RFR: JDK-8143141: Bring in minor build changes from the jigsaw/jake forest

[Erik Joelsson]

New webrev fixing the spaces and removing APP_CLASSPATH completely.

http://cr.openjdk.java.net/~erikj/8143141/webrev.top.03/
http://cr.openjdk.java.net/~erikj/8143141/webrev.jdk.03/

/Erik

On 2015-11-24 12:27, Magnus Ihse Bursie wrote:
> On 2015-11-24 12:13, Erik Joelsson wrote:
>>
>>
>> On 2015-11-24 11:47, Magnus Ihse Bursie wrote:
>>> On 2015-11-18 14:00, Erik Joelsson wrote:
>>>> After posting this, I started thinking and I realized the biggest 
>>>> change here is not needed. I was able to revert all of the 
>>>> GENSRC_OUTPUTDIR, HEADERS_OUTPUTDIR, NATIVE_OUTPUTDIR and 
>>>> MODULES_*_OUTPUTDIR variables in jigsaw/jake and subsequently those 
>>>> changes are not needed here.
>>>>
>>>> New webrevs, only for top and jdk now. The other repos no longer 
>>>> need changes:
>>>> http://cr.openjdk.java.net/~erikj/8143141/webrev.top.02/
>>>> http://cr.openjdk.java.net/~erikj/8143141/webrev.jdk.02/
>>>
>>> Looks good!
>>>
>>> A minor nit:
>>>
>>> In CompileJavaModules.gmk, you could add a space after the comma in:
>>> CLASSPATH := $(foreach d,$(DEPS), $(if $($d_BIN), $($d_BIN), \
>>> and
>>> $(eval $(call SetupJavaCompilation,$(MODULE), \
>>>
>> Sure, will do before pushing.
>>> I also see the need to clarify the logic of 
>>> GendataBreakIterator.gmk, but we can save that for a rainy day. :)
>>>
>>> I'd also like to verify that the removal of the APP_CLASSPATH for 
>>> the two launchers are indeed correct. Where they superfluous?
>>>
>> Yes, we no longer have those jars. All modules are in the module image.
> Does that mean that all references to APP_CLASSPATH are unnecessary? 
> Perhaps, in fact, that the logic for parsing APP_CLASSPATH in 
> LauncherCommon.gmk can be removed?
>
> /Magnus
>