»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: , or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by moderator on 24 November 2009.
masak jnthn: yes, I think I knew that. now I knwo what it's called. 00:00
perigrin you can do that fairly trivially
just derive a new metaclass that does the confliciting metaclasses
00:00 KyleHa joined
perigrin if they can't compose as Roles ... then throw an error and make the programmer manually intervene 00:00
jnthn perigrin: "do that"?
perigrin well in perl6 a Class can be treated as a Role no? 00:01
jnthn perigrin: Metaclass incompat problem, or composition applier compat problem?
perigrin metaclass incompat
sorry
jnthn Yes. I'm just not sure which properly you're proposing a solution to. ;-)
perigrin composition aplier compat ... that's I suspect always gonna be a manual process 00:02
though it may be possible you can use a similar solution there too
masak jnthn: cool AOP example!
jnthn Well, yes. If the trait comp paper brings anything out, it's that being too clever is a Bad Thing.
My main concern is working out how to flag up incompatibilities nicely. 00:03
On the metaclass compat issue
Yes, that's an interesting idea.
perigrin the thing about Clever is that it's additive 00:05
you can always add more clever later once you're sure you have a fallback position that orks
jnthn Well, yeah.
perigrin er "works"
you have no idea how long it took me to figure that out :) 00:06
jnthn For now "just forbid it" seems like the easy fallback position.
(For the applier compat one)
I'm actually also watching Moose's protocol. I read the docs and it appeared it's quite cnjectural for now.
Or at least prone to change.
perigrin yes 00:07
jnthn Do you know if anyone has built alternative application implementations yet?
And how it's panned out?
perigrin Ovid has a variation of them in Cpan
to make them more strict
jnthn I'm sure it'll be the first thing he does with the Perl 6 ones. ;-)
Anyway, the fact that somebody managed to do that is encouraging. 00:08
masak jnthn: IM IN UR S12. tell me again why .accessor shouldn't return the method itself?
jnthn masak: Well maybe it could, that's just not what it's spec'd to do...
perigrin jnthn: the fact someone couldn't do it before Ovid tried was a bug 00:09
masak spec'd where?
jnthn masak: Thus why has-accessor is a better name for the thing we do have.
perigrin :)
jnthn masak: erm, er, spectests are spec too right? ;-)
masak jnthn: do the spectests explicitly say that .accessor is Bool?
jnthn svn.pugscode.org/pugs/t/spec/S12-in...tributes.t
Yes.
masak oh good.
I'll change those as well before I go to bed. 00:10
jnthn Cool.
Wolfman2000 frozen food...it can be useful.
jnthn Wolfman2000: I resorted to that tonight too.
Wolfman2000 I had Penne with White Chicken. Too little chicken
jnthn perigrin: Anyway, does metamodel.pod look something like a good start (obviously, with more filling out to come), or like an epic fail? 00:11
perigrin: If you didn't read it all yet, feel free to do so before answering. :-)
pugs_svn r29188 | masak++ | changed .accessor to .has-accessor on attribute descriptors
perigrin I skimmed it ... I think it looks like a good start
jnthn OK, that's good to hear. 00:12
Wolfman2000 jnthn: I just took a quick look...honestly, I found it a bit hard to understand.
perigrin the AOP example *is* quite nice
Wolfman2000 Then again, I'm somewhat used to this.
jnthn Wolfman2000: I wrote it and I found it hard to understand. :-)
jnthn grabs a beer and blogs 00:14
masak jnthn: hah! -you- wrote that test file! :)
pugs_svn r29189 | masak++ | .accessor -> .has-accessor 00:15
jnthn masak: lol 00:16
masak Wolfman2000: it is hard to understand. I think I grok parts of it because I spent the weekend with jnthn. :)
I'm open to the possibility that .accessor was actually meant to return a method object, or similar. if so, I deserve to have my commits reverted, provided this is also explained in the spec. 00:17
jnthn masak: *nod* 00:18
masak: I'd rather go with has-accessor for now, and leave .accessor free to add that later. 00:19
masak createyourproglang.com/
jnthn: yes, but .accessor would have worked fine as both.
jnthn masak: Yes and no.
masak: We still need to supply has-constructor in the constructor call.
masak nothingmuch tweeted the above URL and wrote '- this is a joke, right?'
has-attribute, surely? 00:20
00:20 rgrau joined
perigrin in my world has-* is a predicate while * itself is an accessor for the thing 00:20
so has-accessor is a predicate and accessor() would be the accessor acceesor
jnthn masak: has-accessor even
masak jnthn: yes, but parameters can be rewired to other attributes.
jnthn :-)
masak: Well, yeah, the thing is that either way, we need to keep hold of the fact that we need to make an accessor around before we actually make one. 00:21
We can just keep it as a private attribute perhaps...
masak nod.
jnthn But I suspect for meta-programming, it may be useful to make it available.
masak jnthn: oh, I see now what you're saying!
we need to keep it around. yeah.
so .has-accessor it is. 00:22
maybe you should even make the parameter be typed Bool.
unless that is known to make life difficult for other HLLs.
jnthn masak: Slight bootstrapping issue. :-) 00:23
masak we can't use types in the MOP? :)
dang. :)
perigrin eh wha?
Wolfman2000 masak: ...why is that something to smile about?
masak Wolfman2000: recursion is fun! 00:24
jnthn masak: We probably need the Attribute meta-object available before we define Bool. ;-)
perigrin masak: where are you seeing that?
jnthn Well, maybe we don't.
:-)
perigrin you don't.
carlin
.oO( To understand recursion you must first understand recursion )
masak and with that, I bid ye a good night.
Wolfman2000 carlin: how profound
jnthn masak: night! 00:25
masak out
jnthn omg mst-ism! :-) 00:26
perigrin jnthn: a Bool wouldn't need to have any attributes
would it?
jnthn perigrin: Yeah, true.
I think.
:-)
Well
Hmm...I guess it needs some storage.
Am I true or am I false...
perigrin right 00:27
hmm
though
00:27 payload joined
perigrin Strange loop :) 00:27
jnthn We could always use the lower-level native types anyway I guess...apart from we don't implement those yet. :-/
Yes. :-)
perigrin Bool.HOW.attributes[0].accessor will need to be lazy :)
cause it returns a Bool :) 00:28
jnthn I can see this is going to be a bit of a headache when I get there, yes.
perigrin the Bootstrap layer is always painful
thats why you want to try to make it as small as possible 00:29
if you look at Class::MOP you'll see lots of places where things get re-defined and are initially defined as hard coded values
jnthn Yeah, I suspect it'll have to be something like that at first. 00:30
dalek vember: 1c9a598 | (Jason Felds)++ | lib/November.pm:
Quickie one-liner from login useful in register.
 00:34
00:39 sjn joined
zaslon loljnthnhazblogged! jnthn++ 'Laying the role meta-foundations': use.perl.org/~JonathanWorthington/j...3?from=rss 01:03
01:04 frew_ joined 01:11 nihiliad joined 01:14 stephenl1 joined 01:25 agentzh joined
pugs_svn r29190 | lwall++ | [CORE] more nametweaking 01:27
r29190 | [STD] better messages on malformed/ambiguous quantifier ranges
r29190 | better message explaining all the things "undef" might really mean
01:29 alanhaggai joined 01:30 ashizawa joined
jnthn -> sleep 01:32
02:18 gfx joined 02:20 orafu joined
pugs_svn r29191 | lwall++ | [STD] catch old use of Object 02:22
r29191 | recommend use of new *.notdef method where appropriate
r29192 | lwall++ | [S32/Basics] add a .notdef convenience method for ease of writing *.notdef matchers 02:23
02:29 JimmyZ joined
JimmyZ Mu-- && ( Obj++ or Object++) 02:32
02:36 Exodist joined 02:37 Exodist joined 02:38 Exodist_ joined 02:40 Exodist joined
JimmyZ ng: .say for 1..4 02:48
p6eval ng bf746a: Missing block at line 1, near ""␤current instr.: 'perl6;HLL;Grammar;panic' pc 519 (src/stage0/HLL-s0.pir:336)␤
Tene see?
colomon JimmyZ: I'm not sure what's causing the grammar panic there, but 1..4 definitely doesn't work yet in ng.
JimmyZ I thought it worked. 02:49
ng: gather
p6eval ng bf746a: Could not find non-existent sub &gather␤current instr.: '_block14' pc 29 (EVAL_1:0)␤
JimmyZ rakudo: gather
p6eval rakudo 7347ec: too few positional arguments: 0 passed, 1 (or more) expected␤in Main (file <unknown>, line <unknown>)␤
colomon Well, it constructs a Range all right, but the Range can't do anything useful yet.
Tene ng: my @a = gather for 1, 2, 3 { take $_ ** 2 }; say @a; 02:50
p6eval ng bf746a: 149␤
JimmyZ ng: gather
p6eval ng bf746a: Could not find non-existent sub &gather␤current instr.: '_block14' pc 29 (EVAL_1:0)␤
JimmyZ ng: .say for 1,4
p6eval ng bf746a: Missing block at line 1, near ""␤current instr.: 'perl6;HLL;Grammar;panic' pc 519 (src/stage0/HLL-s0.pir:336)␤
03:02 synth joined
Wolfman2000 Reminder needed. Do classes only have methods, or can they also have subs in them? 03:13
03:17 hatseflats left 03:31 ihrd left 05:45 gfx joined 06:28 TimToady joined, quietfanatic joined, JimmyZ joined, ssm joined, simcop2387 joined, cognominal_ joined, xinming_ joined, Khisanth joined, QtPlaty[1ireMe] joined, envi^office joined, akl joined, shachaf joined, sbp joined, yves joined, sunnavy joined, nsh joined, literal joined, renormalist joined, dukelet0 joined, Bucciarati joined, silug joined, mj41 joined, jsut|work joined, broquaint joined, astinus joined, itz joined, araujo joined, IRSeekBot joined, rhr joined, Grrrr joined, lisppaste3 joined, ascent joined, Infinoid joined, arnsholt joined, Maddingue joined, mathw joined, yath joined, Juerd joined, athomason joined, gbacon joined, r0bby joined, moritz_ joined, Tene joined, rjh joined, sri_ joined, allbery_b joined, krakan joined, frew joined, lestrrat joined, buu joined, leedo joined, baest joined, constant joined, slbkbs joined, solarion joined, LylePerl joined, wolverian joined, elmex joined, kst` joined, cls_bsd joined, cotto joined, jeremiah joined, cotto_work joined, hugme joined, pnu joined, carlin joined, frettled joined, idemal joined, zostay joined, gfldex joined, stepnem joined, LionMadeOfLions joined, hicx174 joined, meteorjay joined, Helios joined, kolibrie joined, PacoLinux joined, perigrin joined, mdxi joined, tarbo2 joined, s1n joined, zaphar_ps joined, pointme joined, zaslon joined, xomas_ joined, dj_goku joined, xenoterracide joined, szabgab joined, edeca joined, Lorn joined, charsbar_ joined, aindilis joined, avuserow joined, dalek joined, Patterner joined, nothingmuch joined, estrabd_ joined, yahooooo joined, Woodi joined, drbean_ joined, jaldhar joined, Wolfman2000 joined, d^_^b joined, pure1111 joined, astrojp joined, ruoso joined, sjn joined, frew_ joined, orafu joined, Exodist joined, synth joined, pmichaud joined, alanhaggai joined, agentzh joined, KyleHa joined, tylerni7 joined, takadonet joined, RonOreck joined, he_ joined, pjcj joined, huf joined, krunen joined, estrai joined, Jedai joined, jnthn joined, Guest75832 joined, mhsparks joined, omega_ joined, gabiruh joined, viklund joined, felipe joined, Gothmog_ joined, Woody2143 joined, KatrinaTheLamia joined, BinGOs joined, spinclad_ joined, diakopter joined, jrockway joined, jjore joined, p6eval joined, eiro joined, mtve joined, tomaw_ joined
diakopter WELCOME BACK AFTER YOUR UNFORTUNATE NETSPLIT 06:28
o hai
06:28 spinclad joined, dukeleto joined, PZt joined, jantore joined, pugs_svn joined, hatseflats joined, jiing joined, phenny joined
TimToady voice lambdabot 06:28
diakopter the server added them back
JimmyZ TimToady: I also don't like the english computer language with some chinese words
TimToady hard to please, ain't ya?
diakopter Musers... 06:29
JimmyZ welcome
diakopter TimToady: about the caching idea 06:33
06:33 JimmyZ_ joined
TimToady while the language is still mutating, I'd rather not worry about cache coherency 06:34
also, I'd like to concentrate on getting first-run compilation fast
but long haul, it might be useful 06:35
diakopter that's why it's keyed on a hash of the grammar, too
quietfanatic MUtating.
TimToady I knew cats would come into it somewhere...
unless, of course, you pronounce it mootating 06:36
quietfanatic I just regret it's causing so much friction.
06:37 christine joined
diakopter rakudo: musubi # I'm hungry 06:37
p6eval rakudo 7347ec: Could not find non-existent sub musubi␤in Main (file src/gen_setting.pm, line 324)␤
06:45 kaare joined
pugs_svn r29194 | lwall++ | [t/] unify compartmentalized undef and Object concepts into Mu 06:53
r29194 | test suite now contains no mention of undef or Object tokens
r29194 | (expect temporary rakudo regressions till Mu and .notdef are added)
06:58 Su-Shee joined
Su-Shee good morning. 06:58
TimToady howdy
06:59 alester joined
sjohnson hi 07:10
07:11 mjk joined
dalek p-rx: ef7eef0 | tene++ | (3 files):
Basic implementation of CATCH and a test for try and catch.
 07:14
07:42 flip123 joined
flip123 Hello fellows! 07:42
Ready for some more questions?
Tene Hi!
Yes!
flip123 Wow, such enthusiasm that early in the morning ...
With feeds, is there an easy and nice way to feed into "return"? 07:43
As in " ==> my @a; return @a" without the temporary variable?
Tene No idea.
flip123 I tried to write "return map {} ...." and so on, but this was the wrong flow direction and didn't look that good. 07:44
TimToady nobody implements even basic feeds yet, I don't think
flip123 Oh, the basics work ... or at least they seem to. 07:45
That gets me to the next question ...
TimToady but eventually you ought to be able to say ==> return @(*) or some such
flip123 OK, I'll try. Thank you.
TimToady or run it the other way, return <== source
flip123 Well, that didn't look so good in my case. 07:48
Next problem: more often than not I'm getting "Null PMC access" while my P6 scripts get compiled.
I suppose I could run perl6.pbc with parrot tracing on ... but what should I do with the multi-MB of logfile?
I'm not sure that I could identify relevant spots there.
TimToady you'd have to take that up with the rakudofolk
who are mostly asleep right now
flip123 well, I'll try again later .... that's what my "early in the morning" was about.
Tene flip123: The general answer is "try to reduce your script to a minimal case"
flip123 Thank you nonetheless.
Tene: I'm trying to re-write a C program with about 30K lines ...
Tene Which isn't a very helpful answer.
:)
flip123 it's not that easy to identify the bad spot
that's why I'm asking here 07:49
Tene No, it really isn't. Rakudo doesn't have good tools to help with that.
flip123 Mostly I get "file unknown, line unknown" - which doesn't help.
07:49 iblechbot joined
flip123 Well, I'll be back later ... 07:50
07:50 flip123 left 07:54 beggars joined 07:55 kst` joined 08:06 meppl joined 08:09 japhb joined 08:10 ejs joined 08:20 ejs joined 08:24 mberends joined 08:30 IllvilJa joined 08:33 rfordinal joined
mathw Morning 08:37
Tene Hi!
08:44 rgrau joined, jferrero joined 08:45 snearch joined 08:54 rfordinal left 08:55 ashizawa joined 09:03 cognominal joined 09:12 gfx joined 09:15 hp_truth joined 09:17 hanekomu joined 09:18 rgrau joined 09:20 colomon_ joined 09:45 JimmyZ joined
JimmyZ rakudo: (1 ? 'foo' : 'bar').say; 09:46
p6eval rakudo 7347ec: Confused at line 2, near "? 'foo' : "␤in Main (file <unknown>, line <unknown>)␤
mathw rakudo: (1 ?? 'foo' !! 'bar').say; 09:49
p6eval rakudo 7347ec: foo␤
09:50 hanekomu joined
mberends mathw: any news about the $interviews? 09:51
mathw mberends: yes. They've invited me to their office for the final round. 09:53
Which is exciting, and scary. 09:54
mberends great, I'm sure you'd be good for them, the question is just whether they know it too. 09:55
mathw It's all very well talking to them on the phone, but actually going to their office on their expense and meeting them in person to talk about jobs is rather more real.
mberends they're serious if they're making expenses. is it for from where you live?
*far 09:56
mathw mberends: About a hundred and forty miles 10:00
mberends too much for a daily commute then
10:14 xomas joined 10:15 iblechbot joined
mathw mberends: Definitely. This would involve moving. 10:17
One-way travel time is about three hours
However, moving offers certain opportunities, so it's not a bad thing... and it's an amazing job. 10:18
10:27 justatheory joined 10:35 justatheory_ joined 10:44 camenix joined 10:56 edeca left
jnthn oh hai 10:59
11:06 zamolxes joined
mathw oh hai jnthn 11:11
11:15 payload1 joined 11:22 tylerni7 joined 11:24 wollmers joined
dalek vember: c22af29 | masak++ | lib/November.pm:
[November] the strings should be concatenated
 11:27
mathw It is quite quiet in here today
jnthn mathw: aye, it is
mathw: Noisy masak ain't here yet. ;)
mathw lol
mberends was talking but drifted away
jnthn mathw: BTW, good luck for the final bit of the job interview. 11:28
11:29 envi^home joined
mathw jnthn: Thanks. It's actually, I believe, four interviews in one. Should be interesting. 11:40
11:44 payload joined 11:46 alanhaggai joined
jnthn
.oO( junction of interviews )
 11:48
mathw Well, I think you'd have to use S& to combine them 11:50
Because they're not going to happen concurrently
It's fortunate that Larry added that metaoperator, so I have the right terminology to talk about my interviews in Perl 6 11:52
Without having to resort to explicit looping 11:53
mathw -> lunch 11:54
12:05 colomon joined 12:14 flip123 joined
flip123 Hello everybody! 12:14
It's me again, with some fresh questions ... 12:15
#1: what's the best way to diagnose "Null PMC access" while compiling my P6 scripts?
Su-Shee answer them while they're hot! ;)
mberends drifts on 12:16
jnthn flip123: While compiling? That's...ocuh. 12:17
flip123: Do you get a backtrace at all?
flip123 Well, maybe running, but --trace=65535 doesn't help
This here:
Null PMC access in get_pmc_keyed()
in Main (file <unknown>, line <unknown>)
doesn't help that much, does it? 12:18
jnthn Nope.
I normally go and comment out the Rakudo error handler in situations like this.
flip123 This is with "PERL6LIB=lib perl6 --output=/tmp/a.pir --target=pir <input file>"
jnthn Find the line push_eh trap and pop_eh shortly after it in perl6.pir, comment them out and re-build your Rakudo. 12:19
That'll give you the "internal" stack trace instead.
flip123 in the .sub eval method?
jnthn Yeah
flip123 push_eh trap_errors
jnthn right, that one.
and below it a pop_eh
flip123 got them 12:20
but this perl6.pir will be overwritten as soon as the next "git pull" gives me something, right?
Oh well
now compiling
flip123 is happy about 2GB ram for perl6 compilation
jnthn flip123: Well, it'll probably complain next time you git pull about the file being modified. 12:21
flip123 That gives me this: 12:24
Null PMC access in get_pmc_keyed()
current instr.: 'parrot;Parrot;Compiler;load_library' pc 29 (/home/flip/perl/p6_parrot/rakudo/parrot/runtime/parrot/languages/parrot/parrot.pir:16)
called from Sub 'use' pc 17983 (src/builtins/eval.pir:146)
called from Sub 'perl6;Perl6;Grammar;Actions;_block867' pc 330281 (src/gen_setting.pir:68143) (src/gen_setting.pm:429)
called from Sub 'perl6;Perl6;Grammar;Actions;_block704' pc 329125 (src/gen_setting.pir:67763) (src/gen_setting.pm:375)
and a lot more
doesn't help *me* in any way, TBH 12:25
jnthn Means it's a "use" statement that's leading to the crash. 12:26
flip123 Yes, that's what it says. But not any file or line number. And now? Can I help someone diagnose that further? 12:27
12:35 pmurias joined 12:39 ejs1 joined
flip123 says thank you 12:45
12:46 bluescreen joined 12:48 ruoso joined 12:50 ejs2 joined
Wolfman2000 *yawn* morning 12:53
12:57 masak joined
masak good somewhat-later-in-the-day, #perl6. 12:57
Wolfman2000 masak: I'm debating/thinking of putting the owrk I did on my Perl 5 written pastebin onto github. Thoughts? 12:58
masak Wolfman2000: online code is almost always a good thing.
then it has a URL.
Wolfman2000 ...when is it not a good thing? 13:00
masak don't know, really. at worst, it might be a waste of github. but in your case, it sounds like it might come in handy.
your data model, for example. I might want to look at it when I do the Perl 6 version. 13:01
Wolfman2000 masak: The frustrating part is really related to Syntax::Highlight::Perl6. I'm starting to wonder if it would be better to just make our own parser or something...valid code won't parse.
masak Wolfman2000: I think basing the highligter off STD.pm is the way to go. 13:02
Wolfman2000 ...something else is bothering me. November is hosted on github. I have a feeling github doesn't provide web hosting. How does that work?
masak Wolfman2000: you know about feather?
Wolfman2000 masak: I think that's what S::H::P6 does...either that, or it provides its own
masak: Yes, I have an account
masak oh right.
I run it from there.
Wolfman2000: yes, S::H::P6 runs on top of STD.pm. 13:03
Wolfman2000 ...it's just compiling STD.pm that's a pain, that's all. 13:04
masak it's a smaller pain than rolling your own parser... 13:05
Wolfman2000 masak: Then I kind of want S::H::P6 to be updated. I'm unsure if that will take place
masak far as I know, either azawawi or szabgab tends to that module. 13:06
13:10 xomas_ joined
Wolfman2000 ...still torn honestly, but I may be able to go through with it 13:12
mberends good morning/afternoon! both azawawi and szabgab maintain S::H::P6 regularly, but the take a snapshot of STD.pm (which changes almost daily) 13:14
*they take
Wolfman2000 mberends: so on Feather, I should ./gimme5 a lot? 13:15
or at least daily?
mberends Wolfman2000: yes, if the latest changes affect *your* Perl 6 source code
colomon mberends: I thought it was actually quite some time since S::H::P6 updated to match STD.pm... 13:16
mberends colomon: maintenance does tend to go in bursts. It's one thing to change STD.pm, another to render it beautifully 13:17
Wolfman2000 ...and once again, there's a way for me to run ./gimme5, but I'm not doing it right
mberends 'make' in pugs/src/perl6 13:18
Wolfman2000 always forgets that comes first.
okay...once this makes, do I have to copy STD.pm somewhere for it to be recognized by Perl 5?
...also, am I supposed to clear either my Perl5 or Perl6 variables first? 13:19
mberends shouldn't be necessary. you will get a STD.pm5 and STD.pmc file 13:20
13:20 payload joined
Wolfman2000 ...funny...when I clear PERL5LIB beforehand, I don't get any errors on the make 13:20
well, no SORRY errors at least
only error I get is this: rm: cannot remove `CORE.syml.store': No such file or directory But this is ignored 13:21
mberends that error is harmless
masak mberends! \\o/ 13:22
Wolfman2000 PERL5LIB='' make worked: potential difficulties include not being able to locate the DEBUG, NAME, and Stash modules
jeremiah OH HAI
masak jeremiah! \\o/
mberends masak: \\o/
jeremiah masak: :)
mberends !
13:23 colomon joined
Wolfman2000 mberends: my normal PERL5LIB path includes pugs/src/perl6. At this point, Syntax::Highlight::Perl6 should work, correct? 13:23
jeremiah Shouldn't one use PERL*6*LIB for stuff like that? Not that it really matters I suppose. 13:26
But don't other configure scripts look for that var?
mberends Wolfman2000: dunno really, my own scripts all work without PERL5LIB 13:28
jeremiah: the thing is, S::H::P6 is written in Perl 5
jeremiah Ah, okay. :) 13:29
jeremiah goes back to lurking. :)
masak lurking is a time-honored tradition here on #perl6. :) 13:31
mberends "lurked" offline most of the last 2 weeks, reading irclogs ;) 13:32
masak that's pretty advanced lurking. 13:33
you don't even show up on the radar. :)
Su-Shee lurkers anonymous. 13:35
jeremiah Well, in that case, I've been lurking here for years.
Offline lurking of course.
mberends irclog.perlgeek.de is a wonderful thing. even works on a mobile phone # moritz_++ 13:36
when masak++ calls it an addiction, he's right 13:37
masak other people read the newspaper for breakfast. I read the #perl6 backlog. 13:38
mberends :-) ditto
masak overall, much better news. :) 13:39
jeremiah Still a lot of 'die'ing though . . .
mberends news about the future instead of the past
masak jeremiah: yes, but we CATCH some of them in time.
13:39 quantumEd joined
jeremiah I knew you'd TRY something like that 13:40
13:40 bloonix joined
masak it's the LAST time, I promise! 13:40
jeremiah NEXT!
mberends *groan*
Wolfman2000 ...perl puns. great
masak if we had labels, it'd be NEXT joke.
er, or 'next JOKE', rather. 13:41
jeremiah .punning FTW
masak jeremiah: Perl has always been about punning. it's even built into the operators. 13:42
jeremiah Yeah, one suspects that TimToady guy has a sense of humor and knows a bit about language. 13:43
13:44 riffraff joined
masak Perl: taking puns seriously. 13:47
jnthn Srsly? 13:48
:-) 13:49
jeremiah jnthn: \\0/ 13:50
(just greeting you like all the cool kids do.)
sjn o/
\\o
Wolfman2000 jeremiah: the greeting is usually \\o/
sjn \\o/
\\o>
jnthn jeremiah: The cool kids have smaller heads though.
sjn <o/ 13:51
jnthn ;-)
masak \\☺/
jnthn Show-off!
masak jnthn: so I'm 'noisy', am I? :)
masak backlogs
sjn what, no utf-8 hands? :-P
jeremiah lol
jnthn masak: I'd never say such a thing.
;-)
masak sjn: ☜☺☞ 13:52
13:52 BinGOs joined
masak jnthn: I derive some consolation from the fact that I've also been known to write some code. 13:52
jeremiah Those look more like crutches than hands.
mathw mberends: Oh so that's where you'd been.
sjn hey, it's BinGOs! :)
masak BinGOs! \\o/
jeremiah I know him! 13:53
We met at YAPC::EU 2007 I think
BinGOs you crazy people
jeremiah w00t
Wolfman2000 sjn: I'll give you a utf-8 hand. ☝
masak and he talks!
sjn offered BinGOs a free trip to Oslo once, and he declined o_O
masak Wolfman2000: yeah, the ceiling in here should really be repainted, you're right. 13:54
jeremiah Is that you Salve?
sjn <- Salve
jeremiah Aha!
w00t
BinGOs and you'll never let me forget it will you.
jeremiah Sheesh, I know a bunch o folk here.
Drinks for all my friends!
sjn BinGOs: yeah, I'm still a little miffed about that
jnthn \\o/
13:54 slavik joined
BinGOs weeps gently. 13:55
masak hugme: hug BinGOs
hugme hugs BinGOs
slavik ...
13:56 iblechbot joined
jnthn could use a drink right now...after just discovering an SQL query with raw unvalidated unescaped formdata being interpolated into it... *sigh* 14:00
14:01 BinGOs_ joined
masak wonder if a government will ever be overthrown due to an SQL injection. 14:02
14:02 perigrin joined
mathw hugme: hug jnthn 14:02
hugme hugs jnthn
ruoso hopes PHP with register_globals and magic_quotes to be ruled illegal then 14:03
14:03 payload joined
Wolfman2000 ruoso: I believe PHP 6, when it comes, will disable the magic quotes and register globals functions 14:03
Anyway, I'm going semi afk...I want to knock out as much homework as I can now before I think about my fun stuff today. 14:04
14:11 colomon_ joined
pmurias thinks about writing Acme::CGI::magic_quotes 14:11
mathw purpose? 14:13
14:21 colomon joined 14:27 colomon_ joined
pmurias mathw: Acme modules are supposed to be useless 14:34
sjn masak: it's a module that gives you all kinds of Magic quotes (e.g. "Sim sala bim!" or "Abrakadabra!") 14:35
masak sjn: mis-tab? mathw?
sjn thought that was obvious
ah :)
yep :)
masak wants such magic quotes 14:36
mathw ah 14:37
I see
14:39 rgrau` joined, dbrock` joined 14:40 colomon joined 14:48 frew_ joined, colomon joined 14:56 KyleHa joined
moritz_ re 14:59
14:59 colomon joined
pmichaud good morning, #perl6 15:00
jnthn morning, pmichaud 15:02
15:02 payload joined 15:06 ejs joined
masak morning, pm. 15:08
Wolfman2000 is glad that this assignment is at least a semi fun one... 15:09
I get to make a movie with iMovie. :D
15:09 Psyche^ joined 15:32 am0c joined 15:40 jferrero joined 15:41 colomon joined 15:46 nihiliad joined 15:51 payload joined 15:56 astrojp left
frettled pmichaud: good localtime() 16:01
cognominal pmichaud++, jnthn++ # thx for you blogs
pmichaud frettled: good localtime() 16:02
sjohnson rakudo: say localtime().perl 16:07
p6eval rakudo 7347ec: Could not find non-existent sub localtime␤in Main (file src/gen_setting.pm, line 324)␤
moritz_ rakudo: say time 16:08
p6eval rakudo 7347ec: 1259164722.05861␤
sjohnson hello masak good 2 see u 16:09
masak sjohnson: very good to see you too.
sjohnson moritz_: time is now a constant thing?
no more cpan dependencies for simple time? 16:10
rakudo: say time.WHAT
p6eval rakudo 7347ec: Num()␤
sjohnson if so, great
（　｀ー´）
Wolfman2000 rakudo: say time
p6eval rakudo 7347ec: 1259164900.63664␤ 16:11
Wolfman2000 ...right. How DO you format this sucker?
masak Wolfman2000: you might want to check my Temporal-flux project. 16:12
Wolfman2000 masak: right now taking care of my homework
masak right, no problem. 16:13
pmurias homework-- 16:14
16:20 colomon_ joined
Wolfman2000 pmurias: And what do you normally do in the day? 16:21
16:29 snearch_ joined
TimToady hmm, planetsix seems to be hung 16:34
masak was it something I blogged? :/
jnthn masak overloaded it!
TimToady maybe it's slashdotted :)
Wolfman2000 almost done with my movie assignment... 16:35
pmurias Wolfman2000: studing cs
Wolfman2000 pmurias: computer science? Sounds like the path I took. Where are you going? 16:36
pmurias University of Wrocław, Poland 16:37
Wolfman2000 Not familiar with that...symbol between c and a
Looks like a cursive t
Su-Shee it's similar to a w like in "what" 16:38
pmurias Wolfman2000: ł is a modified l 16:39
Wolfman2000 ah 16:40
pmurias it's pronouced the way Su-Shee wrote
Wolfman2000 l = w?
frettled TimToady: perlcabal.org is apparently struggling, too. 16:41
masak: I'm cooking up a new blog post about time handling in Perl 6, I thought I'd briefly compare the spec with Rakudo's current implementation with Temporal-flux's ideas 16:42
16:42 BinGOs joined
Su-Shee pmurias: my personal favorite is the politician Zbigniew Brzeziński to pronounce. ;) 16:42
masak frettled: cool! 16:43
16:43 perigrin_ joined
masak frettled: I'm sure you'll think of things I haven't considered. 16:43
jnthn Su-Shee: Ouch...I just tried and my tongue broke.
:-)
frettled masak: I'll try to keep the general framework for now, not really the specifics. 16:44
Wolfman2000 one version of my movie has been exported... m4v file, 24 megs for about 4 minutes. Is that a decent size?
16:46 cognominal joined
frettled jnthn: Try «ɡvbrdɣvnis» (Kartvelian for «he's plucking us») ;) 16:47
masak very useful phrase.
frettled Yes, now you know what to say when you're in a specific part of Georgia.
Su-Shee cute v-like letter with the sling.. :) 16:48
frettled Su-Shee: there's more cuteness in the article I stole it from: en.wikipedia.org/wiki/Consonant_clu...uistically
But we should be careful, or TimToady will use some of this somewhere in the spec. 16:49
Wolfman2000 frettled: and what's wrong with that?
frettled «not that there's anything wrong with that»
16:49 gbacon left
Su-Shee oh, I don't mind, I'm from the humanities department anyway.. 16:50
frettled oh, the humanity!
jnthn finally finishes the bug-fixing rampage in a codebase he hates hacking on...which he'd managed to procrastinate for 3 months.
Now I can do Rakudo stuff. \\o/ 16:51
16:51 arnsholt joined
masak \\o/ 16:51
16:51 kst` joined
arnsholt masak: Thank you ^^ 16:51
Su-Shee the context stuff was the first thing I understood when I learned programming (with perl.) "of course everything has a context, how couldn't it.."
frettled jnthn: yay
masak arnsholt: we've been waiting all day for you. :) 16:52
arnsholt =D
masak frettled++ # 'he had in his possession a bunchberry plant' 16:54
frettled: that article rocks! :)
16:55 mofino joined 16:57 slavik joined 17:01 ng_feed joined
ng_feed rakudo-ng: jnthn++ 17:01
rakudo-ng: Get us building signatures for subs/methods/role bodies lazily.
rakudo-ng: jnthn++
rakudo-ng: Various tweaks to make sure we're getting as far as properly requesting the role applier.
rakudo-ng: jnthn++
rakudo-ng: First cut of role to class application. Very incomplete, but works for applying methods from a single role, and shows things flow through the overall model correctly.
moritz_ turns out ng_feed died whenever the connection to github was dodgy
ng_feed rakudo-ng: jnthn++
rakudo-ng: Fill out collision and requirement handling a bit. Note that we do it all ourselves - later, we'll track enough to give better errors, but also it'll make it easier to totally drop basing this of Parrot's Role PMC in the future.
rakudo-ng: masak++
rakudo-ng: Merge branch 'ng' of github.com:rakudo/rakudo into ng
rakudo-ng: jnthn++
rakudo-ng: Write Perl6::Metamodel::RoleToClassApplier in NQP instead of PIR. Yay!
rakudo-ng: masak++
rakudo-ng: Merge branch 'ng' of github.com:rakudo/rakudo into ng
rakudo-ng: masak++
rakudo-ng: [docs/metamodel.pod] rw review
rakudo-ng: KyleHa++
rakudo-ng: S14-roles/namespaced.t passes now 17:02
rakudo-ng: (mberends)++
rakudo-ng: [docs/metamodel.pod] minor typo and layout corrections
KyleHa ng_feed playing catch-up?
TimToady combining the ++ lines with the following line would cut the number of lines in about half
moritz_ KyleHa: yes 17:03
TimToady: good idea... but programming is hard, I need to go shopping first :-) 17:05
Wolfman2000 moritz_: what's ng_feed programmed in? 17:06
TimToady programming would be easier if you used a good language :)
moritz_ ng: say (1, 2, 3).join('|')
p6eval ng af3d38: 1|2|3␤
moritz_ Wolfman2000: Perl 5 + POE
Wolfman2000 I've used Perl 5...I don't recall POE.
KyleHa Perl 5 IS a good language! 17:07
moritz_ Wolfman2000: POE is a framework for asynchrone calls
"cooperative threading"
Wolfman2000 I wonder if Perl 6 has that built in 17:08
moritz_ it hasn't
(and shouldn't) 17:09
TimToady well, ==> could be viewed as an event queue
dbrock` does Perl 6 support constructors for roles? 17:11
moritz_ thought ==> was just syntactic sugar for some function calls
TimToady ==> implies threading
just like | implies processes in shell
moritz_ so it's hyper?
TimToady no, it's pipeline 17:12
like the difference between a SIMD machine and a Cell processor
hyper is SIMD
==> is cell
moritz_ dbrock`: since roles can't be instantiated directly, I'm not sure what to make of your questions
*question
dbrock`: you can write a constructor in a role, and the class you compose the role into can use it 17:13
TimToady dbrock`: it supports lazy initializer values with closures that can be called when the role eventually participates in a real object
see WHENCE in the specs 17:14
moritz_ really goes to the shops
masak swimming & 17:15
KyleHa @in ==> map { step1() } ==> map { step2() } 17:16
lambdabot Maybe you meant: index instances instances-importing id wn
17:16 rjh_ joined
KyleHa So is it possible for some calls to step2() to happen before some calls to step1() ? 17:16
17:17 colomon joined
jnthn KyleHa: Yes. 17:17
KyleHa: Well, not for the dependent values. 17:18
KyleHa Right.
Far out.
...and if I want to avoid that, I use "S==>" instead, right?
jnthn KyleHa: But basically, that's start two threads or something like that.
I'm not sure ==> are operators...
std: my @in; @in S==> map { step1() } S==> map { step2() }
p6eval std 29194: [31m===[0mSORRY![31m===[0m␤Can't sequence ==> because sequencer operators are too fiddly at /tmp/yHNAf7im2f line 1:␤------> [32mmy @in; @in S==>[33m⏏[31m map { step1() } S==> map { step2() }[0m␤FAILED 00:01 107m␤ 17:19
KyleHa That answers that, thanks. 8-)
jnthn Last time I was told I was too fiddly was back at primary school. :-)
KyleHa Is there to be a facility somewhere to limit the number of threads one might spawn? Say I have a really long chain of ==>, but I never want more threads than cores. 17:20
jnthn Well, S17 is as good a thread spec as we have at the moment, so if it doesn't say something on that, then the answer is currently unspec. 17:22
KyleHa OK, thanks.
jnthn I'd really like to be able to play with getting threading stuff into Rakudo. It'll almost certainly not happen until next year really though.
17:23 mberends joined
jnthn It'd be really -Ofun to play with that stuff though. :-) 17:23
17:24 zamolxes joined 17:47 cdarroch joined 17:51 ruoso joined
moritz_ rakudo: 1 ==> 2 17:52
p6eval rakudo 7347ec: Feed operators are not yet implemented in Rakudo, sorry␤in sub 17:53
moritz_ is scared of running spectest in master 17:54
17:56 payload joined
jnthn moritz_: What's to fear? :-) 17:56
Wolfman2000 it takes awhile to run perhaps jnthn?
Also, not all of the spec tests pass
moritz_ jnthn: lots of fallout 17:57
from the s/undef//g and s/Object/Mu/g changes 17:58
jnthn moritz_: Ah.
moritz_ so far 10 tests run, 4 are "Dubious"
Wolfman2000 ...ah yes. 無 17:59
Wolfman2000 likes his IME for the most part.
What's next? 祭無? 18:03
bonus points to whoever knows the kanji 18:04
dbrock` moritz_: the "instantiation" of a role is an "object part", and it's often useful to initialize that part before it's used 18:06
jnthn "object part"? 18:07
moritz_ what is an "object part"?
dbrock` a bunch of attributes
jnthn dbrock`: Yes, but you can't instantiate a role. It puns into a class if you try.
18:07 gbacon joined
jnthn dbrock`: Then you get the same semantics as if you'd defined the attributes in the class. 18:08
dbrock` if I have class A does R, and then say A.new, then A's constructor is invoked
I want R's constructor to be invoked as well
TimToady: (thanks, by the way; I'm not ignoring your answer, I just didn't fully understand it) 18:09
jnthn dbrock`: Where "R's constructor" means...?
moritz_ dbrock`: roles are flattened into classes
dbrock`: so any constructor is flattened into the class too 18:10
jnthn dbrock`: Constructors into Perl 6 are nothing more special than methods that happen to call bless.
dbrock` jnthn: a routine that sets up R's part of the object
jnthn *in
dbrock` all the superclass constructors will be called automatically, right?
KyleHa Nonexistent sub Mu, method 'notdef' not found...
jnthn dbrock`: Superclass BUILD methoeds will be called automatically. 18:11
dbrock`: Again: Perl 6 doesn't have constructors that are any more special than a method that calls bless.
18:11 PZt joined
jnthn dbrock`: The name "new" isn't at all special. 18:11
dbrock` okay, so maybe by "constructor" I mean "BUILD method"? 18:12
jnthn dbrock`: Perhaps, yes.
dbrock`: I'm really not sure how they get handled with respect to roles though.
dbrock` I don't really consider A.new to be a constructor; I would maybe call it a factory method or something
jnthn dbrock`: I mean, again, they're just methods, so subject to flattening composition again.
dbrock`: So not quite sure off hand how to do what you're after.
dbrock`: Of course, anything you write as the RHS of a "has $.x = ..." in a role will be run. 18:13
dbrock` okay, so I guess if I put a BUILD method into R, then it will get run automatically upon object construction?
jnthn dbrock`: Not under current semantics.
dbrock` okay
jnthn dbrock`: Under current semantics it'll be ignored iff the class has a BUID.
dbrock` oh, so there can only be one BUILD?
jnthn dbrock`: and conflict with a BUILD provided by other roles if they provide one.
dbrock`: Yeah, that's the issue. It's just a (sub)method 18:14
dbrock` I see
hmm
jnthn hmm indeed. :-)
dbrock` that seems... bad
jnthn wonders what TimToady meant by WHENCE.
dbrock`: Yeah, it'd be nice if we had a good solution for that. 18:15
moritz_ is there a reason why a class can't have two BUILD methods? 18:16
jnthn moritz_: I don't really get the question. 18:17
moritz_: I mean, it's exactly the same as "is there any reason why a class can't have to X methods?"
dbrock` I guess the real question is, "what if a class could have multiple copies of the same method?"
moritz_ jnthn: are there multi-submethods?
jnthn moritz_: There are.
Well 18:18
I guess there are. :-)
Not quite sure we handle those completely right yet.
moritz_ and I think BUILDALL should just call all matching BUILD submethods
jnthn moritz_: I guess if we get people to mark their BUILD as multi or introduce a proto method BUILD in their class, then the ones from roles would merge.
moritz_ (with .* )
jnthn Well, .* needs a bit of care, since it also walks the hierarchy. 18:19
But some mechanism like that, yes.
That's the best I can think of right now.
TimToady see S12:2086
TimToady will be offline for the next 7-8 hours 18:20
jnthn TimToady: I know that section, I'm just not quite clear how exactly it applies here, since we don't really ever do any auto-vivification, that's all. 18:22
TimToady: OK, catch you later. :-)
(We don't auto-viv when composing the role into a class, I meant.)
TimToady BUILD autovivifies role's WHENCE blobs
by incorporating them into the arguments 18:23
jnthn OK...what does attachign a WHENCE block to a role look like
role Foo { has $.x; is WHENCE { x => 42 }; } 18:24
maybe s/is/will/...
TimToady it's not attached to the role, it's just another arg to .bless
FOO{ :x(42) } or some such
*foo 18:25
*Foo
jnthn TimToady: Ah. I think the issue here is that dbrock` wants to write something insdie the role body that will get run at BUILD-time.
TimToady gah
jnthn *inside
Every time, not just on a particular invocation.
If I understand correctly, we'd pass that Foo{ :x(42) } in the call to new/bless?
TimToady at one time we had role's BUILD methods called automatically as part of class's build, not sure where that went 18:26
yes
jnthn OK, that makes sense to me.
It's just not quite the answer to the problem at hand. Calling role's BUILD methods automatically is...well...awkward in that the effect of the role on the class should be over by composition time really...ordinarily, dispatch doesn't go looking inside roles. 18:27
I guess we can work out some way to have them all called though.
TimToady Wolfman2000: I can read the kanji, but that doesn't mean I understand what "festival emptiness" means... 18:28
Wolfman2000 the festival of nothing/festival of emptiness
or the empty festival
jnthn Nobody attended. :-)
Wolfman2000 ...and thus, my joke got hijoked
jnthn++: bravo 18:29
jnthn ng: Nil 18:30
p6eval ng af3d38: ( no output )
jnthn ng: say Nil
p6eval ng af3d38: ␤
TimToady that's okay, I think
moritz_ ng: say Nil.perl
p6eval ng af3d38: ()␤
moritz_ no objetions here either
jnthn erm 18:31
TimToady well, off we go &
jnthn ...the PIR that Nil generates looks...
ng: say Nil.WHAT
p6eval ng af3d38: List()␤
jnthn oh, Nil is a sub...
jnthn was doin it rong. 18:32
dbrock` something is actually called Mu?
Wolfman2000 dbrock`: Mu is the new undef 18:33
dbrock` haha, okay 18:34
awesome
jnthn ng: my @x = 1,2,3; say @x.elems; 18:39
p6eval ng af3d38: 3␤
moritz_ ng: my @x = 1, 2, 3; say @x.perl
that hangs, dang 18:40
p6eval ng af3d38: ( no output )
jnthn :-(
18:41 meteorjay joined 18:45 stephenlb joined 18:49 beggars joined
ng_feed rakudo-ng: jnthn++ 18:51
rakudo-ng: Start to fill out role to role application, including flagging up conflicts. Sort of works, but it appears that some issue in push and another issue prevent the conflict analysis working out properly just yet.
moritz_ jnthn: in master, where is class Object defined? 18:53
dalek p-rx: 7387be6 | tene++ | (3 files):
Add CONTROL blocks
moritz_ I tried to change it to Mu, but I get "Parent isn't a Class." errors from perl6;ClassHOW;onload 18:54
jnthn moritz_: Ouch
moritz_: Well, there's references to it all over the place.
moritz_ $P1 = get_hll_global 'Mu'
$P1 = p6meta.'get_parrotclass'($P1)
addparent $P0, $P1
jnthn It's sometimes known as Perl6Object in lower-level bits too.
moritz_ the addparent dies 18:55
jnthn That's kinda odd.
moritz_ so I suspect that Mu isn't set up properly as a class
hm
jnthn Yeah, it could be a non-fun change. 18:56
moritz_: tbh
moritz_: I'd just do something somewhere on startup that does somethin glike
moritz_ I guess it's called Perl6Object in the low level stuff because Object is already known in parrot?
jnthn $P0 = get_hll_global 'Object'
set_hll_global 'Mu', $P0
Just to alias it.
That should get a bunch of the tests back.
moritz_ I'll try that instead, thanks
jnthn And also, importantly, give the userbase chance to migrate. 18:57
jnthn afk for a little bit 19:00
dalek p-rx: ae0666c | tene++ | t/nqp/44-try-catch.t:
Update test plan
 19:11
kudo: 2eb38b5 | moritz++ | perl6.pir:
alias Object and Mu, as suggested by jnthn++; wins us back quite a few tests
 19:13
szabgab just for the record azawawi is doing S::H::P6 but he seem to be busy with other stuff lately, we have not seen him on #padre for a while 19:14
19:15 slavik joined
Su-Shee hi szabgab. 19:16
19:16 pjcj joined 19:19 gabiruh_ joined, omega_ joined, bigpresh_ joined 19:28 huf joined
dalek kudo: 2fc0a98 | moritz++ | src/setting/Object.pm:
implement Object.notdef

autovivification returning Failure, not Mu.		 19:31
19:34 rfordinal joined
Wolfman2000 phenny: tell Juerd I wonder how easy it is to upgrade Git, or if I should just install a local copy without doing the make install step. 19:36
phenny Wolfman2000: I'll pass that on when Juerd is around.
19:39 IllvilJa joined
moritz_ does a full-ugprade on feather1 19:40
19:41 cognominal_ joined
Wolfman2000 moritz_: I forgot you had control over Feather. Should I disconnect from ssh? 19:41
moritz_ no need
since each ssh connection spawns its own process, the connections survive an upgrade 19:42
19:48 rfordinal left
pmichaud hugme add coke to nqp-rx 19:49
hugme hugs coke. Welcome to nqp-rx!
Juerd The system was recently dist-upgraded, and that included git. Is that already too old?
phenny Juerd: 19:36Z <Wolfman2000> tell Juerd I wonder how easy it is to upgrade Git, or if I should just install a local copy without doing the make install step.
szabgab Su-Shee, rehi 19:50
moritz_ it's 1.6.5.3
now
Wolfman2000 moritz_++: thanks. Thinking of using git for...well, hosting the Perl 6 pastebin running on Perl 5 19:51
moritz_ like github's gists? 19:52
Wolfman2000 moritz_: yeah
Still slightly hesitant, but...I may do so tonight while helping parents out with making desert 19:53
moritz_ Wolfman2000: do you have a basic version running already?
Wolfman2000 of my work? basic, but it works. It's not in a repo though
moritz_ ok
Wolfman2000 not every little thing I want works, but...well, I guess that's where we come in 19:54
dalek p-rx: 294d493 | (Will Coleda)++ | src/stage0/ (3 files):
rebuild stage0 after tene++'s try/CATCH/CONTROL additions.
 20:04
20:04 kst`` joined 20:07 Tene_ joined, r0bby_ joined, omega__ joined 20:08 simcop2387_ joined, arnsholt joined 20:22 pmurias joined
lisppaste3 moritz_ pasted "rakudo master spectest summary" at paste.lisp.org/display/91111 20:34
20:37 colomon joined 20:45 pure1111 joined 20:47 colomon_ joined
jnthn back from nomz 20:50
moritz_: Heh. The paste-bin apparently considers this an emergency. ;-) 20:51
20:52 ruoso joined
moritz_ rakudo: say 4679 / 60 20:59
p6eval rakudo 7347ec: 77.9833333333333␤
moritz_ on 2 cores. That's... not quick
21:02 bluescreen joined 21:04 quantumEd joined
jnthn moritz_: It took 78 minutes? Ouch. :-( 21:04
moritz_ aye.
Tene jnthn: this is the part where you tell moritz that ng is 20x faster. 21:05
jnthn Tene: Yeah...apart from it's not.
arnsholt Sounds about right. IIRC spectest takes about half an hour on my four-core desktop
Wolfman2000 moritz_: I know you claim feather has an updated git, btu when I re-connected to feather, I'm still on the behind version. Is this intentional? 21:13
21:13 gbacon joined 21:14 slavik joined
moritz_ Wolfman2000: which version do you see? 21:18
Wolfman2000 1.5.6.5
moritz_ too
wait 21:19
that's not the version of the git package
Wolfman2000 ...so source is the best way to go.
moritz_ that's because the first git in $PATH is /usr/local/bin/git
Wolfman2000: nope
just use the one in /usr/bin :-)
Wolfman2000 ...and right now bit is...yep
err, git is pointing to usr/local/bin
21:20 masak joined
masak o/ 21:20
Juerd Why *is* there a git in /usr/local?
Wolfman2000 Juerd: you tell me. I don't install stuff on there.
Juerd Does anyone depend on that old one?
Wolfman2000 doesn't.
Juerd Probably was installed there because debian was lagging behind 21:21
masak did this twitterer not sleep because of Perl 6 regular expressions? twitter.com/uasi/status/6058798285 - man, I need to learn Japanese so I won't have to bother other people with translations when Google Translate fails... :)
Wolfman2000 of course, now I forced /usr/bin to come ahead of /usr/local/bin, so I don't know what sort of dangers there are with that.
moritz_ any objections to removing the git in /usr/local/ ?
Wolfman2000 If need be, I can always alias.
Juerd The problem with manual installation to get a newer package, is that it will eventually be outdated again, and there's usually no record of which files to delete to get rid of it. :(
Wolfman2000 moritz_: symlink perhaps?
Juerd moritz_: None from my part. I'd suggest keeping the files around just in case. 21:22
21:22 iblechbot joined
Juerd leto has a git source tree in ~leto/git/git 21:23
Why the hell is my ls sorting case insensitive? :( 21:24
moritz_ moved.
ng_feed rakudo-ng: jnthn++
rakudo-ng: Update some array creations.
Wolfman2000 1.6.5.3
moritz_++: Thanks 21:25
moritz_ to /usr/local/old/bin
Juerd moritz_: How did you determine which files to move?
Same mtime? 21:26
Tene Juerd: lang/locale
moritz_ Juerd: just everything in bin/git*
Juerd Tene: It does this even with LANG=C 21:27
Tene Huh.
Juerd moritz_: There's more, though
moritz_ Juerd: aye, just found it
just also moved share/git* 21:28
21:28 takadonet left
moritz_ only the stringiterator left 21:28
diakopter Juerd: isn't there a way to set up a local repo with a lower priority (that can store one's custom .debs) so that when a newer version is released in the public repos, it becomes available? 21:36
moritz_ diakopter: there is 21:40
but it's a bit fiddly to set up
21:42 compactwater joined
compactwater Quick question: I've installed Perl 6 Rakudo, but I'm lost on how to install dependencies. 21:43
masak compactwater: what kind of dependencies? 21:44
compactwater All the ones for ilbot
moritz.faui2k3.org/en/ilbot
ng_feed rakudo-ng: jnthn++ 21:45
rakudo-ng: Get role composition conflict detection and resolution by having a method provided working again.
21:45 alester joined
masak compactwater: just so we're on the same page here: ilbot and Rakudo are completely distinct. 21:45
one does not depend on the other.
moritz_ right, ilbot is a perl 5 program 21:46
compactwater It says Perl6, doesn't it?
21:46 _eMaX_ joined
masak compactwater: is your goal to have ilbot up and running? 21:46
Tene compactwater: it's a Perl 5 bot that logs the #perl6 IRC channel.
compactwater Oh. That's where I was confused 21:47
masak, yes
masak compactwater: then, as moritz_ said, you don't need Rakudo for that.
compactwater Oh. I feel stupid.
Well, thanks for that, I was way off.
jnthn pmichaud: plz we can haz identifiers in nqp with hyphens in them? :-) 21:48
moritz_ compactwater: it was developed for a Perl 6 channel, not in Perl 6
colomon jnthn: did the ng issue where operators dispatched to the wrong version ever get sorted out?
jnthn colomon: I don't remember hearing that it was sorted out, and the lack of S06-multi tests passing makes me fear not. :-/ 21:49
21:49 Su-Shee left
masak compactwater: now that you have Rakudo installed, you might as well try it out a little. :) 21:49
moritz_ oh, and the spam deletion is bitrotted
frettled chickened out on blogging about Temporal.pm, and wrote another piece instead. Need. More. Time.
compactwater Sadly, I'm not even a Perl programmer. I'm just setting up a quick bot for a network 21:50
Tene frettled: can't you use Temporal.pm to give yourself more time?
jnthn colomon: It got bogged down in not settling on a solution.
masak compactwater: ah. well, best of luck.
frettled Tene: I'll ask my friend Tai about that. ;)
colomon jnthn: seems like that's an important one to land soon.
moritz_ should I move ilbot out of the pugs repo, and into a separate github project? 21:51
jnthn colomon: Agree.
Wolfman2000 moritz_: how many projects are in the pugs repo anyway?
Some separation will probably be good
jnthn colomon: Trouble is, the fix I wanted originally won't fly.
moritz_ Wolfman2000: close to * :-)
21:52 SimonAW joined
Wolfman2000 moritz_: you do NOT have infinite projects on pugs 21:52
moritz_ no, close to
Wolfman2000 I still don't believe that
frettled heh
SimonAW Good evening, gentlemen.
frettled moritz_: I think I may like github, it seems slightly more useful. 21:53
moritz_ one of our math profs used to say "... and the logarithm of infinity is around 5" (regarding the slow growth of the log function)
frettled SimonAW: What you say !! 21:54
SimonAW ;)
Although it's my favorite meme i consider it "common courtesy" ;)
Wolfman2000 alright, afk 21:55
SimonAW I mean at least I didn't say "How are you ~" ;)
frettled SimonAW: yep
SimonAW: Hmm, it looks like I scared all of them into science by picking up that meme. 21:59
moritz_ [off] (irc logging test) 22:00
masak frettled: 'scared into science'? :P
SimonAW Aw, never mind...
frettled heh
masak SimonAW: did someone set up us the bomb?
frettled: I wish I could scare more people into science... 22:01
frettled masak: me, too
jnthn
.oO( how is meme formed? )
frettled ICANHAZMEME, OMGWTFBBQ!
SimonAW oO( OMG, what did I do? )
jnthn SimonAW: Don't worry, it doesn't take much to set off a chain of events like this on #perl6. :-)
masak jnthn: they need to do way instain mother.
jnthn masak: She needs to lady to rest? 22:02
masak jnthn: my pary are with the father. I am truly sorry for your lots.
SimonAW Okay, to get back ontrack: Do I get that right that parrot is still using svn but kind of moving to git?
moritz_ SimonAW: there are tentative plans to move parrot to git after the 2.0 release 22:03
SimonAW Nice, nice...
frettled \\o/
SimonAW Oh, I realise I should be in #parront on irc.perl.org for that question... 22:05
sjohnson parront == chinese knockoff of parrot 22:06
SimonAW sjohnson: Oh, snap!
masak sjohnson: je ne parront francais.
sorry, je ne parront PAS francais. 22:07
moritz_ lol
SimonAW It is late here (11pm)
moritz_ I was just about to correct masak++ :-)
masak SimonAW: here, too!
moritz_: je ne besoin PAS du correction.
moritz_ masak: s/PAS/JAMAIS/ :-) 22:08
masak well, wouldn't go that far...
I have been known to be wrong.
masak re-deflates head 22:09
SimonAW masak: Correction: My writing skills are more susceptible to... lateness(?) ;)
masak SimonAW: ah. quality is a function of time? :)
SimonAW hehe
frettled which Temporal.pm currently does not help us wth 22:10
SimonAW Oh dear. Is this all going right off to the logs?
masak SimonAW: you bet.
frettled Mwaha. Mwahahahaha. MUAWHWWHHAHAAAAA!
masak 'everything you say can, and will be used against you...'
moritz_ SimonAW: yes. But there is so many bullshit in here that it's virtually impossible to find a particular piece of bs in it :-)
masak not if you use my special bs search engine. 22:11
moritz_ built on top of the GGE?
masak naturally.
jnthn /<bs>/
SimonAW GGE?
moritz_ then it's too slow to find anything useful :-)
SimonAW: the Glacial Grammar Engine 22:12
frettled Gary Gygax Enterprises?
masak SimonAW: during the past few months, I've been writing a Perl 6 parser in Perl 6.
frettled SimonAW: and that explains his funny mood
masak s/parser/grammar engine/
SimonAW Heh 22:13
masak tilts his head madly
SimonAW Indeed. It would alter my mood as well if I could write a Perl parser.
frettled Somehow, Freddy Mercury missed that verse in «I'm going slightly mad».
masak 'ven I push dis button, you vill all vaporize...'
whoa. I just had a semi-strong deja vu. have I played mad scientist before here on #perl6? :) 22:14
frettled but then the grammar engine will spend so long handling the command that the vaporization may have temporal correlation with the Sun turning into a red giant.
masak frettled: ingenious, isn't it? 22:15
boil, fools!
frettled masak: yes, yes, quite. Now hand me the beaker with the turquoise bubbly in it, please.
masak pouts and hands over beaker
frettled sips. Mmmmmm. 22:16
masak that's engine grease, you know.
frettled yummy
moritz_ sends a micro black hole from the LHC to swallow masak's button
masak hey! where'd my button go?
moritz_ cackles madly
masak I'm trying to blog here, and you guys are making it very difficult. :P 22:17
moritz_ masak: there's an easy fix... hide your IRC window :-)
frettled masak: if only you'd spaced your blog entries as weekly instead of nearly daily posts, you'd have an arsenal big enough to make mst just give in and ask you for his color already 22:18
haha
Did I manage to say that before or after he logged off?
SimonAW This is weird. I tried to find information on how to checkout parrot but found myself playing tetris on the web a few seconds later. 22:19
diakopter SimonAW: it happens 22:20
www.parrot.org/download\\ 22:21
www.parrot.org/download
ng_feed rakudo-ng: jnthn++
rakudo-ng: Tweak to attribute meta-object API.
rakudo-ng: jnthn++
rakudo-ng: Start to fill out the Attribute meta-object implementation. Untested, beyond the fact that it compiles.
moritz_ it compiles - let's ship it!
SimonAW diakopter: Ya, stumbled upon SVK
jnthn moritz_: I'm almost certain it'll explode on use, but since nothing tries to use it, that's fine. 22:22
;-)
frettled That's the power of science!
</mythbusting>
jnthn It'd be fun to implement the whole metamodel in NQP some day...rather than just parts of it. 22:23
pugs_svn r29195 | moritz++ | [irclog] move sources to github 22:25
frettled jnthn: If it's fun, I'm sure you'll find the time and energy to do it. 22:26
jnthn frettled: It may make some other future things easier... But it's not so needed for ng.
It is vaguely curious how this is shaping up. 22:28
22:28 SmokeMachine joined
frettled Yes, I'm fascinated by the rapid development. 22:29
moritz_ hugme: reload
hugme moritz_: reloaded successfully
22:29 justatheory joined
moritz_ hugme: list projects 22:29
hugme moritz_: I know about book, gge, hugme, ilbot, json, november, nqp-rx, perl6-examples, proto, svg-matchdumper, svg-plot, temporal-flux-perl6syn, tufte, web
diakopter i'll call success when hugme gets kicked for flooding on `list projects` 22:32
masak: you there? watching the irclogs? 22:34
jnthn I'm in ur irclogs, whatchin ur memes.
frettled diakopter: enter a comment in his blog, he'll notice ;) 22:36
diakopter wait, does GGE *actually* stand for glacial? 22:38
as in glacially slow?
moritz_ yes
jnthn diakopter: Yes really!
jnthn was at the talk and saw the beautiful glacier picture 22:39
diakopter is it really that slow?
moritz_ diakopter: why don't you try it? :-)
diakopter well...
jnthn moritz_: Lack of time? ;-) 22:40
frettled Don't tell us - you started a process three days ago, and it hasn't finished yet?
diakopter I'm busy [re-]writing my own dynamic grammar engine in javascript...(jsmeta)
moritz_ running the optable tests took less than 10 minutes on my machine :-)
diakopter .. optimizing for Perl 6, which just means ripping out all the memoization and left-recursion-evasion 22:41
it already supports parse-time grammar mutation/defintion
and rule 'categories'
it's, like, fast, and stuff 22:42
moritz_ rakudo: say Bool.defined
22:42 ShaneC joined
p6eval rakudo 7347ec: 0␤ 22:42
diakopter whereas jsmeta uses a trampoline and js method lookups, the new thing uses opcodes/switch/case and a much more efficient node-cloning technique 22:43
pmurias: remind me to show you
or now
pmurias diakopter: hi
frettled Hmm, where did zaslon go?
diakopter pmurias: are you doing object cloning (or object prototype-derivation)?
pmurias of what?
diakopter js objects 22:44
22:44 astrojp joined
pmurias i'm not doing object cloning (i'm still unsure about the question) 22:46
22:46 huf_ joined
pmurias diakopter: you mean how is the js object representing a perl6 object created? 22:47
dalek kudo: f151334 | moritz++ | src/setting/ (2 files):
get rid of undef in the setting
 22:53
22:56 masak joined
masak diakopter: I'm here. 22:57
Tene masak: no you aren't!
masak diakopter: GGE hasn't bloomed out yet to show its full slowness. I have high hopes, though.
diakopter: I'll come back to you when I have hard figures.
22:59 patspam joined 23:00 zeke_ joined
jnthn watched with despair at just how glacial GGE was at BPW. 23:00
masak jnthn: that was mainly compilation.
jnthn: for such a small regex, running is pretty OK. 23:01
jnthn masak: Ah...OK, that's not quite so bad then.
masak jnthn: no. but I really got to track down that compilation bug.
lisppaste3 masak pasted "jnthn: ever seen anything like this?" at paste.lisp.org/display/91118 23:05
jnthn masak: No. 23:06
That looks...screwed up.
masak ah! good news! 23:07
23:07 mikehh joined
moritz_ masak: what's the return status of the compilation? 23:07
echo $?
jnthn masak: Does --target=past give anything interesting?
frettled masak: why does that remind me of Farnsworth in Futurama?
masak jnthn: I managed to minimize the whole file to this 'my $var = 42; my $var = 5; say $var'
jnthn: let me check.
jnthn masak: A dupe var causes it?! 23:08
masak moritz_: 0
jnthn: uh huh.
jnthn bbq!
masak this needs fixing.
jnthn er, 'bbq!
Well, yes.
So do the other n-hundred tickets in RT.
This one certainly ranks high on wtf factor though. 23:09
masak jnthn: good news is, if I find and remove the origin of the warnings, I solve the problem. :)
jnthn rakudo: my $var = 42; my $var = 5; say $var
p6eval rakudo 7347ec: Redeclaration of variable $var␤5␤
jnthn ng: my $var = 42; my $var = 5; say $var 23:10
p6eval ng c8b9b4: Redeclaration of symbol $var at line 1, near " = 5; say "␤current instr.: 'perl6;HLL;Grammar;panic' pc 519 (src/stage0/HLL-s0.pir:336)␤
masak jnthn: oh, it runs fine.
jnthn Heh. ng panics!
So I can't see if it's "already fixed" without fixing that first... :-(
masak submits rakudobug
frettled «Results 1 - 10 of about 26,500 for "masak submits rakudobug". (0.38 seconds)» 23:11
masak frettled: lol
ng_feed rakudo-ng: jnthn++ 23:12
rakudo-ng: Use the compose method on Attribute to do accessor method generation now, rather than doing it in the actions. This means that we're now implementing the Attribute meta-object API from metamodel.pod and using it.
23:12 mikehh joined
jnthn The upshot of the above is that we've much of what we need for custom accessor generation now. 23:13
frettled nice! 23:14
Good night!
moritz_ good night
jnthn o/ 23:15
moritz_ \\o
jnthn That was to frettled, not 'cus I'm going. ;-)
masak I am. 'night. 23:16
frettled: night!
sjohnson what is everyone's favourite non-ascii open and close bracket for doing giant qq/ type strings? 23:18
moritz_ «...»
Tene ⦃ ⦄ 23:27
Juerd diakopter: Not that I know of, but you could use a version number just below the expected one that Debian will use. e.g. 1.19999 if the actual number is 1.2. 23:30
23:40 xenoterracide joined
diakopter pmurias: no, I mean cloning nodes in the interpreter 23:45
or cloning any js object in the runtime 23:46
I'm just curious if your runtime does anything like that, because I have a very efficient way to do it
Juerd: moritz_ replied that there is (a way) 23:47
Juerd diakopter: I'd be very surprised. I only know of several ways to *prevent* such upgrades. 23:50
diakopter Juerd: ? wiki.debian.org/HowToSetupADebianRepository 23:55
pmurias diakopter: the code a compiled js switch is kept seperate from the data
Juerd It's not clear to me how that has anything to do with lowering priority 23:56
Good night for now
zz
pmurias diakopter: mildew-js is very very unoptimised atm 23:57
diakopter ok... have you committed lately?
or worked on it lately?
pmurias i think i last comitted yesterday 23:58
multis now work

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!