"Tuesday at 20:30 UTC"
Set by moderator on 20 September 2010.
02:35 bluescreen left 03:15 whiteknight left 05:28 tcurtis joined 06:11 tcurtis left 06:26 bacek left 07:16 spinclad left 07:22 bacek joined 07:34 bacek left 07:40 bacek joined 09:53 bacek left 10:06 bacek joined 12:43 particle1 joined 12:49 particle left 13:01 bluescreen joined 13:22 PerlPilot joined 13:27 PerlJam left 13:48 bacek_ joined 13:52 bacek left 15:10 PerlPilot left, PerlJam joined 16:13 mariano joined 16:14 mariano left 16:21 particle joined 16:26 particle1 left 17:29 tcurtis joined 18:31 PerlJam left 19:00 Coke joined 19:06 Coke left 19:47 atrodo joined 19:49 whiteknight joined
whiteknight WHAT I DID: 19:56
* More work on the embedding API. Getting close to a demonstrable "product". Hoping to unveil the first draft this week
* Having trouble teasing apart some of the .h files to include the things we want in the API, but not include a bunch of crap that users shouldn't see
* Writing lots of blog posts, and getting lots of feedback
* Lots of pre-release testing. Submitted at least a dozen smoke reports in various platform configurations
WHAT I WILL DO:
* Looking to get the embedding API fork building and elicit some more community feedback
* Will support the gsoc_nci branch merge however I can.
* Going to start looking at gsoc_instrument branch again. I would like to get that fixed up and merged to master before 2.11
* Need to do some last-minute stuff for GCI. Already heard from at least one student and would like to start actively attracting more
* Working on a draft of adding i18n support to Parrot. Would like to get that draft finished and sent off to the rest of the architecture team for initial review
WHAT I AM BLOCKING ON:
* Only 24 hours in a day? Damnit.
19:57 kid51 joined
kid51 kid51's reportDONE: 19:57
DONE: 19:58
* First git commits (small refactorings and testing of new config steps)
* Correct some configuration problems in gsoc_nci branch
* trac.parrot.org/parrot/ticket/1851
WORKING ON:
* trac.parrot.org/parrot/ticket/532: headerizer refactor
* trac.parrot.org/parrot/ticket/855: debugging/profiling options
EOR
whiteknight Addendum: I've take all the partcl-related tickets for myself, and am going to start triaging them aggressively. I would like the world to be kinder to partcl by 2.11 19:59
kid51 Addendum: Like others, did pre-release testing, plus some testing of gsoc_nci branch 20:00
Util # Done: 20:05
* Tested branch gsoc_nci
* Merged branch pbc_to_exe_direct_output
# Plan to do:
* More pbc_to_exe improvements
# Blockers:
* None
.end
q1q
tcurtis I am finishing up releasing Parrot 2.10.0. In this coming week, I will probably still be lacking tuits for actually working on Parrot. 20:08
20:09 mariano__ joined 20:10 bluescreen left 20:12 mariano__ is now known as bluescreen 20:19 mikehh joined 20:21 allison joined 20:29 NotFound joined
NotFound What I did: 20:31
-parrot
* Added a few test
-winxed
* Some refactoring
What I will do:
No plan,
EOR
mikehh What I did since my last report:
* building and testing parrot on amd64/i386, with gcc/g++
* some git related fixes
* some branch testing
* pre/post release testing - up to smoke/fulltest
+ blocking on some $work related issues
What I intend to do in the next week:
* testing and fixing
* get something workable for GCI
.eor 20:32
whiteknight hello 20:37
cotto_work hello
Util Hello
mikehh hi there
NotFound hola
Util What happened to the channel topic?
20:37 dukeleto joined
whiteknight who's running this show? is chromatic around? 20:37
dukeleto i can run it, if no one else is volunteering. 20:38
whiteknight damnit, beat me to it
dukeleto++
dukeleto We just released 2.10.0, yes?
whiteknight yessir 20:39
mikehh yes, tests ok for me
whiteknight tcurtis++ on that
dukeleto tcurtis++ for going from GSoC student to release manager in a very short time.
mikehh tcurtis++
allison tcurtis++
Util tcurtis++
cotto_work tcurtis++
mikehh especially as it is our first git based relesae 20:40
dukeleto dreams of Unicorns riding a DAG into the sunset
NotFound There is karma collection in this channel?
dukeleto Did we accomplish last weeks goals?
cotto_work not angry unicorns, I hope
dukeleto NotFound: no, but ilbot2 is our historian 20:41
what were last weeks goals?
NotFound Invisible Flying Unicorns
whiteknight We got most of the git migration issues sorted out. We got the release out. Much of this week was feature-freeze in master 20:43
we have a few branches either merged already, or imminent
20:44 PerlJam joined
whiteknight I don't remember what other goals we had 20:44
dukeleto I just merge gsoc_nci. 20:45
merged, even. Sometimes i merge in the present tense.
moderator Vision for 2.0: Production Users | Don't forget to post closed tickets in your report. | Note: This channel is only for our Tuesday status meetings (at 20:30 UTC); you probably want #parrot instead. | irclog: irc.pugscode.org/ 20:45
whiteknight okay, that's another big one down. Yippie! 20:46
okay, nobody else is talking about goals from last week. Let's move on. 20:47
dukeleto What should our goals for this week be?
whiteknight 5 minutes ago, I would have said the gsoc_nci merge
dukeleto whiteknight: i love doing things before you tell me to do them :) 20:48
whiteknight did chromatic's gc branch merge?
dukeleto I think Google Code-In is a priority for community building.
perlbuzz.com/2010/11/google-code-in...ities.html
whiteknight gc_ms2_sf doesn't look merged. That should be priority. I think we were seeing 5-10% improvements for most workloads 20:49
20:49 bacek_ is now known as bacek
bacek ~~ 20:49
dukeleto We stand to gain many young developers into our communities by thinking hard about what we want to get accomplished with Google Code-In.
This program will have the smartest 13-18 yr olds in the world applying. We are already getting emails from very smart students. 20:50
kid51 What, specifically, do we have to do in next 1-2 weeks re Google Code-In? (I applied my name as potential mentor to certain tasks.)
cotto_work dukeleto: are there any that aren't going to parrot-dev or parrot-users?
mikehh hopefully another tcurtis or two
dukeleto cotto_work: i think whiteknight got a direct email that he forwarded to me 20:51
cotto_work The more we see that there's a real interest, the more we'll be encouraged to make our gci tasks accessible
NotFound Someone added my name as a mentor for spanish translations.
dukeleto I haven't added all our tasks to the Code-In site yet. That is where the students can actually sign up for them.
NotFound: you can blame me :)
NotFound: if you don't want to, i can remove it.
NotFound: these are mostly "potential mentors". I have to add it to the actual Code-In site for it to be official. 20:52
I don't want to take up #ps with lots of talk about Code-In, but we should definitely discuss it on parrot-dev
NotFound dukeleto: according my experience, spanish translations aren't very useful and people waste its time discussing about minutiae.
dukeleto NotFound: we just want various web pages in a few languges, so that developers that don't speak english can learn about parrot 20:53
NotFound: we don't care about the minutiae, we just want something that is grammatically correct and doesn't sound ridiculous (such as what comes out of Google Translate)
NotFound I suppose I can help with that. 20:54
dukeleto Ideas for code priorities this week?
NotFound: thanks!
whiteknight dukeleto: what, specifically, is the most important thing that we can do for GCI?
brainstorm new ideas?
dukeleto We need to think of our process of welcoming young people into our community. What info do we point them towards? What is the best way to make them feel welcome? 20:55
whiteknight: more tasks are also of course good, and improving the descriptions of our current tasks.
whiteknight okay
in terms of code priorities this week, I would like to get chromatic's gc_ms2_sf branch reviewed, well-tested, and merged 20:56
I think we're also going to need to throw more developer effort behind the generational_gc branch, to help bacek out
that branch becomes higher and higher priority every day 20:57
dukeleto Wishing for ponies that understand the GC doesn't make it happen. 20:58
whiteknight it's not a matter of understanding it at this point. We need testing
dukeleto whiteknight: can you write a short intro for people not familiar with the GC? Some kind of "this is enough to make you useful for a code review" document
whiteknight and, if possible, people willing to open up a debugger if necessary 20:59
I can do that, yes. I'll ship it out to parrot-dev
dukeleto whiteknight: testing is fine. code review is arguably more important, though
whiteknight++
whiteknight if we could get a handful of test reports on gc_ms2_sf and generational_gc branches, we would be in good shape going forward 21:00
bacek gen_gc is on hold 21:01
I do want to rewrite it with new algo.
Last 2 commits in branch are describing it.
cotto_work bacek: isn't it close to working?
bacek it's slow.
dukeleto well, since i just merged gsoc_nci, testing master should be a priority as well.
whiteknight bacek: what help do you need with that rewrite? 21:02
so let's skip generational_gc branch for now. We *do* need testing in gc_ms2_sf for anybody who is able to do that
bacek whiteknight, proof read of new algo. 21:03
I've got another branch gc_ms2_faster_list also
will be ready to merge in few days.
whiteknight bacek: okay, let us know and we can get testing on that. I'll read the new algorithm over and see what I can see 21:04
bacek whiteknight, deal
whiteknight anybody seen cotto? Does anybody know the status of his opmap_aware_pmcs branch? 21:05
cotto_work whiteknight: I'm stuck debugging
whiteknight you need help with that?
cotto_work sure, especially from people with some familiarities with packfiles 21:06
whiteknight cotto_work: can you write up a quick summary of the branch to give interested devs a crash-course in it?
cotto_work but it could be more obvious goof too
10-4
whiteknight cotto++
dukeleto So do we have goals for the week?
mikehh merge baranches, after testing of course and GCI 21:07
concentration on gc as much as possible 21:08
whiteknight +1 21:10
dukeleto ok. And evidently fix the 2.10.0 release tarball, because it is broken.
Questions?
whiteknight if we can get one or two big branches merged, that would be nice
dukeleto whiteknight: too vague.
whiteknight that's the way I like it
cotto_work Util had a question 21:11
Util I have re-written pbc_to_exe.pir in NQP, to see if it would be easier to work with. 21:12
It is! (~40% reduction in SLOC)
Any objections in principle to pbc_to_exe.nqp ?
NotFound Speed? Memory comsuption? 21:13
cotto_work speed is my concern
Util Speed went from 2 seconds (for perl6.pbc) to 6 seconds. Acceptable to me.
Memory use has not been checked, but should be resonable 21:14
(I use Q:PIR for one critical section) 21:15
cotto_work not too bad, given that it won't be run too many times during the rakudo build, though several internal parrot tools use it.
dukeleto nqp performance is platform-dependent.
Util: which platform did you see the jump from 2-6 seconds?
whiteknight If it's more maintainable and more readable, I think it's worth-while to do it
sorear kid51 might hate you, if the memory use increased much 21:16
can it still run with less than 192 MB working set?
Util sorear: kid51 has made me mindful of such restrictions; I will keep them in mind.
cotto_work kid51++ for that
Util dukeleto: my MacBook
dukeleto I like deleting code, but that file is a bottleneck
How many times do we call pbc_to_exe in a build ? 21:17
A few dozen for various tools?
that adds a lot to our build time
Util 13 files get places in Rakudo's bin dir on install 21:18
whiteknight It's more motivation for us to start radically speeding up NQP
Util Also, most are not as large as perl6.pbc, so will run faster.
whiteknight Util: is the code public? There might be some things we could do (dynpmcs/dynops) to increase the speed 21:19
dukeleto I have no problem *adding* a NQP port of pbc_to_exe, but I am -1 to deleting/replacing pbc_to_exe until NQP is faster
NotFound Did we really need the pbc_to_exe fakecutable? Will not be better to use its pbc? 21:20
dukeleto having both around will let us do benchmarks with them
Util I will start a branch, and put out the call for benchmarking when I think it is time.
whiteknight I like that
cotto_work I wouldn't require it to be faster, but at least more comparable.
dukeleto Util: +1
cotto_work +1
whiteknight +1
dukeleto cotto_work: yes, i meant NQP needs to get faster, so that the performance hit is less
cotto_work gotcha
dukeleto Sounds good. Next question?
If there are no more questions, we might have a wrap. 21:24
We need to fix the 2.10.0 tarball, if anybody has tuits, please volunteer in #parrot.
whiteknight good week, team! 21:28
moderator Priority for this week: GC branches(test and merge), Google Code-In | Don't forget to post closed tickets in your report. | Note: This channel is only for our Tuesday status meetings (at 20:30 UTC); you probably want #parrot instead. | irclog: irc.pugscode.org/ 21:29
21:30 NotFound left
dukeleto it is a wrap. Please tip your server and have a nice day. 21:30
21:30 dukeleto left 21:32 atrodo left 21:33 kid51 left 21:51 spinclad joined 21:58 whiteknight left 22:28 bacek left 23:00 tcurtis left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!