Vision for 2.0: Production Users | Don't forget to post closed tickets in your report. | Note: This channel is only for our Tuesday status meetings; you probably want #parrot instead. | irclog: irc.pugscode.org/
Set by moderator on 2 February 2010.
00:10 eternaleye joined 00:57 cotto joined 01:37 cotto joined 05:12 rin1024 joined 13:18 plobsing joined
plobsing What I Did: 13:24
* worked on TT1015 branch (wrong approach, but learned) 13:25
* pushed for shallow clone semantics for clone vtable
* created a patch for PDD17
* saw zavolaj, got inspired
* started bikeshedding^W redesigning NCI signatures
What I Plan:
* put the proposal for NCI signatures to the list
* create proposals to improve pain points in NCI (UnManagedStruct, h2inc)
* appropriate deprecations so NCI changes can move ahead
* if I can get approval, deprecate deep cloning behaviour in any and all core PMCs
* define a small API for frame builders
EOR
16:01 eggyknap joined, eggyknap left 16:35 Coke joined
Coke Wiki: 16:36
- wikify branch_status.pl for trac.parrot.org/parrot/wiki/BranchDescriptions
- added faces.pl for trac.parrot.org/parrot/wiki/ParrotFaces (is your face not there? add your email and pic to gravatar.com)
- faces.pl will warn about duplicate emails in CREDITS (we have one)
Build:
- Remove or mark some -Wunused code.
- reduce 'corevm'/'coretest' back down.
Trac:
- closed 5 tickets.
- 657 Active Tickets left
- 2.2 has 22 active tickets (missed)
.
18:13 NotFound joined
NotFound 2010-mar-16 18:23
What I did:
- parrot
* Just some testing.
- Winxed
* Implemented some pir directives and other minor improvements in pirado.
What I will do:
No plan
EOR
japhb DONE: 18:31
* Metadata updates, c/o fperrad++ 18:32
* Add setup.nqp support to Plumage
* Answered lots of Plumage questions in #parrot and #perl6
BLOCKED:
* Convert Plumage build system to distutils; blocked on distutils feature needed from fperrad 18:33
* NEXT:
* More Plumage work
* Help with HLL interop as requested
EOR 18:34
18:47 whiteknight joined
whiteknight WHAT I DID 18:56
* Twiddled around in the PCC branch. Mostly following progress, didn't make any substantive changes myself
* Fixed installation on kakapo, and added some new test methods
* Updated the Parrot-Linear-Algebra test suite to use Kakapo and its unit test functionality. Test coverage is dramatically increased, and inherited test classes mean we can write once and test multiple types immediately.
* Fixed some bugs in PLA, added some functionality, cleaned up code, and trying to standardize on common behaviors.
* Some exploratory hacking and a few commits on the fix_hll_mmd and tt389_fix branches, though didn't get either workable before the release.
WHAT I WILL DO
* PLA and PDS are going to track the 2.2 release, the 2.3 release, and then only supported releases (as opposed to tracking trunk).
* PDS will be updated to use Kakapo for it's test suite.
* Will be adding an iterator type to PLA to iterate over elements in a matrix from PIR
* Continue to follow the pcc_hackathon_6Mar10, fix_hll_mmd, and tt389_fix branches. Hope to get them all working, tested, and merged soon.
WHAT I AM BLOCKING ON
* Nothing
I will not be at the #ps meeting today.
18:56 allison joined
allison Last week: 19:04
- Worked on little language (camle) using nqp-rx.
- Swapped my parrot-dev email so it goes straight to my inbox, seems to make it easier to be responsive.
- Caught up on mailing list discussions.
Blockers:
- Spent the weekend buried in coursework.
- Traveling this week.
Next week:
- Fix remaining failures in PCC branch, prepare for merge now that 2.2 is out.
EOR
(and may be a bit late for the meeting)
Coke win 2 19:17
spinclad lose 1 19:18
cotto_work #did: 19:45
* kicked out the 2.2.0 release with numerous typos
* removed cpu_ret
* some opsc hacking
#will do:
#closed TTs:
#eor
Tene did: lots of talking, no actual work at all. 19:51
KTHXBAI
20:01 allison joined 20:11 allison_ joined 20:20 chromatic joined 20:26 bacek joined
bacek Done: 20:27
- Update ops_pct branch to properly handle new Grammar by cotto++
- Some work on pcc_hackatron branch
Todo:
- Implement more runcores in ops_pct.
Util # Done: 20:30
* Had lunch with ash (John Harrison), discussing his NQ-NQP project, and GSOC prospects.
* Turned up a bit-rotted chunk of debugging code. TT to follow.
# Plan to do:
* Help with TT#1489.
.end
chromatic I worked on the pcc_hackathon branch and the TT #389 branch.
I did a few optimizations. 20:31
I plan to do a few more optimizations related to constant STRING handling (hopefully for 2.3) and TT #389.
next?
cotto_work helo 20:32
allison hi
Util Hello
bacek aloha
japhb o/
chromatic hello 20:33
Let's review last week.
How are we doing for HLL bugs?
NotFound hola
chromatic TT #389 should get fixed very soon. whiteknight and I are close. 20:34
Util Rakudo is not building for me on Win32. TT1489.
On feather, too
chromatic bacek and I looked into that. We need to figure out a different STRING compacting strategy. 20:36
whiteknight or "any" string compacting strategy
chromatic It should be reasonably easy to switch over to using sized pools. 20:37
That might even be faster.
Util Is this achievable before the next Rakudo release? 20:38
chromatic It's possible, I suppose.
Any volunteers to look at it?
Sounds like a no. 20:39
Other HLL bugs? 20:40
Bueller? 20:41
whiteknight Bueller?
no other HLL bugs that are pressing 20:42
chromatic How about the corevm target? How's that work going?
whiteknight none that I can think of off the top of my head
20:42 allison_ joined
whiteknight that target is fixed, I don't think there has been any testing 20:42
(not that we expected it)
chromatic The PCC Hackathon branch should be able to land almost any time.
bacek chromatic, op return_info is broken. TT#1511 20:43
chromatic ~almost~ any time
bacek :) 20:44
chromatic Let's move on to new priorities for the coming week.
We have a month before 2.3, which is a Big Important Release for Rakudo Star.
Rakudo needs 1) bugfixes and 2) performance. 20:45
Suggestions?
whiteknight dual focus on both would be nice
Coke do we care about compilation time?
whiteknight though nebulous
Coke (the build is SLOOOOOOOOW)
whiteknight slow build is fine. End users don't see that 20:46
cotto_work I'll need to go afk in a few minutes, but a roadmap review of the 2.2 items should happen at some point ( trac.parrot.org/parrot/query?group=...estone=2.2 )
particle sure would have been nice to have rakudo building with parrot 2.2 before it was released... 20:47
chromatic Does it not build, or do you mean the memory use problem?
Util I suspect that a great deal of the slowness is the same memory issue halting the build on other platforms.
particle it doesn't build for me with msvc on windows
i have 4GB ram
Coke particle: ISTR it was broken with 2.1 20:48
particle you recall correctly.
chromatic Suppose I tell someone exactly how to fix it. 20:49
Coke so, fixing the build for rakudo should be a priority, before speed.
chromatic Who will fix it?
Coke ... if you know what needs to be done, I can be a code monkey.
whiteknight I can look into it
particle i have some tuits this weekend
chromatic The easy fix is using fixed_size_pools for the actual STRING storage (not the headers, but their contents). 20:50
The harder fix is using multiple pools so they don't always grow.
Coke chromatic: is that the memory issue or the winders build issue?
chromatic Should be both. 20:51
I think that's the same issue.
particle rakudo builds against parrot releases. can we release 2.2.1 when fixed?
cotto_work If it's figured out tonight, I can do the 2.2.1 release. 20:52
or tomorrow night
chromatic Other priorities for 2.3? 20:53
I'd like to get the STRING constant mess figured out, so I can write up a task list for that this week.
Coke as long as we can deal with the rakudo-documentation issues of "which release are we using?" that we had after 2.1.1.
20:54 allison_ joined
chromatic Any hopes of the sweep-free GC for 2.3? 20:54
whiteknight if you and I have time to throw it together 20:55
bacek chromatic, unlikely...
Coke TT #1427 needs to go in DEP.pod if we're going to rip it out.
whiteknight I can focus in that direction if nothing else is more important
chromatic My guess is HLL bugs are more important.
If we can fix freeze/thaw/clone, Rakudo might get more benefit.
whiteknight freeze/thaw might be nearly done. clone is a bigger problem. plobsing has been posting mails to the list 20:56
particle make sure to ask jnthn about rakudo's needs, he's in change while pmichaud is away
chromatic Defaulting to shallow clones should help a lot.
whiteknight but we have to make sure that doesn't break anything 20:57
chromatic particle, can you talk to jnthn? Do you have time to liase? 20:58
particle not today, but yes.
i can do that tomorrow.
chromatic liaise, sorry... missing i
I'll set the topic in #parrot.
Question time!
Coke Note: 20:59
I opened a ticket with osu about the caching issues I was having with www.parrot.org and trac.parrot.org - everything seems fixed now; if you were having odd caching issues when editing pages, should be resolved.
whiteknight has to leave. Later 21:00
chromatic Other comments or questions? 21:01
allison I have started updating the online roadmap, would we like to review the 2.3 goals quickly? 21:03
and see that they match what we really intend for 2.3?
bacek chromatic, I can hack strings to use Fixed_Size_Pools in few hours. May be over weekend. 21:05
chromatic Yes, let's review. 21:06
bacek, whiteknight said he'd poke at them later today too.
allison 1) [DEPRECATED] RetContinuation PMC 21:07
is that in the plans for 2.3?
or move out?
chromatic Not a priority from my perspective. 21:08
allison me neither
move out
2) GC, currently listed as "generational copying/compacting, ordered destruction"
seems far too ambitious
chromatic Sweep free is a possibility.
Coke (retcon) if it's in the plans /ever/ we should add it to the DEP.pod. 21:09
chromatic We'd have to get it ready to land in about two weeks though.
Coke (otherwise we're stuck until 2.6)
allison Coke: I think it's already in DEP.pod, but could you check?
Coke it's not.
not by number, anyway.
allison Coke: okay, ticket is #1427 21:10
chromatic: is sweep-free worthy of a priority this month? 21:11
Coke ... yes, I know. Ok, I'll add it later today.
chromatic I'm not making sweep free a personal priority. I think we'll get more benefit from improved constant STRINGs. 21:12
I'm sure we can land that improvement, as well as more PCC optimizations.
allison will leave it in for 2.3, move the other GC tasks later 21:13
3) subroutine leave semantics/exit handlers
this is a priority for Rakudo, as I understand it
chromatic Tene was working on that.
Tene Yeah, I was. 21:14
Life's very hectic for me right now, though. :(
allison Tene: better to hand it off or move it out? 21:15
Tene Where does that compare with HLL interop, as a rakudo priority?
allison: hand it off.
chromatic As I understand it, it's higher than interop.
Tene nods.
allison we can check with patrick/jonathan to be sure
Tene allison: if I do end up doing any work on it, I'll make a branch and put it there, but don't count on it.
allison Tene: sounds good 21:16
any volunteers to work with Tene?
chromatic If the work's documented, I can poke at it some in my copious free time. 21:17
21:17 snl joined
allison we can talk on list and see what's left 21:17
might be a weekend hackathon subject
4) Core Libraries/Plumage 21:18
seems there's been good progress there
japhb I'm back working on Plumage, I'm stuck on something I need from fperrad, but:
allison is it ready for a Parrot release? 21:19
japhb I'm looking for any "I'd really like this by 2.3" features, so I can prioritize my work
I don't think it's ready to ship with Parrot right now, but I'd like to try for 2.3. About a 50% chance of success there.
So anybody with any requests, ping me in #parrot or send me an email. 21:20
allison sounds like a good goal
and not a high risk if we push it out
japhb exactly
allison that's all for 2.3
chromatic Besides the inevitable bug fixes and performance improvements. 21:21
allison indeed 21:22
do we want to have a pre-release hackathon
bug fixes and testing?
chromatic Definitely.
Other comments or questions? 21:26
Let's call this done then. 21:35
21:37 bacek left 21:40 NotFound left 22:09 eternaleye joined 22:12 Whiteknight joined 22:16 PacoLinux left 23:58 eternaleye 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!