Volker - thanks for getting this code out so quickly - and an especial thanks for publishing the build instructions and build logs too. I'll give the code a build spin on my machine and let you know what happens On 25 Jun 2012, at 16:56, Volker Simonis wrote:
Hi gentleman,
as attentive readers of this list may have noticed, we've pushed today initial changesets to our ppc-aix-port forest which make it possible to build a full (although non-functional) JDK on Linux/PPC64.
We've tried (and we will try in the future) to be as non-intrusive as possible with regards to shared code and other existing platforms in order to simplify an eventual merge of our port into the main line. The current version is not very exciting by itself (after all, it doesn't run:) but it will serve as a foundation for the implementation of the C++-Interpreter in the next weeks and month. We also plan to push another changeset soon, which will enable the same, stubbed-out interpreter build of the HotSpot on AIX such that our colleges can start with the class library build on AIX.
Some basic build instructions are in the new README-ppc.html file (http://hg.openjdk.java.net/ppc-aix-port/jdk7u/raw-file/tip/README-ppc.html). As we move forward we expect to add more information to this file. Most of the currently needed extra build variables (like HOTSPOT_TARGET=all_debugcore CC_INTERP=true OPENJDK=true CORE_BUILD=true) are necessary because of the HotSpot build which currently only supports a "CORE" (i.e. interpreter only) version. Others like the settings for RMIC and IDLJ are necessary because we use another than the original Oracle JDK for bootstrapping.
Build logs of the initial version on Linux/PPC64 have been made available here:
http://cr.openjdk.java.net/~simonis/ppc-aix-port/build-logs/
And there's one more peculiarity: in order to build the complete JDK, we had to transplant "7177701: error: Filling jar message during javax/imageio/metadata/IIOMetadataFormatImpl compilation" from "jdk8/tl/langtools" to work around http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7177701. This may lead to a duplicate bug ID in the langtools repository in the future if we push from 7u-dev. On the other hand we don't expect to do many changes in the langtools repository at all, so we may just overwrite our langtools repository with the one from 7u-dev once it contains this change.
Please feel free to comment or report any problems to the mailing list.
Regards, Volker