Perl 6 - the future is here, just unevenly distributed

IRC log for #phasers, 2011-02-01

| Channels | #phasers index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
06:58 sorear pre-report:
06:58 sorear DID:
06:59 sorear - merged a patch from moritz++ adding spectest support to niecza
06:59 sorear - improved internals documentation a wee bit
06:59 sorear - released v2, the first bootstrapped release
07:00 sorear - fixed an insideous bug with the new map optimization
07:00 sorear WILL DO:
07:00 sorear - study 6model in much greater detail
07:00 sorear - steal as much as I can
07:00 sorear BLOCKERS:
07:01 diakopter []
07:01 sorear - $reallife is placing a tuit demand
07:01 sorear EOR
07:01 sorear (will miss next 15 #phasers)
07:04 * diakopter wonders how out-of-phase you'll e
15:07 PerlJam left #phasers
15:07 PerlJam joined #phasers
18:12 colomon joined #phasers
18:27 PerlJam Apparently there was a some mutual unspoken agreement to not do #phasers today?  :)
18:37 colomon It's not for another 23 minutes.
18:38 PerlJam Hrm.
18:38 PerlJam my clock is an hour off
18:38 PerlJam That's okay, I never really know when #phasers is anyway unless someone mentions it.
18:38 masak joined #phasers
18:39 colomon left #phasers
18:39 colomon joined #phasers
18:40 colomon I meant to mention it on #perl6, but keep getting distracted.
18:40 * colomon has a 2.5 year old child running around who cannot decide if he is Winnie the Pooh or Prof. Harold Hill.
18:52 moritz_ Winnie the Prof!
18:54 tadzik joined #phasers
18:56 diakopter pre-report: DID: WILL DO: * BLOCKERS: EOR
18:56 moritz_ hm
18:56 * moritz_ tries to remember what he did
18:57 masak pre-report: more time-constrained by $work than usual. at least I've blogged a bit in the past week. plan to get a Yapsi release out the door tonight. EOR
18:57 moritz_ ah, right. Wrote a command line parser in nqp-rx/nom, to be usable in HLL::Compiler, Rakudo or other components (not feature complete, of course): https://github.com/moritz/nqp-cl-parser
18:58 moritz_ apart from that, trolled #perl6, #parrot and #parrot-dev
18:58 moritz_ blocking on: tuits, sleep.
18:58 moritz_ I want to find the motivation to further hack on book. Currently motivation nowhere near in sight
18:59 moritz_ plan to do: dunno
18:59 moritz_ EOR
18:59 masak +1 on motivation for book.
18:59 Util pre-report: working on .dmg for R* EOR
18:59 masak I have ideas for book now and then, but never get a round tuit.
19:00 colomon o/
19:01 Util \o
19:01 masak o/
19:01 moritz_ it's #phasers time now
19:01 jnthn o/
19:01 tadzik o/
19:01 masak it's that phase of the week again.
19:02 moritz_ any on-time reports? jnthn maybe?
19:02 * jnthn 's meeting ended just in time for £phasers
19:02 * PerlJam has his clock fixed now too
19:02 jnthn er, #phasers
19:02 jnthn moritz_: Give me a moment... :)
19:02 colomon I'll go. :)
19:02 colomon Played a lot with zavolaj++.
19:03 colomon biggest thing was Math::BigInt.
19:03 colomon three blog posts to go with those.
19:03 PerlJam colomon++
19:03 masak colomon++
19:03 moritz_ colomon++ indeed
19:03 colomon plan: clean up Math::BigInt to make it easier for others to use.  Implement Math::FatRat.  Talk to pmichaud about getting some sort of BigInt working in Rakudo.
19:03 colomon EOR
19:03 jnthn colomon++
19:04 * jnthn can go
19:04 jnthn No pre-typed report...but not done a lot this week...
19:04 jnthn * Refactored REPR API a bit in prep for native types support
19:05 jnthn * Got enough done so native types are usable from PIR
19:05 jnthn * Got Cursor to use them
19:05 jnthn * Quite a good test case for their robustness ;)
19:05 jnthn * Talked with pmichaud++ about how to get them supported from HLL level; have a plan
19:05 jnthn * Design work on roles in 6model
19:06 jnthn * Uncommited patches to get us creating meta-objects at compile time in NQP as a first step towards the roles bits
19:06 jnthn Blockers
19:07 jnthn * $dayjob a little more demanding than normal this week, but that's a temporary glitch. (Am away from home, teaching a course. Next week, things are back to normal.)
19:07 jnthn Plans
19:07 jnthn * Land roles in nqp-rx
19:07 jnthn * Branch Rakudo and start hacking on new meta-objects for it
19:07 jnthn * Worry about lots of things
19:07 jnthn EOR
19:07 masak jnthn++
19:07 moritz_ jnthn++ indeed
19:08 colomon jnthn++
19:08 moritz_ "note done a lot", but still about a screen full of report :-)
19:08 moritz_ tadzik: have a report?
19:08 jnthn :)
19:08 jnthn What was done was worth talking about. ;)
19:09 moritz_ I know [Coke]++ has had a tour through RT
19:09 moritz_ or anybody else got a report?
19:09 pmichaud I do.
19:09 moritz_ \o
19:09 tadzik moritz_: kinda
19:09 moritz_ then go
19:09 moritz_ let's say pmichaud first
19:09 pmichaud I developed NQP plans with jnthn++ and others, then wrote an article on NQP Roadmap
19:10 pmichaud I also participated heavily in PDS 2010 and wrote an article on Rakudo needs from Parrot for 2011
19:10 jnthn (phone, bbs)
19:10 pmichaud This week (as in the next couple of days) I'm planning to move nqp-nom out of the nqp-rx repository and into its own repo
19:10 pmichaud then I'm going to write up a few benchmarking utilities so we can do some reasonable speed comparisons over time
19:10 pmichaud end-of-report
19:11 colomon pmichaud++
19:11 moritz_ pmichaud++
19:11 moritz_ tadzik?
19:11 masak pmichaud++
19:11 tadzik well, not much to say
19:11 [Coke] o/
19:12 moritz_ ok
19:12 moritz_ I've got a question, mainly for pmichaud and jnthn
19:12 tadzik DID: Encouraged masak++ to move his modules to the new infrastucture, >50% of modules.perl6.org is now available. Thought a bit more about keeping the module database, downloading multiple files from multiple sources is a farce with a current performance of rakudo
19:13 [Coke] I've a note.
19:13 tadzik PLANS: When BLOCKERS are gone away, I plan to begin merging the new ecosystem, and possibly launch some modules-related webservices on feather (testing, rsses, etc)
19:14 tadzik BLOCKERS: Exam session on the Uni
19:14 tadzik EOR
19:14 moritz_ tadzik++
19:14 moritz_ [Coke]: please go ahead
19:14 masak tadzik++
19:15 tadzik oh, I've done nothing :)
19:15 colomon tadzik: I would have thought downloading multiple files from multiple sources is a task where Rakudo's slowness isn't a big hinderance?
19:17 tadzik colomon: it suprised me too. Iterating over an array and launching wget every times takes ages, almost literally. For a few times I encountered a github downtime during the update, and had to start over again
19:17 jnthn (back)
19:17 tadzik may be the JSON parsing those files though
19:18 moritz_ might be... it's not optimized for speed
19:18 moritz_ anyway, my question
19:18 moritz_ the new command line parser has an API that's independent of HLL::Compiler
19:19 moritz_ and quite a bit different (because I didn't grok HLL::Compiler's mixing of stages and options)
19:19 moritz_ so, how should I proceed? add it as a separate class to nqp, and then write a thin wrapper in HLL::Compiler?
19:19 pmichaud I'm happy for HLL::Compiler's current staging mechanism to go away.
19:20 moritz_ fwiw, it doesn't do any staging
19:20 moritz_ just ordinary command line parsing
19:20 pmichaud I've never liked the current system of staging
19:20 pmichaud anyway, separate class in NQP is fine with me, hopefully under the HLL:: namespace
19:20 pmichaud HLL:: is the set of utility stuff for hll implementors
19:20 moritz_ with a few options suitable for compilers, for example the option to ignore any options after -e or the first argument
19:20 pmichaud thus HLL::Grammar and HLL::Compiler
19:21 moritz_ pmichaud: names can be easily adopted
19:22 moritz_ HL::CommandLine::Parser and HLL::CommandLine::Result or so
19:23 moritz_ so, I'll add it to nqp, and see what else we'll do with it.
19:23 moritz_ EOQ.
19:23 pmichaud +1
19:24 jnthn moritz_++
19:24 jnthn moritz_: Does it just take an array of strings and work from those?
19:24 pmichaud fwiw, my idea for a command line parser was to have an array of signature-specifications
19:25 jnthn <scared look>
19:25 moritz_ jnthn: it's a class; it takes a list of strings that specify options
19:25 jnthn :)
19:25 moritz_ and you can also add some methods to change behaviour
19:25 pmichaud well, since that's what Rakudo ends up doing with MAIN as a multi
19:25 jnthn moritz_: OK, but I meant more does it process @*ARGS, or argv, or whatever the raw array of things ends up coming in as?
19:25 moritz_ jnthn: just an array coming in
19:25 jnthn moritz_: Great. :)
19:26 moritz_ and it doesn't destroy that array, as other parsers do
19:26 pmichaud +1
19:26 moritz_ it returns an object that has a hash for options, and an array for arguments
19:26 pmichaud sounds a lot like a match object :)
19:26 moritz_ pmichaud: don't even think of it :-)
19:26 pmichaud or some other form of Capture
19:27 moritz_ currently it's has-a, not is-a
19:27 jnthn .oO( He who writes a grammar engine thinks of everything as match objects :-) )
19:27 pmichaud .oO( He simply follows in the lead of the language designer. )
19:27 jnthn :P
19:27 pmichaud .oO( who also tends to think of things in terms of Captures these days :)
19:27 PerlJam more like "pattern matching is fundamental"  :)
19:27 pmichaud Exactly
19:27 jnthn Darn, why does Pm always make such good points. :)
19:27 pmichaud that's kind of a motto for Perl 6 -- it's all based on pattern matching of various types
19:30 jnthn pmichaud: What do you plan to work on after the nqp shuffling?
19:30 jnthn pmichaud: I'm thinking if that's iminent I'll wait for it before making the rakudo/nom branch. :)
19:30 pmichaud jnthn: I suggest waiting, yes.
19:31 jnthn OK, will do.
19:31 pmichaud because rakudo/nom really needs to be built on NQP_REVISION and not PARROT_REVISION
19:31 jnthn Right.
19:31 * jnthn wonders if pmichaud++ might set that bit up... ;-)
19:31 pmichaud after nqp shuffling I'm still planning to focus on benchmarking and working on nqp itself
19:31 jnthn I promise to write awesome meta-objects in return. ;-)
19:32 jnthn <- not good at config-y stuff :)
19:32 jnthn Anyways, I'll be finishing up the roles bits in nqp/nom in the immediate coming days.
19:37 Util q1q
19:38 moritz_ Util: I think you can go rigth ahead
19:39 Util Does Rakudo (or R*) come with any demo code?
19:39 Util i.e. "Hey end-user, run this and see what happens, then look at the actual code if you care to see how we did it."
19:39 PerlJam not exactly
19:40 PerlJam It comes with the Using Perl 6 book which has some code they can try out
19:40 PerlJam There is no other tutorial though
19:40 Util Working on the .dmg for R*, and hoping for someone to do Win32 and Linux, I see a need for demos that can be run without even needing to know how to open a text editor; already coded in individual files (though the book is a great include in itself)
19:42 Util I will write down my thoughts and a few examples, and post to the ML.
19:42 [Coke] sorry, $DAYJOBing, missed my chance. ;)
19:43 [Coke] before I forget: we have a tag on RT for tickets that need testing: "testneeded".
19:43 masak \o/
19:43 jnthn I thought that tag was called "moritz" ;-)
19:44 jnthn testneeded++ :)
19:44 masak and a URL to show a list of tickets marked with that tag. it's in the #perl6 backlog of today.
19:44 jnthn oh, cool
19:44 jnthn Maybe we could link it from the how to help page on perl6.org?
19:44 jnthn "Want to write tests? Here's some inspiration of what we need!"
19:44 masak +1
19:45 colomon +1
19:47 PerlJam +*
19:56 [Coke] I was happy to see that several people jumped in writing tests after I identified a few tickets that were otherwise closable.
20:02 masak #phaser ended for today?
20:04 moritz_ seems like
20:19 tylercurtis joined #phasers
20:41 masak left #phasers
23:40 tylercurtis left #phasers

| Channels | #phasers index | Today | | Search | Google Search | Plain-Text | summary