First of all, let me apologize for posting this week’s post so late in the day. I’ve spent all day helping friends renovate the apartment they want to move into until the end of the month.
The PLPW in Poznan, Poland, has taken place over the weekend. Here’s a link to the list of talks again. Filip Sergot announced, that the talk recordings will be up on youtube soon. Here’s the slides of tadzik’s talk about game development in Perl 6 and his lightning talk about the lovely rakudobrew utility.
The Czech Perl Workshop 2014 is starting real soon. Here’s a link to the talks list.
Other things I’ve seen:
- There’s now a NativeCall distribution on CPAN for Perl 6, but Panda doesn’t know how to use CPAN yet.
- The patch to Rakudo’s source code to make slangs work a bit better is now merged into master. FROGGS’ v5 module will soon be Panda-installable, too, to make use of that. I suspect we’ll be putting v5 into Rakudo Star, too.
- Bitwise operations on bignums were fixed on MoarVM thanks to skids.
- ymd-from-daycount was broken on Parrot due to a broken implementation of bitwise shifts. Now there’s a workaround in place, thanks to dwarring.
- FROGGS fixed my-scoped regexes inside of classes.
- FROGGS also finished his work on loop labels for NQP and Rakudo on all three backends. The patches will be reviewed before merging.
- Thanks to significant work by lizmat, $*KERNEL, $*DISTRO, and $*VM are now full-blown objects.
- lizmat implemented subtests for Rakudo’s Test.pm to make TODOing tests more robust and also to make test output prettier.
- jnthn made NQP able to turn simple for loops into much simpler code, which should result in a speed gain. I didn’t measure it yet, though. This helps on all backends.
- On top of that, $_ can now also be turned into a local instead of lexical variable on all back-ends, giving more optimization (and specialization) opportunities.
- There’s also sundry improvements in spesh.
- Nicholas Clark improved the handling of Serialization Context IDs that belong to objects.
- If I understand correctly, the bug where MoarVM incorrectly turned nums into strings was fixed. Turns out it was forgetting about the ‘e’ inside scientific notation.
- donaldh made sure nqp::seekfh and nqp::readfh work correctly on JVM and Parrot and added tests.
- flussence prettied up the HTML source of perl6.org.
Since I’ve been busy all day today and yesterday and will be doing the same tomorrow and perhaps even the day after that, I’ve probably missed some interesting things and will not be able to immediately fix it. Sorry about that!
Anyway, I wish you all a pleasant week :)