Portability

[Stephen Weeks]

On Wed, 2009-10-07 at 21:17 +0100, Philip Herron wrote:
> Hey guys,
> 
> Since my last thread, i am quite interested in parrot now, i've been
> playing with it a little i've stopped my gcc work to concentrate on
> this and my programming language. But taking a peek around not sure
> where to get invlovled i mean i've been looking at some bug fixes but
> i am not sure where i would be more usefull. I hold access to the gnu
> compile farm and wondering do you do much testing on more random
> architecures like alpha or mips or sparc? Maybe thats were i could be
> usefull. But if you have work in garbage collection or misc internals;
> or i saw someone working on an LLVM backend sounds pretty cool. Any
> suggestions would be great.
> 
> --Phil

Welcome to the project, Phil!

Smoke reports from less-common architectures would be very nice.  The
main problem there is availability of developers with access to those
architectures with knowledge and interest in fixing problems on those
architectures.  It would certainly be useful to have, though.

For GC and JIT, you'll want to talk to WhiteKnight, who iirc has been
working on plans for both of those.  We have *a* GC system, but last I
heard it needed to be generalized to be pluggable.  I don't remember if
that was finished or not.  After it's pluggable, we really want to have
some better GC algorithms implemented.  Our current GC is less awesome
than we would like.  WhiteKnight's blog has information about his GC
plans as well.  For JIT, we currently have a declared goal of a
pluggable JIT system, with llvm as our first backend.  We currently do
not even have a declared design for a pluggable JIT, but I believe
WhiteKnight has significant thoughts about the subject, and probably a
few others.  It would be great to see work start on that.  Again, a good
place to start is WhiteKnight's blog.

http://wknight8111.blogspot.com/search/label/Parrot

The other internals project that needs a lot of attention soon is a
replacement for our PIR compiler.  There's a partially-completed PIR
compiler called "pirc" in compilers/pirc/ in the Parrot SVN repo that
has a much nicer design than our current aging behemoth imcc, but it's
unfinished.  If you're looking for a project without a current champion,
that wouldn't be a bad choice, but might have a heavy startup cost.

This mail isn't meant to be exhaustive, but I think it covers the
most-significant areas you might be interested in.

If you're going to want commit access to the SVN repo, which I expect
you will, you'll want to first submit some patches that are accepted
into the repo, then submit a CLA (committer licensing agreement) to the
Parrot Foundation assigning copyright to them, or something.  I don't
remember exactly what it says. ;)  I look forward to seeing your
patches, and hearing what you choose to work on.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://lists.parrot.org/pipermail/parrot-dev/attachments/20091007/720c55ac/attachment.pgp>