harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jimmy,Jing Lv" <firep...@gmail.com>
Subject Re: [Proposal] Harmony Select Milestone One
Date Thu, 27 May 2010 09:52:03 GMT
Hi,

      Just study a little in the Hadoop project - it seems it mainly focus
on Linux platform and windows is not suggested. I guess for Hadoop we can
test on Linux only, save time for tuning it best :)

2010/5/26 Jimmy,Jing Lv <firepure@gmail.com>

> Hi,
>
>
> 2010/5/25 Jim Yu <junjie0122@gmail.com>
>
> 2010/5/25 Jimmy,Jing Lv <firepure@gmail.com>
>>
>> > Hi Everyone,
>> >
>> >      We are very close to publish Harmony5 M14 and Harmony6 M2, what
>> great
>> > achievements since we created Harmony project!
>> >
>> >      Now we may think about something different, satisfy more
>> requirements.
>> > Different application requirement need different JRE - it may differ
>> very
>> > much in class-lib modules, providers, and VMs.  Harmony select, say, a
>> > smaller, more flexible harmony runtime, may be a good candidate in such
>> > environments. We may not follow Oracle/Sun exactly in this runtime,
>> allow
>> > more innovations, different modules selections,  and different
>> > functionalities to meet different requirements. We may also call for
>> more
>> > input from real Java applications, developers, especially hot
>> open-source
>> > projects who are the potential customers of Harmony.
>> >
>> >      We've already created an initial target in Harmony to build a
>> Select
>> > binary. We remove some client-side related modules (like AWT, swing etc)
>> to
>> > reduce the size while keep the main functionality for server side
>> > applications. This is a startup, very far from the finial versions.
>> However
>> > I believe we can start some milestones with defined criterion like
>> standard
>> > Harmony.
>> >
>> >      For the first milestone of Harmony-Select, I suggest we'd have
>> > following criterion:
>> > 1. build pass on both windows/linux platforms;
>>
>> 2. pass all Harmony unit testcases in the selected module on these
>> > platforms;
>>
>>
>> I agree we should pass them. But I suspect some of our tests will rely on
>> the
>> modules that are not in the scope of Harmony Select. I remember some tests
>> of beans module will call awt classes...
>>
>>
> Agreed, depart these testcases should be a pre-work for running testcases.
>
>
>>  >
>>
>> 3. Choose Apache Hadoop as a real Application test, pass its common test
>> on
>> > these platforms for the first milestone.
>> >
>>
>> Do you mean the startup tests for Hadoop Common? So far as I know, Hadoop
>> Common
>> can startup with one of three modes(standalone, pseudo-distributed and
>> full
>> distributed).
>>  In full distributed mode, we can involve many machines with different
>> platforms. So even
>> if Harmony Select can work well to support this mode with two machines, we
>> can't ensure
>> it can work well to support dozens of machines.
>>
>> I agree that Hadoop is a good server application for Harmony Select
>> testing
>> and we can
>> try with it to see its capability. But seems it is not so easy to define
>> our
>> milestone criteria
>>  based on it ;-)
>>
>>
> Yes I understand. We can refine the definition, e.g, pass 2 modes first as
> the  target of milestone one. I suggest we still keep the criteria on real
> application so that to prove Harmony Select is a workable runtime ;)
>
>
>> >
>> >      Wait for any kindly suggestions, opinions and comments.
>> >
>> > --
>> >
>> > Best Regards!
>> >
>> > Jimmy, Jing Lv
>> >
>>
>>
>>
>> --
>> Best Regards,
>> Jim, Jun Jie Yu
>>
>
>
>
> --
>
> Best Regards!
>
> Jimmy, Jing Lv
>



-- 

Best Regards!

Jimmy, Jing Lv
China Software Development Lab, IBM

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message