[rfc][icedtea-web] run on any open/oracle jdk (and ibm jdk!)

[Andrew Azores]

On 05/20/2014 09:10 AM, Jiri Vanek wrote:
>
> On 05/17/2014 12:32 PM, Jiri Vanek wrote:
>> On 05/16/2014 04:24 PM, Omair Majid wrote:
>>> * Jiri Vanek <jvanek at redhat.com> [2014-05-16 08:25]:
>>>> On 05/16/2014 12:24 AM, Omair Majid wrote:
>>>> heheh :)) Somebody should review it ;)
>>>
> RFC! :)
>>>
>>>>>> On 05/13/2014 08:43 PM, Jiri Vanek wrote:
>>>>>>> This is probably not final version of the patch - as it break all
>>>>>>> tests using PluginAppletMocks - Its easy - tests do not run on
>>>>>>> bootclassapth, so the Ancestor of my Access class is in different
>>>>>>> classlaoder.
>>>
>>> Are you planning on fixing this first?
>> I wonted, and i would like to. Its work in progress, and may longer 
>> then I thought.
>>
>> Between first verion of patch - where was this original statement - 
>> and second version, where I
>> wrote a bit more, several things happened:
>>
>> I tried to force to load sveral explicit classes directly in java 
>> code by bootclasslaoder, however,
>> not successfully.
>> So I tried to change xbootclasspath parameter. It didn`t help 
>> either!  However, same change on
>> commandline usage outside makefile helped :-/. But caused 7other 
>> tests to fail (the parser family)
>>
>> So I ended in no op for Friday. There is still few ways to try but - 
>> as we (did we?) agreed - it do
>> not need to be blocker for patch itself. The patch can lurk for a 
>> while in Head before tests are fixed.
>>
>>
>> So answer is yes. But may happen taht with longer delay :(
>>
>> Thanx for your thought!
>>
>> J.
>>
>
> This patch have all the tests fixed.
> The changes in netx itself are 100% same as before. Only makefiel.am 
> and two tests have changed.
>
> This patch should be (and will be befor I push) split into three parts.
> 1) the patch to netx itself, which alows us to run with 
> usptreamopenjdk, oracle jdk and ibm jdk
> 2) the fixes to tests and makefile which adapt to need of 
> xbootclassapth instead of classapth
> 3) added the semicolons behind every CLASSPATH variable which was 
> modified/added/used in 2
>
> Well tbh, I do not fully understand why 2 is needed. But At least I 
> can guess. But I have absolutely
> no idea why the semicolon (3) have so huge impact.... Any hints welcomed.
>
> J.
>
> Also for record - I have attempted to upstream the minor private->pkg 
> private patch:
> http://mail.openjdk.java.net/pipermail/jdk9-dev/2014-May/000726.html
> Still no reply.
>
>
>
>

Well, I don't really feel fully qualified to review this, and I haven't 
run any tests with the patch applied - but looking over the patch it 
seems fine, and it builds fine and some light manual testing looks 
normal at least. So, +0.5?

Thanks,

-- 
Andrew A