[oe] [meta-java][PATCH] openjdk-8: add recipes for openjdk-8 and openjre-8

[Jens Rehsack]

> Am 05.11.2015 um 14:44 schrieb Richard Leitner <richard.leitner at skidata.com>:
> 
> On 11/04/2015 06:57 PM, Jens Rehsack wrote:
>> Sure? Didn't you "bitbake openjdk-8"? Or is it because virtual/javac-native now points
>> to openjdk-8-native and there is no known way for me to let a user sanely pick the
>> beloved one on his/her own?
>> 
>> But openjdk-7 should not depend on virtual/javac-native -- so some enlightening is
>> required ...
> 
> I'm sure I built openjdk-7 ;-)
> 
> Following PREFERRED_PROVIDES were defined:
> PREFERRED_PROVIDER_virtual/java-initial-native = "cacao-initial-native"
> PREFERRED_PROVIDER_virtual/java-native = "jamvm-native"
> PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native"
> PREFERRED_PROVIDER_openjdk-7-jre = "openjdk-7-jre"
> 
> Any idea?

Yes. Say bye to "PREFERRED_PROVIDER_virtual/java-native" - since they're insane.

1) openjdk-7 must not use such a virtual provider, but a well known bootstrap path
   DEPENDS = " ... icedtea7-native ... "
   do a "bitbake -g openjdk-7" to figure out which bootstrap-component requires virtual/java-native or virtual/javac-native
2) When several recipes in bootstrap path define the same provider, bitbake refuses install, because one can only one particular package satisfying such a provider

I have with such a limitation no idea how to let people choose the desired provider for virtual/javac-native between jamvm/ecj, icedtea7-native or openjdk-8-native.

But this is stated in the commit message of both patches you had applied and is mentioned in the README as well. Please read the provided documentation, no worries that you cant' debug problems later on ...

Cheers
-- 
Jens Rehsack - rehsack at gmail.com