[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-hotspot-dev
Subject:    Review request: White box testing API for HotSpot
From:       Alan.Bateman () oracle ! com (Alan Bateman)
Date:       2011-12-15 10:14:49
Message-ID: 4EE9C899.1050002 () oracle ! com
[Download RAW message or body]

On 15/12/2011 06:34, David Holmes wrote:
> On 15/12/2011 1:56 AM, Alan Bateman wrote:
>> On 14/12/2011 01:42, John Pampuch wrote:
>>> Will the modularity changes coming in JDK 8 affect this?
>> Yes as the class path, boot class path etc. should go away in modules
>> mode. The likely approach would be to have the wb API installed as a
>> module where it could be used in both legacy/classic mode and when
>> running in modules mode.
>
> But in that case there is no bootclasspath and so the current validity 
> check of being "on the boot classpath" becomes invalid. So what would 
> replace that? Does the concept of the boot loader disappear?
Yes, this should go away but code in the platform modules that except 
the the class loader to be null will have to continue to see null and a 
new method introduced to return the module class loader. More generally 
for white box testing is there will need need to be a mechanism to allow 
tests to access types and internals aren't exported by modules.

-Alan

[prev in list] [next in list] [prev in thread] [next in thread]