RFR: JDK-8230649: Make jpackage tool an experimental feature

[Kevin Rushforth]

In that case, it may be better to issue any warnings about it being 
experimental directly from jpackage instead of relying on the incubating 
modules.

-- Kevin


On 9/15/2019 1:05 PM, Andy Herrick wrote:
> yes - result of this change is as you suggest, everything shows this 
> warning, so we will need further discussions as to what it might mean 
> to make jpackage "experimental".
>
> /Andy
>
> On 9/15/2019 12:35 PM, Alan Bateman wrote:
>> On 15/09/2019 14:58, Andy Herrick wrote:
>>> Please review the jpackage fix for bug [1] at [2].
>>>
>>> This is a fix for the JDK-8200758-branch branch of the open sandbox 
>>> repository (jpackage).
>>>
>>> This is the at least the first small set of changes that need to be 
>>> make to make jpackage an experimental package.
>>>
>>>  - Add flags to the build so the module jdk.jpackage is built as an 
>>> experimental module.
>>>
>>>  - Modify test programs to tolerate the warning emitted when 
>>> jpackage is run.
>> I think this change will need discussion. Can you provide a summary 
>> on what you mean by "experimental package"? I remember seeing Mark's 
>> comment go by where he suggested that the tool should be an 
>> experimental feature but I'm not sure if this translates to a warning 
>> or a configure option.
>>
>> I see the JIRA issue references the JEP for Incubating Modules but 
>> I'm not sure that it makes sense here as jdk.jpackage doesn't export 
>> an API and will eagerly participate in service binding because it 
>> `provides java.util.spi.ToolProvider`. There are subtle issues around 
>> incubating modules that want to provide services that were not worked 
>> out in the JDK 9 time frame. In this case, java.base uses 
>> ToolProvider so jdk.jpackage will be resolved when it is observable. 
>> I assum `java -version` will print a warning and that will not be 
>> welcomed.
>>
>> -Alan
>>