GSoC 2011 Proposal: Bytecode Emitters for POST
benji at silverinsanity.com
Sun Apr 3 16:36:42 UTC 2011
Are you referring to the nqp_pct branch? Since that was linked to from the task description on the GSoC ideas page, I was under the impression that it was more of a rewrite of PCT using NQP than the ability for POST to emit PBC. Looking over it now with what I learned writing my proposal, I see that it does include a to_pbc method towards the end of POST::Compiler.
From the way you talk about it, I take it that the branch is incomplete. Which leads me to ask several questions:
* I don't see any examples of use of "newPOST" or any tests for it, is there documentation for it somewhere?
* Does PAST in that branch do anything useful, or does it call code that no longer functions?
* What would need to be done to make it functional?
* Is that work appropriate for a GSoC project?
I will try to find time to write a proposal for the LALR grammar task that I was originally talking about doing. Unfortunately, it took me most of a day to write my original proposal and I have a significant amount of schoolwork to complete. Might I suggest that for next year tasks that are already mostly complete do not get added to the ideas list?
~~ Brian Gernhardt
On Apr 3, 2011, at 5:41 AM, Vasily Chekalkin wrote:
> Actually, there is next good proposal for emitting PBC from POST -
> make PAST emit "newPOST" and finally glue everyting together. In the
> perfect world - including PIRATE for parsing <inline> nodes on fly.
> On Sun, Apr 3, 2011 at 6:28 PM, Jonathan "Duke" Leto <jonathan at leto.net> wrote:
>> Don't let this discourage you. I still highly recommend that you
>> submit another proposal idea.
>> On Sun, Apr 3, 2011 at 12:11 AM, Vasily Chekalkin <bacek at bacek.com> wrote:
>>> 1. It's really good plan.
>>> 2. Unfortunatelly all of this is already implemented (mostly) in
>>> PIRATE (parsing and emitting newPOST for PIR) and pct_pbc branch (for
>>> "library to emit PBC from POST directly").
More information about the parrot-dev