Step-by-step OpenJDK on OS X?

[Ben Evans]

Hi Kurt,

On Sat, Jan 31, 2009 at 11:10 PM, Kurt Miller <kurt at intricatesoftware.com>wrote:

>
> Ben Evans wrote:
> > On Sat, Jan 31, 2009 at 1:15 PM, Kurt Miller <kurt at intricatesoftware.com
> > <mailto:kurt at intricatesoftware.com>> wrote:
> >
> >     Ben Evans wrote:
> >     > Hi Charlie,
> >     >
> >     > My understanding is that the current bsd-port is still broken for
> OS X
> >     > and that we expect it to be unbroken sometime soon.
> >
> >     To unbreak the OS X build you can issue the following commands which
> >     will revert the minor breakage in the tree right now. But first
> enable
> >     the mq extension in ~/.hgrc to enable the hg strip command:
> >
> >     [extensions]
> >     hgext.mq <http://hgext.mq> =
> >
> >     $ cd bsd-port
> >     $ cd hotspot
> >     $ hg strip 32826d5f1894
> >     $ cd ../jdk
> >     $ hg strip 97139295a172
> >
> >     Temporally DEVTOOLS_PATH needs to be set to point an existing
> directory.
> >     It doesn't matter what dir since OS X doesn't use it, just something
> to
> >     satisfy the sanity check.
> >
> >
> > OK.
> >
> > I'm having trouble with the ALT_JDK_IMPORT_PATH sanity check. My make
> > line seems basically identical to yours - except that you don't set
> > ALT_JDK_IMPORT_PATH.
> >
> > If I try building without that set, I get tons of warnings and then fail
> > immediately with this:
> >
> > # Running javac:
> > Check_ALT_JDK_IMPORT_PATH/bin/javac -J-XX:ThreadStackSize=768 -J-client
> > -J-Xmx896m -J-Xms128m -J-XX:PermSize=32m -J-XX:MaxPermSize=160m -source
> > 1.5 -target 5 -encoding ascii -classpath
> > /usr/local/soylatte16-i386-1.0.3/lib/tools.jar -sourcepath
> >
> /Users/boxcat/projects/jdk7-2009-01-30/sources/build/bsd-i586/corba/gensrc:../../../src/solaris/classes:../../../src/share/classes
> > -d
> >
> /Users/boxcat/projects/jdk7-2009-01-30/sources/build/bsd-i586/corba/classes
> >
> @/Users/boxcat/projects/jdk7-2009-01-30/sources/build/bsd-i586/corba/tmp/sun/javax.transaction.xa/.classes.list
> > /bin/sh: Check_ALT_JDK_IMPORT_PATH/bin/javac: No such file or directory
> > make[4]: *** [.compile.classlist] Error 127
> >
> > Could you post the output of set in your build shell so I can try and
> > spot any differences in the way our environments are setup?
>
> I'm sorry I only have OS X on my PowerPC PowerBook. Currently I don't
> have an OS X system capable of doing a build. :(
>
> You can safely ignore the warning about ALT_JDK_IMPORT_PATH not being
> set.
>
> Perhaps try starting the build like this to ensure your environment
> is clean:
>
> env -i PATH=${PATH} make \
> ALT_BOOTDIR=/usr/local/soylatte16-i386-1.0.3 \
> ALT_BINARY_PLUGS_PATH=/usr/local/lib/jdk-7-icedtea-plugs \
> ALT_FREETYPE_HEADERS_PATH=/usr/X11R6/include \
> ALT_FREETYPE_LIB_PATH=/usr/X11R6/lib \
> ANT_HOME=/usr/share/ant \
> NO_DOCS=true \
> HOTSPOT_BUILD_JOBS=1 \
> DEVTOOLS_PATH=/usr/bin
>

That seems to be a bit happier now - thanks.

Thanks also to Xiaobin for helping me track down the source of the
environment problems.

For the record, in case anyone else has problems - looks like Fink installed
an ant which conflicted with the build - so check to make sure that /sw and
things aren't on your PATH and ensure you unset ANT_HOME.

My build now gets much further, but fails because
build/bsd-i586/hotspot/outputdir/bsd_i486_compiler2/product/test_gamma
doesn't exist - I have to create this as an empty file, and set it
executable, then rerun the make command. It hasn't hit it yet, but I'm
betting that the same will be required for compiler1 in a few minutes time.

Can anyone spare any cycles to theorise as to why I still have problems with
test_gamma when others don't?

Thanks,

Ben
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/bsd-port-dev/attachments/20090201/36c4fe66/attachment.html