RFR: Initial code for a JSON-based module descriptor.

[David Bosschaert]

On 11 May 2012 04:25, Shi Jun Zhang <zhangshj at linux.vnet.ibm.com> wrote:
> On 5/9/2012 6:24 PM, David Bosschaert wrote:
>>
>> Hi all,
>>
>> I have updated my changes and hopefully this addresses the issues Shi
>> Jun pointed out.
>>
>> * I abandoned the use of java.util.Scanner - I ran into that exception
>> too.
>> * Moved handling of module-info.json in ModuleClassLoader down to the
>> Library level.
>> * Moved multiple charAt() calls to local variable in JSONParser.
>> * There are unit tests in _JSONParser and
>> module-info-reader.sh/_ModuleInfoReader.
>> * A simple system test can be found in hello-json.sh.
>>
>> I didn't change all of the module-info.class handling code that exists
>> in other places yet, I'd like to take an incremental approach to this
>> so this is just the beginning, which gets the basic hello-json.sh
>> testcase going. I'd like to address the other places that need
>> updating for JSON module descriptors in separate commits.
>>
>> You can find my updated changes here:
>> http://cr.openjdk.java.net/~davidb/mijson/webrev.02/
>>
>> Best regards,
>>
>> David
>>
>> On 27 April 2012 13:45, David Bosschaert<david.bosschaert at gmail.com>
>>  wrote:
>>>
>>> On 24 April 2012 09:59, Shi Jun Zhang<zhangshj at linux.vnet.ibm.com>
>>>  wrote:
>>>>
>>>> Another comment, java.util.jar.JarFile.getModuleInfo() also needs to be
>>>> modified accordingly.
>>>>
>>>>
>>>> --
>>>> Regards,
>>>>
>>>> Shi Jun Zhang
>>>
>>> Thanks, yes and there are even more. I found ModuleInfo readers in
>>> org.openjdk.jigsaw, sun.tools.jar and com.sun.classanalyzer. And there
>>> might even be more. Ultimately they all need to be modified to deal
>>> with the JSON definition.
>>>
>>> Best regards,
>>>
>>> David
>
> Hi David,
>
> This patch looks good to me.
>
>
> --
> Regards,
>
> Shi Jun Zhang

Thanks - I'll commit it today.

Cheers,

David