Obsolete jdk15, jdk16, linux-sun-jdk15, linux-sun-jdk16

Mon Mar 31 18:07:06 PDT 2008

Aryeh M. Friedman wrote:
> Greg Lewis wrote:
>> On Mon, Mar 31, 2008 at 04:03:55PM -0400, Robert Huff wrote:
>>  
>>> Aryeh M. Friedman writes:
>>>    
>>>>  >>> Are there any plans to update jdk ports in accordance with
>>>>  >>> recently updated versions (1.5.0_15, 1.6.0_05)? (since all
>>>>  >>> previous are already not available for download officially)
>>>>  >>>   >> A side note diablo also needs to be rebuilt to get rid of 
>>>> kse references.
>>>>  >
>>>>        > Very unlikely to happen unless someone wants to fund 
>>>> this.  You can't just
>>>>  > rebuild it, you've got to rebuild it and then run it through the 
>>>> JCK tests
>>>>  > again.  Time is much more likely to be invested in getting a 
>>>> Diablo release
>>>>  > for 1.6 done rather than updating the 1.5 release unless someone 
>>>> is paid
>>>>  > to do it.
>>>>
>>>>  Valid SoC project?  If so I will make an application for it. (I
>>>>  already applied for ports 2.0 but will make this my alt)
>>>>       
>>>     As I understand things, the obstacle is not "rebuild it" but
>>> "run it through the JCK tests again".  The latter takes time, it
>>> takes (five digits ?? of) money, and it takes one or more Official 
>>> People babysitting the process.
>>>     Diablo-1.5 happened in part because a third party thought it
>>> worth their effort to fund the effort.  Given a viable successor is
>>> already in the field, it would seem to ake sense for them to fund
>>> that ... if they fund anything at all.
>>>     
>>
>> The monetary cost of doing it I was speaking of was to pay someone to 
>> do it
>> (the Foundation has previously done this as contract work).  If the
>> Foundation were willing, a volunteer could do this.  The volunteer would
>> need to be ready for a significant time investment and be familiar
>> with the JDK build process and code.  All the current porting volunteer
>> time being spent on Diablo is being spent on 1.6.  The next most 
>> pressing
>> project that is ongoing has to do with OpenJDK, which will heopfully 
>> make
>> these binary releases obsolete to some extent.  Given that, its unlikely
>> that one of the current porting team is going to look at updating the
>> current 1.5 binary.
>>
>> I don't know if this is a valid SoC project or not since I'm not sure 
>> what
>> qualifies as one.  If you have to be able to show code at the end of it
>> then it isn't, since the code would be restricted.
>>
>>   
> My understanding is I just need to show a deliverable of some kind 
> (binary is fine).   I am familiar with the JDK build process but not 
> the code (but do have a solid course work background in VM's and 
> Compilor Theory as well I have read the JVM spec as well as the class 
> format spec as well as 25 years experience in general coding before 
> returning to school).  Since the deadline for appiling for SoC is 
> tonight I will file the application immediately (I can always revise 
> it).    Some things I will have to  guess at.   I will post a copy of 
> the application as a reply to this thread.   The way I understand it 
> someone like Greg would need to be my "mentor".
> _______________________________________________
> freebsd-java at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-java
> To unsubscribe, send any mail to "freebsd-java-unsubscribe at freebsd.org"
>
Here is the text of the application:

Name: Aryeh M. Friedman
Email: aryeh.friedman at gmail.com
Phone: +1 845 558 4829

Project Title: Diablo 1.5 porting to post-KSE kernels

Possible Mentor: Greg Lewis

Benefits to the FreeBSD Community:

    As outlined by Greg Lewis in order to port Diablo 1.5 to kernels 
that don't have KSE (and other more modern IPC/threading mechanisms).  
Requires the following:

     * Sigificant time commitment to bring the code upto date

     * Needs to pass the JCK test suite

Currently due to KSE's removal the only way to "boot strap" a native JDK 
is to install Diablo but only after making the approriate mappings in 
/etc/libmnap.conf which are largely undocumented and dependant on the OS 
version and build platform.

This means that a working JDK (required for Open-Office, Eclipse, Ant, 
Eclipse, etc. as well any working Java developer) on 7 (latest patch cvs 
version) or 8-CURRENT one needs to perform some rather undocumented and 
less then intutive manual adjustments.   The goal of this project is to 
rebuild/port Diablo 1.5 to not require these manual procedures.

Deliverables:

    A. Binaries for all supported platforms

    B. Record of the JCK test results for all supported platforms

    Note: I can only do i386 and AMD86 directly

Project Schedule:

    Total Time Comittment: 8 Weeks

    Can start: 6/1/08

    Key Milestones and deadlines:

        Milestone                    Deadline
        --------------------------------------------------------
        Build of Diablo 1.5 on 8-Current AMD64        Week 4
        Above passes JCK Tests                Week 6
        Rebuild and test for all supported platforms    Week 8

Availability:
    30 to 40 hrs/week avaible

    I have 1 (maybe 2) classes this summer.

Bio:

    I am a experienced UNIX developer (self taught) who has worked as a 
CS researcher and product developer (see resume, 
http://www.flosoft-systems.com/resume note it is about 4 years out of 
date).  I returned to school for two reasons:

    1. More and more employers insist on a BA/MA min. for even an interview

    2. Be able to rewrite some of my research in more formal form in 
pursuit of my PhD.

    I feel I am the person to do this project for the following reasons:

        1. I have used FreeBSD as my primary (only) Java development 
platform for 5 years thus am familure with the user level interface to 
the JVM Wrapper system as well use of the JDK 1.5 command line

        2. Broad experience in multiple languages and systems level 
coding has given me a view point that would be hard to find in a 
non-returning student.

    Summery of Programming Experience:

        * 25 Years of general coding (as a hobby and professionally)

        * 20 Years of userland coding experience with BSD

        * 5 Years experience with Java

        * Have been involved as a coder and/or designer in many projects 
that exceed 20k lines of code

        * Single developer on several 10+k line of code projects

    Other experience:

        * Techinical manager of 2 program teams
        * 20 Years of work in and/or arround open source projects

    Most recent/relivent course work:

        * Theory of Computation (aka Intro to Theortical CS at the grad 
level) (currently enrolled)
        * Compilor Theory (self taught)
        * Advanced Data Structures (including Graph Theory) (A)
        * Data structures (A-)
        * Alogrithm design/ananylsis (A)
        * Operating system development (B+)
        * Oracle DBA/RDBMS theory (A)

        Cumm. GPA: 3.75
        Total credits: 116 (96 from prior learning credit)

    Current professional interests:

        * Operating systems
        * Transport protocols
        * Optimal search algorithms (subuject of a pending patent)
        * Social/political effects of FOSS
        * Alternative licensing schemes that allow for better blending 
of FOSS and commerical projects
        * Wide scale use of persitent object spaces
        * Data representation
        * Multi-language/paradigm compilors
        * Agile methods