Priorities for this week: irclog.perlgeek.de/parrotsketch/201...#i_3126985 | 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/parrotsketch/today
Set by moderator on 16 January 2011.
00:59 lucian_ left 02:46 whiteknight left 08:16 lucian joined 08:23 lucian left 09:52 contingencyplan left 10:06 mikehh joined 12:52 bluescreen joined 14:35 lucian joined 16:07 contingencyplan joined 16:19 atrodo joined 17:32 kid51 joined
kid51 kid51's report 17:32
* DONE
** PaFo business
** Several phone conversations with other devs to take pulse of project post-PDS 17:33
** Blogged on results of PDS, need for us to be more user-centric: www.parrot.org/content/feedback-jan...eds-served
** Wasn't at parrotsketch last week, but prior to that performed cage-cleaning which led to closing of TT #207, TT #1775, TT #1739, TT #1454
* DOING
** TT #1988 and TT #2000: code cleanup under lib/Parrot/Pmc2c/*. Will close tomorrow.
** TT #1954: make docs target (not yet ready)
** Looking at some other code cleanup under lib/Parrot/Pmc2c/ suggested by bacek 17:34
* WILL DO
** Talking with potential contributors to Python-on-Parrot; expect blog post later tonight
** Review some PaFo legal/financial docs
EOR
17:34 kid51 left 18:10 ilbot2 joined
moderator Priorities for this week: irclog.perlgeek.de/parrotsketch/201...#i_3126985 | 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/parrotsketch/today
19:29 pmichaud joined 19:44 NotFound joined 19:54 whiteknight joined 19:58 bluescreen left 20:00 bluescreen joined
whiteknight WHAT I DID: 20:00
* More work on the imcc cleanup/refactor work. Stuck on bugs. Looking for some volunteer debugging help.
* Worked on Parrot#, including many fixes/cleanups/refactors, an improved test suite, and wrappers for several PMC types
* Trying to fix/debug parrot-instrument. We're getting closer to making it build and work correctly, but it's pretty well bitrotten in some places
* Created an extending function to wrap char** arrays up as *StringArrays
WHAT I WILL DO:
* Several things to blog about
* Need to fix the PLA test suite so I can cut a release of that
* Keep working on the IMCC branch, hope to get it merged in before 3.2, which requires debugging and testing
EOR
NotFound 2011-feb-08: 20:02
What I did:
-parrot
* Fixed a problem with strings passed to nci not null terminated
* Rearranged initialization and freeze/thaw of Array types
* Created experimental function Parrot_ext_try
* Minor fixes
* Added and fixed some tests
-winxed:
* Several refactors, cleaning and fixes
* Better messages for features unsupported in stage 0
* Added predef __FUNCTION__
* return and yield now can have argument modifiers
* Added operators delete, %= and ^
* Improved operators += and -=
* Option --noan now works in all stages and with both drivers
* Allowed usage of null in argument lists.
* A lor more of eye candy in example fly
What I will do:
Play a bit with a C++ backend for winxed
EOR
20:07 kid51 joined
kid51 kid51 supplemental: 20:08
At the tail end of this blog post: www.parrot.org/content/feedback-jan...eds-served ...
I make suggestions for a timeline over the next two-and-a-half months re:
translating objectives discussed at recent PDS -- particularly "make HLLs faster" objectives -- into steps with certain deadlines. 20:09
It would be helpful if people read that and let us know whether the suggested timelines are realistic.
EOsupplementalR
mikehh What I did since my last report: 20:14
* building and testing parrot on amd64/i386, with gcc/g++
* some fixes
* branch testing and fixes
* testing Rakudo and Winxed on latest parrot
What I intend to do in the next week:
* testing and fixing
* prepare and release 3.1 on Feb 15
.eor
cotto_work *did: 20:19
- closed some tickets
- talked with kid51 about the state of the project
- submitted an OSCON talk
- worked with whiteknight++ to get gsoc-instrument building again
-- passes some tests, segfaults for others
- M0 roadmap progress
-- started figuring out coordination for the Lorito team - weekly concall sounds like it'll work
-- started a branch for the spec - m0-spec (not pushed yet)
- profiling runcore progress
-- none 20:20
*will do:
- put something into the m0-spec branch, M0 concall
- profiling runcore cleanup
- more rfc review, more ticket closing
- try to finish resurrecting gsoc-instrument
*blockers:
- none
*eor
dukeleto What I did: * Submitted 3 OSCON talks * closed some TT's * merged tt1331-osx-conf-fix branch * applied a patch from benabik++ * other stuff i probably forgot about 20:23
* What I will do: * Attempt to get lorito stuff out of my head and into my branch
* Blockers: * Life, The Universe and Everything
dukeleto hates hates hates his pasting functionality 20:24
EOR
Util # Done: 20:26
No Parrot work; Perl 6 only.
# Plan to do:
Misc Perl 6, but especially .dmg for Rakudo Star
# 7-day ticket report:
2 closed: done
1 closed: duplicate
16 closed: fixed
1 closed: invalid
10 closed: wontfix
7 new
.end
pmichaud did: closed tickets, nqp and rakudo discussions 20:27
plan: more nqp stuff, more articles 20:28
.end
20:31 benabik joined
cotto_work hello, all 20:31
mikehh hi there
NotFound Hola
dukeleto hola 20:32
kid51 hello
cotto_work How'd we do on last week's goals?
bacek aloha 20:33
cotto_work I know we got the ticket count down thanks to many people's efforts.
bacek My report: work on gen_gc. Reaching stable state quickly.
Util Hello
cotto_work The other goal was to get more eyeballs on whiteknight's imcc branch(es) 20:34
How did that go? 20:35
whiteknight plobsing put some eyes on it, but wasn't able to duplicate bugs on his machine
other than that, no progress. Still running into some weird memory corruption issues
disappears with -G, but I don't know if GC causes it or just exposes it
cotto_work ok. Any other thoughts on the past week? 20:36
mikehh master seems pretty stable at the moment
kid51 q1q 20:37
mikehh any branch merges before Saturday
then we can have a nice bugfest
cotto_work mikehh: thanks for taking the 3.1 release
kid51 I tested generational_gc today for the first time in a long time; got errors in 3 tests
much fewer than previously
mikehh I think I fixed the codetest errorsa there 20:38
cotto_work kid51: go ahead with your question if that wasn't it.
bacek kid51, 2 gc 1 ro? They are "expected"
kid51 bacek: Yes, I think that's what they were.
NotFound generational_gc build winxed well, pass its test and run the hardest programs I have
benabik I've tested gen_gc a bit. I get those and a couple checkdepend. Result segfaults building Rakudo. 20:39
cotto_work It's good to see that one hll is happy with it.
bacek Unfortunately rakudo explodes. Even more since 2 days ago...
kid51 As I mentioned in my pre-post, on my parrot.org blog I posted recommendations for a timeline for formulating our next set of Roadmap Goals.
It would be good to get people's reactions to that. 20:40
bottom of: www.parrot.org/content/feedback-jan...eds-served
20:40 nwellnhof joined
kid51 (That was my question) 20:40
mikehh looked good to me
q1q 20:41
whiteknight goes afk for a minute 20:42
kid51 Any other reactions to suggested dates?
mikehh could we push it forward?
cotto_work Of those goals, the only one that's not being worked on (afaik) is serialization. 20:43
If nothing else, we should form a goal around that. The other items may not need dedicated roadmap goals, but we should monitor how they're going to ensure that they're making progress. 20:44
The timeline doesn't look bad. 20:45
kid51 mikehh: There's nothing to stop us from beginning work on any goal before those dates. 20:47
If we complete work on any of those objectives before, say, April 15, they don't have to go on the roadmap.
cotto_work kid51: that sounds like a good deadline. 20:48
kid51 But if we expect that bulk of work will take place between 3.3 and 3.6, then that objective should get a roadmap goal and a team.
kid51 satisfied; who's next?
cotto_work No other questions were queued. Does anyone want to ask something? 20:49
kid51 mikehh had ?
cotto_work mikehh: apologies for missing that. Go ahead.
mikehh I tried to get html_cleanup_2 working (locally) but there are a lot of weird dependencies 20:50
We should be able to remove mose of lib/Parrot/Docs
I think kid51's eliminate_make_docs branch needs to be sorted first
most
kid51 mikehh: Yes, we need to fix that branch. +1 to getting rid of as much of lib/Parrot/Docs as possible. 20:51
cotto_work Removing unused code is always welcome. 20:52
mikehh lib/Parrot/DocsSection should go but it seems to be used in tests etc 20:53
lib/Parrot/Docs/Section
cotto_work There's a good chance the tests are no longer needed. Do you need help digging in?
kid51 I'll try to look at eliminate_make_docs this weekend 20:54
mikehh related to that I have a local branch to add more html docs, I will push it later and hopefully include it in the release
kid51 cotto_work: I suspect mikehh and I can handle these things between us 20:55
cotto_work kid51: ok. Any other questions?
kid51 has another ? 20:56
mikehh just remember tested branched can be pushed until Saturday
cotto_work kid51: go for it.
kid51 Let's assume that people are already working on the "make HLLs faster" objectives identified by pmichaud on parrot-dev last week. 20:57
It would be good to know who is working on which objective already?
i.e., match people to items 20:58
so that we can ask about progress in next week's #ps
:-)
cotto_work ok. I'm on profiling.
kid51 Who is working on GC?
bacek We need faster pcc... Currently it creates at least 2 gcable per call. 20:59
kid51, me :)
kid51 bacek: Naturally!
bacek, cotto: Anyone working with you on those objectives?
cotto_work profiling, no
bacek no
kid51 "Create .pbc files directly from a Parrot program. " Anyone on that? 21:00
bacek kid51, me
kid51 Anyone else working on any of the 4 items mentioned here? groups.google.com/group/parrot-dev/...096a993525
bacek It's on hold. Waiting for "newPCT". Will work with pmichaud on implementing it. 21:01
pmichaud bacek: iiuc, the packfile format isn't going to be stable enough for Rakudo to rely upon, though.
bacek pmichaud, it will be hidden in POST::Compiler.
kid51 So we have cotto working on profiling; bacek working on GC and pbcfiles (with the latter on hold)
bacek same as now
kid51 Is there anyone else who would like to partner with cotto or bacek on these objectives? 21:02
pmichaud okay, if POST::Compiler is committed to staying up-to-date and being an official Parrot API, I'm okay with it.
see message from lists.parrot.org/pipermail/parrot-d...5446.html, though.
bacek pmichaud, something like this.
whiteknight I'll partner with bacek as soon as my IMCC work is sorted out 21:03
cotto_work The profiling work would divide pretty naturally so that another person could help with it without us stepping on each other's toes.
whiteknight which could be sped along if people are willing to help debug it...
bacek pmichaud, Packfile PMCs will hide implementation details.
pmichaud bacek: that's not what whiteknight++'s message says 21:04
bacek pmichaud, "Once available, this API will remain stable, even if the underlying file format is not.". Packfile _PMCs_ is API :) 21:05
21:06 allison joined
NotFound That means packfile pmc will have assembler-alike functionality? 21:06
bacek NotFound, erm... Why?
NotFound Packfile and/or Opcode PMCs will allow to write stable code generators? 21:07
bacek NotFound, I hope so. I'm working on it. 21:08
whiteknight pmichaud: yes, that's what I mean. the packfile PMCs will be the API, in whole or in part
pmichaud whiteknight: okay. 21:09
bacek: okay
thanks :)
NotFound If the file format is stable but we should to rewrite things every time opcode numbers, calling conventions... change, will not be very stable.
bacek pmichaud, pirate was my proof of concept. It's doable :) 21:10
pmichaud "doable" and "supported" aren't always identical :)
I'm not concerned about whether it can be done (I know that it can); I'm more interested in will it be done, when, and with what level of support? ;-) 21:11
bacek NotFound, "opcode" is encapsulated in OpLib. Calling conventions are pretty stable now.
21:11 benabik left
whiteknight pmichaud: Parrot won't be doing anything to the current IMCC-based toolchain until we have a doable and supported alternative 21:11
bacek pmichaud, at least covered by deprecation policy.
whiteknight it's still early to be worrying about it
pmichaud whiteknight: well, it's early to be worrying about it, but not too early for nqp which is doing new codegen steps
whiteknight ah, okay. I see your point 21:12
pmichaud nqp is expecting to redo major portions of pct.... so in that sense, it's not too early
we can target imcc, or we can see about targeting whatever comes next if it's going to be available and stable-ish
NotFound whiteknight: that may be problem. Will be hard to think that a imcc replacement is fully supported until is used instead of imcc in most uses.
pmichaud I'd rather not develop things twice
bacek pmichaud, you can target newPOST nodes. POST::Compiler.pbc already functional.
whiteknight NotFound: not so. We're supporting the new embedding API and the old embedding API now, until the old one is deprecated and removed 21:13
We can support multiple alternatives while we migrate from one to the other
bacek pmichaud, one chunk of work left - annotations. But it's few days work.
NotFound whiteknight: there are levels of support. The thing used to build most tools is the one that gets faster notice if something breaks, which means full support. 21:14
pmichaud so, how stable is OpLib?
whiteknight that's fine. not worth arguing about it
OpLib has been pretty stable I think
pmichaud in the sense of "we don't expect its api to change much?" 21:15
whiteknight right. the existing API shouldn't change much
pmichaud wfm, then
+1
carry on :)
whiteknight it's sort of anemic though. Probably needs more features
so if the interface change, it will be through additions
cotto_work What works will continue to work. What's not there may at some point start being there. 21:16
bacek afk # dayjob, will backlog
mikehh Any thoughts on GSoC 2011 - I haven't seen any posts on or linked to the wiki page I set up 21:20
Coke ... what timezone is bacek in now that it's DAYJOB time? ;)
cotto_work syndney, australia 21:21
mikehh SYD: Wed, 08:21 EST
kid51 mikehh: I'll put something up there re timing of GSOC projects in relation to our release dates, etc. -- but that will only refer to process, not to content 21:22
whiteknight has to decommute. Later 21:23
cotto_work Does anyone have specific goals for the week?
mikehh Release
cotto_work ok 21:24
21:24 whiteknight left
mikehh I'd like a bugfets/documentation hackathon over the weekend 21:24
bugfest
cotto_work +1. Having <500 tickets for 3.1 would be really nice. 21:25
here's what I have
GOAL 1: close 23 tickets by next #ps (get total down to <=500) 21:26
GOAL 2: monitor progress of Rakudo's needs (speed, gc, profiling, newPOST, serialization)
GOAL 3: no merges after Saturday in preparation for 3.1
GOAL 4: test HLLs after Saturday, fix bugs as needed
dukeleto sounds reasonable
kid51 +1 21:27
mikehh +1
cotto_work Alright. 21:28
Let's call it a wrap.
21:28 kid51 left, dukeleto left 21:30 bluescreen left 21:38 plobsing joined 21:42 atrodo left 22:04 lucian_ joined 22:07 lucian left 22:40 lucian_ left 23:00 plobsing left 23:07 plobsing joined 23:17 whiteknight joined 23:19 plobsing left 23:26 NotFound left 23:30 nwellnhof left 23:40 lucian joined 23:41 plobsing joined 23:47 lucian 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!