weekly meeting at 18:30 UTC, Tuesday | logged at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 30 September 2008.
02:01 japhb joined 03:06 japhb joined 14:00 masak joined 15:10 wknight8111 joined 15:25 davidfetter joined 15:42 pmichaud joined 17:26 masak joined 18:09 cotto joined 18:22 rdice joined 18:26 allison joined 18:27 barney joined 18:28 moritz joined
masak hi 18:31
moritz helo
18:31 Coke joined
allison ola 18:31
barney hi
Coke ohaiyo gozaimasu
cotto sup 18:32
18:32 NotFound joined
spinclad
.oO( allo )
18:33
moritz so who plays master of ceremonies today?
allison no chromatic? 18:34
Coke?
18:35 chromatic joined
chromatic allison, please start. 18:35
allison - The first Parrot Dev Summit was a great success. 18:36
- Worked with Jerry on testing/fixing the new I/O architecture on Windows. Now passes all tests.
- Finished porting/debugging buffering for reads and writes.
- Adding the final few methods to the FileHandle PMC, preparing to merge I/O branch right after the release today.
EOR
chromatic barney?
barney Applied some patches.
Fiddled with Pipp. 18:37
Got stuck because I misunderstood PCT.
.eor
chromatic I applied some patches and fixed some bugs.
Per discussion, I've decided that we can release this version with a check for Storable 2.12, not 2.18. If things break, we can get more creative. 18:38
I've fixed almost all bugs in fulltest, but found a big issue with classes and bytecode that I probably can't fix before the release (make testr fails one test).
More reports welcome.
Otherwise, I'll do the release tonight, so please keep filling in NEWS and other documentation, and make only low-impact commits.
Coke?
Coke . 18:39
ripped more code out of parrot
got most of ::tcl::mathop working in partcl. (~100 more passing tests)
partcl is slow. hellllp.
working on implementing [load foo.pbc] in partcl. Worst bit is providing a way to get at the top level parrot namespace from partcl. (which has a fake NS root at parrot's ['tcl'] NS.
.
chromatic cotto?
cotto * getting a grip on PCT by refactoring Pipp's PHP tag handling
* am fixing #59810 by removing seed arg from key_hash_* functions (will consult list before committing, of course) 18:40
* applied various and sundried patches to PIRify tests, make vim-install happy on windows and fix URLs in generated Makefiles
.eor
chromatic japhb?
masak?
masak * contributing to Rakudo by finding and reporting bugs 18:41
* crossing my fingers for the imminent lex unbreaking
that pmichaud++'s working on
* Rakudo is cool already, and getting cooler
* working on November, gaining speed in the last few days
* having a hackathon right now
EOR
chromatic moritz?
moritz * usual testing and patching
* Here's my offer to the rakudo hackers: if you fix a bug for which there's a ticket in RT, you can assign it to me if tests are needed. I can't promise anything, but I'll do my best. (jonathan++ for doing that first)
.end
chromatic NotFound?
NotFound Very few these last weeks. A few fixes and some updates from deprecated pir syntax 18:42
eor
chromatic particle?
particle ~ at parrot developer summit, we created a roadmap for parrot 1.0 and beyond, available at trac.parrot.org/parrot/wiki/ParrotRoadmap
~ notably, parrot 1.0 is due for release on 17MAR2009
~ this release aims to achieve a stable api for hll developers
~ started work on some roadmap tasks
~ fixed bugs, applied patches, created/updated tickets
.end 18:43
chromatic pmichaud?
pmichaud ** Rakudo spectests (r32819): 220 files, 4576 passing, 41 failing
== Miscellaneous
: attended Parrot Developer Summit
: updated NEWS, DEPRECATED
: removed get_array and get_hash from Capture/Capture_PIR (deprecated)
== Parrot stuff
: did a lot of work on lex2 branch; very close
: need to fix context refcounts in Parrot (working on that now)
== PCT stuff
: make sure 'callmethod' nodes quote method names
: added 'add_method' method to P6object metaclasses
== Rakudo stuff
: updated ROADMAP
: cleaned up Num.ACCEPTS
: added .PARROT method (this will likely move to P6object) 18:44
: fixed ObjectRef semantics for PGE::Match objects
: Parrot Strings now auto-promote to Rakudo Str on assignment
: Rakudo now maps Parrot BigInt as Rakudo Int
: made moritz++ 's offer about rt tickets into rakudo official policy
1 Question
.end
chromatic tewk_?
cotto 1 question
chromatic wknight8111?
wknight8111 *Calling conventions branch 18:45
**Renamed Parrot_PCCINVOKE to Parrot_pcc_invoke_method_from_c_args
**Got a few functions unified and refactored with help from allison++
**Been writing more documentation about it all
**Preparing/planning the next steps in this project (ready to merge it in later tonight)
EOR
chromatic Did I miss anyone?
Okay. Question time. pmichaud. 18:46
pmichaud as part of the lex2 branch we no longer need the Closure PMC. IMCC doesn't generate it, and if updated it doesn't do anything different from a normal Sub PMC. Shall I remove it (and mark it as deprecated in today's release)? 18:47
chromatic Remove on the branch and mark as deprecated on trunk (just to clarify). 18:48
particle can you check with rgrjr first?
i don't know of anyone else who might be using Closure
barney Can capturing of :outer vars be turned off for 'Sub' ?
Coke mark it deprecated. if we undeprecate it later, that's fine. 18:49
(A QWESTON. I HAZ WON.) 18:50
chromatic cotto has the next question.
cotto Post 1.0, what will the process of building and running a out-of-tree Parrot-based compiler look like?
Especially, will languages need to be downloaded into Parrot's languages dir, or is something else planned?
chromatic If make install doesn't work by 1.0, it's not 1.0.
cotto and once make install works, the languages can live wherever they want? 18:51
Coke at 1.0, no languages should be in core.
allison cotto: yes, they can live anywhere
Coke IMO.
allison Coke: that's the general idea
Coke "you don't have to go home but you can't stay here."
cotto coke++ 18:52
eoq
pmichaud at 1.0, only example languages will live in core.
wknight8111 ...if you call that livin'
18:52 Tene joined
pmichaud we may still have languages, but most "real" languages will be in their own repos. 18:52
Coke (see also partcl for an example of how to live outside of core. =-)
allison Coke: some may still be in the repository, but we won't include them in the tarball
Coke allison;reasonable.
allison Coke: will ship them in separate tarballs
Coke: but, we encourage moving out
especially encourage using an existing public infrastructure like google code or sourceforge, for greater visibility 18:53
Tene What? But if I move out, I'll have to get a job. When I live here I don't have to pay rent.
spinclad
.oO( rent? great idea! take a note, Marge! )
18:54
pmichaud barney: (capture of outer vars on #parrot) 18:55
chromatic coke, you had a question. 18:56
Coke "why is partcl so slow?" 18:57
chromatic You're doing it wrong?
Coke OOOOH, I have a followup question.
which is, "which bits am I doing wrong?", but I know that's a hard one to answer with the current toolset. 18:58
chromatic Sadly, yes. 18:59
Are there other questions?
cotto Are we there yet? 19:00
masak cotto: March 17. 19:01
Tene Was there discussion of how we're going to advertise our new release plan?
cotto sorry, not intended to be serious
masak cotto: :)
chromatic Tene, not much.
allison Tene: we'll make the information public, but not make a big deal out of it 19:02
moritz should I write something for perlmonks.org
Coke I suggest picking Andy Lester and Josh McAdams brains.
allison Tene: the big deal is after releasing 1.0
Tene Okay.
allison making a big deal about a *plan* to release 1.0 isn't going to have much impact when people have been waiting for it for years, just sounds like more promises 19:03
Tene nods.
I'm done, then.
particle i'm converting the roadmap to a table
so we can add columns like ticket number
chromatic Last chance for questions. 19:04
pmichaud any idea when we _should_ start using trac for tickets? 19:05
perhaps when we're able to move the parrotbug@perl.org address?
allison pmichaud: as soon as I get the parrot-tickets@lists.parrot.org set up
pmichaud excellent. 19:06
masak we're switching to trac? why?
pmichaud is there a ticket for that? ;-)
Coke yes, actually.
allison pmichaud: it's in my hiveminder queue too
pmichaud my hiveminder queue keeps getting bigger. That's probably not a good sign.
Coke I don't think we have to shut off RT yet; we can gently push people towards the new system.
(but let's not dawdle once the new one is turned on.) 19:07
allison masak: to move off perl.org infrastructure, and for better integration with SVN and the development wiki
pmichaud one of our 1.0 milestones was that tickets would be closed/migrated in RT.
also, thus far Trac seems much quicker than RT.
masak allison, pmichaud: ok.
pmichaud (also, note that rakudo tickets will remain in RT, since they're Perl)
allison pmichaud: that could just be better hardware, I don't know the general performance characteristics of the two 19:08
pmichaud allison: good point.
allison pmichaud: we are on screaming fast hardware at OSU OSL
masak pmichaud: good. I like RT.
Coke is out.
19:08 Coke left
pmichaud from the user perspective, though, trac is faster than perl.org's RT instance :-) 19:08
barney and Trac can be ported to pynie
allison pmichaud: true
pmichaud okay, I'm done. Back to working on lex2. 19:09
masak \\o/
chromatic Let's call it a week then. 19:11
moritz ciao
19:11 moritz left 19:13 cotto left, allison left 19:17 masak left 19:57 chromatic left 23:19 davidfetter joined 23:42 Whiteknight joined