RFR [11]: Fix post-merge build problem: Shenandoah should be disabled with Zero

[Aleksey Shipilev]

Thanks, pushed. Respinning tests...

-Aleksey

On 3/6/19 6:19 PM, zgu at redhat.com wrote:
> Looks good.
> 
> -Zhengyu
> 
> On Wed, 2019-03-06 at 18:12 +0100, Aleksey Shipilev wrote:
>> Build-test shows I did a little merge mistake in sh/jdk11: I forgot
>> to add shenandoahgc to the new
>> block that disables Zero for all unsupported GCs.
>>
>> Here is the rectification:
>>
>> diff -r 3f6d0b57c622 make/autoconf/hotspot.m4
>> --- a/make/autoconf/hotspot.m4  Tue Mar 05 20:27:45 2019 +0100
>> +++ b/make/autoconf/hotspot.m4  Wed Mar 06 18:00:55 2019 +0100
>> @@ -356,11 +356,11 @@
>>      AC_MSG_RESULT([no, platform not supported])
>>    fi
>>
>>    # Disable unsupported GCs for Zero
>>    if HOTSPOT_CHECK_JVM_VARIANT(zero); then
>> -    DISABLED_JVM_FEATURES="$DISABLED_JVM_FEATURES epsilongc g1gc
>> zgc"
>> +    DISABLED_JVM_FEATURES="$DISABLED_JVM_FEATURES epsilongc g1gc
>> shenandoahgc zgc"
>>    fi
>>
>>    # Turn on additional features based on other parts of configure
>>    if test "x$INCLUDE_DTRACE" = "xtrue"; then
>>      JVM_FEATURES="$JVM_FEATURES dtrace"
>>
>> Testing: local x86_64-zero build
>>
>> -Aleksey
>>