out for the BSD port in the IcedTea build.  We
already do:

if !USE_CVMI
ICEDTEA_PATCHES += \
        patches/icedtea-javafiles.patch
endif

so if it's needed, I don't have an objection to a similar USE_BSD block.

I'm currently trying to build the BSD tree on linux-amd64 and am confused by:

http://hg.openjdk.java.net/bsd-port/bsd-port/jdk/diff/4fef42eb432c/make/common/internal/BinaryPlugs.gmk

Why were these additions made?  They aren't in either tl or the icedtea forests.

I spotted this because icedtea-copy-plugs.patch won't apply.  I may
just drop the copy-plugs patch altogether unless it's actually being
used; I think it may be dead.
-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8