harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Senaka Fernando" <senaka...@gmail.com>
Subject Re: GSOC 2008: harmony-gc-5
Date Sat, 29 Mar 2008 16:47:49 GMT
Hi Alexei,

I was reading into gc_gen. Actually it seems that you are referring to an
archived gc, I believe you meant gc_v4 here, isn't it?

Regards,
Senaka


On Sat, Mar 29, 2008 at 7:48 PM, Xiao-Feng Li <xiaofeng.li@gmail.com> wrote:

> On Fri, Mar 28, 2008 at 7:16 PM, Senaka Fernando <senakafdo@gmail.com>
> wrote:
> > Hi Alexei,
> >
> >  Thanks.
> >
> >  Regards,
> >  Senaka
> >
>
> Hi, Senaka, I am catching up the emails. The discussions below were
> interesting. You can take the mark-sweep GC of Harmony as the starting
> point. It's a parallel GC, and can optional compact the heap under
> serious heap fragmentation. You are encouraged to apply for the
> project. There could be multiple applicants though...
>
> Thanks,
> xiaofeng
>
> >  On Fri, Mar 28, 2008 at 3:24 PM, Alexei Fedotov <
> alexei.fedotov@gmail.com>
> >
> >
> > wrote:
> >
> >  > Senaka,
> >  > Thanks for your enthusiasm and good luck.
> >  >
> >  > It turns out that Xiao Feng is silent due to his vacations. I've
> asked
> >  > a guy who works with him to remind him about official mentor
> >  > registration. Hope he either is already registered or will do it
> soon.
> >  >
> >  > Thanks.
> >  >
> >  > On Fri, Mar 28, 2008 at 10:39 AM, Senaka Fernando <
> senakafdo@gmail.com>
> >  > wrote:
> >  > > Hi Alexei,
> >  > >
> >  > >
> >  > >  On Fri, Mar 28, 2008 at 3:15 AM, Alexei Fedotov <
> >  > alexei.fedotov@gmail.com>
> >  > >  wrote:
> >  > >
> >  > >
> >  > >  > Parrot community have a pretty interesting and resourceful test
> base
> >  > >  > including tests which run regular expressions against their code
> >  > which
> >  > >  > fail when the code style is not good. The only stress test they
> have
> >  > >  > is fortunately a test against GC [1], so it may be a good start.
> >  > >  > Generally this would be a good idea to address these questions
> to
> >  > >  > their list and ask them for cross-mentoring. Building good
> relations
> >  > >  > and understanding their code of conduct would ease acceptance of
> your
> >  > >  > work later.
> >  > >
> >  > >
> >  > >  +1, for the thought. I will start this discussion on their list,
> too.
> >  > >
> >  > >
> >  > >
> >  > >  >
> >  > >  > As for the porting layer, Maria tried merging Harmony and Parrot
> >  > >  > porting layers and faced difficulties adjusting our hythread
> layer
> >  > >  > based on port layer based on apr layer based on OS layer to
> parrot
> >  > >
> >  > >
> >  > >  I'm familiar with apr source so I can use it as a base from which
> i
> >  > >  investigate the port layer. I believe that a ground-up approach
> could
> >  > rather
> >  > >  be appealing.
> >  > >
> >  > >
> >  > >
> >  > >  >
> >  > >  > based on source configure-like preprocessing. Another challenge
> is a
> >  > >  > pure C nature of Parrot code which makes embedding our sources
> >  > tainted
> >  > >  > with C++ painful. For me it seems feasible to build GC into a
> >  > separate
> >  > >
> >  > >
> >  > >  There is the possibility of Class Loading to solve this issue.
> This
> >  > requires
> >  > >  defining an exportable C interface so that Parrot wouldn't need to
> >  > bother
> >  > >  about the Harmony source.
> >  > >
> >  > >
> >  > >
> >  > >  >
> >  > >  > dynamic library using Harmony build system, and plug it into
> Parrot
> >  > >  > using a binary DLL interface you have to tune.
> >  > >
> >  > >
> >  > >  I'm interested in taking this project. This sounds very
> interesting.
> >  > >
> >  > >
> >  > >
> >  > >
> >  > >  >
> >  > >  >
> >  > >  > [1] parrot/t/stress/gc.t
> >  > >  >
> >  > >  > On Thu, Mar 27, 2008 at 11:21 PM, Senaka Fernando <
> >  > senakafdo@gmail.com>
> >  > >  > wrote:
> >  > >  > > Hi Alexei,
> >  > >  > >
> >  > >  > >  On Fri, Mar 28, 2008 at 1:17 AM, Alexei Fedotov <
> >  > >  > alexei.fedotov@gmail.com>
> >  > >  > >  wrote:
> >  > >  > >
> >  > >  > >  > Hello Senaka,
> >  > >  > >
> >  > >  > >
> >  > >  > >  Thanks for the prompt reply.
> >  > >  > >
> >  > >  > >
> >  > >  > >  >
> >  > >  > >  > Welcome to the community. If I would set expectations
for
> >  > >  > >  > harmony-gc-5, I would suggest that Parrot VM test base
> passes on
> >  > >  > >  > Parrot with XFL garbage collector plugged in.
> >  > >  > >
> >  > >  > >
> >  > >  > >  Since there is a test base I believe I that it would be much
> >  > easier to
> >  > >  > track
> >  > >  > >  the flow of my work and understand the expectation. Do you
> believe
> >  > that
> >  > >  > the
> >  > >  > >  Parrot VM test base would cover all the necessary
> requirements so
> >  > that
> >  > >  > I can
> >  > >  > >  take a look of what is expected?
> >  > >  > >
> >  > >  > >  Also, I would like to know whether there are any JIRA issues,
> test
> >  > >  > branches,
> >  > >  > >  etc. or any similar initiatives that you've taken regarding
> this
> >  > >  > feature.
> >  > >  > >
> >  > >  > >
> >  > >  > >
> >  > >  > >  > This includes subtasks
> >  > >  > >  > such as unplugging Harmony GC from DRLVM, adjusting
GC
> interface
> >  > and
> >  > >  > >  > code to work with object layout coded in Parrot PMCs*
> instead of
> >  > >  > >  > layouts accessed via java object virtual tables and
> resolving
> >  > porting
> >  > >  > >  > layer and linking conflicts.
> >  > >  > >
> >  > >  > >
> >  > >  > >  I can rather target these tasks as objectives. Btw, what
> typically
> >  > >  > >  constitutes the porting layer?
> >  > >  > >
> >  > >  > >
> >  > >  > >
> >  > >  > >  >
> >  > >  > >  >
> >  > >  > >  > Thank you for your interest!
> >  > >  > >  >
> >  > >  > >  > * Which are very different from Apache PMCs.
> >  > >  > >
> >  > >  > >
> >  > >  > >  :)
> >  > >  > >
> >  > >  > >  Regards,
> >  > >  > >  Senaka
> >  > >  > >
> >  > >  > >
> >  > >  > >
> >  > >  > >
> >  > >  > >  >
> >  > >  > >  > On Thu, Mar 27, 2008 at 9:55 PM, Senaka Fernando <
> >  > senakafdo@gmail.com
> >  > >  > >
> >  > >  > >  > wrote:
> >  > >  > >  > > Hi Xiaofeng,
> >  > >  > >  > >
> >  > >  > >  > >  I'm a 3rd year university student with a C/C++
> background and
> >  > also
> >  > >  > a
> >  > >  > >  > >  reasonable amount of knowledge in Perl and Java.
I'm
> >  > interested in
> >  > >  > >  > working
> >  > >  > >  > >  on the "*Build a garbage collector for C/C++ programs
on
> the
> >  > top
> >  > >  > of
> >  > >  > >  > Harmony"
> >  > >  > >  > >  project (harmony-gc-5). I would like to know your
> >  > expectations
> >  > >  > >  > regarding
> >  > >  > >  > >  this.
> >  > >  > >  > >
> >  > >  > >  > >  I have read through the Markmail archives of the
Harmony
> dev
> >  > list,
> >  > >  > and
> >  > >  > >  > also
> >  > >  > >  > >  the discussion at [1]. And, would like to know
a bit
> more on
> >  > this
> >  > >  > >  > project.
> >  > >  > >  > >
> >  > >  > >  > >  [1]
> >  > http://www.perl6.org/perl6-internals/2007-10/msg00091.html
> >  > >  > >  > >
> >  > >  > >  > >  Thanks,
> >  > >  > >  > >
> >  > >  > >  > >  Senaka Fernando,
> >  > >  > >  > >  Dept. of Computer Science & Engineering,
> >  > >  > >  > >  University of Moratuwa,
> >  > >  > >  > >  Sri Lanka
> >  > >  > >  > >  *
> >  > >  > >  > >
> >  > >  > >  >
> >  > >  > >  >
> >  > >  > >  >
> >  > >  > >  > --
> >  > >  > >  > With best regards,
> >  > >  > >  > Alexei
> >  > >  > >  >
> >  > >  > >
> >  > >  >
> >  > >  >
> >  > >  >
> >  > >  > --
> >  > >  > With best regards,
> >  > >  > Alexei
> >  > >  >
> >  > >
> >  >
> >  >
> >  >
> >  > --
> >  > With best regards,
> >  > Alexei
> >  >
> >
>
>
>
> --
> http://xiao-feng.blogspot.com
>

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