Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2011-05-17

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

All times shown according to UTC.

Time Nick Message
01:33 lucian_ left #parrotsketch
01:36 whiteknight left #parrotsketch
05:39 dafrito left #parrotsketch
08:28 contingencyplan left #parrotsketch
12:10 whiteknight joined #parrotsketch
12:16 bluescreen joined #parrotsketch
12:20 bluescreen left #parrotsketch
12:29 lucian joined #parrotsketch
12:35 bluescreen joined #parrotsketch
13:21 lucian left #parrotsketch
13:35 contingencyplan joined #parrotsketch
14:10 PacoLinux left #parrotsketch
14:29 bluescreen left #parrotsketch
14:43 bluescreen joined #parrotsketch
15:03 spinclad left #parrotsketch
15:59 particle left #parrotsketch
15:59 particle joined #parrotsketch
16:17 bluescreen left #parrotsketch
16:17 bluescreen joined #parrotsketch
16:18 particle left #parrotsketch
16:22 lucian joined #parrotsketch
16:27 particle joined #parrotsketch
16:38 particle left #parrotsketch
16:49 darbelo joined #parrotsketch
17:07 particle1 joined #parrotsketch
17:13 particle1 is now known as particle
17:44 darbelo left #parrotsketch
17:52 tcurtis Reporting early so I can nap until #ps.
17:52 tcurtis Did:
17:53 tcurtis * Gave a lightning talk about my project at the Chicago GSoC meetup.
17:53 tcurtis * Was surprised by how many people had heard of Parrot.
17:53 tcurtis * Little else.
17:53 tcurtis Will do:
17:54 tcurtis * Read papers.
17:54 tcurtis * Blog.
17:54 tcurtis * Talk to darbelo again.
17:54 tcurtis * Sketch LALR grammar object representation.
17:54 tcurtis Blockers:
17:55 tcurtis * ICPC tryouts
17:55 tcurtis EOR
18:21 whiteknight WHAT I DID:
18:21 whiteknight * Tested like a madman to help ensure a good release. gerd++ on another smooth one.
18:21 whiteknight * Worked on the NCI issues with cotto++, plobsing++ and pmichaud++.
18:21 whiteknight * Attended PDS (kid51++ and others for setting that up)
18:21 whiteknight * Expanded, Cleaned up, documented, and tested the Rosella Memoize library. It's now "stable".
18:21 whiteknight WHAT I WILL DO:
18:21 whiteknight * See if I can break ground on some new profiling tools, or at least put out some design ideas
18:21 whiteknight * More work on Rosella, trying to get several libraries improved in various ways
18:21 whiteknight EOR
18:27 pmichaud What I did this past week:
18:27 pmichaud - more benchmarks of Rakudo and Parrot
18:27 pmichaud - built a simple sub-level profiler for Rakudo
18:27 pmichaud - patched rakudo to get ~15% speed improvements
18:27 pmichaud - helped restore NCI functionality for zavolaj and others
18:27 pmichaud - created NCI::Utils::ncifunc as a replacement wrapper for dlfunc
18:27 pmichaud - drafted a "how Parrot is doing" message for PDS
18:27 pmichaud - attended PDS
18:27 pmichaud - drafted a new relationship manager policy for Parrot/Rakudo
18:27 pmichaud Rakudo-specific improvements:
18:27 pmichaud - added infix:<gcd> and infix:<lcm> operators
18:27 pmichaud - discussed changes to .perl method
18:27 pmichaud - added "config.status" to Configure.pl
18:27 pmichaud - fixed NaN relational comparisons
18:27 pmichaud - fixed bugs in --makefile-timing flag
18:27 pmichaud - refactored Mu.item to be much faster
18:28 pmichaud What I plan to do this next week:
18:28 pmichaud - more improvements to rakbench and profiling tools
18:28 pmichaud - more Rakudo performance fixes
18:28 pmichaud - create an interim Star release (likely but not definite yet)
18:28 pmichaud - reviews of reports, nom, nqp
18:28 pmichaud EOR
18:38 darbelo joined #parrotsketch
18:50 darbelo left #parrotsketch
19:04 darbelo joined #parrotsketch
19:11 NotFound joined #parrotsketch
19:22 NotFound What I did:
19:22 NotFound -parrot
19:22 NotFound * Minor fixes
19:22 NotFound * Started using the gcc compile farm for testing,
19:22 NotFound fixed the build in sparc/OpenBSD while on it.
19:22 NotFound -winxed
19:23 NotFound * Fixed some problems with operand types in some operators,
19:23 NotFound diagnosing more errors.
19:23 NotFound * Added more scope searching abilities, avoiding the need for
19:23 NotFound using and using static declarations.
19:23 NotFound * Improver 'using' statement both at function and a namespace level.
19:23 NotFound * Added 'using namespace' statement.
19:23 NotFound What I will do:
19:23 NotFound * Use the compiler farm for testing more platforms.
19:23 NotFound * More improvements to winxed scope search.
19:23 NotFound EOR
19:34 ShaneC joined #parrotsketch
19:36 darbelo left #parrotsketch
19:36 darbelo joined #parrotsketch
19:36 particle left #parrotsketch
19:37 mikehh What I did since my last report:
19:37 mikehh * building and testing parrot on amd64/i386, with gcc/g++
19:37 mikehh * some fixes
19:37 mikehh * building and testing rakudo and winxed against latest parrot
19:37 mikehh What I intend to do in the next week:
19:37 mikehh * testing and fixing
19:37 mikehh * start going through "skipped" tests and see which now pass and which can be "TODO"ed
19:37 mikehh * work to eliminate warnings in the build
19:37 mikehh .eor
19:48 particle joined #parrotsketch
19:54 darbelo left #parrotsketch
19:54 kid51 joined #parrotsketch
19:55 particle left #parrotsketch
20:01 particle joined #parrotsketch
20:09 ShaneC left #parrotsketch
20:11 bubaflub joined #parrotsketch
20:13 darbelo joined #parrotsketch
20:14 kid51 kid51's reportDONE
20:14 kid51 DONE
20:15 kid51 * Participated in Parrot Developer Summit Sat May 14 2011
20:15 kid51 ** Prepared agenda
20:15 kid51 ** Posted summary on parrot-dev
20:15 kid51 * Infrastructure: communicated with OSU OSL re mailing problems on Sat/Sun
20:15 kid51 * Cage-cleaning: closed these tickets:
20:15 kid51 ** http://trac.parrot.org/parrot/ticket/2104
20:15 kid51 ** http://trac.parrot.org/parrot/ticket/2105
20:15 kid51 * Testing: Extensive pre-release testing particularly on Darwin/PPC
20:15 kid51 ** On Darwin/PPC, we are experiencing inconsistent test results, i.e. files that will FAIL during 'make test' but PASS during 'prove' or vice versa.
20:15 kid51 ** No consistency:  The (typically single) failing file will change from one commit to the next.
20:15 kid51 ** Suspect this is GC-related.
20:16 kid51 ** But haven't filed a TT because this is practically non-reproducible.
20:16 kid51 WILL DO
20:16 kid51 * YAPC
20:16 kid51 ** Beginning to prepare talk on Parrot community and process for YAPC::NA::2011.
20:16 kid51 ** Make your conference and hotel reservations before May 20 to get discount.
20:16 kid51 ** Make your plane reservations now.
20:16 kid51 * Tickets: Intend to close:
20:16 kid51 ** http://trac.parrot.org/parrot/ticket/1722
20:16 kid51 EOR
20:17 soh_cah_toa joined #parrotsketch
20:21 cotto_work *did:
20:21 cotto_work - participated in PDS, seemed to go well
20:21 cotto_work -- blogged: http://reparrot.blogspot.com/2011/05/thoughts-on-pds.html
20:21 cotto_work - made sure that 3.4.0 didn't escape without some kind of nci 't' support
20:21 cotto_work -- a solution was found and pmuichaud seems to be happy
20:22 cotto_work -- let's not do that again
20:22 cotto_work - M0 progress:
20:22 cotto_work -- mostly thinking
20:22 cotto_work -- scheduled meeting with allison and dukeleto for the 28th
20:22 cotto_work - GSoC progress:
20:22 cotto_work -- nothing
20:22 cotto_work *will do:
20:22 cotto_work - M0 hacking, gsoc
20:22 cotto_work *eor
20:22 pmichaud clarification:  "satisfied."  Not "happy."
20:31 PacoLinux joined #parrotsketch
20:32 kid51 hello
20:32 Util 7-day ticket report:
20:32 Util 2 closed: done
20:32 Util 6 closed: fixed
20:32 Util 8 new
20:32 Util 2 reopened
20:32 Util .end
20:32 cotto_work hello, all
20:32 NotFound Hola
20:32 Util hello
20:32 mikehh hi there
20:32 soh_cah_toa hello
20:33 cotto_work How'd our goals go this week?
20:33 cotto_work GOAL 1: Find someone to lead testing infrastrcture planning
20:33 cotto_work GOAL 2: Contact all GSoC students
20:34 soh_cah_toa oh yeah, i still have to meet w/ dukeleto
20:35 mikehh Release went out ok
20:35 * kid51 sees only 4 posts by GSOCers on parrot.org; I thought we had 7 students
20:35 cotto_work Has anyone heard from the GObject student?
20:35 Util NotFound said he would try the GCC compile farm, but no mention of overall leading.
20:36 * cotto_work is trying to find his name
20:36 soh_cah_toa i'll mirror my blog after #ps
20:36 NotFound Util: ? I mentioned it in the report and in #parrot several times.
20:36 dukeleto hola
20:36 dukeleto our GObject gsoc student is MIA
20:36 cotto_work cgaertner
20:36 dukeleto Christoph Gaertner
20:37 kid51 soh_cah_toa: You already have one post on parrot.org, so I was counting you as one of the four.  but mirroring is necessary if you're primary blog is elsewhere.
20:37 soh_cah_toa kid51: oh, okay. will do
20:37 dukeleto kid51: i told my student (bubaflub++) that he could write his blog post in the next week, since he is graduating today
20:37 Util NotFound: so you did. Apologies; late to meeting and still backscrolling.
20:37 NotFound I'm doing make test right now in Debian/ppc64 in the farm, by the way.
20:37 kid51 That leaves cgaertner and lucian, correct?
20:38 whiteknight left #parrotsketch
20:38 lucian kid51: yeah
20:38 lucian sorry, i've only just demo-ed my dissertation today
20:39 cotto_work lucian's here, so he's not a risk. ;)
20:39 kid51 okay; a short, introductory post is all we need now.  1 paragraph sufficient
20:39 lucian apparently UK uni finishes very late
20:39 lucian yeah, will do. tomorrow, likely
20:40 tadzik lucian: you should have seen PL Uni...
20:40 darbelo left #parrotsketch
20:42 darbelo joined #parrotsketch
20:42 cotto_work any questions?
20:43 kid51 Questions ... in general? or in some specific area?
20:43 cotto_work kid51: in general
20:43 kid51 So there was a lot of talk during PDS and on #parrot about benchmark testing of Parrot+Rakudo.
20:44 kid51 Does anyone know who pledged to do what on whose servers?
20:45 kid51 My impression is that a lot was to be done on pmichaud's machines ...
20:45 kid51 ... which makes sense, since he's the primary consumer of benchmarking reports.
20:46 cotto_work The intent is that those be the canonical machines we use to decide whether performance has improved.
20:46 kid51 okay.  And is there someone from the Parrot side who is the point person on this?
20:48 cotto_work I don't recall having someone volunteer for that role.
20:49 kid51 Well, I will write him to see what more specifically he needs from us.
20:49 pmichaud I'm here
20:49 pmichaud I'll continue benchmarking if only because I need to keep track of rakudo performance changes as well
20:49 pmichaud so it's easy for me to simply publish results
20:49 kid51 What, if anything, do you need from us on this (other than better code)?
20:49 pmichaud I'm expecting to (next couple of days) clean up the scripts a bit so that others can do similar tests as well
20:50 kid51 Can you post to parrot-dev when that's ready?
20:50 pmichaud sure
20:50 pmichaud I don't have any other specific needs on this, other than to know if there are things I can be testing that would help Parrot. (more)
20:50 pmichaud bacek++ has already been using the test suite to help tune his gc work
20:51 mikehh I am quite happy to help here, but don't necessarily want to be a pointy type person
20:51 pmichaud later today I expect to have a benchmark of rakudo-2011.04 running on Parrot 3.4.0
20:51 pmichaud that should give us a really good idea of how parrot performance changed from 3.3.0 to 3.4.0
20:52 kid51 I have another question.
20:52 kid51 cotto_work: From your blog post, I gather that you and whiteknight will be the core of the Profiling team -- correct?
20:53 cotto_work kid51: yes but I haven't had a chance to coordinate with him yet.
20:53 kid51 When you do, if there's any part of the task in which you can rope in a 3rd or a 4th, that would be good (more)
20:53 kid51 ... because you 2 are already committed to work in many different parts of the project.
20:55 cotto_work Volunteers are welcome.
20:55 cotto_work The amount of work is unclear.  There could be some very low-hanging fruit that gets up 95% of the way to something usable with 5 hours work, or it could take much longer.
20:59 cotto_work any other questions or thoughts, about PDS or something else from the past week?
21:00 kid51 Make your YAPC OSCON and FOSSCON travel and presentation plans now!
21:01 cotto_work good idea
21:04 mikehh oh yes - we need something like smolder to keep track of tests, any ideas there
21:04 mikehh the current implementation seems broken
21:05 cotto_work mikehh: dukeleto's been using jitterbug
21:05 cotto_work http://jitterbug.leto.net:3000/project/parrot
21:06 mikehh that seems to give build results, we also need test results
21:07 mikehh particularly failures
21:07 cotto_work It did.  I suspect it's a temporary bug.
21:09 mikehh we need some way of reporting test failures on multiple platforms, so we can look at them in more detail
21:10 mikehh hopefully something like taptinder reports frpm the gcc compiler farm and others (windoze too)
21:10 dukeleto cotto_work: parrot found some bugs in jitterbug that i haven't had time to fix yet
21:10 mikehh from
21:10 cotto_work If anyone wants to help and knows Perl 5, I'm sure dukeleto could put you to work.
21:10 mikehh well let me know what to do and I''l give it a try
21:10 dukeleto https://github.com/franckcuny/jitterbug/issues/70 is what is blocking parrot on jitterbug
21:10 dukeleto feel free to fix any other issues, too :)
21:14 kid51 left #parrotsketch
21:15 * dukeleto hopes to get parrot+jitterbug working again soon, but can give no time estimates
21:17 dukeleto any gsoc students have questions?
21:18 soh_cah_toa nope
21:22 cotto_work do any non-students have questions?
21:23 mikehh I think we need to post the weekly goals to the Parrto Developer wiki, = I can do that if you want
21:24 dukeleto i guess i have a question
21:24 cotto_work mikehh: thanks.
21:24 cotto_work dukeleto: go ahead
21:24 * mikehh my typing is getting really bad, probably need some sleep
21:24 dukeleto currently, we have a lot of important docs only in our repo or only on the wiki page. This makes finding them hard.
21:24 dukeleto we need to improve this situation
21:24 mikehh yes
21:25 dukeleto one example is deprecation info. api.yaml stores the basics, but all the actually important details are only on the wiki
21:25 dukeleto this seems suboptimal
21:25 dukeleto For example, if i have a parrot git repo and no net access, i can't figure out deprecation stuff. That sucks.
21:25 Util docs.parrot.org should be exporting the HTML-rendered docs from the repo.
21:26 dukeleto Util: sure. I am mostly talking about info that is only in the wiki
21:26 Util ok
21:26 soh_cah_toa left #parrotsketch
21:26 dukeleto I am +1 to moving *all* deprecation data into our repo, and then we can link to the deprecation info on github instead of the wiki
21:27 mikehh or both
21:27 dukeleto mikehh: we want to minimize needing to keep things in sync, because, invariably, they will fall out of sync
21:27 dukeleto we need *one* canonical place for deprecation data. Storing some of it parrot.git and some in our wiki is making the information less accessible.
21:28 mikehh that also suggests moving the wiki to github
21:29 mikehh a lot of our documentation is out of sync (also our tests and other things too)
21:30 mikehh I intended to start working on the tests in the next week
21:31 soh_cah_toa joined #parrotsketch
21:31 eternaleye_ joined #parrotsketch
21:32 cotto_work Consolidating all the deprecation data is a good idea.
21:33 cotto_work mikehh: were you volunteering?
21:33 cotto_work (no pressure; just clarifying)
21:34 mikehh I'll do some work on it, as much as I can work out what is needed
21:34 cotto_work mikehh: thank you
21:35 dukeleto mikehh: i don't want to convert all wiki pages to docs in our repo, only deprecation data, just to clarify
21:35 cotto_work dukeleto: yes
21:35 cotto_work What goals should we set for the week?
21:36 mikehh dukeleto: yeah, but it might be a good idea to move other things to github, let me think on it
21:36 eternaleye left #parrotsketch
21:38 dukeleto cotto_work: performance improvement would be nice, but wishing for them doesn't make them appear
21:39 mikehh do we have any branches, ready for merging/testing after the release
21:40 cotto_work I want to get the Select dynpmc into master soon.  Apparently some Rakudo folks want it.
21:42 cotto_work other goals?
21:42 cotto_work I have:
21:42 cotto_work GOAL 1: Merge/review Select dynpmc (cotto)
21:42 cotto_work GOAL 2: Contact all GSoC students
21:42 cotto_work GOAL 3: Consolidate deprecations data into api.yaml (mikehh)
21:42 dukeleto cotto_work: not sure all the info should go into api.yaml, but i agree with the rest
21:42 dukeleto cotto_work: we may want to store the detailed explanations in a directory of POD files
21:43 dukeleto cotto_work: like docs/deprecations/deprecations_for_parrot_x_y_z.pod or somesuch
21:43 cotto_work dukeleto: that's also an option.  I'll leave it to mikehh++ to decide what makes the most sense.
21:43 mikehh we would need to keep that up-to-date
21:44 mikehh I'll have a look at what we have and then discuss it on #parrot or in parrot-dev
21:44 cotto_work mikehh: sounds good
21:46 dukeleto mikehh: sounds good. we definitely need to ask for more feedback
21:49 cotto_work Is there anything else before we call it a wrap?
21:52 cotto_work Let's call it a wrap.
21:52 NotFound left #parrotsketch
21:54 bluescreen left #parrotsketch
22:10 dukeleto left #parrotsketch
22:11 soh_cah_toa left #parrotsketch
22:18 bubaflub left #parrotsketch
22:28 darbelo is now known as me
22:28 me is now known as darbelo
22:30 whiteknight joined #parrotsketch
22:37 pmichaud left #parrotsketch
22:42 bubaflub joined #parrotsketch
23:11 bubaflub left #parrotsketch
23:20 darbelo_ joined #parrotsketch
23:20 darbelo left #parrotsketch
23:20 darbelo_ is now known as darbelo
23:29 lucian left #parrotsketch
23:57 darbelo left #parrotsketch

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