Bug fix: Debug builds fail

[Edward Nevill]

Hi Joseph,

Thanks for fixing this. Did you want to try pushing this and see if push
is working for you now?

All the best,
Ed.

On Tue, 2016-01-05 at 00:22 +0000, Joseph Joyce wrote:
> Hi Ed,
> Thanks for testing this. I get the same result as you. I didn't test
> actually running the build as the debug builds worked previously -
> sorry.
> 
> I've made a new patch with the build fixes and a new fix to stop the
> assertion failing. It combines both store_check_part_1 and
> store_check_part_2 into a single method as has happened in the aarch64
> version.
> 
> I've tested it (by running jedit) across the release, fastdebug and
> slowdebug builds (both with and without -XX:+UseCondCardMark) and it
> appears to work ok.
> 
> Also while jedit starts successfully there are a couple of other bugs
> I've noticed such as entering text crashes it, as too does clicking on
> the About jEdit option in the help menu. These affect both debug and
> release builds. I'm going to look into these tomorrow. 
> 
> Thanks
> Joseph
> 
> 
> On Wed, 2015-12-30 at 21:40 +0000, Edward Nevill wrote:
> > Hi Joseph,
> > 
> > This patch looks good and fixes the debug build, thanks for doing
> > this. Would you like to try pushing it, and see if the problems you
> > were having pushing before persist.
> > 
> > Although it now builds, I get assert failures straight off with just
> > the java command on its own. Is this expected? FYI: The assert
> > failure I get is
> > 
> > 
> > # A fatal error has been detected by the Java Runtime Environment:
> > #
> > #  Internal Error
> > (/work/ed/aarch32/jdk9/hotspot/src/cpu/aarch32/vm/macroAssembler_aarc
> > h32.cpp:1712), pid=28250, tid=28251
> > #  assert(bs->kind() == BarrierSet::CardTableModRef) failed: Wrong
> > barrier set kind
> > 
> > All the best,
> > Ed.
> > 
> > On Wed, 2015-12-30 at 17:46 +0000, Joseph Joyce wrote:
> > > Hi all,
> > > The build fails on both fastdebug and slowdebug builds from an
> > > invalid
> > > conversion from ‘InstanceKlass**’ to ‘Klass**’ in
> > > methodHandles_aarch32.cpp.
> > > This patch fixes that and also some warnings about possibly
> > > unitialised
> > > values.
> > > 
> > > Thanks
> > > Joseph
> > 
> >