RFR: JDK-8188012 Nashorn build targets version 9 source

[Magnus Ihse Bursie]

On 2017-09-27 10:55, David Holmes wrote:
> On 27/09/2017 6:30 PM, Magnus Ihse Bursie wrote:
>>
>> The nashorn java code requires a somewhat convoluted compilation. As 
>> a result, we use a special java compiler setup, 
>> GENERATE_NEWBYTECODE_DEBUG. This explicitly lists -source 9 -target 
>> 9, which generates this warning:
>>
>> warning: [options] bootstrap class path not set in conjunction with 
>> -source 1.9
>>
>> when building in jdk10. I assume this is just a mistake, and that it 
>> should really target 10. (Or, to always be in sync, perhaps even 
>> $(VERSION_MAJOR)).
>
> Seems like a reasonable assumption. Though one has to wonder why we 
> even set the source/target? That would only be necessary if we needed 
> to force building to an older level. I think these could just be removed.

Historically, we have told javac explicitly which source/target version 
to use, instead of relying on the default. (This is not the only place 
we have this hard-coded in.)

We should change this behavior. I've created 
https://bugs.openjdk.java.net/browse/JDK-8188015 to track this.

/Magnus



>
> Cheers,
> David
>
>
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8188012
>> Patch inline:
>> diff --git a/make/BuildNashorn.gmk b/make/BuildNashorn.gmk
>> --- a/make/BuildNashorn.gmk
>> +++ b/make/BuildNashorn.gmk
>> @@ -41,7 +41,7 @@
>>   $(eval $(call SetupJavaCompiler, GENERATE_NEWBYTECODE_DEBUG, \
>>       JVM := $(JAVA_JAVAC), \
>>       JAVAC := $(NEW_JAVAC), \
>> -    FLAGS := -g -source 9 -target 9 --upgrade-module-path 
>> "$(JDK_OUTPUTDIR)/modules/" \
>> +    FLAGS := -g -source 10 -target 10 --upgrade-module-path 
>> "$(JDK_OUTPUTDIR)/modules/" \
>>            --system none --module-source-path $(call 
>> GetModuleSrcPath), \
>>       SERVER_DIR := $(SJAVAC_SERVER_DIR), \
>>       SERVER_JVM := $(SJAVAC_SERVER_JAVA)))
>>
>> /Magnus