weekly meeting at 18:30 UTC, Tuesday | logged at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 22 November 2008.
04:14 jimmy joined
jimmy My mind is that parrot can support documentation i18n officially. Ask cotto for more information. 04:16
04:17 jimmy left 07:36 jimmy joined, jimmy left 07:43 spinclad joined 10:02 particle joined 13:04 davidfetter joined 14:05 masak joined 14:45 jimmy joined
jimmy irclog? 14:45
moderator IRCLOG:irclog.perlgeek.de/parrotsketch/today 14:46
moderator Parrotsktech Every Tuesday. irclog at:irclog.perlgeek.de/parrotsketch/today 14:48
jimmy left, jimmy joined 14:49 jimmy left, jimmy joined
moderator Parrotsktech Every Tuesday | irc log:irclog.perlgeek.de/parrotsketch/today 14:50
moderator Parrotsktech Every Tuesday | irc log at irclog.perlgeek.de/parrotsketch/today
jimmy left, jimmy joined, jimmy left 14:51 jimmy joined 14:52 jimmy left 16:42 jhorwitz joined 17:59 Wknight8111 joined 18:06 Infinoid joined
Infinoid *lurk* 18:07
PerlJam you lurk loudly. :) 18:15
Wknight8111 he's just practicing 18:18
Infinoid has not yet attained his stealth merit badge
PerlJam this is the most talking I've done here ever I think :) 18:19
18:29 masak joined, allison joined 18:30 barney joined, cotto joined, kj joined, Coke joined
Coke ~~ 18:31
masak =$*IN 18:32
Tene here maybe in 10 minutes.
cotto hi
barney Hello
kj hello 18:33
jhorwitz hola
allison hi
Coke Alrightee. I can take the gavel, I suppose.
reverscibetical; Wknight8111?
I presume TimToady is just lurking... tewk? 18:34
Wknight8111 * Fumbled around more on the calling_conventions branch, didn't make any progress
* Putting it on the back-burner for now
* Did some more update work on [Book]
EOR
Coke Thanks, wk.
Wknight8111 no prob, coke
Coke I know working on the book is a thankless task, so i won't thank you. =-) 18:35
(too late). Tewk?
tewk subid loading is working.
18:35 chromatic joined
tewk see parrot backlog for issues and shortcuts. 18:35
helped kj get basicbc generation working in pirc. 18:36
EOR
kj tewk++
tewk subid lookup, still not checked in though.
Coke ok. just don't get hit by a bus.
tene?
spinclad? 18:38
(barring them, perljam?) 18:39
PerlJam just lurking as usual
Coke fair enough. partcl?
(particle, in case he's waiting for a beep) 18:40
Wknight8111 the second-half of the alphabet is always so depressing around here. So few people with names M-Z 18:41
tewk particle: was headed to the dentist I believe
Coke masak?
masak * reported bugs as usual
* helped track down test failures
* lots and lots of November hacking
* Rakudo feels must stabler now, after the lexbug
EOR
Coke Danke. kj?
kj @tasks = (<<'PIRC', <<'OTHER', <<'QUESTIONS) 18:42
+ basic bytecode emitting is working; working on a bcg api-like thingie.
PIRC
+ spent the weekend in RT, trying to close tickets that I understand
OTHER
oh there's a syntax errror there. In any case, eor
Coke mmhehe.
jhorwitz?
jhorwitz orked on mod_parrot roadmap and architecture docs. posted on parrot wiki. release 0.5 is one milestone away. 18:43
^w
for threaded MPMs, interpreters are now bound to individual connections so we reuse the same interpreter throughout a request
made a tweak to allow HLL modules with no Apache directives. this let me upgrade mod_pipp to the latest architecture. while i was in there i added support for POST data now that mod_parrot handles that for us.
working on porting november to mod_perl6 -- one last hurdle to overcome.
EOR
Coke spif. japhb?
masak jhorwitz++
Coke integral ? 18:44
Infinoid? 18:45
Infinoid Lurklurklurk.
Coke davidfetter ?
davidfetter ENOWORK :(
cotto same 18:46
Coke urk.
cotto q 1 q
Coke contingencyplan ?
18:47 pmichaud joined
Coke my report: 18:47
puts {half-heartedly poked at partcl, but getting parrot GC segfaults in trunk. [trac 10]};
puts {some RT cleanups. We're now at 12 active trac tickets; 587 active RTs. (up to 91 stalled RTs, though.)}
exit;
chromatic ?
pmichaud arrives, belatedly. 18:48
chromatic I profiled startup and moved some runtime creations to compile time; that sped things up slightly.
There's more to do there, specifically related to CONST_STRING.
I've gone through the Coverity reports and fixed or marked invalid all open tickets that they reported. They should move us to tier2 (harder bugs) soon.
I fixed a couple of memory leaks. One was in IMCC macros (horrible code) and the other in dynops. 18:49
I think there's still something wrong in the macro code, as kid51 has Darwin/PPC failures. I'll look at that.
I cleaned up a couple of failures in Rakudo I created during that process.
Otherwise, I've been going through the SKIPs/TODOs in our tests and documenting them.
I'll write in my journal about the process. It's really easy to get other people involved. 18:50
Mostly we just need to review them for applicability and make sure that every one has a ticket.
EOR
cotto q 1 more q
Coke barney?
barney [Pipp]
Added support for a return value for user defined functions.
Started on 'require_once'.
Applied some suggestions and patches from jimmy++
Bought book 'Extending and Embedding PHP'.
Will be busy in the next three weeks.
jhorwitz++
.eor
Coke pmichaud? #I'll let allison keep her a-list spot. 18:51
pmichaud == Parrot stuff
: built some test cases for :subid and tewk++'s updates
== PCT stuff
: allows Integer/Float/String to appear directly in PAST
== Rakudo stuff
: fixed outer handling
: added :g(lobal) to .subst method
: added multi-bracket quotes and texas quotes 18:52
: fixed prefix:<^> as in ^@list
: fixed infix:<Z>
: updated spectest reporting
EOR
queue 1 item
Coke danke; allison?
allison - Lots of small fixes on the I/O branch, resolving the final few failing tests. Have two more substantial bugs there.
- Ripped out the old I/O system from the branch, preparing for merge.
- Looked at the calling_conventions branch. We have one set of failures in the PCT tests (apparently context-related) before we can merge.
EOR
jhorwitz pmichaud++ # :g 18:53
Coke alright. Anyone who hasn't reported who wants to chime in before we move to questions?
Tene I haven't done anything. I'm thinking of rewriting pheme wit hPCT this week.
Thinking is cheap, though.
.eor
Coke puts {half-heartedly poked at partcl, but getting parrot GC segfaults in trunk. [trac 10]}; 18:54
puts {some RT cleanups. We're now at 12 active trac tickets; 587 active RTs. (up to 91 stalled RTs, though.)}
exit;
contingencyplan integral japhb particle samlh spinclad TimToady?
whoops.
Coke needs a "you already pasted that into this window. Are you sure you want to do it again? filter.)
chromatic That's a weird exit value.
Coke it's tcl, man. 18:55
Alright, popping the question stack: pmichaud?
pmichaud trac.parrot.org/parrot/wiki/ParrotRoadmap has the things we expect to have done for the Dec release
are we on track for each of those items, or do we need to see about adjusting?
the dec release is two weeks from today. 18:56
chromatic nan/inf bothers me somewhat. I'm not sure what that will take.
The exceptions milestone also seems risky.
pmichaud I'm particularly curious about opcode set/review/deprecation, should we have a special online item for that?
or a call to the mailing list for suggestions? 18:57
Coke perhaps a combo wiki page/mailing list thread for that.
allison chromatic: the nan/inf was just a matter of updating the PDD based on the list discussion
chromatic It's probably a short Perl program which parses all opcodes and then skims everything in languages/ and t/ to find frequency.
pmichaud I'll start a thread for it. 18:58
but I did want to see if others have issues. Note that pct, loop+exceptions is somewhat blocked on resumable exceptions.
allison I beleive the pdd24 was just a review of the pdd to make sure it matches the current state of reality 18:59
pmichaud (that may be working now -- but we need to check.)
Coke discovers his name is on some line items.
Tene pmichaud: I can work on that today/tonight
pmichaud in particular, I think at this point we really need to have a "lead person" listed for each item targeted for the Dec release.
otherwise we'll start falling behind. 19:00
allison pmichaud: agreed
Tene That's a great idea.
pmichaud I'll take "opcode set".
Coke if there's any gaps at days end, I'll send out an email looking for volunteers.
Tene I'll take whatever I'm assigned.
Wknight8111 if I can get the calling_conventions thing resolved, I'll add my name to a line or two
allison Tene: how about exceptions debugging, since you were already started on that? (feel free to pull in others to answer questions or pinch-hit on pieces of the code) 19:01
Coke excellent. We'll revisit this on and off over the next week to make sure we stay on target.
pmichaud yes, we just need a "chief worrier" for the task -- that person doesn't necessarily have to do the work.
who is doing the dec release? 19:02
Wknight8111 me
and let's hope it isn't a disaster!
chromatic Read through it now.
allison pdd14 and pdd24 fall to me if no one else volunteers, but I'm very aware of the fact that it would mean I have more on my plate than can probably be finished by one person
chromatic You can do most of it beforehand. 19:03
I'll do pdd14 (numbers).
Tene What needs to be done with PDD24?
Coke what does the pdd involve?
Wknight8111 trust me, I've read things and read them again. And practiced.
allison chromatic: excellent, it's already started
Tene: pdd24 is just a review "does this match reality?" 19:04
pmichaud anyway, that's it for my q; but q another for the end (time permitting)
Coke ok. let's followup details on the assignments in #parrot;
Tene I'll do pdd24.
allison Tene: for places that don't match, flag with a question and we can decide if reality should change or the PDD should change
Coke let's move onto one of cotto's questions.
allison Tene: excellent
cotto What plans do we have for allowing translation of Parrot's documentation?
(no other questions)
Coke cotto: allowing? "feel free?"
allison cotto: the more translations the merrier 19:05
chromatic Let's tie that in with marketing.
Coke chromatic: +1
chromatic "We're getting ready to release. We want Parrot docs in your language. Can you help us help you?"
allison cotto: but, I think the deeper question, is what infrastructure we'll set up for hosting translations
Tene Who's going to edit the roadmap to add these assignments?
cotto allison, yes
Coke tene: you can edit it and claim your item(s) 19:06
Tene checks to see if he has an account.
Coke a more fundamental question is what's the infrastructure for our canonical english documentation.
allison cotto: we don't have that set up yet, will set up when we start getting translation volunteers
Coke throw particle under that bus; I know he wanted to localize parrot itself. 19:07
allison cotto: I expect we'll take the infrastructure from an existing project, like Ubuntu/OpenOffice/Mozilla
Infinoid ah, JIT.
allison cotto: no need to reinvent the wheel on translations
cotto it'd be nice to have something in place for when the first person volunteers 19:08
allison Coke: yes, we need substantial cleanups to our English documentation
cotto: yes, it would, but then we'll have a better idea what we need once we start getting volunteers
Coke I'd throw our published authors under that bus if they weren't busy writing code.
chromatic and books
allison cotto: at the moment we'd just be shooting in the dark
cotto ok
Coke cotto: please open a trac ticket so we don't lose the task. 19:09
cotto will do
allison cotto: I don't expect we'll get 100 volunteers right at the start, more likely 1 or 2
cotto eoq
Coke ok, next. kj, you had a question?
kj Yes. regarding ticket 48737
.namespace <ident> /.endnamespace <ident>
the first attempt to implement scopes. 19:10
I had replied on that ticket, as the feature is deprecated.
My question then is:
is this feature totally useless, or can it be used as a cheap way of implementing scope ?
I'm fine either way
allison kj: totally useless
kj but just want to make sure we're not throwing out something
allison (unfortunately)
kj I was thinking about langauges that don't need closures 19:11
and therefor no lexpads
(such as C)
allison kj: we do need a lightweight scope feature, but that's not it
kj ok
that's clear then :-)
I'll contineu with ripping it out
.eoq
allison if you have thoughts on what we need, though, make a note and we'll add it to the feature request list 19:12
kj well it seems to me it's jsut straightforward now to generate PIR
with nested scopes, using these .namespace identifiers.
Coke I'm adding a Q to the stack.
kj but yeah, guess this is not it. Maybe it's a 1.5 or 2.0 feature anyway 19:13
not so important, compiler writers can implement it themselves as well.
Coke kj: all set?
kj yesh.
Coke pmichaud ?
pmichaud any word on when pdd30 might land? I know it's slated for the Jan release, but rakudo is rapidly nearing the point where it might be useful to move it out of Parrot trunk. 19:14
i.e., Parrot trunk changes are causing Rakudo breakages, so perhaps we need to de-sync them.
allison pmichaud: what particular pieces do you need from pdd30?
pmichaud We can do it before pdd30 is ready, but I see getting "make install" to work in January to be very important.
Wknight8111 agreed 19:15
pmichaud allison: mostly just assurance that what's in pdd30 is what we're going with (or fairly close to it)
allison pmichaud: the branch is unlikely to land as a whole in it's current state (for one thing it's unmergable in it's current state), so we'll have to pull out pieces and apply them to a new branch
(pmichaud: I tried your technique for recovering branches, but it was too far gone even for that) 19:16
pmichaud I'll report back next week what specifics on what I'm looking for in pdd30/make install
s/what/with/ # can't type anymore :-(
allison pmichaud: thanks
and, anyone who wants to volunteer for a little branch recovery work is welcome (yet another task blocking on me, which is one more too many) 19:17
pmichaud eoq 19:18
allison cotto: I thought we had translations on the long-term roadmap, but don't see it now, mark the ticket you create for the 2.5 milestone (portability) 19:20
cotto will do
Coke my question is more of a pleading regarding trac.parrot.org/parrot/ticket/10 : i'd like to get that resolved before the december release, so I can continue to target release-latest for partcl.
So if anyone can spare some eyeballs on that, I'd appreciate it. 19:21
.eoq
chromatic Whew, wasn't my checkin.
Coke, if you have a PMC which extends Sub, that's the place to look.
allison Coke: thanks for the instructions on how to build Partcl, that helps
Coke I have an object that extends Sub.
allison Coke: so it's a high-level class that subclasses the low-level Sub? 19:22
Coke yes.
but we can move that to #parrot.
allison ah.... there are known dragons there
yes
pmichaud I'd like to see that example. 19:23
I also have a todo for something with "vtable invoke" that I owe allison.
(it's in my hiveminder queue.)
Coke Anyone gets blocked on something, keep in touch. 19:24
If no one has anything else...
cotto infastructure
ignore
Coke ... it's a week.
~~
allison thanks, Coke!
masak Coke++ # emceeing
19:25 kj left 19:28 chromatic left 19:29 jhorwitz left
pmichaud Coke++ 19:30
19:30 pmichaud left 20:16 Wknight8111 left 20:37 allison left 21:09 Infinoid left 21:17 Coke left 22:00 cotto left 22:25 japhb joined 22:33 davidfetter 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!