Priorities for this week: irclog.perlgeek.de/parrotsketch/201...#i_3798168 | Post closed tickets in your report. | Note: This channel is for our Tuesday status meetings (at 20:30 UTC); you probably want #parrot instead. | irclog: irclog.perlgeek.de/
Set by moderator on 24 May 2011.
00:06 lucian left 00:12 contingencyplan left 00:14 contingencyplan joined 01:20 whiteknight left 09:28 contingencyplan left 09:40 contingencyplan joined 09:52 contingencyplan left 12:57 bluescreen joined 14:25 lucian joined 15:26 lucian left 15:38 lucian joined 15:59 lucian_ joined 16:03 lucian left 16:44 mikehh joined 17:07 lucian joined 17:12 lucian_ left 17:33 lucian_ joined 17:37 lucian left 18:15 benabik joined, lucian joined 18:19 lucian_ left 18:22 lucian_ joined 18:25 lucian left 18:28 NotFound joined 18:33 contingencyplan joined 18:39 bluescreen left, bluescreen joined 19:05 kid51 joined
kid51 kid51's report 19:05
* DONE
** Took and closed these Trac tickets: TT #1979; TT #2025; TT #2116; TT #2117; TT #2119.
** Cage-cleaned; closed TT #2036; TT #2127.
** (But, with creation of new tickets, we're still at approx 520 tickets this week.)
** Drafted YAPC::NA::2011 talk.
** Provided cotto off-line feedback on github pull requests as patch references. 19:06
* WILL BE DOING
** Contemplating working on TT #849.
** More work on YAPC presentation.
* EOR
benabik DID: 19:15
- Met w/ Buffalo.PM, failed to find ride
- Continued my war on Q:PIR, got about 2/3rds of it done
- Mostly caught up with pmichaud's PCT changes in master
- Started tracking future TODOs in github issues: github.com/benabik/parrot/issues
WILL DO: 19:16
- Finish removing Q:PIR blocks from PAST
- Figure out how to use the new POST classes
- Blog
- See if I can carpool with anyone in Toronto.PM (or just rent a car)
SEMI-BLOCKERS:
- Strange goto logic
- Lack of bacek
- pmichaud changing PCT in master
- pmichaud back-porting pir::const::*
EOR
19:23 atrodo joined 19:30 soh_cah_toa joined 19:53 bubaflub joined
NotFound What I did: 19:55
-parrot
* Merged the with_winxed branch, a winxed snapshot now lives in master.
* Added the get_main method to the Eval PMC.
* Added the Eval get_pointer vtable to access its underlying packfile.
-winxed
* Added the write-to-output feature that was lost during refactorings,
its absence was causing bootstrap to eat too much memory.
* Use get_main to look for the main sub in the installed driver.
What I will do:
Figuring the better way to do basic tests of the winxed snapshot.
EOR
soh_cah_toa REPORT 20:01
PARROT:
* Started working on TT #2134
GSOC:
* Basic design is just about finished
* Today I will be implementing the CLI
* Blogged (parrot.org/content/progress-last)
TODO:
* Begin writing unit tests (I forsee this within the next few days)
* Have TT #2134 closed by next meeting 20:02
* Begin writing tutorial if I can get tests to pass
* Blog
EOR
benabik ADDENDUM:
DID: - blogged: www.parrot.org/content/gsoc-4-unsteady-ground
SEMI-BLOCKER: - NQP + multiple return = ??
EOA
tcurtis DID: 20:15
* Finished finals
* Moved out of dorm for the summer
* Almost done writing blog post, finally
WILL DO:
* Finish blog
* Actually generate the parsing tables for LR(0) grammars
* Hopefully get started on the LALR(k) parser tables
* Decide whether to go to YAPC::NA
EOR
bubaflub DID: 20:28
GSoC:
Blogged - www.parrot.org/content/parrot-gmp-g...-and-tests
Finished Winxed class with documentation
Handle strings properly 20:29
Begin writing tests
WILL DO:
Finish test suite, more docs, tutorial
EOR
Util # Done: Many new Perl 6 solutions on RosettaCode. Details at: irclog.perlgeek.de/phasers/2011-06-14#i_3926119
# Plan to do: TT#919 and TT#1302
# Blockers: $WORK has picked up.
.end
mikehh What I did since my last report:
* building and testing parrot on amd64/i386, with gcc/g++
* some fixes
* building and testing rakudo and winxed against latest parrot
* cleared up a bunch of compiler warnings in the build
especially src/pmc/structview.pmc which was a major problem
as line numbers did not correspond (due to macros I think)
What I intend to do in the next week:
* testing and fixing
.eor
cotto_work *did: 20:30
- M0 progress:
-- cleaned up a few loose ends in the spec
-- started a todo list: gist.github.com/1019986
-- wrote up rough draft intro to M0: gist.github.com/1022340
-- got the calling conventions test working 100%
- GSoC progress:
-- helped soh_cah_toa++ with various questions and build issues
*will do:
- gsoc as needed
- revamp parrot talk for OS Bridge
- M0 fibonacci example
*eor
q1q
mikehh q1q
benabik q1q
cotto_work hi 20:31
tcurtis Hello.
benabik Yellow.
soh_cah_toa hey
Util Hello
NotFound Hola
bubaflub hello 20:32
cotto_work How'd this past week go? kid51 seems to have gone on a ticket rampage.
mikehh hi there
cotto_work no takers? 20:34
Util kid51++
tcurtis I got very little done, I'm afraid. 20:35
NotFound with_winxed merging seems to have been harmless.
cotto_work ok. any comments on issues raised during the week?
NotFound: yes
mikehh we also reduced the number of warnings in the build
cotto_work that seems to be going well.
benabik I made progress, but had underestimated the sheer size of PAST::Compiler. Plus pmichaud keeps changing it out from under me. ;-) 20:36
cotto_work He's sneaky.
kid51 Down to 6 warnings during 'make' with gcc 20:37
mikehh I tested winxed, both with the version included and building from the repo
cotto_work Let's move to questions then.
mine: After a bit of discussion with kid51, it sounds like he's ok with using pull requests in addition to patches attached to trac. Any objection to making this official?
Util +1 20:38
kid51 +1 if that '.diff' hack continues to work; +0 otherwise
cotto_work alright. I'll continue to look for official documentation that says it's a long-term supported feature. 20:40
eqo
mikehh: go ahead
*eoq
mikehh With regard to tickets/tests - when a ticket is closed, we really need to
include an appropriate test to cover the problem that causes the failure
with the ticket so that it does not happen again and we have a regression
test for it.
What I really propose is that we do NOT finally close the ticket unless we
have an appropriate test (if that is applicable)
cotto_work mikehh: I remember a discussion on this and suspect that that's our official policy. 20:41
NotFound Even better, add the test before the fix. 20:42
cotto_work at any rate, it's appropriate and recommended to reopen any tickets that get closed without tests for which reasonable tests can be written.
20:42 Searle joined
mikehh NotFound: of course 20:42
cotto_work q1q
actually, unqueue that
mikehh I have a couple of tickets that had problems, that look as though they gone away, but have no idea what caused the problem in the first place 20:44
s/they/the problem has/
cotto_work mikehh: feel free to reopen tickets as needed. We need to be sure that problems aren't reappearing. 20:46
NotFound Closing tickets is a mean, not a goal by itself.
mikehh what I need is test tuits
20:46 bluescreen left 20:47 bluescreen joined
cotto_work Closing closeable tickets is good because it reduces the noise in the ticket queue and makes it more approachable. 20:47
mikehh one of my current TODO's id to reexamine all skipped tests
is
cotto_work mikehh: does that answer what you need answered? 20:48
mikehh gah, getting used to the new keyboard
NotFound Yeah, even just figuring a way to change to todo will be better than let them skipped.
mikehh just really wanted to bring it up, we need more test coverage, especially where we had failures 20:49
cotto_work ok. Thanks for bringing it up. It is important to avoid regressions. 20:50
benabik: go ahead
20:50 particle left
benabik I haven't seen bacek in a while, and hadn't talked to him much when he was around. IIRC, PerlJam was my backup mentor... Most of my Qs have been answered by various folks on #parrot, but I wanted to be sure there wasn't some administrative stuff I wasn't doing by not talking to bacek. :-/ 20:51
cotto_work benabik: if you can't catch him, we can't blame you for not doing so.
keep trying and hopefully others can answer your questions 20:52
mikehh benabik: that is also why we have backup mentors
benabik I'll poke PerlJam with a stick. I haven't been blocking on anything, really, just don't want my midterm eval to say "never talked to mentor". :-) 20:53
eoq 20:54
mikehh benabik: you might try bacek_at_work, don't know if he is checking that though
cotto_work He's been quiet recently.
benabik He had mentioned that $job or $life was busy when last I had talked to him. 20:55
mikehh he did mention that he had a major project coming up
cotto_work q1q
benabik Although talking to bacek soon would help. I'm getting to the point where I need to figure out his code instead of pmichaud's. :-D 20:56
EOQ (for real this time?)
mikehh benabik: jut remember feel free to ask 20:57
just
benabik mikehh: I haven't been too shy so far, I think. I even sometimes talk to myself on #parrot trying to figure things out. :-D 20:58
cotto_work asking myself questions on #parrot is surprisingly helpful
my question: It's getting really old to be missing kid51, dukeleto and whiteknight at these meetings. I'm starting a doodle to find a better time and sending it to parrot-dev. What times of day don't work for people here? 20:59
Util is full adaptable (modulo U.S. sleep hours) 21:00
kid51 cotto_work: I recommend starting by asking whiteknight and dukeleto what times work for them, as it is their absence that hurts us the most
benabik mostly ditto, although $life does happen in US/Eastern evenings.
cotto_work kid51: good idea. 21:01
kid51 ... at least, limiting ourselves to devs in Western hemisphere (unlike bacek)
21:01 lucian joined
cotto_work kid51++ for ticket rampaging, btw 21:01
kid51 With devs around the globe, no time will work for everyone, so you have to figure out how to maximize chances your most active devs can attend. 21:02
cotto_work Right. We can do better than what we have now.
21:03 particle joined
kid51 And, once we get a better time, pre-posting discussion topics on parrot-dev is highly recommended (like the github pull thing) 21:03
tcurtis has no time at which he absolutely cannot attend.
21:04 lucian_ left, Searle left
mikehh I am pretty flexible since I work from home and have clients in Chicago and Southern California and elsewhere 21:04
cotto_work I'll start with team leads and post the common availability to parrot-dev in a doodle.
kid51 Until last year, I could not login to #ps from $job at all. Now I can, but I mostly get distracted with $job issues.
cotto_work Does anyone have suggestions for goals or other points of discussion? 21:06
kid51 They're paying me; y'all aren't. :-)
mikehh we must also consider our devs in Europe and elsewhere, but maybe we can vary the time to fit others in
kid51 mikehh: To what extent are our devs in Europe in contact with one another? 21:07
cotto_work mikehh: certainly. Europe is a nice place.
kid51 E.g., have you ever met any other parrot devs F2F?
mikehh no idea, I mostly contact people in #parrot and #perl6 etc 21:08
kid51 We often seem to be on a "1 parrot dev per nation" quota in europe ...
... which makes imagining a Parrot dev gathering difficult. 21:09
mikehh don't think I have met any of the devs face-to-face
kid51 I believe some folks may go to YAPC::EU, which this year is in Latvia
NotFound A bit too far to me.
cotto_work I'll be there 21:11
It seems we're slowing down. Should we set any collective goals for the week? 21:14
mikehh We have a major release next week don't we? 21:15
cotto_work not a major release, but a normal release
mikehh oh yeah, the .6 release is the major one 21:16
cotto_work I'll probably call a code freeze for the weekend. In the meantime, testing HLLs is a good idea
NotFound It's Major with capital M, because it will be the first with Winxed Inside(tm) ;)
cotto_work NotFound: you have Winxed trademarked? ;] 21:17
NotFound cotto_work: I'd like it, but won't pay for the privilige.
mikehh good for a bugfest over the weekend 21:18
was over in #phasers earlier (on freenode) we should have more representation there 21:19
cotto_work mikehh: good idea for those with tuits 21:20
21:20 bluescreen left
cotto_work I'd like to see fewer than 500 open bugs for 3.5.0. 21:21
mikehh just got the impression that perl6 devs are not too happy with parrot
and more tests
cotto_work here's what I have: 21:22
GOAL 1: solid 3.5.0 release (no bugs or regressions)
GOAL 2: get bug count below 500
I agree that we need someone in #phasers. When does it happen? 21:23
mikehh 20h00 GMT today
cotto_work 30 before #ps
mikehh an hour and a half before #ps 21:24
cotto_work ok
benabik I tend to be in #phasers, mostly out of curiosity.
mikehh sorry must be 19h00 GMT, I am on BST
cotto_work anyone who's there, please summarize what's going on when you can 21:25
mikehh benabik: yeah, saw toy there
you
tcurtis lurks but doesn't always read.
mikehh irclog.perlgeek.de/phasers/today 21:26
kid51 What is purpose of #phasers?
cotto_work I need to attend to $dayjob. Feel free to continue discussion, but let's call the official part of the meeting over.
tadzik kid51: #ps for Perl 6 folks
mikehh pretty much the same as #ps I think, for perl6 21:27
21:28 kid51 left 21:43 bubaflub left 22:10 benabik left 23:00 whiteknight joined
tcurtis whiteknight: now's fine. 23:39
whiteknight awesome. I read your blog post 23:40
it's good most of your distractions are disappearing
tcurtis whiteknight: yes. I'm quite sorry about the lack of getting much done in the last couple of weeks, though. 23:42
whiteknight nothing to be sorry about, we just have to figure out how far off schedule you are, and what to do about that 23:43
do you have a link handy to your timeline?
23:43 lucian left 23:44 kid51 joined
whiteknight nevermind, I got it 23:44
tcurtis www.google-melange.com/gsoc/proposa...ercurtis/1 # in case anyone's reading along
whiteknight okay, so do you have the algorithm digraph built? Have you started with any testing? 23:45
tcurtis No. I haven't started on that. I haven't yet done the LR(0) parsing tables part. 23:47
whiteknight okay
okay, so you have the classes for the grammar written?
or, at least prototyped?
(I'm not trying to prod, I'm just trying to get an idea of status) 23:48
tcurtis whiteknight: yes. Currently, the only changes I expect to that are to replace my ad-hoc array copying for .nonterminals/.terminals with a Rosella.Proxy and to do something similar for .rule() 23:49
whiteknight oh, is that what you're going to be using the immutable stuff for? 23:50
tcurtis And those are both things I feel are okay to put off until later.
Yes.
whiteknight yeah, that doesn't seem like a high priority task. Leave a TODO note in the code, but don't worry about performance right now 23:51
23:54 particle1 joined
tcurtis So, getting the LR(0) parsing tables done is my immediate priority, followed by catching up on the generalization to LALR. 23:54
23:54 particle left
tcurtis Now that I've read more of the background material, I suspect that that generalization will probably not be too difficult, so I think I can likely catch up a bit there. 23:55
whiteknight okay, don't underestimate. You're probably not gong to make up the time lost 23:56
I mean, just focus on doing the tasks, not doing them AND making up tme 23:59
time

Powered by Cro and the Raku® Programming Language.

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

Thank you!