The Perl 6 design team met by phone on 22 April 2009. Allison, Jerry, Will, Patrick, Nicholas, and chromatic attended.
Allison:
- worked mainly on the book this week
- reviewing patches, preparing for the release
- did more work on the calling conventions branch
- you have to update all of the last scattered bits throughout the cold
- gave a talk to the SF LUG last night
- good questions there
Jerry:
- worked on GSoC
- TPF has nine slots this year
- two projects are for Parrot and one is for Perl 6
- we originally had 10, but one student had duplicate proposals into another organization
- all of the mentors and students are excited and ready to go
- we're looking forward to the start of coding in May
- getting everyone set up with the proper credentials and prerequisites for commits
- will have a good amount of time to the Perl 6 command line in the near future
Patrick:
- spent the last week at the Nordic Perl Workshop and a subsequent hackathon
- the workshop was excellent and well-organized
- a two-day track for Perl 6 and Parrot
- quite a bit of enthusiasm from attendees about Perl 6 and Parrot
- the initial guesses about attendance were wrong; we had to switch to bigger rooms for the Perl 6 stuff
- LinPro sponsored the three-day hackathon
- it went extremely well
- Larry and I attended Gabor's Perl 6 introduction
- I wanted to see where people would have trouble with Rakudo and the like
- other people worked on various modules in Perl 6
- we now have socket support in Rakudo
- people can write web servers and clients in pure Perl 6, no PIR
- quite a few people asked for something to do
- we gave them challenges that they might not be able to complete
- most of them all worked out though
- we found lots of little bigs
- we found some conceptual problems
- we implemented a lot of new features
- I worked with Gabor on integration of syntax highlighting of Rakudo in Padre
- PCT now provides the information Padre needs
- any PCT-based language could quickly get syntax highlighting through Padre
- that works out very well
- it took some time to get the design right, but Gabor took it from there and started doing amazing things
- Jonathan and I cornered Larry on some issues
- we kept asking him with hard questions and frustrated him with our ideas
- at one point, he started banging his shoe on the table to get our attention
- the
prefix:= operator is gone -- iterate an iterator operator
Jerry:
- I'm going to miss the prefix fish operator!
Patrick:
- I'm going to write about it
- when we were trying to nail down laziness in Perl 6 and its meaning, we realized that prefix:
<=> was trying to solve two different problems
- they're contextually related
- but multidispatch makes getting context information more difficult
- people in the tutorial found the syntax ugly, especially in a scalar assignment
- reading a single line from a POD filehandle was even uglier:
my $x = =$=podhandle;
- it's now
.get for a single item and .lines to read in list context
- that makes the code much easier to read in general
- the fish operator is now
lines()
- it's more readable for people who don't know Perl
- as Masak has previously noted, whenever someone mentions the need for something (the desire for golf or obfuscation), he says "There will be modules."
- we managed to put Larry's ideas about binding in place
- that's been difficult in Parrot with PCT and PMCs as currently designed
- I think things look better and simpler for us now
- Larry seemed declined to support some things we thought we might have to support, in particular certain binding features
- we'll clean up the binding syntax in Rakudo over the next couple of months
- Jonathan and I spent one evening reviewing and updating the Copenhagen roadmap
- I'll turn that into some readable prose
- we'll have tha sometime this week
- Jonathan finished his grant review and is waiting review from his grant manager
c:
- oh, I know what that means...
- I'll get to that
Patrick:
- he's busy working on his next grant proposal
- we spent a good amount of time clarifying things so he could finish up his grant
- that largely happened at the hackathon
- we made some PGE improvements
- one participant was interested in quantifiers, so he did some of the hard parts of figuring out how to parse regexp quantifiers
- fixed that in a few hours one day
- a feature for people writing parsers
- that closed a couple of RTs
- I'll post in detail about the hackathon happenings in the next couple of days
- it was very productive
- lots of good energy from the participants
- we had at least five or six people just hacking with Rakudo Perl
- someone put together a pretty good XML parser
- we also talked quite a bit about module distribution for Perl 6
- no concrete plans, just broad brushstrokes
- we're looking at something more distributed, like Git and GitHub
- Masak has something in the works with his Proto module
Nicholas:
- CPAN solves uploading and mirroring
- it doesn't solve the distro problem as much
- mapping uploads to different people and such
- how do you lay things out on disk and figure out how to load them?
Patrick:
- we didn't get to that level of detail
- we're keeping it in mind as we go on
- it'll come up sometime over the summer
- very likely a YAPC::EU topic
- working on Rakudo's release
- expect to get that done in the next few hours
- a little early, as we have several pending patches
- no reason to hold the release in that case
c:
Patrick:
- ~3190 new passing tests since the last release
Jerry:
- let's see you do that next month!
c:
Patrick:
- 2100 were Unicode tests
- that's still over 1000 tests in the past month
- several people asked how to port the stuff I did from Rakudo back into PGE
- even without Unicode, it was a significant month
- we'll have trouble matching this month's progress
- we're passing tests people have already written
- but people are adding new tests
- the size of the test suite is growing
- we're catching up
- I'm almost to the point where we should look not at absolute numbers but at the percentage of the test suite
Nicholas:
- that number's more meaningful externally
- are you above or below 50%?
Patrick:
- we're above 60%
- 100% is a receding number though
- some parts of the spec are woefully undertested
- the next big chunk is fixing named argument passing
- we can't handle all of those in the right way all of the time
Will:
- Parrot 1.1 came out yesterday
- general station keeping with coding standards tests and ticket wrangling
- I think I fixed an IMCC bug which blocked lexicals with Unicode names
- was a blocker for Rakudo
- hope to steal some of Allison's energy to get Partcl building against an installed Parrot
c:
- merged in some of the header cleanup work
- have a lot left to go
- will be editing the Parrot book in hopes of publishing it very soon
Will:
- the PASM chapter has a lot of bad code in it
Allison:
