weekly meeting at 18:30 UTC, Tuesday | logged at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 30 September 2008.
01:16 particle1 joined 01:55 japhb joined 02:31 davidfetter joined 05:57 TimToady joined 06:31 particle joined 06:57 davidfetter joined 08:26 contingencyplan joined 10:03 davidfetter_ joined 10:30 jimmy joined 15:00 jhorwitz joined 16:51 pmichaud joined 17:31 Coke joined
Coke I guess it's time. =-) 17:31
it's a US holiday today, and some folks still haven't recovered from the US DST switch.
How's it going? 17:32
pmichaud parrotsketch in 60
Coke aha! so it is. I've -over- adjusted. Also, I blame japhb!
(ah. that's my paris clock, not my utc clock.)
japhb Coke: yep, all my fault. :-) 17:42
Is the OpenGL examples conversion from PIR to Perl 6 interesting here?
PerlJam japhb: probably only in as much as it related to changing parrot in some way. 17:46
japhb PerlJam: that's what I thought. Then I will probably trigger some reports another day, when the various problems I plan to report get fixed. ;-) 17:47
nr as usual then.
PerlJam you had no problems with NCI? 17:51
japhb PerlJam: oh, several 17:52
I just haven't had time to put in bugs for them
And I also found interesting ways to crash and segfault, too.
PerlJam you could still report those problems in a general sense and see if someone has the inclination to look into it.
japhb Hmmm, not a bad thought. Probably worth collecting my thoughts, then. 17:54
18:17 allison joined 18:19 masak joined 18:22 barney joined 18:29 jonathan joined 18:31 chromatic joined
masak sal' 18:31
chromatic morning
allison yo 18:32
barney howdy
jonathan ahoj
japhb morning
particle !
chromatic Shall we begin? allison? 18:33
allison - Spent the week on I/O work.
- Reading from files is working fine, but printing wasn't working without buffering, so I'm adding that.
- More MMD debugging. 18:34
- Looking forward to the dev summit this weekend.
EOR
chromatic barney?
barney Applied some patches.
Got stuck with disappearing 'our' vars in NQP actions. Trying to provide a test case.
.eor
chromatic I fixed some bugs.
I'm trying to remove as much global state as possible from IMCC, in the hopes that that'll magically solve the re-entrant compile for :immediate subs problem.
So far it hasn't.
I'm happy to look at any other blockers, as always. 18:35
Also I helped Coke remove type IDs from PIR.
Coke?
Coke typeids are gone from pir in trunk ...
converted a lot of .return -> .tailcall ; (nearly done)
trying to cleanup as much DEPRECATED as possible. 18:36
.end
chromatic japhb?
japhb yep,
While converting the OpenGL PIR examples to P6, came across some bugs
jhorwitz arrives late 18:37
japhb Several are Rarkudo, but others are (or appear to be) Parrot
There's still the known problem with i386 NCI JIT that Tewk is working on, he appears to be making slow progress there.
Also, if you exit() from a callback, it's a crasher.
In my MAIN() routine, if I don't put a 'return;' after 'glutMainLoop()', it crashes instantly, even though glutMainLoop() never returns. 18:38
I report that here, because it's not clear if it's a Rakudo bug, or a Parrot calling conventions problem,
pmichaud (report it as a rakudobug for now, since that's where it manifests) 18:39
japhb because the crash error is that there was 1 arg expected, but 0 sent. Which sounds like a problem with an expected return value from the main program.
pmichaud could be a lexicals/outer issue 18:40
japhb I also had a segfault (no backtrace), but I have not been able to recreate that this morning (I forgot the sequence) ... still working to recreate that.
EOR for Parrot -- the Rakudo boys get more later.
chromatic jhorwitz?
jhorwitz using november as a real-world test case for ModPerl6::Registry. finding lots of missing pieces needed for CGI compatibility and implementing them. november needs to be tweaked slightly for mod_perl compatibility, but it all works surprisingly well.
it's becoming harder to port mod_perl6's implementation of interpolated namespaces to HEAD. deprecated the patch and added a fudged 'call_sub_with_namespace' sub in PIR to mod_perl6. this should tide me over til interpolated namespaces are implemented in rakudo. 18:41
posted suggestions for PDS agenda items on the wiki. please have a look, add more.
EOR
chromatic masak?
masak * Been reporting bugs, applying patches and generally enjoying myself.
* Learned a bit more PIR, implemented Perl 6's .fmt in Rakudo.
EOR
chromatic particle?
jonathan J! J! It comes before M!
particle ~ reviewed container branch pre-merge, fixed bugs and added docs there
~ unfudged a number of spectests that passed after the merge
~ created the %*VM var, and exposed parrot's config in %*VM<config>
~ pds2008 planning continues, just a few final details before it starts
.end
chromatic pmichaud? 18:42
pmichaud (jonathan: I think we're going in order of increasing importance... how's that? ;-)
** Rakudo spectests (r32548): 216 files, 4478 passing, 68 failing
== Parrot stuff
: worked on lexical autoclose semantics
== Rakudo stuff
: added .Str and get_string semantics (#60350)
: container/value semantics (with help from jonathan++, particle++, others)
: fixed bug in double-quoted strings returning non-Str 18:43
: closing brace in regex/token (#57874)
: allow pointy blocks for if/unless/while/until
: refactor constructs to use <xblock>
: refactor and correct Junction semantics (#60168)
: corrected assignment operators (e.g., += ) for protoobjects (#60452)
EOR
chromatic jonathan?
jonathan * Spent Thursday and today hacking on Rakudo
* Thursday was mostly taken up doing a big refactor of container stuff with pmichaud - went smoothly and seems quite an improvement
* Found and fixed a key bug in candidate sorting for the Perl 6 MMD algorithm - should be a big improvement
* Got the is also trait working for adding methods, which should be the last blocker to pre-compiled PIR modules
* Need to have a re-think of how we're doing the traits stuff, after discussion with TimToady yesterday...
* Got it so that Num type now accpets Int too
* Been fixing various other bits, applying the odd patch...been happy to apply some of late from (Chris Dolan)++
* Seems I've got a first cut of the MAIN sub working - the rest is for particle to worry about...about to ci that ;-)
* Spoke on Perl 6 at Twin City Perl Workshop at weekend - and with that, my speaking for the year is over
* Looking forward to being at Parrot Developer Summit at the weekend and seeing/hacking folks
EOR
chromatic tewk?
japhb pmichaud: when did you work on the += stuff? += and ||= were broken for me yesterday. 18:44
18:45 PacoLinux joined
chromatic Looks like no tewk; let's go to question time, which I see japhb has already started. 18:46
japhb Sorry.
chromatic Just for that, you have to wake pmichaud up. I suggest an airhorn. 18:47
Very Texas.
japhb chromatic: pinged him in #parrot
pmichaud sorry, was busy looking up the answer.
japhb ah 18:48
pmichaud r32496
yesterday (about 25 hours ago)
I suspect ||= isn't implemented.
at least, I didn't see it.
japhb hmmm, += should have been there by the time I was working on it, then. 18:49
Fair enough re: ||=
chromatic Other questions? 18:50
Coke I have one for allison.
allison ok 18:51
japhb pmichaud: When I try += in Rakudo, I get a crash with the error: Multiple Dispatch: No suitable candidate found for 'i_add', with signature 'PP'
current instr.: 'infix:+=' pc 10987 (src/gen_builtins.pir:6919)
18:51 rurban joined
pmichaud japhb: let's look at this in #parrot 18:52
japhb pmichaud: will do
chromatic Coke, go ahead.
Coke integer type ids: the ticket was basically to remove the user-visible bits (but didn't say so explicitly). Is the plan to also remove all the C-level appearances? only remove them from things that are in PARROT_API ?
allison for now (and I mean until parrot 3.0 or 4.0) we keep them in C 18:53
PARROT_API isn't a reliable boundary, because many things are flagged as PARROT_API that shouldn't be 18:54
Coke no, but it -should- be reliable. =-)
allison yes, that's one of our "path to 1.0" items
Coke ok. so, in that case, the current version of 'remove integer type ids' is closable.
allison cleaning up what is and isn't marked as part of the API
chromatic The question is, should someone embedding Parrot be able to say PMC *my_pmc = pmc_new(interp, enum_class_LOOK_A_TYPE_ID); ? 18:55
Coke and if we're that far out on the internals, we don't need another ticket right now.
allison yes, close it
Coke allison: VTABLE_type is currently marked as deprecated. Does that count as C?
since you can't directly invoke a vtable, I'd say so. So I think we can reject that ticket, and not worry about VTABLE_morph
allison chromatic: no, they shouldn't, but they don't have another interface yet (a pmc_new_str or something like that) 18:56
Coke: yes, we keep VTABLE_type for now
reject the ticket
Coke ok. I'll reject that ticket and cleanup DEPRECATED.
hopefully we can wrap up a few more before release, too.
particle allison: perhaps it's time to rename the current pmc_new to pmc_new_typeid
Coke I encourage language folks in repo to test test test. we've been deleting a lot of stuff.
allison particle: good idea 18:57
chromatic Next question? 18:58
rurban I'm late. Just wanted to say: started testing the upcoming cygwin-1.7 with gcc-4.3 18:59
particle rurban: did you package 0.8 for cygwin?
rurban not yet. no time at all.
particle ok, might as well wait for 0.8.1
allison: how's the debian packaging going? 19:00
pmichaud japhb (and others): the += bug you're seeing is RT #60036
allison I was just wondering that myself, smash has been doing it lately, but he's busy with school
japhb pmichaud: thx
allison so, 0.8.0 probably didn't happen, but I can do 0.8.1 19:01
pmichaud we could potentially try to work around it in rakudo.
particle ok, let's push the packaging discussion to pds
pmichaud particle: add to agenda, perhaps?
particle yes, please 19:02
i'm too busy with work now, somebody please do it?
pmichaud done.
19:04 cotto joined
Coke hey, zarquon. 19:04
chromatic cotto, any report? 19:05
cotto just misc. Pipp grammar and actions work
eor 19:06
chromatic Thanks.
Any other questions?
Let's call it a week. 19:08
See you this weekend.
rurban bye
barney tschüss 19:09
19:10 chromatic left, Coke left, PacoLinux left 19:11 jonathan left 19:13 rurban left 19:19 jhorwitz left 19:22 cotto left 19:52 Wknight8111 joined, Wknight8111 left 21:31 rdice joined 21:40 rdice joined 23:22 Whiteknight joined