RFC: Patch to enable adapter_opt_spread jsr 292 adapter code in zero

[Andrew Dinn]

Hi Chris,

I installed this patch and tested it on ARM. It failed initially but
only because of a bug introduced by the recent optimization patch (ed's
big patch). Once I fixed that I got the same result as you i.e. all but
13 of the 55 tests completed successfully on ARM.

regards,


Andrew Dinn
-----------

On 06/06/12 14:19, Chris Phillips wrote:
> Hi
> Attached is  an update to this patch with further fixes.
> And the summary from current 64 bit testing is attached below below.
> (I will build and run the sametest on arm later this am)
> 
> On 04/06/12 08:15 AM, Chris Phillips wrote:
>> Hi
>> This patch provides support  for jsr 292 in zero.
>> It is against :
>>   http://icedtea.classpath.org/hg/release/icedtea7-forest-2.1/hotspot
>> but will eventually apply to upstream also, ( after fixing other
>> issues to
>> allow testing). It enables a portion of jsr 292 logic that was not
>> functional
>> in the c++ Interpreter. It passes a large portion of the internal
>> dynamic invoke test suite (test/java/lang/invoke) although there are
>> still
>> 4 failing tests and 10 "NYI" (mostly involving Ricochet frames).
>>
>> It has been vetted once in a slightly altered form internally but I would
>> appreciate a few more eyes, would like to apply so qa testing can
>> continue later today.
>>
>> Chris
>>
>>
> Attached is  an update to this patch with further fixes.
> And the summary from current 64 bit testing is attached below below.
> (I will build and run the same test on arm later this am, also will provide
> patches and Makefile changes for testing with zero shortly.)
> 
> Chris