Parrotsketch Every Tuesday @ 18:30 UTC | IRC Log at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 26 April 2009.
01:02 eternaleye joined 04:45 japhb joined 12:09 rdice joined 16:03 rin1024 joined 16:53 jhorwitz joined 17:07 NotFound joined 17:25 PacoLinux joined 17:30 jonathan joined 17:43 Util joined 17:53 pmichaud joined 18:06 fperrad joined 18:08 Infinoid joined 18:13 particle- joined 18:16 barney joined
pmichaud time for #ps ? 18:33
jonathan I make it so.
moritz good.
fperrad hello
NotFound hola
barney hi
Util hi 18:34
moritz let's go in alphabetical order of those who said hi 18:35
barney?
barney nothing to report
.eor 18:36
moritz fperrad?
fperrad * release 1.1.0
* packages @ parrotwin32.sourceforge.net/
* set up a VMware with Xubuntu 9.04
recipes available on github.com/fperrad/misc/
Note: 'git-gui' was deprecated in favor of 'git gui'
EOR
moritz jonathan?
ETIMEOUT. NotFound? 18:37
jonathan Grants 18:38
* Submitted final report for my now-completed Hague grant...
* ...and immediately afterwards submitted an application for another one.
* Got plenty of free time in May - hope to get lots of Rakudo progress in.
Rakudo
* NPW Hackathon was very productive, but also left me exhausted, plus I had lots of non-Perl 6 tasks so didn't get much done last week
* Rakudo day today; will do a second one later this week too in order to catch up
* Implemented constant scope declarator today - or at least, enough to get it basically working
* Fixed subtypes based on user-defined classes and roles
* Currently working on improving type-check-related error messages.
Parrot
* Need to write "set_addr op considered dangerous" style post to p6i at some point; we have at least one and probably many (quite possibly exploitable) segfaults out of it.
.end
(sorry for slowness, was JIT-writing it!)
moritz it seems we have a bad net split right now
NotFound? 18:39
NotFound * Just some fixing and cleaning
EOR
moritz pmichaud?
pmichaud * Rakudo now passing 10931 spectests
* Issued Rakudo release #16 ("Bratislava")
* Added tools/dev/create_language.pl, a rewrite of mk_language_shell.pl based on experience in building Rakudo outside the Parrot tree.
* NQP: Added ability to do method calls without parens (to support ".ast")
* PGE: At NPW, added <item> ** <sep> quantification with help from Martin Kjeldsen
* PGE: added underscore as a match for <alpha>, per Perl 6 specification
* PGE: allow whitespaces around terms in \\c[...], \\x[...], etc.
* PCT: Added a parse tree dumper format for Padre
* PCT: Cleaned up .loadinit handling for PAST::Block :compiler(...) nodes
* P6object: Cleaned up stringification of protoobjects to include parens (TT #168)
* Currently writing up status reports, roadmaps, and blog posts for Rakudo 18:40
.end
moritz Util?
Util Tried building the Parrot book into PDF, hoping to create a `make` target to automate the process. Ran into trouble on several different fronts, then ran out of Tuits. Will resume later this week.
EOR
moritz does anybody else here have a report?
pmichaud moritz? ;-) 18:41
moritz my report is: trying to grok Perl 6 regexes and action methods. mostly failing. EOR.
pmichaud I'd be very glad to see where people are having difficulty with regexes and action methods
since one of my big presentations for the summer (yapc/oscon) is likely to be about perl 6 regexes and grammars 18:42
18:42 darbelo joined, allison joined, chromatic joined, Coke joined, particle- joined, Infinoid joined, jhorwitz joined, rin1024 joined, rdice joined, japhb joined, contingencyplan joined, eggyknap joined, samlh joined, particle2 joined, cj joined, TimToady joined, cotto joined, Tene joined, mmpf joined, tewk_ joined, Tweety joined, spinclad joined
chromatic jonathan? 18:42
pmichaud uh oh, looks like two separate #ps's were taking place on independent netsplits :-)
moritz chromatic: jonathan reported on the other half of the netsplit - see the IRC logs
jhorwitz netsplit-- 18:43
jonathan moritz: I think tweety was on the other half of the netsplit.
Let me re-paste...
moritz jonathan: yes, but irclog.perlgeek.de/parrotsketch/today was on ours
Infinoid merges your IRC branches
pmichaud can I see the reports on the other side somewhere... ? 18:44
jonathan Grants
* Submitted final report for my now-completed Hague grant...
* ...and immediately afterwards submitted an application for another one.
* Got plenty of free time in May - hope to get lots of Rakudo progress in.
Rakudo
* NPW Hackathon was very productive, but also left me exhausted, plus I had lots of non-Perl 6 tasks so didn't get much done last week
* Rakudo day today; will do a second one later this week too in order to catch up
* Implemented constant scope declarator today - or at least, enough to get it basically working
* Fixed subtypes based on user-defined classes and roles
* Currently working on improving type-check-related error messages.
Parrot
* Need to write "set_addr op considered dangerous" style post to p6i at some point; we have at least one and probably many segfaults as a result.
.end
moritz: ah, I'm on both now. :-)
chromatic particle-? 18:45
particle- setting up parrot.org to accept donations
.end 18:46
chromatic Tene?
particle- oh, various yapc::na/parrot virtual machine workshop logistics, too
q1c
Tene PAST:
* Lots of work using Perl 6, not much Parrot work
* Tried to migrate rakudo partially to .HLL to help work with PMC HLL issues. I didn't even make it that far. Got lots of segfaults. Stalled.
* Made a simple example of .HLL PMC failure.
* Harassed jhorwitz about some mod_parrot fails.
* I keep thinking about packaging and distribution.
FUTURE:
* More work using Perl 6 18:47
* Harassing jhorwitz about mod_perl6 fails, which might be related to some other PMC issues... still investigating
* PCT Presentation at Penguicon the coming weekend... still need to write my talk. :)
KTHXBAI
chromatic tewk_?
Util?
Util Tried building the Parrot book into PDF, hoping to create a `make` target to automate the process. Ran into trouble on several different fronts, then ran out of Tuits. Will resume later this week. 18:48
EOR
chromatic Did I miss anyone?
rdice ping?
pmichaud pong
chromatic Go ahead, rdice.
rdice Thanks.
Re: TPF - processing jworthington's recently completed and happily-successful Hague grant, 18:49
and setting up the rfc process for his new one. (both are now on the news.perlfoundation.org blog)
I have also put an rfc re: TPF employing me for the next 6 months online at the blog too. I expect the TPF Board will be making a decision on this matter in the next few days.
Personally: I'm working on some Parrot NCI stuff that will make a certain big C++ library available to Rakudo. I'm just scratching the surface and it's all baffling and stuff, but I am optimistic. :-) 18:50
.end
jonathan rdice++ # dealing with my grant stuff quickly - thanks! 18:51
chromatic Anyone else?
pmichaud pong
chromatic go ahead
pmichaud * Rakudo now passing 10931 spectests
* Issued Rakudo release #16 ("Bratislava")
* Added tools/dev/create_language.pl, a rewrite of mk_language_shell.pl based on experience in building Rakudo outside the Parrot tree.
* NQP: Added ability to do method calls without parens (to support ".ast")
* PGE: At NPW, added <item> ** <sep> quantification with help from Martin Kjeldsen
* PGE: added underscore as a match for <alpha>, per Perl 6 specification
* PGE: allow whitespaces around terms in \\c[...], \\x[...], etc.
* PCT: Added a parse tree dumper format for Padre
* PCT: Cleaned up .loadinit handling for PAST::Block :compiler(...) nodes 18:52
* P6object: Cleaned up stringification of protoobjects to include parens (TT #168)
* Currently writing up status reports, roadmaps, and blog posts for Rakudo
.end (again :-)
chromatic Anyone else? 18:53
Question time.
Coke <me>
18:53 schmalbe joined
chromatic Coke? 18:54
Coke what should we do with tickets for old milestones on trac? leave them at the old milestone? auto-bump them to the next one? force a discussion to figure out where they really go? 18:55
(I vote NO on prop 2.)
allison The default is auto-bump them to the next milestone
chromatic They do no good stuck to old milestones.
allison actually, it's impossible to keep them attached to a completed milestone 18:56
Coke if we auto-bump them, they are about as likely to get finished for the next release.
Infinoid At least they stay in our faces.
Coke ... if your face is full of things, you can't see anything.
particle- (force a discussion)++
allison right, so I've been reviewing them as I mark the milestone as completed
some of them I bump, some of them I unmark for a milestone 18:57
I though I added a step to the release manager guide on closing the milestone
Coke so, "review", then. note: we have 35/50 un-solved tickets on the 1.1 milestone.
(many of which are deprecated items that need to be ripped out asap) 18:58
allison Coke: okay, anything critical should be kept for the next milestone
chromatic We're not very good at prioritizing our work in these weekly meetings. 18:59
allison that's what the weekly roadmap review is for
if "remove deprecated items" is a priority, it should be a roadmap item 19:00
Coke priority and roadmap are not necessarily the same thing.
chromatic Could we committers loosely agree to work on one high priority item every week? 19:02
Coke I cannot commit to that, no.
chromatic By "loosely agree" I mean "You can't tell volunteers what to do, but if we all did a little bit on this item, we'd make a lot of progress."
Depending on available time, energy, and resources, of course. 19:03
allison with your loose definition of "loosely agree" yes 19:04
particle- i can commit do devoting my parrot dev time to high priority items only.
chromatic Free electrons? NotFound? Tene? Whiteknight? 19:05
Tene Eh?
chromatic Infinoid?
allison how to make the priority for the week public?
Tene I have some time this week maybe.
chromatic I suppose we just pick one every week. This week, removing deprecated items?
Infinoid I can't agree to any guaranteed tuits at the moment, switching jobs here. 19:07
allison chromatic: makes sense
Coke: from that list of tickets, were there 3 deprecated items you would especially like removed soon?
Coke perhaps the MMd one from six months ago. 19:08
but no.
chromatic It's more like "If I have time and resources to work on Parrot this week, I will make ___ a priority."
Coke I have no rankings in particular.
allison chromatic: How about putting that on the top of the Parrot wiki? 19:09
chromatic Okay.
allison "If you have time and resources to work on Parrot this week, make removing deprecated items a priority." 19:10
or something
chromatic Other questions for discussion? 19:11
allison Who is closing out the milestone for 1.1? 19:12
particle- comment, for presenters/attendees of parrot vm workshop: trac.parrot.org/parrot/wiki/Parrot...hop%202009
wiki's are never correct, but please try to make this wiki more correct 19:13
NotFound chromatic: I'm not sure how much time I can spend on parrot next weeks, I'll try.
fperrad put the weekly priority on the irc topic 19:15
allison okay, I'll close the milestone and move all tickets to 1.2 (I'll post a mailing list message on any questionable ones)
jkeenan wanted me to mention that tickets marked as ownership 'rurban' are available for reassignment 19:16
Coke on a related note, if you own a ticket and aren't going to do anything with it, give it up.
allison good point 19:17
have we hit roadmap review?
Coke (roadmap review) in an effort to kill trac.parrot.org/parrot/wiki/ParrotRoadmap, I opened tickets for all the roadmap items for 1.2 19:18
allison I just saw that, many thanks 19:19
Coke there are 3 roadmap items on 1.1 that were not closed out.
(and not listed on ParrotRoadmap)
allison ah, we were only reviewing ParrotRoadmap 19:20
oh, the first two were rescheduled on the wiki page, but not changed on the tickets 19:21
(a good sign we need to get rid of the wiki page)
Coke #236, #237, #240
allison #236 was rescheduled to 1.4
changing now
I'm pretty sure #237 was landed a couple of releases ago 19:23
pmichaud #88 is close to landed -- just needs a final review. (That one's mine anyway.) 19:24
allison #240 was landed last month, but kept the ticket as an ongoing task 19:25
Coke ongoing tasks should probably not have tickets.
allison probably not, but the ticket tracked which files had been reviewed
we could move that to a wiki page
Coke if there are files left to review, then it's not done. if it's something where it has to be done all over every so often... 19:26
allison the task is "review todo/skip tests"
we'll always have some todo/skip tests
pmichaud I think the original purpose of that ticket was to try to eliminate todo/skip before 1.0. But that ended up being unrealistic. 19:27
allison it's a task of periodically reviewing them to see if they're still needed, remove as many as possible
chromatic It's a lot of busy work.
allison it's essentially the same thing as bug resolution
pmichaud I think the ticket can be eliminated. 19:28
particle- prepend [CAGE], remove milestone
allison technically, we should have a ticket to track each todo/skip test
Coke (meta tickets)--
allison one general ticket for all todo/skip tests isn't useful
particle- or, if removing ticket, add to cage cleaners wiki page
allison it's not a meta ticket, it's a ticket for the bug that meant we had to todo/skip the test 19:29
either a todo/skip is there because there's a bug it's avoiding, or the test should be removed
ah, it looks like the list of files in the ticket was created with a simple search of the t/ directory 19:31
Coke and even includes a pod file! whee. =-)
allison so, easily repeatable. I'll close the ticket.
Tene: you're listed as the owner of the PIR docs, did we finish that? 19:32
I remember marking it as landed in the wiki
Tene I have no idea.
pmichaud last time I looked I thought that pdd19 looked reasonably complete. 19:33
if there are specific items that need doing, perhaps we need tickets for those specific ones.
allison okay, I'll close the general ticket
That brings us to the 1.2 roadmap items 19:34
- packfile pmcs (jonathan)
should we change the owner to bacek? 19:35
trac.parrot.org/parrot/ticket/504
jonathan allison: bacek would seem to be working on those at the moment. 19:37
allison okay, changed
- Clone and Modify in DynPMC VTABLE Initialization (chromatic) trac.parrot.org/parrot/ticket/528 19:38
any update?
and, should this stay as a milestone for 1.2?
chromatic I'll try to get it for 1.2. 19:39
I think the infrastructure is related to the vtable swap milestone too.
allison okay
- remove optional features from default config, refactor config probes trac.parrot.org/parrot/ticket/565 19:40
any updates? can we assign an owner for this task?
should we remove the task from the 1.2 roadmap? 19:41
I suspect that's a question for jim keenan, so will take it to the mailing list 19:42
- export conventions (cross hll) trac.parrot.org/parrot/ticket/566
what needs to be done here, and do we have an owner? 19:43
Tene looks
allison The next two tasks seem to be connected to that one. 19:44
Tene wow, trac is slow
Tene still waiting on 566 to load
allison - pdd31-hll interop trac.parrot.org/parrot/ticket/567
- hll interop trac.parrot.org/parrot/ticket/568 19:45
Tene export conventions is storing appropriate items in ::EXPORT::ALL
iirc
allison I'll take the pdd31 task (launch it out of draft) 19:46
Tene I'm not sure which of those tickets it fits in, maybe this one, but we need to get library import/export into PCT
allison Tene: could you add details to the ticket?
Tene well, into a standard interface in compiler objects
pmichaud would know better which ticket that goes in 19:47
Also, we still haven' tbeen able to get rakud ointo .HLL 'Perl6'
allison the first one sounds like it is the ticket
Tene If we can't at least do that, that's a big problem
And I've been the only person trying to work on that. I'm still not sure what the problems are. 19:48
allison I got your message with the details there, I'll put that on my priority list for this week
Tene allison: that email was bogus, just ofund out.
please ignore / delete it.
allison ah, okay
do you have updated info on the problem?
pmichaud I'll see if I can get library semantics into PCT this week.
Tene trying to work on it now with jonathan
pmichaud that's nicely high priority. 19:49
Tene but, no
allison okay 19:50
Should I put patrick/tene as owner of the HLL export conventions ticket?
or leave it unowned?
pmichaud I'm fine with any of those. 19:51
Tene I'm okay with it
allison trac should let me assign multiple owners
Tene yay firefox crashed
pmichaud a ticket with multiple owners doesn't have any :-) 19:52
allison going with pmichaud, with the understanding that it's dually owned)
pmichaud: yeah, probably true
okay, last roadmap item
- vtable swap (e.g. bignum) trac.parrot.org/parrot/ticket/569 19:53
any updates? are we planning to do this for 1.2?
chromatic, should I assign you as owner? 19:54
Tene I really really want to get rakudo into a .HLL, so I can start really working on inter-hll library loading.
allison Tene: yes, I consider that a critical task for 1.2
Tene The library export stuff and maybe the PMC stuff are the only big issues, iirc
chromatic I'll try to get to vtable swap for 1.2, but I'm not sure if it or dynpmc vtable cloning will make it. One probably will. The other might.
allison yes, I imagine only one per release 19:55
is there a dependency ordering?
does one depend on the other? 19:56
chromatic No.
They may use similar APIs though and refactoring of the vtable functions.
allison okay, then which is simpler (since the first one you do will involve some API refactoring)? 19:57
chromatic Not sure.
allison which would you rather work on first?
chromatic Let me try dynpmc vtables first.
allison okay, I'll bump the other one to 1.3
that's it for the milestone review 19:58
I'm going to run through the rest of the roadmap wiki and convert it to tickets
it's too confusing to have two locations to look for the roadmap
Coke +1 20:00
pmichaud +1
allison sounds like that's a wrap. Thanks everybody! 20:02
20:02 Coke left, fperrad left
pmichaud (perhaps someone should paste the earlier reports so they're logged) 20:02
20:02 Util left, particle- left 20:03 chromatic left
Infinoid Can do. 20:04
11:35 < chromatic> allison?
11:35 < allison> Parrot:
11:35 < allison> - Gave Parrot talks at the SFO Linux user's group and LinuxFest Nortwest.
11:35 < allison> - Reviewed Partcl to collect a list of build problems, and fixed one. 20:05
11:35 < allison> - Made some changes to language shell generator based on experience giving the talk.
11:35 < allison> Pynie:
11:35 < allison> - Set up a google group mirror for pynie-dev list.
11:36 < allison> EOR
11:36 * Tene is here
11:36 < chromatic> barney?
11:37 * jhorwitz arrives late
11:37 < chromatic> I merged in some of the source file location cleanups. SVN merge is not my friend.
11:37 < chromatic> I added a couple of optimizations to hash key lookups and Key STRINGification. PGE should be ~5% faster, as is anything that uses hash STRING lookups.
11:38 < chromatic> I ponder the utility of a get_string_ro VTABLE entry, as creating COW STRING headers is surprisingly expensive amortized over all of our uses. (Either that, or rewrite Parrot such that escape
analysis is feasible.)
11:38 < chromatic> Coke?
11:38 < Coke> parrot - docs.parrot.org updated for 1.1
11:38 < Coke> partcl - minor updates (allison++), but still no build. attempted to recruit others (cotto++) for specific fixes.
11:38 < Coke> yapc - not going this year. :|
11:38 < Coke> .eor
11:39 < chromatic> cotto?
11:39 < cotto> Parrot: 20:06
11:39 < cotto> * worked with bacek++ on his PCT-based PMC parser
11:39 < cotto> * also went on a malformed PMC-fixing binge
11:39 < cotto> * took a stab at fixing partcl, will need more stabbing
11:39 < cotto> GSoC:
11:39 < cotto> - nobody packages decnumber as a shared library
11:39 < cotto> * darbelo started a Google Code project to host this, added me and whiteknight
11:39 < cotto> * I started a Google Group (groups.google.com/group/decnum-dynpmcs) for discussion (feel free to join)
11:39 < cotto> .end
11:39 < chromatic> fperrad?
11:40 < chromatic> Infinoid?
11:40 < Infinoid> * Made c2str stop using Math::BigInt. This sped up c2str by over 50%, or about 10 seconds, and sped up my overall build times by 8%.
11:40 < Infinoid> * I later discovered that the hash value it was computing wasn't used anyway, so I ended up ripping out even more code.
11:40 < Infinoid> * That ripped out code is in a separate commit (r38357), and can be easily reverted if someone wants to use it later.
11:41 < Infinoid> * Finally found a workaround for the libgmp "ohnoes it crashed" popup dialog I was getting during configure on mingw.
11:41 < Infinoid> * Worked with he++ for a couple of final fixes for netbsd-alpha.
11:41 < Infinoid> * Worked with LylePerl++ to find a HLL realclean bug on mingw, updated create_language.pl with the resulting fix. 20:07
11:41 < Infinoid> * Some patch mongering and bug triaging.
11:41 < Infinoid> 1;
11:41 < chromatic> jhorwitz?
11:41 < jhorwitz> getting back on track.
11:41 < jhorwitz> fixing random mod_parrot bugs
11:41 < jhorwitz> running into lots of issues precompiling mod_perl6 due to how mod_parrot registers its classes. had some discussion in trac w/ pmichaud that i need to follow up on. for now i've worked around the problem.
11:42 < jhorwitz> s/trac/rt
11:42 -!- Netsplit magnet.llarian.net <-> magnet.shadowcat.co.uk quits: jonathan, pmichaud, PacoLinux, NotFound, barney, ilbot2, moritz, integral, Util, ascent, (+3 more, use /NETSPLIT to show all of them)
11:42 < jhorwitz> starting to remove mod_perl6 from mod_parrot.
11:42 < jhorwitz> EOR
11:42 -!- Netsplit over, joins: barney, fperrad, pmichaud, Util, jonathan, PacoLinux, NotFound, ilbot2, moritz, ascent (+3 more)
(end)
20:08 Infinoid left, PacoLinux left 20:09 NotFound left 20:13 pmichaud left 20:16 darbelo left 20:55 jonathan left 21:44 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!