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.
09:12 particle1 joined 09:16 particle left 12:05 whiteknight joined 12:34 contingencyplan left 12:35 contingencyplan joined 13:17 particle1 is now known as particle 15:12 particle1 joined 15:16 particle left 15:35 particle1 is now known as particle 16:40 particle1 joined 16:43 particle left 16:48 lucian joined 16:52 particle1 is now known as particle 18:05 benabik joined 18:29 lucian left, kid51 joined
kid51 kid51's report 18:30
* DOING
** Looking into these Trac tickets: TT #1682; TT #1979; TT #2116; TT #2117; TT #2118 (tt2118/check_depend branch). More eyeballs welcome on all of those.
** Planning on attending both YAPC::NA in AVL and FOSSCON in PHL.
EOR
18:32 lucian joined 19:00 lucian left 19:01 lucian joined 19:18 NotFound joined
NotFound What I did: 19:19
-parrot
* Created the with_winxed branch.
-winxed
* Some special purpose scope search in stage 0 to allow namespacing
of stage 1.
* Some more improvements in scope search.
* More tests.
* Enclose stage 1 classes and functions in namespace Winxed.Compiler.
* Added option to send output directly to a handle and use it in the
non installed driver to reduce bootstraping memory needs.
* Minor fixes.
What I will do:
No plan
EOR
benabik DID: 19:23
- (Ab)used my commit bit. nqp_pct now updated in parrot/parrot
- Got my code to pass codetest
- Removed the last bit of Q:PIR from PAST::Node and children
- Blogged: www.parrot.org/content/gsoc-3-my-go...-full-goto
19:23 atrodo joined
benabik WILL DO: 19:23
- Continue my war on Q:PIR
- Go to Buffalo BarCamp to meet people from Buffalo.PM
-- Hopefully will find someone to carpool with
BLOCKERS:
- goto, goto, goto
- my $old := $*VAR; my $*VAR := $new # Doesn't work!
EOR
lucian What I did: 19:27
- significantly refactored object system
- improved syntax
- paved the way for interop
- added more tests (but broke older tests)
- fixed bugs
- attribute resolution
- object creation
- type instantiation
What I'll do:
- finish essential object system bits
- functions
- exceptions
- fix __getattribute__ bugs
- implement more types
- retarget compiler to use object system
EOR
19:39 bluescreen joined 19:45 smash joined 19:50 kid51 left 20:09 mikehh joined
Util # Done: 20:15
* Previewed YAPC::NA::2011 talk at Atlanta.pm
* Added Perl 6 solutions for 2 more RosettaCode tasks.
# Plan to do:
* Commit code for TT#919
# 7-day ticket report:
1 closed: done
2 closed: fixed
3 new
.end
tcurtis exec -out cat /tmp/ps-report 20:17
oops
20:19 soh_cah_toa joined
tcurtis DID: 20:21
* Took most of my finals
* Did some refactoring of the grammar class
* Read some more
WILL DO:
* blog
* write me a shiny LR(0) parser generator
* finish finals, pack
EOR
soh_cah_toa REPORT 20:23
* Added support for non-pbc files
* Minor additions like switches, help, etc.
* Began work w/ breakpoints using program counter instead
* Blogged
TODO:
* Begin writing a tutorial
* List source command
* Blog 20:24
EOR
cotto_work *did: 20:28
- was asked to write a chapter about Parrot for volume 2 of AOSA ( www.aosabook.org )
- M0 progress:
-- write a bunch of tests for math ops
--- if someone wants to jump in, the boolean ops need similar tests
-- started writing example/test for calling conventions
-- also made a few spec updates to support calling conventions manipulation
- GSoC progress:
-- helped soh_cah_toa++ with various questions and build issues
*will do:
- m0: implement *working* calling conventions test
- plan out the AOSA chapter
- gsoc help as needed 20:29
*eor
q1q
20:29 kid51 joined, whiteknight left
cotto_work hello 20:30
20:30 bubaflub joined, bubaflub left, bubaflub joined
mikehh hi there 20:30
bubaflub hello 20:31
soh_cah_toa ahoy hoy
kid51 Hello
cotto_work How'd the past week go? What was great? What wasn't?
I didn't have any goals written down for the week, but general comments are still welcome. 20:32
kid51 It appears that several of the GSOC projects are underway, as evidenced by ... 20:33
... conversations on #parrot;
... several of the GSOC students now have commit bits and have begun committing.
We have approximately the same number of open TTs as we have had for the past month. 20:34
521 the way I count (which is slightly different from Util's but trends similarly) 20:35
cotto_work It's been a while since we took a shot at reducing that number.
kid51 q1q 20:36
cotto_work Let's make it a goal to get that back down. 510 would be reasonable.
kid51 I would like to extend that a bit. 20:37
By the time of the next supported release (3.6, July 19, for which I happen to be RM), I would like to have the number of BZ tickets below 500 ...
NotFound Hola
mikehh we really want to get that below 500, then 400 etc 20:38
kid51 ... and I would like to eliminate most of the warnings we are getting during 'make'.
Util +1 to warning reduction as a goal
mikehh kid51: been working on it - still some to go :-}
cotto_work I like that. 20:39
kid51 mikehh: Do you have any insight into all the 'structview' warnings during 'make'?
mikehh kid51: yeah - have got a good idea on most of them, how to fix it - hmmnnn 20:40
actually I think I can do most of them without breaking anything
kid51 mikehh: Can you reach out to whoever worked on the respective source code files? plobsing?
mikehh yup mostly plobsing++ 20:41
cotto_work That PMC has some unusual code. It's not surprising that it'd produce warnings.
kid51 k. Not this week, but maybe in the next, I'll post to parrot-dev a request about the warnings and describe what remains to be done.
cotto_work It also needs more tests. 20:42
Unfortunately it's not especially easy to understand as-is.
kid51: was that your question?
kid51 Not so much a question as a topic. 20:43
benabik q1 idea 20:44
cotto_work benabik: fire away
kid51 I have a 20-minute slot speaking at YAPC in < 3 weeks about Parrot as a project.
I am facing writer's block on that ...
... mostly because the Parrot Foundation side of what I do has been so depressing. 20:45
benabik Warnings and build errors are VERY easy to miss in a noisy build. Git has a few tricks that make "gcc -o file.o -DDEFINE -Wwarning ....." into just "GCC file.o"
Would a patch that makes the normal build less full of noise be welcome?
kid51 Well, I'll stop for the other discussion to -proceed.
benabik (Git's Makefile, not git the program)
cotto_work benabik: we used to have a perl script that did something similar, but we removed it in the interest if making the build more transparent. 20:46
thanks for clarification
benabik github.com/git/git/blob/master/Makefile#L1598
kid51 benabik: Actually, if you grep the output of 'make' for 'warning:' -- note the colon -- you get exactly what you need.
cotto_work perhaps a make target that makes warnings more apparent is in order 20:47
mikehh I usually do a build to a file so I can peruse it a leisure (date && time make world 2>&1 | tee make_world.3_4_0-169.32.log) or something like that 20:48
benabik I could invert the way git does it. Verbose by default, quiet on `make Q=1`... I generally find that it makes any irregularity in build stand out far more.
But if something similar has already been pulled out, that's fine. EOQ 20:49
cotto_work kid51: back to you 20:50
kid51 So I suspect now that my talk will take the format of: "10-20 questions which participants and leaders in OS projects should ask themselves"
I am willing to include suggestions for such questions from the members of this project (not here; send email). 20:51
(I haven't formulated these thoughts much beyond this point.) 20:52
mikehh kid51: what specifically depresses you about PaFo
kid51 The internal functioning of the board has been suboptimal -- but that's not really appropriate for #ps.
mikehh kid51: 'k 20:53
kid51 The legal entities an OS project establishes for itself are *some* of the questions an OS project should consider -- but are nonetheless a minority of such questions. 20:54
So, if people want to send me their thoughts about questions that are important for OS projects -- as human communities -- to consider, that would be appreciated. 20:55
EOQ
soh_cah_toa kid51: i don't think i have your email address? what is it?
kid51 For here, let's use: jkeenan at cpan dot org 20:56
soh_cah_toa great
kid51 (I'll reply from the real one, which is probably in CREDITS.)
soh_cah_toa oh yeah, CREDITS
cotto_work my q: How do you all feel about officially considering github pull requests equivalent to attached patches for trac tickets, i.e. submitting a ticket for with a pull request is equivalent to submitting one with a patch? 20:57
kid51 cotto_work: As you might suspect, I'm not in favor of this. 20:58
One more thing I (or anyone else) has to learn in order to participate in the project.
Speaking personally, with the limited amount of time I have available to contribute to the project, ... 20:59
... I don't have the time to learn all the advanced git (or, really, github) stuff that other people do.
lucian kid51: afaik, you can download patches from github forks
cotto_work My view there isn't anything to learn. Just look at the pull request and click "diff" to see the most current state of the request. 21:00
kid51 has to attend a meeting at $job; sorry
cotto_work eoq then 21:02
any other questions? We'll have to consider that one unresolved. 21:03
Util No objections on my part # on pull req
soh_cah_toa i have a question 21:04
cotto_work I suspect that kid51 is the most likely to object, but I don't want to override his concerns without giving him a say. I'll take it to parrot-dev.
soh_cah_toa: go ahead
soh_cah_toa i've noticed that the copyright notice at the top of source files still read "2001-2010." now, i'm no expert in copyright law but shouldn't these be updated to 2011 since...well, it's 2011?
for some at least. not all of them 21:05
mikehh If I modify a file I usually remenber to update the copyright (not always unfortunately)
cotto_work soh_cah_toa: mikehh's approach is what we've been doing.
mikehh so the copyright reflects the last time the file was updated 21:06
soh_cah_toa i see, i'll have to remember that 21:07
cotto_work If there's not a compelling reason, it's a lot of churn without any benefit.
Util you cannot claim copyright as of every year that something is in existence, just years of "publication" with changes.
(IANAL) 21:08
lucian Util: so a release would mark that?
i guess publication is fuzzy when source is available
soh_cah_toa right
mikehh wrt changes, certainly
Util I would think so, but you would only change/add the date if the file had changed as well. 21:09
soh_cah_toa alright then 21:10
cotto_work Does anyone have other questions or suggestions for goals? 21:11
lucian cotto_work: is this a time to report on 6model investigation promised last #ps? 21:12
cotto_work lucian: sure
mikehh work on tickets, especially older ones, fix warnings for next release as far as possible, check skipped tests
lucian from what i've seen, it's a rather traditional MOP
benabik q1q
lucian if you implement an object system, all your classes/objects will be instances of some host system's classes 21:13
'class' is a loose term here, any collection of structures and callables could be used
atm, the host system is nqo
nqp
and almost everything about the object system you implement is customisable, much more so than parrot's Class/Object 21:16
cotto_work lucian: eor? 21:19
lucian cotto_work: pretty much 21:20
if anyone has questions, you can ask. but don't have high hopes
i guess i could add that i'm not using it atm, but what i do have is quite similar in design 21:21
cotto_work The goal I have listed for the week is to reduce our ticket count to 510. 21:25
any others?
21:26 bubaflub left
mikehh fix compiler warnings 21:26
cotto_work mikehh: can you make that more concrete? 21:28
mikehh try and have a clean make for parrot with as few warnings as possible for the next release
NotFound q1q 21:29
cotto_work NotFound: go ahead
NotFound Merge with_winxed branch? 21:30
mikehh NotFound: looks good to me
cotto_work Has anyone tried it? 21:32
NotFound You mean other than me? I think lucian is using it for his project.
cotto_work the included version?
lucian no 21:33
i use separate winxed
NotFound Then I don't know.
lucian in my defence, building parrot here takes a bit too long
NotFound I've built and used it in i386 and amd64 21:34
mikehh NotFound: I did some tests on the branch, but I have the same platforms as you do 21:35
lucian will take one for the team and build it on arm soon
NotFound I can use some gcc farm machines, but the interesting ones are painfullly slow. 21:36
lucian my gf likes Hayden Panettiere, can't say i don't agree 21:37
bah, wrong channel
lucian apologises profusely
mikehh lucian: happens to us all :-}
NotFound I don't see a need for a wider testing, tough. Is just compiling pir to pbc and a bit of make. 21:38
cotto_work NotFound: I'm fine with merging with_winxed into master.
any objections?
kid51 is back
NotFound BTW ticket for the branch is TT #2127 21:40
If someone not attending now has any problem or suggestion, please comment in the ticket. 21:41
21:41 bluescreen left
cotto_work NotFound: wfm. Give it a few days and if nobody objects, merge it. 21:41
NotFound Ok
eoq 21:42
mikehh I think benabik had a question
cotto_work benabik: go ahead
benabik Is anyone maintaining nqp-rx? I know jnth's moved onto nom and such.
cotto_work benabik: nqp (new) is the recommended replacement. I believe there are few compatibility breaks. 21:43
lucian notes that nqp(new) comes with 6model 21:44
benabik cotto_work: Is new NQP going to become parrot-nqp soon, or if I find minor bugs/enhancements for -rx, should they be directed somewhere?
cotto_work benabik: that's a good question
sorear benabik: AIUI, nqp-rx has been forked and is maintained by #parrot now, but its days are numbered, most Parrot hackers have moved to Winxed 21:45
mikehh the last changes in nqp-rx were 3 months ago 21:46
lucian sorear: i have observed the same 21:47
benabik Well, I'll bring up patched in #parrot then. 21:50
cotto_work benabik: ok.
kid51: do you have tuits to resume the github pull request discussion? 21:53
21:56 lucian_ joined
cotto_work I genuinely think that using them means less work for us Parrot developers and don't understand why an attached patch would be preferable. 21:58
kid51 got called back into $job discussion 21:59
lucian_ not to be rude, but i think kid51 just needs to read a few tutorials
kid51 recommends cotto post something to parrot-dev against which we can provide feedback 22:00
lucian_ well, i was rude. also, the 'hub' cli tool helps
22:00 lucian left, lucian_ is now known as lucian 22:01 kid51 left
cotto_work let's call it a wrap. I'll move the github discussion to parrot-dev 22:01
mikehh goals? 22:02
cotto_work 510 tickets by this time next week 22:08
22:35 soh_cah_toa left 23:04 whiteknight joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!