Friday November 30, 2007
Perl 6 Design Minutes for 14 November 2007
The Perl 6 design team met by phone on 14 November 2007. Larry, Patrick,
Nicholas, and Jesse attended.
- Still trying to work out the longest token matching stuff.
- Going a little slow because I've been fighting lack of sleep from tendinitis and medications, and having to upgrade infrastructure and e-mail.
- New spam filter caused a DOS attack - Installed ClamAV and new SpamAssassin rules, then someone sent about 100 identical messages which completely bogged down SpamAssassin.
- I had to write a blacklist thing that went through lastlog and blacklist those IPs
- Got the Fedora upgrade and several other things done, waiting for my brain to come back.
- Is your hardware holding up?
- USB being a little flaky, but doesn't matter as I don't use USB much. The space bar is gimpy
- Doing stuff that is optimised for my right hand, ie doesn't need as much typing.
- Also thinking about stuff people have sent me, such as Damian.
- In terms of real Perl 6 work, primarily just thinking
- Things are gelling better, or still in the stewing phase?
- I know I could do it if I had a whole brain, but I don't have enough of a brain to fit it in right now.
- Got the source to egrep, and all sorts of things that do this sort of thing
- Want a perl specific solution to it -- egrep just tells you line matches, doesn't tell you how it matches.
- Want the algorithm specified in Perl 6, even if the eventual compiler does it some other way
- Don't know how fast it will be, but the algorithm is designed to examine each character only once, so shouldn't be too slow.
- Get it correct, then make it fast.
- Hard to explain to computer what I mean.
- If only we had a good language for that
- Ended up with infrastructure problems too.
- For work, I had to reinstall the OS to get Adobe Captivate.
- Turns out that the Parrot testing facility doesn't work with the latest Test mods from CPAN.
- Possibly. It's been refactored somehow?
- - More re-implemented from scratch.
- First hint was that the smoke tests didn't run.
- Got help on IRC; we tracked it down to a missing dependency.
- Fixed that.
- Now make test isn't running
- Just filed a ticket; I expect someone from #parrot will pick it up and have it solved within an hour
- Right now I think that it's because harness is passing an option to parrot that parrot doesn't understand.
- One of the authors of Test::Harness 3 is Andy Armstrong, AndyA; you may be able to find him on IRC.
- Talked about a bunch of stuff last week
- I still need to format my part.
- Hoped to have new NPQ checked in before I blogged it, but not going to hold the blog post up for it.
- One way or another it will go out tonight: list message, use.perl journal entry, anywhere else you can think.
- For a use.perl journal entry, there's a tickbox, which submits the entry to the moderation queue for the front page.
- Last week talked to Patrick and several of the other authors about the state of various implementations [of Perl 6], to (not quite) get a blog post together.
- Haven't done the step of blog post, but everything else is done and is in public.
- In the IRC logs of #perl6, so not "published" but is Google-able.
- because I want to refer to it in a post to perl6-compiler.
Nicholas asked about state and
START blocks, which Larry
answered with a concise explanation about context, which we agreed would be
better written out in e-mail to perl6-language), given that vocally it's a
tongue twister of symbols.