Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2011-02-01

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

All times shown according to UTC.

Time Nick Message
01:26 chromatic left #parrotsketch
02:16 whiteknight left #parrotsketch
02:43 bluescreen left #parrotsketch
05:42 allison left #parrotsketch
06:06 particle1 joined #parrotsketch
06:09 particle left #parrotsketch
08:10 allison joined #parrotsketch
08:52 contingencyplan left #parrotsketch
14:32 bluescreen joined #parrotsketch
15:07 PerlJam left #parrotsketch
15:07 PerlJam joined #parrotsketch
15:16 bluescreen left #parrotsketch
15:24 bluescreen joined #parrotsketch
16:09 bluescreen left #parrotsketch
17:08 particle1 is now known as particle
17:24 bluescreen joined #parrotsketch
17:53 sheant joined #parrotsketch
18:41 bluescreen left #parrotsketch
18:50 NotFound joined #parrotsketch
18:52 chromatic joined #parrotsketch
19:32 contingencyplan joined #parrotsketch
20:01 Util # Done:
20:01 Util * Participated in PDS.
20:01 Util * Email discussion on NQP programs within the Parrot build.
20:01 Util * Working on OS X packaging (.dmg) for Rakudo Star.
20:01 Util # Plan to do:
20:01 Util * Finish .dmg for this month's R*
20:01 Util # Blockers:
20:01 Util * Tuit supply low, but growing.
20:01 Util # 7-day ticket report:
20:01 Util 3  closed: done
20:01 Util 13  closed: fixed
20:01 Util 3  closed: wontfix
20:01 Util 5  new
20:01 Util 2  reopened
20:01 Util .end
20:19 tcurtis joined #parrotsketch
20:24 cotto_work *did:
20:24 cotto_work - closed some tickets, reviewed others, rejected most ideas in MakeEveryPMCAnObject
20:24 cotto_work -- many ideas will be similar in M0+6model, so it's not an outright rejection
20:24 cotto_work - PDS
20:24 cotto_work -- great success, very active discussion, lots of follow-up on #parrot and parrot-dev
20:24 cotto_work -- need to make sure ideas don't get lost
20:24 cotto_work -- I volunteered to help take a flamethrower to imcc, publicly committed to M0 roadmap goal
20:24 cotto_work - M0 roadmap progress
20:24 cotto_work -- have a concrete goal by 3.3 (April): runnable demos using structs, arrays, hashes and functions with arguments
20:24 cotto_work -- dukeleto++, atrodo++, chromatic++ and I will be working a spec and implementation between now and then
20:24 cotto_work - started working on making the profiling runcore more useful
20:24 cotto_work -- working on adding direct callgrind output so the post-processing step can go away
20:24 cotto_work -- next step is figuring out why kcachegrind misinterprets current output of pprof2cg
20:24 cotto_work -- reverse-engineering ftw
20:24 cotto_work - thought about how to deal with nqp's new direction
20:24 cotto_work -- my current thinking is that we'll include it as an external project (same as nqp-rx now) and maintain our own "use Parrot;" layer internally
20:24 cotto_work -- other options exist and are worth considering
20:24 cotto_work *will do:
20:24 cotto_work - work with dukeleto and atrodo to produce a spec for M0
20:24 cotto_work - make the profiling runcore less wonky with kcachegrind, add directl callgrind-compatible output
20:24 cotto_work - more rfc review, more ticket closing
20:24 cotto_work *blockers:
20:24 cotto_work - none
20:24 cotto_work *eor
20:24 cotto_work NOTE: There's a PaFo concall 30m after the beginning of #ps, so people on the board of directors won't be around then.
20:25 nwellnhof joined #parrotsketch
20:25 NotFound 2011-feb-01:
20:25 NotFound What I did (last two weeks):
20:25 NotFound -parrot
20:25 NotFound * minor fixes
20:25 NotFound * avoided the 'might be clobbered' warnings
20:25 NotFound * rearranged C code in some test to avoid uninformative
20:25 NotFound lack of coverage reports
20:25 NotFound -winxed:
20:25 NotFound * Several refactors, fixes, code cleaning and nano optimizations
20:25 NotFound * Tailcall optimization applied in more cases
20:25 NotFound * Set exit values and use them in compiler stages and drivers
20:25 NotFound * Operators << and >>
20:25 NotFound * --nowarn option
20:25 NotFound * handle_types_except modifier in try
20:25 NotFound * Syntax for use of labels in pirops
20:26 NotFound * predefined symbol __FUNCTION__
20:26 NotFound What I will do:
20:26 NotFound No plan
20:26 NotFound EOR
20:26 nwellnhof what i did (in 2011)
20:26 nwellnhof - a lot of unicode and IO related work
20:26 nwellnhof plans
20:26 whiteknight joined #parrotsketch
20:26 nwellnhof - finish full Unicode support for OS dynpmc
20:27 nwellnhof - get gc_dynamic_threshold merged finally
20:27 nwellnhof .eor
20:27 whiteknight DID: More hacking on the imcc work. Getting very close, but seeing some weird PCC/GC bugs now that I need to debug. Also, looked at bacek's new GC algorithm quite closely. PDS and talked with Rakudo folks a lot.
20:27 whiteknight WILL DO: Keep working on IMCC till it's done. Then, GC.
20:27 whiteknight EOR
20:28 tcurtis This was another week of me not doing Parrot-related work. Hopefully next week won't be, but school is a bit of a blocker.
20:28 tadzik joined #parrotsketch
20:31 cotto_work hello\
20:32 whiteknight hello
20:32 Util Addendum: Worked with Coverity; scans working! chromatic++ for jumping in to fix a Parrot bug that was causing about half of the defects. .end
20:32 Util hello
20:33 tcurtis Hi.
20:33 cotto_work How'd we do on last week's goals?
20:33 cotto_work q2q
20:34 whiteknight no idea. What were the goals?
20:34 cotto_work GOAL 1: fix ipv6 merge-related failures, review code (all)
20:34 cotto_work GOAL 2: test branches - imcc_interfaces (whiteknight), annotations-tree (cotto)
20:34 cotto_work GOAL 3: PDS prep - roadmap goals (kid51, dukeleto, cotto, whiteknight)
20:34 cotto_work GOAL 4: concerns to address before 3.3 in April (all)
20:34 cotto_work ipv6 seemed to get some love
20:35 cotto_work I got swamped and didn't end up reviewing annotations-tree
20:35 whiteknight I've been working like a madman on the IMCC stuff. Not ready to merge yet
20:35 cotto_work yeah, really
20:35 whiteknight probably won't be ready this week either, unless other people are interested in getting involved with it
20:35 whiteknight PDS went very well, it seems
20:35 cotto_work whiteknight: do you have a current task list?
20:36 cotto_work I know one exists, but I don't know if it's current
20:36 whiteknight cotto_work: All primary development is done now. In the debugging stage
20:36 whiteknight so tasklist looks like this: 1. Make it work
20:37 whiteknight I'll put out a note to the list tonight with current status
20:37 cotto_work great
20:38 cotto_work any thoughts before we move to questions?
20:39 cotto_work alright then, question time
20:39 cotto_work Q: I propose another couple weeks of emphasizing closing bugs.  We're at 546 two weeks from 3.1.  I'd like to try to get down to 523 by next week and 500 by the release.
20:40 whiteknight +1
20:40 whiteknight hard, but doable
20:41 cotto_work with 6 people working on it, that 4 tickets per person per week and we're there
20:41 pmichaud there's a good number that kid51++ pointed out to me that I can work on closing
20:41 cotto_work excellent
20:42 pmichaud looks like about 20 or so pge and pct related tickets
20:42 pmichaud so I can close many of them as 'wontfix' or explain where we stand on them
20:42 cotto_work pmichaud: thanks
20:42 whiteknight for those, if you aren't going to fix them or aren't going to fix them under the auspices of the parrot umbrella, dump them
20:43 pmichaud Exactly.
20:44 cotto_work any other volunteers?
20:44 whiteknight I'll try to close several, especially IMCC-related as my branch matures
20:44 cotto_work everyone's welcome to play along at home
20:44 nwellnhof I can finish at least one deprecation, and maybe some other bugs.
20:45 cotto_work next question
20:45 NotFound Late hello
20:45 cotto_work What are people's thoughts on how Parrot should adapt to the new nqp that pmichaud will be working on?
20:45 whiteknight that's a pretty big topic to discuss at #ps methinks
20:45 pmichaud my suggestion: wait a week and see where the nqp is actually at
20:46 cotto_work wfm
20:46 cotto_work pmichaud: how long do you think it'll take nqp to get to where nqp-rx is now?
20:46 chromatic I'm -1 on even waiting.
20:46 pmichaud it's pretty close already.  the bigger question is how long it'll take for rakudo to move over to it
20:46 * cotto_work keeps forgetting that nqp-rx/nom is nqp
20:47 pmichaud yes,I plan to do the repo moves in the next day or two
20:48 pmichaud anyway, my suggestion is give it a week to see the progress of this week, but if waiting isn't an option, then start discussing a new system soon :)
20:48 cotto_work chromatic: what's your suggestion?
20:48 chromatic Parrot should host its own language.
20:48 cotto_work chromatic: separate from nqp?
20:48 whiteknight and what is it, a frozen/forked version of NQP?
20:49 whiteknight nqp-rx?
20:49 chromatic We've been through enough churn from PGE/NGE to NQP to NQP-rx and now yet another rewrite of NQP with increased scope, I'd like to jump off of that treadmill somehow.
20:49 chromatic Sorry, TGE not NGE.
20:50 whiteknight right, I think we all want something a little bit different. Problem is finding a solution that everybody is fine with
20:51 cotto_work I see why you say that, but I don't like the idea of having to maintain a language.  There is charm in having something completely under our control.
20:51 whiteknight IMCC is completely under our control now. Cut that out and we have space to add another compiler
20:51 whiteknight especially if it's minimal enough
20:51 chromatic I don't like the idea of having to maintain a language either, but given the choice of that or telling our users "Upgrade to a new compiler infrastructure" every year... I know which one I pick.
20:52 whiteknight let's hash this out a bit more. We obviously have more thinking and talking to do than is going to fit in a #ps
20:53 pmichaud I don't know why one would expect that a new infrastructure would be any more stable in this regard, fwiw.
20:53 whiteknight maybe we can put together some kind of a poll on parrot.org, once we have a list of options
20:53 Util whiteknight: IMCC is being replaced, not removed.
20:53 chromatic I don't trust the new NQP's scope creep.
20:53 chromatic Maybe I'm wrong.
20:53 pmichaud and we've proved many times over that IMCC really isn't suited for HLL developers.  Something higher is needed.
20:54 PerlJam whiteknight: I mention again as I did yesterday that perhaps someone should write down the required characteristics of this language so that comparison can be made with existing languages or, if none are worthy, it's a start on a spec for a new language.
20:55 NotFound pmichaud: counterexample: winxed
20:55 whiteknight right. We're not going to do all that at #ps
20:55 pmichaud NotFound: would you use winxed to write a hll compiler?
20:55 pmichaud NotFound: example -- draft a version of ops2c in winxed
20:55 whiteknight we can still use NQP to write HLL compilers. it's not like that option disappears
20:55 pmichaud chromatic's point is about toolsets for hll writers, I think.
20:56 chromatic It's about the whole stack and the upgrade tax.
20:56 NotFound pmichaud: maybe, but the point is that winxed itself targets IMCC
20:56 whiteknight We don't want to be stuck in the mindset that we need exactly one language to fit the needs of everybody
20:56 pmichaud "Telling our users 'Upgrade to a new compiler infrastructure'"  is definitely talking about hll compiler writers, not parrot internals.
20:56 pmichaud NotFound: I think the topic is hll compiler development on parrot.
20:57 pmichaud not necessarily 'what should replace IMCC'
20:57 whiteknight let's move this discussion back to #parrot
20:57 pmichaud wfm
20:57 cotto_work ok.  Are there any other questions?
20:58 NotFound pmichaud: my point is that PCT is not the only way to write HLL compilers.
20:59 cotto_work I've got closing tickets as a goal.  other suggestions?
21:02 whiteknight I would like to get some more eyes on my IMCC branch
21:02 whiteknight but I can't promise that will be ready to review/merge by the end of the week
21:06 cotto_work our goals then:
21:06 cotto_work GOAL 1: close 23 tickets by next #ps
21:06 cotto_work GOAL 2: more eyeballs on the IMCC branch
21:06 cotto_work Unless someone has closing comments, that's a wrap.
21:07 nwellnhof left #parrotsketch
21:08 whiteknight wrap
21:08 chromatic left #parrotsketch
21:09 NotFound left #parrotsketch
21:10 whiteknight left #parrotsketch
21:14 plobsing joined #parrotsketch
21:18 dukeleto joined #parrotsketch
21:20 dukeleto over?
21:20 tadzik yeah
21:20 cotto_work it's over
21:30 pmichaud left #parrotsketch
21:41 plobsing left #parrotsketch
23:31 plobsing joined #parrotsketch
23:35 plobsing left #parrotsketch
23:40 tcurtis left #parrotsketch
23:51 whiteknight joined #parrotsketch

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