"Tuesday at 20:30 UTC"
Set by moderator on 20 September 2010.
00:24 whiteknight joined 00:44 particle joined 03:03 whiteknight left 03:05 bluescreen left 05:01 eternaleye left, eternaleye joined 09:11 bacek left 10:13 bacek joined 10:39 contingencyplan left 11:43 kid51 joined
kid51 Parrotsketchers: Remember that 20:30 UTC arrives one hour earlier in most of North America starting today: 3:30pm EST, 12:30pm PST. 11:46
kid51's report 11:57
DONE
* Closed, re-opened and re-closed trac.parrot.org/parrot/ticket/1841 adding 'osvers' key-value pair to configuration data.
* Closed trac.parrot.org/parrot/ticket/396 "32/64 bit data on smolder reports"
* Worked on several older ticket:
* trac.parrot.org/parrot/ticket/855 "profiling options too specific to GCC"
* trac.parrot.org/parrot/ticket/532 "headerizer refactor" 11:58
* trac.parrot.org/parrot/ticket/1415 "parrot macport"
WILL DO
* Will continue to work on above tickets
* Won't have time for Parrot work until Wed eve; will probably be absorbed with transition to git rest of week
REQUESTS
* Need feedback on trac.parrot.org/parrot/ticket/1840 "config step to detect RTEMS"
* trac.parrot.org/parrot/ticket/1844: Can someone add basic git instructions to README? (These could be brief but also point to more complete discussion on Parrot wiki.)
And thanks to dukeleto++, mikehh++, cotto++ and all other who spent many hours on version control system transition. 11:59
END
12:10 particle1 joined 12:13 particle left 12:29 bluescreen joined 13:24 kid51 left 14:12 dukeleto joined 14:31 bluescreen left, bluescreen joined 14:32 bluescreen left, bluescreen joined 15:19 plobsing_ joined
plobsing_ What I Did: 15:26
* worked towards gsoc_nci merge (ugly sync w/ master, codingstd)
* complain (TT #1842)
* support for parrot-libjit-fb (yay I haz a user)
What I Plan:
* gsoc_nci merge sometime *this* week (testing much appreciated, ask me how)
* concrete action plan for TT #1842 (pending decision in #ps)
EOR 15:27
15:41 atrodo joined 15:47 NotFound joined 15:50 whiteknight joined
whiteknight WHAT I DID: 15:59
* Cheered from the sidelines for the git migration. Kudos to all involved.
* Created a fork on github, started prototyping the new embed API. Coming to a point where I will want wide feedback, especially on a few key issues
* Blogged, and got my new blog feed up on planet parrot
* Merged in the stdhandle_meths branch. I had intended to do it weeks ago but it fell through the cracks. Done now.
* Created a small handful of GCI tasks, have more ideas, but need time to jam them in.
WHAT I WILL DO:
* Continue working on the new API
* Planning to create a new project to test it. The xchat-parrot project I started doesn't seem the right place for it.
* Create more GCI tasks
WHAT I AM BLOCKING ON:
* Teething baby. Seriously, the poor kid has like 8 teeth coming in simultaneously.
EOR
16:17 darbelo joined 16:26 plobsing_ left
dukeleto Did: 16:30
* Worked on GCI stuff. Still need to do a lot more.
* Converted us to Git. Had to do it again because I accidentally gave commits from GeJ++ to gerd++
This is now fixed and all is well. You need a "git pull --rebase --tags" if you have the old repo.
Or you can clone again.
* Currently at an Evolutionary Biology Hackathon, and talked to many people about Parrot and Perl 6,
fixing FUD.
* Looked at some of our failing tests for Parrot::Revision. I think Parrot::Revision is a big ball of overcomplication.
It needs to be fixed with fire.
Will do:
* Make our org page for GCI
* Add more tasks to GCI
* Write blog posts about GCI
* Write a "how to submit patches" section for our git_workflow.pod, requested by allison
Blocking on:
* Wanting to be utterly lazy, but failing at that.
.EOR
16:45 particle1 is now known as particle 16:55 contingencyplan joined 17:03 dukeleto left 17:05 dukeleto joined 17:11 dukeleto left 17:13 dukeleto joined 17:24 dukeleto left 17:37 dukeleto joined, mikehh joined
mikehh What I did since my last report: 17:48
* building and testing parrot on amd64/i386, with gcc/g++
* some fixes
* branch testing and fixing
* did some work on GCI - needs to do more
* got my git migration working - trying to sort out problem tests
What I intend to do in the next week:
* testing and fixing
* get something workable for GCI
.eor
18:11 moritz joined
moritz I have more of a question than a report 18:12
I have a local patch for parrot that changes the 'revision' config to git-describe output instead of a SHA1
moritz@trudi:~/p6/rakudo/parrot>./parrot_config revision
RELEASE_2_9_1-204-gc762e42
this has the advantage of letting us (ie the rakudo and other HLL devs) compare revisions without having the parrot git repo available 18:13
any objections to me pushing the commit?
EOQ.
P.S.: patch at nopaste.snit.ch/25448 18:18
18:47 stilgar joined
dukeleto moritz: let's talk about that question in #parrot 19:03
moritz dukeleto: wfm
cotto_work *did: 19:29
- helped dukeleto knock out the git migration
- tried and failed to notifty parrot-dev due to the latter being broken
- worked with osuosl to get trac working with the github trac plugin
- udpated trac to include the history dukeleto rewrote
- everything seems to be working now 19:30
*will do:
- help clean up the leftovers from the git move
- get back into opmap_aware_pmcs
- raise the bus number for Lorito's overall design
*blockers:
*eor
19:48 chromatic joined
chromatic I'm still poking at the GC MS2 tuning branch. A full Rakudo spectest run on 64- and 32-bit platforms would make me more comfortable about merging. 19:50
NotFound What I did (last two weeks): 20:03
-parrot:
* Added a few test
What I will do:
Get used to git.
EOR
Util moritz: +1 to push "git describe" patch. 20:14
# Done:
* No pure-Parrot work this week.
* Added 14 Perl 6 solutions to RosettaCode
* Expanded RosettaCode Quaternion solution to full Perl6 module.
# Plan to do:
* Contribute Advent calender ideas.
* No pure-Parrot work next week. Pbc_to_exe after that.
.end
20:18 tcurtis joined 20:27 kid51 joined
whiteknight 'ello 20:33
kid51 Hello, #parrotsketch
chromatic hello
mikehh hi there
darbelo Hi
cotto_work hi
chromatic I hear our migration worked.
whiteknight yessir. Wonderfully
cotto_work no limbs were lost
chromatic Now Rakudo needs some way to track head. 20:34
kid51 Have we identified loose ends and the people who are going to tie them up?
dukeleto hola
Blood was shed in the migration, but I have recovered.
kid51: are you volunteering? ;)
mikehh seems to be working - test seem to pass now, wqith a bit of fiddling 20:35
kid51 No, because I won't have any time to spend on that until next Monday at the earliest
mikehh with
dukeleto Parrot::Revision is currently slightly broken and needs to be cleaned with fire.
kid51: well, then. No fun for you.
kid51 I can look at Parrot::Revision but only if it's non-urgent
dukeleto kid51: i am most likely going to delete it, so it isn't urgent
mikehh I've got 3 tests TODOed at the moment
dukeleto mikehh: i just fiddled with those :)
mikehh: git pull --rebase :) 20:36
chromatic Can I suggest an alternate approach for Rakudo?
dukeleto chromatic: alternate approach to what?
chromatic Compare timestamps instead of SHAs.
kid51 IIRC we used Parrot::Revision to generate data used in 'parrot --version'
mikehh kid51: yes
dukeleto chromatic: we are probably going to compare git-describe strings. But dates could work for most cases. But they will fail when it comes to corner cases in version numbering. 20:37
NotFound Sorry, got distracted. Hola.
chromatic We don't need to compare version numbers when working with trunk.
whiteknight We need to capture all relevant data, and leave it up to the HLL devs to decide how to use it in comparisons 20:38
dukeleto chromatic: i have a plan.
chromatic My concern is overcomplication.
dukeleto Tarballs can only depend on version numbers. If the parrot repo is available, then sha1's or describe strings can be used
chromatic: i worry about that as well. 20:39
chromatic Separating tarballs from head is a good start.
dukeleto so tarballs will use the 'version' config key, repos can use the 'sha1' or 'git_describe' config key 20:40
chromatic Anything else we need to schedule/prioritize regarding the migration?
dukeleto I just pushed a commit that makes Parrot::Revision always return "r50000" to tell old HLLs "this parrot is too new for you"
i will store the git sha1 in a 'sha1' config key 20:41
kid51 Has the Subversion repository been made read-only?
mikehh I kinda feel a bit lost at the moment - it is surprising how much I relied on emails and dalek to work out what has been happening
cotto_work kid51: yes
dukeleto kid51: yep, a while ago
darbelo dukeleto: No HLL cheks for that. They check for a parrot *too old*. 20:42
mikehh at r49810
dukeleto darbelo: hmm. I suspect you are right. My thinking was backwards.
kid51 mikehh: I agree with you about the "lost" feeling
dukeleto darbelo: i guess i should be returning 0 instead
kid51 mikehh: What I've come to realize is that there are a vast number of administrative tasks needed to keep this project going. 20:43
dukeleto why do y'all feel lost?
mikehh kid51: yeah definately
dukeleto i understand that some things, outside of our control, conspired to make the git migratin more difficult
like parrot-dev going down
kid51 Because I like to know what's being committed, when the wiki is changing, etc.
dukeleto I can't do anything about that.
kid51: git log -p
dalek went down: That was out of our control as well. sorear++ just fixed it 20:44
kid51 dukeleto: I haven't had time to begin working with the github repo, so git log -p is a ways off for me.
mikehh dukeleto: was just saying I looked at the emails and dalek stuff in #parrot a lot
kid51 dukeleto: But is sorear's fix *still* working?
dukeleto kid51: well, let me know when you catch up to the future. There is also github rss feeds that you can subscribe to. 20:45
kid51: if you want to volunteer to check, go ahead
Complaining without volunteering isn't very useful.
kid51 dukeleto: I don't find these comments helpful.
whiteknight no fighting
mikehh dukeleto: not complainin' just observin'
dukeleto It was a general statement, i am not trying to say that kid51 or mikehh are complaining. 20:46
chromatic Shall we move on?
dukeleto But as someone who spent dozens of hours in the last few days on the migration, I am hearing more complaints than thanks.
And that doesn't make me feel good.
kid51 dukeleto: The problem is that we need people for these administrative/infrastructural tasks who are not also heavily involved in coding.
mikehh dukeleto: ++
dukeleto kid51: yes, i agree. 20:47
mikehh or should I say dukeleto++
NotFound dukeleto: thanks!
cotto_work We do plan on making parrot-commits work again. We might just need a breather.
whiteknight it's a matter of deciding how they should work. We could have github post-commit emails being sent out immediately if that's what we wanted
dukeleto Just to clarify: I do not think kid51 or mikehh were complaining, i was just expressing a general feeling.
whiteknight if we want a format or features they don't offer, that's different 20:48
cotto_work Seeing that a commit happened and which files it touched is nice (which is what github offers), but having the contents of those diffs is much nicer.
whiteknight cotto_work: yeah, but if those diffs are huge, getting them via email is lousy 20:49
20:49 darbelo left
cotto_work I suppose we could easily enough enable github's default emails until something better is possible. 20:49
20:49 darbelo joined
chromatic Do we have a volunteer to take that on? 20:50
cotto_work whiteknight: we can have separate lists if you prefer. It's a separate list so it won't be sent to anyone who didn't sign up for it.
er, separate lists for what github sends out by default and for the full diffs
whiteknight cotto_work: if we don't have a custom solution in place soon, the standard github emails do seem like a nice stopgap
cotto_work I'll get that going.
whiteknight who knows, after experiencing them, people might prefer to more brief emails 20:51
dukeleto cotto_work: it should be clicking one button in the github admin interface
cotto_work using parrot-commits for now
dukeleto: and any permissions issues with osuosl
mikehh I just find that the fact that some parrot mailing lists went down and dalek gave up just when we were migrating is just another case of Murphey's Law
whiteknight yes, it was a particularly bad confluence of events
the osuosl guys don't seem to know the root cause of the list failures, but they are watching it closely 20:52
20:52 nwellnhof joined
dukeleto so, what is our business for this #ps? 20:53
most of this conversation could be in #parrot
whiteknight We have 2.10 coming up in a week. tcurtis is on the chopping block for it
mikehh 'k - I think chromatic should merge gc_ms2_sf 20:54
whiteknight he's been busy, we should probably check that he's still up for it
chromatic Are there any features we must merge for 2.10?
whiteknight mikehh: Agreed
chromatic or deprecated features to remove?
dukeleto gsoc_nci is *so close*
cotto_work We need a volunteer to work through the git-based release process to make sure it's workable.
chromatic Oh yes, gsoc_nci.
whiteknight I need to get in touch with shockwave, I think I have all the embedding interface stuff he requires for his work. It's a very very small subset of things I will be adding
so we're good on that front for 2.10
I can put some eyes on gsoc_nci in the next few days, probably 20:55
dukeleto Since I wrote the docs for a git-based release, I am not a good person to test it. But I will help whoever does it.
plobsing said he needed testing on gsoc_nci
chromatic Do we have a volunteer to manage the gsoc_nci merge? 20:56
kid51 Re deprecations:
chromatic I'll work on gc_ms2_sf merge after I hear about whiteknight's testing tomorrow.
kid51 nwellnhof: I think the branch mentioned here is ready to merge: trac.parrot.org/parrot/ticket/1778
Correct?
nwellnhof yes, i was just waiting for the git switch to happen. 20:57
mikehh chromatic: I tested on i386 and amd64 (g++ with --optimize) taests same as in trunk
but a bit faster, especially rakudo build (10% improvement) 20:58
kid51 Also re deprecations: If we can get feedback from plobsing/luben we can address this one: trac.parrot.org/parrot/ticket/1655
whiteknight chromatic: I will try to push that up to tonight then, if you're waiting on little ole me
chromatic Please do, thanks.
Confirmation from two other people increases my confidence.
whiteknight plobsing was looking at gsoc_nci too. I'm sure he'll have his hands on it
chromatic Other organizational or scheduling issues to discuss today? 21:00
whiteknight none for me
chromatic Question time then. 21:01
Any questions? 21:02
mikehh what about GCI?
dukeleto I need help with it.
dukeleto volunteers for too much.
Students will start grabbing tasks Nov 22nd. 21:03
21:03 allison joined
mikehh I've got some stuff to do there and am working on it as well 21:03
dukeleto needs to write a blog post about GCI. Really needs to
whiteknight I can offer a little help here and there 21:04
dukeleto: can you outline what help you need, or what roles specifically you need other people to fill?
mikehh so we have less than 2 weeks to get things ready 21:05
dukeleto whiteknight: i need people to volunteer as mentors for tasks 21:06
dukeleto will not be mentoring any tasks. Admin'ing is work enough.
whiteknight I can mentor almost any task that doesn't involve translating 21:07
I barely speak english
mikehh I'll help as I can and will amke myself available for any #parrot related questions
make
dukeleto mikehh++ whiteknight++ 21:08
mikehh but my sentiments regarding translations agfree with whiteknight
agree 21:09
dukeleto we already have translation mentors. They just might not know it yet.
kid51 prints out GCI wiki page to look at tasks
mikehh we probably need to push for othewrs there, and clearly define what is to be translated 21:10
others, my typing had started to go to pieces 21:11
chromatic Is that sufficient committer-power to handle GCI?
Util lives next door to a T-shirt screen-printer, and should have time to mentor. 21:12
dukeleto Util++
chromatic Anything else to discuss today?
dukeleto we should focus on getting rakudo to talk to parrot.git before 2.10 21:13
i am doing some of the parrot side, moritz is doing the rakudo side
mikehh and taptinder needs to move to git
smoke seems to work ok 21:14
although most of the smokes seem to be still tied to svn at r49810 21:15
whiteknight I think there are some smoke bots that haven't gotten the git memo yet
mikehh I seem to remember one bot that was running parrot 0.9.0 for a long time 21:16
chromatic Should we gather a list of contact points for various smokers? 21:17
whiteknight yes
dukeleto Can we deprecate Parrot::Revision ? 21:18
does any external project use it? Does it need a dep cycle ?
whiteknight I say "no" 21:19
tcurtis On the subject of 2.10, yes, I am still up for being release manager.
whiteknight tcurtis++
dukeleto tcurtis++
cotto_work tcurtis++
dukeleto tcurtis: you are a brave man :)
chromatic Anything else for today? 21:20
dukeleto One thing. 21:21
People should try to use branches more often now, for anything "bigger than a breadbox"
cotto_work I'm sure that will come more naturally as the pain of svn branches fades from memory.
dukeleto cotto_work: indeed. 21:22
cotto_work I've blocked it out, but I'm sure other people can remember it.
mikehh cotto_work: :-}
cotto_work +1 to more branching
chromatic Is that in the committer guide?
whiteknight if not, it will be
chromatic Can someone mail the highlights of the changes to parrot-dev? 21:23
dukeleto highlights?
github.com/parrot/parrot/blob/mast...rkflow.pod
chromatic What's new that Git provides that we should consider doing differently.
dukeleto chromatic: that is a good question. 21:24
dukeleto will pontificate on it 21:25
is it a wrap?
whiteknight wrap it 21:26
chromatic Wrap it. Thanks, everyone.
21:26 dukeleto left 21:27 darbelo left 21:29 NotFound left 21:30 kid51 left, whiteknight left 21:35 atrodo left 21:36 bluescreen left 21:45 bacek left 22:03 tcurtis left 22:08 nwellnhof left 22:45 PerlJam left 22:50 whiteknight joined 23:07 bluescreen 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!