This week’s “most commits across all repositories” award doubtlessly goes to lizmat, who’s been speccing, implementing, and writing tests for supply operations left and right! Here’s what else is new:
- FROGGS completed support for loop labels on both NQP and Perl 6 on both the Parrot and MoarVM backends. The code is still in the “loop_labels” branches in each of the repositories, but with it you can assign a label to different kinds of loops and then supply that label to next/last/…
- dwarring added yet more advent calendar tests to the test suite.
- jnthn gave MoarVM a Repr that can be used to differentiate between a low-level null pointer and a null object. This is then used to make the allocation of $!, $/ and $_ lazy. We have to allocate these for routines, blocks, loops etc, but the user doesn’t use them in every case.
- The bytecode specializer of MoarVM is currently learning a few more tricks that make it more useful for Perl 6 code; Currently it only really benefits NQP code. That improves compile speed and build times, but not really run speed.
- jnthn has created a release candidate for Rakudo Star 2014.04 – It may be released and announced this evening or tomorrow or something. Watch rakudo.org or subscribe to the perl6-announce mailing list for the announcement.
- Thanks to donaldh, rakudo-jvm now also builds the IO::Socket::Async portion of the Core Setting.
- colomon wrote a bit on tadzik’s rakudobrew on his blog.
- Currently, the “community bonding period” of GSoC is happening.
- For Parrot’s GSoC project to optimize method signatures, the student is improving the handling of GC Write Barriers.
- brrt, who is going to implement a dynasm based JIT for MoarVM, has been toying with dynasm and learning the ins and outs of spesh.
I’m still waiting for the recordings from NLPW2014 to surface so I can hear the audio track that goes with jnthn’s slides I linked to in my last post.
I hope y’all have a lovely week :) The weather where I’m at is certainly pretty lovely!