»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
00:03 lue left 00:15 lue joined
TimToady .tell lizmat These days Nil is considered the best way to indicate "missing information" 00:29
yoleaux TimToady: I'll pass your message to lizmat.
00:30 raiph joined
dalek rl6-roast-data: c0eca35 | coke++ | / (5 files):
today (automated commit)
 00:31
[Coke] rakudo.parrot failing 62 tests (2nd day running)
00:33 SmokeMachine joined
[Coke] can't get the test that's abort to abort when running it by hand 00:35
00:35 zacts joined
zacts can rakudo also implement scheme or haskell? 00:36
TimToady .tell jnthn If the two sides are semantically different (utf8 vs buf8), it should fail; if they're merely unit-sized differently (blob8 vs blob16) ops like ~ still fail but integer-based ops like ~& are allowed ot upgrade both sides to 16 bits
yoleaux TimToady: I'll pass your message to jnthn.
zacts and if so, could you combine scheme/haskell with perl6?
[Coke] zacts: it's plausible, sure. 00:37
TimToady zacts: anything is possible, but I sincerely doubt you'll find anyone willing to implement Haskell on top of Perl 6
Scheme maybe
[Coke] (note that if you want to run perl6 on haskell, there's also pugs, though it's showing its age.)
zacts ah, ok cool
[Coke] TimToady: I got the impression he was saying it the other way 'round.
zacts well, pugs is also interesting, but yeah
well, even scheme would be cool. 00:38
TimToady there are certainly other ways of running two languages together than implementing one in terms of the other
zacts TimToady: no I don't mean implementing haskell or scheme with the perl6 language 00:39
I guess I my understanding was that parrot could also implement other languages
TimToady well, rakudo is perl6, so your first question more or less asked that
zacts sorry I typed my question wrong
s/rakudo/parrot/
TimToady ah, well, that's a bird of a different color 00:40
zacts hehe punny
TimToady certainly the intent was to allow the implementation of many dynamic languages on top of parrot, but the vision exceeded the reality in a number of respects 00:41
zacts ok, interesting 00:43
TimToady there is always this danger with bottom-up design that you will analyze the problem into the wrong abstractions for your eventual users, especially when you're not sure who your users are going to be 00:53
it's a vast oversimplification to say that that is what happened with parrot, of course, but there's some truth there 00:54
from the parrort perspective, it didn't help that their users didn't really know what they wanted at the beginning either :) 00:55
00:57 benabik joined 00:58 BenGoldberg left
zacts TimToady: do you think that parrot will be successful for perl6 nonetheless? 01:00
TimToady there are many definitions of "success" 01:01
by some of them, it has already succeeded
by others, it probably never will
zacts hehe ok 01:02
TimToady unless parrot can redefine itself to be the VM of choice for at least one popular language
perigrin and define "popular" in a way that people agree with it 01:03
well a majority of people agree with it.
[Coke] Certainly it's been critical for getting rakudo to where it is today. 01:04
TimToady doesn't even have to be a majority, as long as enough employers are willing to hire
[Coke]: yes, in that respect it was a big success
zacts so why didn't pugs evolve further?
was it too slow? 01:05
was it a prototype?
TimToady no champion
well, most people don't set out to write a prototype :)
zacts ok
geekosaur pugs was pretty much a one-person project. that person got ill and nobody else has been able to decipher the code :)
TimToady but everything is a prototype for what comes after :)
geekosaur (several of us have tried...) 01:06
zacts I see
perigrin geekosaur: if it were anybody but Audrey you were talking about I'd argue that point
zacts so in general would you say that parrot is more well designed than the internals of perl5? 01:07
I've heard that perl5 source code is nutty to read..
01:07 BenGoldberg joined
TimToady oddly, rakudo on parrot has a similar problem, but instead of things being hidden in the abstractions, they're hidden in a very large codebase, with many operations that interact mysteriously with each other 01:07
zacts: s/ to read// 01:08
geekosaur shrugs
I don't claim to be a Haskell expert. I do well enough with xmonad, but that;s about 5% Haskell and 95% X11 and I'm fine with the latter :) 01:09
TimToady zacts: the perl5 source code was successful insofar as it was the preferred VM for a popular language :)
perigrin TimToady: s/ source code//; # probably fair too.
geekosaur it's not difficult to confuse me with complex Haskell code
and I seem to be one of the better Haskell folks here, which leaves us kinda stuck with respect to pugs 01:10
perigrin geekosaur: I do pretty well with Perl and Audrey can lose me quickly there.
geekosaur every so often I trawl in #haskell to see if anyone is interested in helping out, sadly no bites as yet
zacts so was the person who initially designed pugs a mad genius? 01:12
TimToady well, pugs was a brilliant hack, but haskell programmers tend not to think of themselves as hackers...
zacts: everyone here is a mad genius in their own way :) 01:13
zacts :-)
TimToady: and you are a mad genius in multiple ways timtoady.
as your nick seems to imply 01:14
TimToady yes, well, sometimes madness is its own reward...
zacts :D 01:15
perigrin "We're all mad here. I'm mad, you're mad ... "
colomon That passage is one of my little boy's favorite passages. He likes to play Luke Skywalker (him) and the Chessire Cat (me). 01:17
ack, spelling 01:18
[Coke] what is the set version of this:
r: say 16 - (1|2|3|4);
camelia rakudo 4a977d: OUTPUT«any(15, 14, 13, 12)␤»
perigrin colomon: my daughter was (recently) reading it to my son.
[Coke] r: say 16 >>-<< set<1 2 3 4>; 01:19
camelia rakudo 4a977d: OUTPUT«Lists on both side of non-dwimmy hyperop of infix:<-> are not of the same length␤left: 1 elements, right: 4 elements␤ in sub hyper at src/gen/CORE.setting:14608␤ in sub hyper at src/gen/CORE.setting:14596␤ in block at src/gen/CORE.setting:14587␤ in block at /…
[Coke] r: say 16 <<-<< set<1 2 3 4>;
camelia rakudo 4a977d: OUTPUT«16 16 16 16␤»
colomon [Coke]: there isn't really a set version of that
TimToady r: say set 16 X- 1,2,3,4; 01:20
camelia rakudo 4a977d: OUTPUT«set(15, 14, 13, 12)␤»
TimToady that's about as close as you can get
but sets don't know how to subtract members
*subtract from
n: say 16 <<-<< set<1 2 3 4>; # this may tell you more what happened here 01:22
camelia niecza v24-88-g1f87209: OUTPUT«16 16 16 16␤»
TimToady or not :) 01:23
std: say 16 <<-<< set<1 2 3 4>; # this may tell you more what happened here
camelia std c2215f0: OUTPUT«===SORRY!===␤The 'set' listop may not be called without arguments (please use () or whitespace to clarify) at /tmp/LNdRh_dm6U line 1:␤------> say 16 <<-<< set⏏<1 2 3 4>; # this may tell you more wha␤Potential difficulties:␤ Use …
TimToady r: say set<1>.WHAT
camelia rakudo 4a977d: OUTPUT«(Bool)␤»
TimToady r: say set<1> 01:24
camelia rakudo 4a977d: OUTPUT«False␤»
TimToady basically, that means set()<1>
which is why std disallows it 01:25
r: say 16 <<-<< set <1 2 3 4>;
camelia rakudo 4a977d: OUTPUT«("1" => 15, "2" => 15, "3" => 15, "4" => 15).hash␤»
[Coke] r: say (18 X- 3,5).sort.uniq.grep({*<7}) 01:26
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤Whitespace required before < operator␤at /tmp/cLE4PGHNP0:1␤------> say (18 X- 3,5).sort.uniq.grep({*<7})⏏<EOL>␤ expecting any of:␤ postfix␤»
colomon it's taking the set as a hash?
TimToady there's that darn hash subscripting again :)
colomon: seems like 01:27
[Coke] r: say (18 X- 3,5).sort.uniq.grep: {*<7}
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤Whitespace required before < operator␤at /tmp/a2We8BzwMa:1␤------> say (18 X- 3,5).sort.uniq.grep: {*<7}⏏<EOL>␤ expecting any of:␤ postfix␤»
TimToady when all else fails, read the error message
colomon TimToady: that's quite unfortunate in this case
TimToady r: say 16 <<-<< set(<1 2 3 4>).list; 01:28
camelia rakudo 4a977d: OUTPUT«15 14 13 12␤»
TimToady it seems unfortunate that hyper would choose the hash nature over the list nature in this case, yes
seems to violate the notion that hyper is based on Iterability, and that's more of a list concept than a hash concept 01:29
[Coke] rakudo.jvm's REPL gets very upset about that missing whitespace. goes into an infinite loop complaining. 01:30
TimToady r: my &fromx = 16 - *; say set(<1 2 3 4>)».from16 01:31
camelia rakudo 4a977d: OUTPUT«No such method 'from16' for invocant of type 'Bool'␤ in method dispatch:<hyper> at src/gen/CORE.setting:1119␤ in block at /tmp/4z88Jv0TUQ:1␤␤»
TimToady r: my &fromx = 16 - *; say set(<1 2 3 4>)».&from16
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤Undeclared name:␤ &from16 used at line 1. Did you mean '&fromx'?␤␤»
TimToady r: my &from16 = 16 - *; say set(<1 2 3 4>)».&from16
camelia rakudo 4a977d: OUTPUT«("1" => 15, "2" => 15, "3" => 15, "4" => 15).hash␤»
TimToady still seems buggy
colomon is hyper supposed to work on hash keys or values? 01:32
TimToady for an ordinary hash, the values 01:33
colomon until we spec a better way, that's what's happening here
TimToady r: my &from16 = 16 - *; say { :a(1), :b(2), :c(3), :d(4) }».&from16 01:34
camelia rakudo 4a977d: OUTPUT«("a" => 15, "b" => 14, "c" => 13, "d" => 12).hash␤»
TimToady like that
16 - True, yeah
colomon r: my &from16 = 16 - *; say { :a(True), :b(True), :c(True), d:(True)) }».&from16
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤No such method 'has_compile_time_value' for invocant of type 'NQPMu'␤»
colomon r: my &from16 = 16 - *; say { :a(True), :b(True), :c(True), d:(True) }».&from16 01:35
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤No such method 'has_compile_time_value' for invocant of type 'NQPMu'␤»
TimToady d:
interesting failure though
r: foo:(42)
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤No such method 'has_compile_time_value' for invocant of type 'NQPMu'␤»
colomon r: my &from16 = 16 - *; say { :a(True), :b(True), :c(True), :d(True) }».&from16 01:36
camelia rakudo 4a977d: OUTPUT«("a" => 15, "b" => 15, "c" => 15, "d" => 15).hash␤»
TimToady nr: for set(<a b c>) { .WHAT.say } 01:38
camelia niecza v24-88-g1f87209: OUTPUT«(Set)␤»
..rakudo 4a977d: OUTPUT«(Str)␤(Str)␤(Str)␤»
TimToady nr: for set(<a b c>) »~» 'x'{ .WHAT.say } 01:40
camelia rakudo 4a977d: OUTPUT«===SORRY!===␤Missing block␤at /tmp/dP2CAaSTH0:1␤------> for set(<a b c>) »~» 'x'{ .WHAT.say }⏏<EOL>␤ expecting any of:␤ method arguments␤ postfix␤ statement end␤ statement modifier␤ statemen…
..niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Missing block at /tmp/Ku0mvx1lqR line 1 (EOF):␤------> for set(<a b c>) »~» 'x'{ .WHAT.say }⏏<EOL>␤␤Parse failed␤␤»
TimToady nr: for set(<a b c>) »~» 'x' { .WHAT.say }
camelia rakudo 4a977d: OUTPUT«(Pair)␤(Pair)␤(Pair)␤»
..niecza v24-88-g1f87209: OUTPUT«(Str)␤»
TimToady nr: for set(<a b c>) X~ 'x' { .WHAT.say }
camelia niecza v24-88-g1f87209: OUTPUT«(Str)␤» 01:41
..rakudo 4a977d: OUTPUT«(Str)␤(Str)␤(Str)␤»
diakopter heh
TimToady there is a bit of...variability...
diakopter which magnifying glass are we focusing on camelia today
TimToady the one of "what does it mean to iterate a set?" 01:42
seems hypers disagree with lists on that subject, and niecza disagrees with rakudo 01:43
n: say set(<a b c>) X~ 'x' 01:44
camelia niecza v24-88-g1f87209: OUTPUT«a b cx␤»
zacts TimToady: does pugs support tail recursion of subroutine calls?
TimToady n: say set(<a b c>)[] X~ 'x'
camelia niecza v24-88-g1f87209: OUTPUT«ax bx cx␤»
TimToady zacts: well, of course, but I don't know if it does the corresponding optimization :) 01:45
01:45 raiph left
zacts yeah, I mean does it optimize? 01:45
so, probably not..
TimToady I didn't say that... 01:46
I said I dunno
zacts ok
TimToady is not enough of a solipsist to assume that if he doesn't know something, it doesn't exist...
diakopter I didn't think that existed 01:47
TimToady well, there are ways to test that...
though camelia may timeout before you can tell the difference 01:48
diakopter the universe may cycle a few googolplex times
.seen japhb 01:49
yoleaux I saw japhb 13 Jun 2013 16:24Z in #perl6: <japhb> diakopter, pong
diakopter !
TimToady oddly, modern FP thinking is moving away from defining everything in terms of recursion over a list
diakopter soon computers will have more cores than instructions any one program could ever run 01:50
TimToady recursion implies sequentiality that is not desirable in parallel computations
So languages that optimize for tail recursion are, as it were, barking down the wrong (branch of) the tree. 01:52
diakopter "do you have the latest exacore?" "no, a prototype zottacore.."
er. 01:53
zetta
zetta, yotta, who can tell the difference?
TimToady s['of) the'] = 'of the)'
yoda the diffence tell can! 01:54
*difference
or you could ask a Difference Engine
zacts TimToady: interesting 01:55
so, what programming styles lend well to parallel programming?
TimToady depends on the kind of parallelism 01:59
02:00 skids joined
TimToady pipelines, MIMD, SIMD, mapreduce... 02:00
zacts so SICP stuff, but without focusing so much on recursion?
mapreduce makes me think *SICP* for some reason
I'm going to be starting that book
02:01 raiph joined
zacts not to start a long discussion about this.. 02:01
TimToady mapreduce mostly makes me think of Google :)
grid parallelism is yet another style 02:02
hierarchical parallel relationships... 02:03
zacts interesting
TimToady parallelism where everything depends on everything else (the hardest kind)
zacts so is perl6 being designed a lot with parallel computing in mind?
TimToady certainly, for some defininition of "in mind"
==> feeds are pipelines, hypers are SIMD, race context (we hope) can do mapreducey stuff 02:04
MIMD is just basic threads
everthing depending on everything else is probably a complete message-sending crossbar of some sort 02:05
zacts like an async message passing system?
TimToady yes 02:06
that also fits better into NUMA systems
hypers are good for feeding vectors through your GPU (or your Cray:)
zacts Cray is the supercomputer right? 02:07
TimToady gather/take is good for demand-driven pipelining
yes, I'm thinking of the original Crays in terms of their built-in vector processing 02:08
really just a form of pipelining, but it did speed up SIMD calculations on vectors by more than an order of magnitude
14 times, if recall 02:09
zacts ah ok
TimToady or maybe that was just how many stages there were in the pipeline...
diakopter imagines aliens using base 900, "... by more than an order of magnitude!" "ZOMG"
zacts interesting stuff
TimToady diakopter: that's very nearly "but what's three orders of magnitude among friends?" 02:10
zacts will perl6 be easy to install even if you don't know perl? for example, I write an irc client in perl. my users don't know how to install perl modules.
TimToady zacts: that also is one of our goals 02:13
but then, we've got lots and lots of goals...
diakopter the penalty kickoff is in its eleventyeth round
zacts ok, cool. well gtg.
TimToady zacts: it may well depend on which VM you want to run your irc client on... 02:14
nice talkin'
02:49 orafu joined 02:52 benabik left
sorear o/ 03:10
TimToady \o 03:13
03:18 arlinius left 03:30 BenGoldberg left 04:02 arlinius joined 04:24 benabik joined 04:28 Psyche^_ joined 04:43 Guest65022 joined
Guest65022 hi all 04:44
eloooooo 04:45
diakopter hi
04:45 Guest65022 left
diakopter ima troll biscuit 04:45
05:04 xilo left 05:39 kaare__ joined 05:40 risou_awy is now known as risou 05:44 domidumont joined 05:48 spider-mario joined 05:52 risou is now known as risou_awy 06:04 domidumont left, domidumont joined 06:05 zby_home joined
zacts so perl6 will also be a compiled language? 06:05
if you want that? 06:06
I would love to be able to compile perl programs, and distribute binaries. mainly so perl newbies don't have to install a lot of perl stuff to run my program. 06:07
06:14 fridim__ left
sorear yeah, we're pretty big on supporting precompilation here 06:19
zacts neato 06:23
sorear admittedly it's largely out of necessity 06:33
06:34 raiph left
avuserow zacts: if you're interested in parallelism, have you taken a look at languages like Rust, Go, or Erlang? 06:49
they certainly have some interesting ideas about what the future there might be
06:49 spider-mario left
zacts avuserow: no, I'll check those out. I've never heard of rust, but I have heard of go and erlang. 06:50
but, right now I'm focusing on perl5 and scheme
i'm just curious as to what is out there.. neat
avuserow yeah rust is a newcomer that I'm interested in. By the mozilla folks. My take-away is that it's a C-style functional language that's good for systems and parallel programming 06:51
zacts interesting 06:52
06:54 spider-mario joined 07:43 domidumont left 07:46 zacts left 07:56 risou_awy is now known as risou
dalek kudo/nom: 364e2a3 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Make .infinite return Nil rather than Bool/Mu when it doesn't know

  <TimToady> .tell lizmat These days Nil is considered the best way to indicate "missing information"		 08:05
lizmat good *, #perl6! 08:07
yoleaux 00:29Z <TimToady> lizmat: These days Nil is considered the best way to indicate "missing information"
lizmat :-)
08:09 risou is now known as risou_awy 08:13 pupoque joined 08:15 rindolf joined 08:23 lue left 08:37 lue joined 08:47 berekuk_ left 08:48 berekuk joined
lizmat wonders whether "report_innevitable_dispatch_failure" is misspelt deliberately 09:01
09:04 zacts joined
jnthn lizmat: no :P 09:23
yoleaux 20 Jul 2013 23:35Z <lizmat> jnthn: it feels to me that in some cases List.infinite is being called on a Iterable object
20 Jul 2013 23:36Z <lizmat> jnthn: when judging by the code, it should call Iterablle.infinite
20 Jul 2013 23:36Z <lizmat> jnthn: could this have to do with the fact that Iterable.infinite returns Mu ?
00:36Z <TimToady> jnthn: If the two sides are semantically different (utf8 vs buf8), it should fail; if they're merely unit-sized differently (blob8 vs blob16) ops like ~ still fail but integer-based ops like ~& are allowed ot upgrade both sides to 16 bits
lizmat jnthn: o/ 09:24
jnthn: my feeling was wrong
apparently, List.infinite is called with an Array object 09:26
r: class A { has $!a=42; method a { $!a } }; class B is A { }; say B.new.a # which basically is this situation 09:27
camelia rakudo 364e2a: OUTPUT«42␤»
lizmat so I'm not sure why I'm getting "Can not get attribute '$!infinite' declared in class 'List' with this object" 09:28
when apparently it is being called with an Array object
jnthn Me either, there must be something more to it...
lizmat ah…. I think I know
jnthn Oh...wait a moment...
Are the rest of the attributes for the class not declared in BOOTSTRAP? 09:29
lizmat the attribute should be known at class compose time, so should b
indeed
jnthn Right
Well, class compose time can happen more than once
But representation compose can only ever happen once.
lizmat indeed :-) 09:31
09:32 fgomez left
lizmat new error! Nominal type check failed for parameter 'lhs'; expected Any but got Mu instead :-) 09:37
ah, I see 09:38
$!infinite shouldn't be Mu
09:47 Rotwang joined 09:52 wk joined, wk is now known as Guest61163
lizmat jnthn: would it be an idea to make gen-cat.pl smarter so that it wouldn't need "use" to load Perl6::BOOTSTRAP? 10:24
jnthn lizmat: I don't see how that could possibly work.
lizmat aka, copy some of the code currently in load_module verbatim into CORE.setting ?
jnthn That's incredibly unlikely to work 10:25
lizmat I mean, gen-cat.pl *knows* where Perl6/BOOTSTRAP lives
but if it would work, we wouldn't need "use" in CORE.setting
jnthn You don't even have *any* types or meta-objects to hand before you "use Perl6::BOOTSTRAP".
lizmat so we could write it all of "use" in Perl6 10:26
jnthn Want to write a string literal? Forget it, no Str yet. etc.
lizmat hmmm…. 10:27
jnthn As I've mentioned before, we don't have to have "use" mapped to the real/ultimate module loader at this point.
lizmat indeed
ok, scratch that idea then :-) 10:28
jnthn But trying to eliminate the use for code is probably asking for a world of pain.
I mean, Block doesn't even exist yet, so you can't write a BEGIN block! :)
lizmat :-)
point taken
jnthn I guess an alternative is to recognize the --setting=NULL case and do something magical in comp_unit to load Perl6::BOOTSTRAP. 10:29
That is, the "use" is implicit
lizmat would there be other use cases for --setting=NULL ?
jnthn Writing a completely custom setting, perhaps. But that's a pretty crazy thing to do. 10:30
dalek kudo/nom: a1e237b | (Elizabeth Mattijsen)++ | src/Perl6/Optimizer.nqp:
Scratch misspelling OCD itch
lizmat adding a --setting=BOOTSTRAP would also work, and prevent overloading =NULL usage, I guess 10:31
hmmm… that seems like a viable way of moving forward 10:32
remove the "use Perl6::BOOTSTRAP" from traits, and make it implicit with --setting=BOOTSTRAP 10:33
jnthn But Perl6::BOOTSTRAP needs to be use'd, not be a setting...
I think doing it off --NULL is sensible enough.
lizmat ok
will mull this over in the next hours :-) Thanks for the fb 10:34
10:34 Gruber is now known as Grrrr 10:42 logie left 10:54 rindolf left 11:31 logie joined 11:39 logie left 12:24 jaldhar joined 12:26 jaldhar left 12:29 jaldhar joined, census joined 12:34 dmol joined 12:41 birdwindupbird joined
dalek p: de37343 | jnthn++ | src/vm/ (2 files):
Make nqp::encode take a buffer, not create it.
 12:56
kudo/rebuf: 763cfe4 | jnthn++ | src/Perl6/Grammar.nqp:
Fix multi sort so Blob ~~ String is true.

Fixes the ambiguity issues that kept any of the Blob operators from being reached.		 12:57
kudo/rebuf: da24001 | jnthn++ | src/Perl6/Metamodel/CurriedRoleHOW.nqp:
Fix a thinko.
kudo/rebuf: 4c60347 | jnthn++ | src/core/Str.pm:
encode method should do the buffer creation.

This is 'cus we may need to pun a role to do so, which is best done here, not inside of some op.
13:04 FOAD left 13:05 FOAD joined, dmol left 13:19 colomon left 13:22 benabik left 13:23 eternaleye left, colomon joined 13:28 SamuraiJack_ joined, eternaleye joined 13:34 benabik joined
lizmat jnthn: FWIW, I've given up on adding an $!infinite attribute: the overhead seems to be more than repeatedly calling multiple methods over and over again 13:35
spectest typically used 1.5% more CPU with the patch 13:36
13:38 BenGoldberg joined
jnthn lizmat: Well, guess this is why measurement matters... :) 13:38
lizmat still no solution as to why (1..Inf).uniq.infinite doesn't know it's infinite 13:39
13:39 PacoAir left 13:40 PacoAir joined
jnthn lizmat: I think that infiniteness is generally not contaigous at the moment... 13:42
lizmat: I'm not sure about the details; Pm is the right person to know about this stuff.
13:44 ajr joined
jnthn s/know/ask/ 13:44
13:44 ajr is now known as Guest28823 13:48 benabik left 13:54 benabik joined
lizmat is stupid 13:58
"nqp::istype(nqp::atpos($!rest, 0), Iterable)" only fires if $!rest is Iterable, and Iterable.infinite is by definition Nil 13:59
I guess in the end, this is a case of NYI
14:02 Ben_Goldberg joined, rindolf joined
lizmat grrr of course, ListIter is also Iterable 14:03
14:03 zakalwe left 14:04 BenGoldberg left
Ben_Goldberg Perhaps there should be an explicit way of finding out if the length is unknown (method unknownlength, perhaps)? 14:04
14:04 zakalwe joined, zakalwe left, zakalwe joined
lizmat as opposed to infinite, you mean? 14:05
the thing is, I *do* think the information is hidden on there somewhere
the thing is that on @a=(1..Inf), @a's Iterrator is of type Range 14:06
that apparently is a simplified version of the one that is created for @a.uniq
*Iterator 14:07
14:07 cognominal joined
Ben_Goldberg A list made by gather might be infinite, or it might be finite... there only way to know from the outside is to solve the halting problem! 14:07
By letting each iterable type define it's own length_is_unknown method, you can avoid having to deal with it from the outside 14:10
Or, I suppose, just treat *every* list as infinite
lizmat come to think of it, I guess you could argue that you can never tell 14:11
Ben_Goldberg Plus, finite but very large lists, e.g., (1..1e7), should probably be treated as infinite for most purposes 14:13
lizmat r: my @a := ( 1 xx Inf ) # infinite list with a finite uniq
camelia rakudo a1e237: OUTPUT«Cannot coerce Inf or NaN to an Int␤ in method Numeric at src/gen/CORE.setting:10559␤ in sub infix:<==> at src/gen/CORE.setting:3333␤ in sub infix:<==> at src/gen/CORE.setting:3331␤ in sub infix:<xx> at src/gen/CORE.setting:6790␤ in sub infix:<xx> at src/gen/CO…
lizmat n: my @a := ( 1 xx Inf ) # infinite list with a finite uniq
camelia niecza v24-88-g1f87209: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/K_qw8G7ixj line 1:␤------> my ⏏@a := ( 1 xx Inf ) # infinite list with␤␤»
lizmat n: my @a := ( 1 xx Inf ); say @a # infinite list with a finite uniq
camelia niecza v24-88-g1f87209: OUTPUT«(timeout)»
jnthn But we don't statically know it's finite, in general :) 14:14
14:15 ggoebel2 left
Ben_Goldberg n: my @one = (1); my @inf := @one xx Inf; my @u := @inf.uniq; @u[0].say; @u[1].say 14:16
14:16 Rix left
camelia niecza v24-88-g1f87209: OUTPUT«(timeout)1␤» 14:16
Ben_Goldberg Should @u really be considered infinite? After all, looping forever (instead of returning a value) isn't the same thing as returning an empty list 14:18
Err... I meant, "Should @u really be considered finite?" 14:20
14:23 Rix joined, kaare__ left 14:24 Bzek joined, Bzek left
dalek kudo/rebuf: 3300512 | jnthn++ | src/core/IO.pm:
Correct a comment.
 14:25
kudo/rebuf: 28ebf0f | jnthn++ | src/core/IO/Socket.pm:
Update IO::Socket for updated Buf.
kudo/rebuf: 75d3e57 | jnthn++ | src/core/IO.pm:
Update IO::Handle for Buf changes.
kudo/rebuf: b1383a5 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION.
ast/rebuf: e2c0079 | jnthn++ | S (5 files):
Various test updates for Buf changes.

Rakudo has become more closely aligned with the spec in various ways, and these tests were caught out in the process.		 14:29
kudo/rebuf: 310b13d | jnthn++ | src/core/Buf.pm:
Some exception pre-decls we need.
 14:32
GlitchMr prn: while True { prompt '?' }
dalek kudo/rebuf: ba6da77 | (Solomon Foster)++ | src/core/Cool.pm:
Initialize sprintf argument handler when needed.
rakudo/rebuf: 45d4473 | (Elizabeth Mattijsen)++ | src/core/Any.pm:
rakudo/rebuf: Make {@keys}:k use slighlty less CPU
14:32 dalek left
camelia niecza v24-88-g1f87209: OUTPUT«(timeout)???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????… 14:33
..pugs: OUTPUT«*** No such subroutine: "&prompt"␤ at /tmp/NGleTXYWWW line 1, column 14-25␤»
..rakudo a1e237: OUTPUT«(timeout)????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????…
GlitchMr Excess Flood...
diakopter u killed kenny
jnthn Merge heuristic fail
14:33 census left, dalek joined, ChanServ sets mode: +v dalek
jnthn lizmat, or anybody else: If you have a moment to kick it off, I'd appreciate a run of the spectests for Rakudo on Parrot in the rebuf branch, using also the rebuf branch of the spectests. 14:34
14:36 pupoque left
jnthn (I changed IO related things so it's good to get conf from another platform/machine :)) 14:37
14:37 xilo joined
GlitchMr std: say "Process ID is $$." 14:42
camelia std c2215f0: OUTPUT«===SORRY!===␤Unsupported use of $. variable; in Perl 6 please use the filehandle's .line method at /tmp/f1tsCYfelA line 1:␤------> say "Process ID is $$.⏏"␤Parse failed␤FAILED 00:00 42m␤»
GlitchMr ok...
std: say "Process ID is ", $$, "."
camelia std c2215f0: OUTPUT«===SORRY!===␤Non-declarative sigil is missing its name at /tmp/DJphWgw7Hk line 1:␤------> say "Process ID is ", $⏏$, "."␤Use of uninitialized value $first in string eq at STD.pm line 66215.␤Use of uninitialized value $first in str…
GlitchMr I guess it cannot detect invalid use of $$ ($*PID).
std: $>; 14:44
camelia std c2215f0: OUTPUT«===SORRY!===␤Unsupported use of $> variable; in Perl 6 please use $*EUID at /tmp/PzKxvAGJTu line 1:␤------> $>⏏;␤Parse failed␤FAILED 00:00 41m␤»
14:47 FROGGS[mobile] joined
FROGGS[mobile] o/ 14:47
jnthn o/ FROGGS[mobile] 14:51
14:51 pupoque joined
jnthn
.oO( is the FROGGSmobile like the batmobile? )
 14:51
colomon better, it has irc 14:53
14:54 pupoque left
jnthn thinks he has rebuf to the point where it makes things spec-ier overall, doesn't bust spectest on Parrot, and improves spectest on JVM. 14:55
FROGGS[mobile] FROGGSmobile are the cars you've seen in Frogger 14:57
JimmyZ can't understand why utf8/16/32 and Buf have "is repr('VMArray')", since Blob has too ... 14:59
15:00 logie joined
jnthn JimmyZ: 'is repr' on a role only applies to if the role is punned. 15:04
It can't really have compositional meaning, given you can compose multiple roles... 15:05
15:06 pupoque joined 15:08 pupoque_ joined 15:09 ggoebel joined
jnthn Looks like we'll get another 50 spectests on Rakudo JVM once rebuf is merged. 15:10
15:10 pupoque left, pupoque_ left
JimmyZ nice 15:10
15:10 pupoque_ joined
jnthn And that's without .read and .write being implemented yet 15:11
Away for a bit 15:13
15:14 risou_awy is now known as risou, benabik left, benabik joined
Ulti zacts have you tried the PAR stuff in Perl5 search.cpan.org/~rschupp/PAR-1.007/...torial.pod you mentioned a lot about distribution/packaging 15:15
I am looking forward to just doing --target=jar in Rakudo though :) 15:16
15:16 raiph joined, pupoque_ left
zacts Ulti: I'll check it out thanks 15:17
15:17 pupoque_ joined 15:19 pupoque_ is now known as chpoque, eternaleye left 15:20 chpoque is now known as pupoque_, logie left, pupoque joined 15:23 eternaleye joined
flussence jnthn: Files=746, Tests=26411, 1415 wallclock secs ( 8.66 usr 0.52 sys + 5114.43 cusr 159.50 csys = 5283.11 CPU) 15:30
1 fail in S16-filehandles, 5 in S32-io/socket 15:31
(and a bunch of fails in S05-mass/ which is normal icu screwup on my machine...)
15:33 pupoque_ left
flussence oh, this site's still working! smolder.parrot.org/app/projects/rep...ails/43540 15:34
diakopter FROGGS[mobile]: lolz
15:41 pupoque_ joined 15:45 ajr joined, Guest28823 left, ajr is now known as Guest19079 15:49 domidumont joined, raiph left 15:50 raiph joined 15:52 Guest19079 is now known as ajr_ 15:53 FROGGS[mobile] left, pmurias joined
pmurias jnthn: how does $node<foo> or $capture<foo> work? 15:54
jnthn: ahh, found it, possitional_delegate 15:55
15:57 rindolf left
moritz \o 15:57
moritz back from traveling
diakopter moritz: wb
GlitchMr prn: say atan2 0, -0e0 16:01
camelia rakudo a1e237: OUTPUT«3.14159265358979␤»
..niecza v24-88-g1f87209: OUTPUT«3.1415926535897931␤»
..pugs: OUTPUT«*** No such subroutine: "&atan2"␤ at /tmp/Foo4Gi8XZp line 1, column 5 - line 2, column 1␤»
GlitchMr rn: say atan2 0, '-0e0' 16:02
camelia niecza v24-88-g1f87209: OUTPUT«3.1415926535897931␤»
..rakudo a1e237: OUTPUT«0␤»
timotimo lo moritz
16:02 pupoque_ left
GlitchMr I wonder, which implementation is wrong in this case? 16:02
timotimo what are you up to this nice and warm sunday?
16:04 colomon left, FROGGS[mobile] joined
moritz GlitchMr: isn't atan periodic in pi? 16:05
GlitchMr: if so, none of them needs to be wrong 16:06
GlitchMr I'm not talking about atan2 itself.
I'm talking about '-0e0'
moritz prn: say +'-0e0'
camelia rakudo a1e237, niecza v24-88-g1f87209, pugs: OUTPUT«0␤»
GlitchMr In niecza, atan2 sees -0.
But in Rakudo, it sees 0.
prn: say (+'-0e0').perl
camelia rakudo a1e237, niecza v24-88-g1f87209: OUTPUT«0e0␤»
..pugs: OUTPUT«0.0␤»
GlitchMr 0e0... ok
moritz doesn't see the difference 16:07
GlitchMr Yet, atan2 sees it
rn: say atan2 0, -0e0
camelia rakudo a1e237: OUTPUT«3.14159265358979␤»
..niecza v24-88-g1f87209: OUTPUT«3.1415926535897931␤»
GlitchMr rn: say atan2 0, 0e0
camelia rakudo a1e237, niecza v24-88-g1f87209: OUTPUT«0␤»
GlitchMr I'm really confused
Rakudo numerizes that to 0e0, but Niecza numerizes that to -0e0 16:09
rn: printf "%g %g %g\n", 0e0, -0e0, +'-0e0' 16:10
camelia rakudo a1e237: OUTPUT«0 -0 0␤»
..niecza v24-88-g1f87209: OUTPUT«0 0 0␤»
16:12 pupoque_ joined
dalek p: 05574c2 | (Pawel Murias)++ | t/nqp/73-delegation.t:
Add a test for 'is positional_delegate' and 'is associative_delegate'.
 16:15
GlitchMr rn: say <-> 16:19
camelia niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Malformed parameter at /tmp/cKH8ZL1FNJ line 1 (EOF):␤------> say <->⏏<EOL>␤␤Parse failed␤␤»
..rakudo a1e237: OUTPUT«===SORRY!===␤Malformed parameter␤at /tmp/hByY2Kw3We:1␤------> say <->⏏<EOL>␤ expecting any of:␤ argument list␤ prefix or term␤ prefix or meta-prefix␤ parameterized block␤»
GlitchMr std: say <-> 16:20
camelia std c2215f0: OUTPUT«ok 00:00 43m␤»
16:21 Ben_Goldberg left 16:22 kylecazar joined 16:24 BenGoldberg joined
GlitchMr std: /\(([^()]*|(?R))\)/ 16:25
camelia std c2215f0: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/R3VOcbTBs7 line 1:␤------> /\(([^(⏏)]*|(?R))\)/␤ expecting any of:␤ quantifier␤ regex atom␤ sigmaybe␤Parse failed␤FAILED 00:00 42m␤»
16:25 pupoque_ left, pupoque is now known as pupoque_ 16:26 colomon joined, risou is now known as risou_awy
colomon rn: say atan2 0, (1/-Inf) 16:26
camelia rakudo a1e237: OUTPUT«3.14159265358979␤»
..niecza v24-88-g1f87209: OUTPUT«3.1415926535897931␤»
colomon rn: say -0e0 16:27
camelia rakudo a1e237: OUTPUT«-0␤»
..niecza v24-88-g1f87209: OUTPUT«0␤»
16:27 raiph left
colomon …. okay, that's the opposite of what I was expecting. 16:27
afk # off to parents' 16:28
16:28 colomon left 16:30 pupoque joined, kaare__ joined 16:31 kylecazar left 16:34 pupoque left, pupoque joined
dalek p: 38e207a | (Pawel Murias)++ | t/nqp/73-delegation.t:
Test that positional_delegate and associative_delegate are visible through the metamodel.
 16:37
16:37 Rotwang left 16:38 fridim__ joined 16:41 rindolf joined 16:55 Rotwang joined 16:56 _ilbot left 16:57 ajr_ left 16:58 _ilbot joined
jnthn flussence: The S16 one seems unrelated; the socket one is more concerning, but passes for me, so I'm not too sure what I can do... 17:14
17:18 FROGGS[mobile] left 17:30 Rotwang left 17:33 birdwindupbird left 17:37 fgomez joined
pmurias jnthn: is positional_delegate works by setting a bit in the repr? 17:49
s/setting a bit/marking things up/
18:00 BenGoldberg left, grondilu joined 18:01 domidumont left 18:02 BenGoldberg joined 18:04 dagurval joined
diakopter moritz: even if it's periodic in pi, it requires all programmers who type atan2 to know that to interpret the result properly 18:14
.. because the domain of whatever it's fed to may be bigger
pmurias jnthn: sanity check, if an class has an attribute with 'is positional_delegate', the repr will have pos_del_slot set to something different then -1 18:16
jnthn pmurias: Stores the slot to find the positional delegate in repr data...
GlitchMr std: ().^WHAT
camelia std c2215f0: OUTPUT«ok 00:00 42m␤»
jnthn pmurias: Yes
GlitchMr rn: ().^WHAT
jnthn -1 indicates "no, there isn't one"
camelia rakudo a1e237: OUTPUT«===SORRY!===␤Cannot use .^ on a non-identifier method call␤at /tmp/rLjnMJIKdZ:1␤------> ().^WHAT⏏<EOL>␤ expecting any of:␤ method arguments␤»
..niecza v24-88-g1f87209: OUTPUT«Unhandled exception: Unable to resolve method WHAT in type ClassHOW␤ at /tmp/bBMlNIJUf0 line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4579 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4580 (module-CORE @ 576) ␤ at /h…
jnthn It's configured at REPR compose time.
GlitchMr What is Rakudo talking about?
pmurias jnthn: and the version is serialization format is >= 3 in vaguely modern nqp, so the pos_del_slot is serialized 18:17
jnthn yeah, 3 is pretty old
pmurias so we should remove the if statement now? 18:22
if statement = version check
jnthn pmurias: Yeah, if the minimum version it claims to support in serialization.c is higher 18:23
Which I think it is
GlitchMr std: &infix:<+>; &infix:<&&> 18:26
camelia std c2215f0: OUTPUT«ok 00:00 44m␤»
GlitchMr Just wondering, how can I declare an operator that works like &&?
(I assume macros could be helpful, but...)
pmurias jnthn: the minimum version seems to be 1 18:28
jnthn pmurias: Well, so long as we don't bump it past whatever is in stage0... :) 18:29
GlitchMr ok, I think I managed to do it using macro infix:<!!!>($a, $b) { quasi { {{{$a || $b}}} } } 18:30
jnthn GlitchMr: yeah, that looks right 18:32
GlitchMr Well, it's {{{$a}}} || {{{$b}}}
But seems to work
18:33 sdo joined
sdo hello 18:33
GlitchMr Hello
sdo I have a simple question
GlitchMr Just ask
sdo I went on rakudo website (official) and look to buy a TShirt. I could not find any place. Is it possible to buy a TShirt of Rakudo star? 18:35
grondilu " "Saying Java is a nice programming language because it works on all platforms is like saying anal sex is nice because it works on all genders." <- lol 18:37
grondilu hope this was not too nsfw. It was so funny I couldn't help writing it here. 18:38
18:38 cognominal left, zakalwe left, orafu left, skids left, hugme left, Exodist left, kst left, mikemol_ left, atrodo left, jercos left, diakopter left, cotto left, risou_awy left, zamolxes left
sdo why not oral sex 18:38
timotimo because in order to insult java, the author of the joke has to come up with something "icky" or "disgusting" 18:39
mst sdo: because heteronormative culture hasn't made men irrationally scared of oral sex compromising their masculinity
18:40 dalek left, dalek joined, ChanServ sets mode: +v dalek
GlitchMr sdo: rakudo.spreadshirt.de/ 18:40
(however, it's just standard Rakudo, not Rakudo Star) 18:41
(not that Rakudo Star has logo)
sdo ok 18:42
thx
18:42 rindolf left
sdo I check 18:42
18:47 zakalwe joined, jercos joined, Exodist joined, kst joined, orafu joined, atrodo joined, mikemol joined, zakalwe left, zakalwe joined, orafu left, orafu joined 18:48 zamolxes joined, hugme joined, ChanServ sets mode: +v hugme, diakopter joined, skids joined 18:49 fgomez left, risou_awy joined, sdo left 18:51 cotto joined 18:55 pupoque_ left 18:57 fgomez joined 19:10 lue left
moritz diakopter: whoever uses a function should have at least basic familiarity with it 19:12
19:12 wk joined, Ben_Goldberg joined, wk is now known as Guest71337 19:13 pjcj_ joined, jtpalmer joined 19:14 jercos_ joined, zby_home__ joined, prammer_ joined, gabriel joined 19:15 Teratogen joined
moritz diakopter: i'm not saying that it actualyl is correct, but the periodicity is something that needs to be taken into account when checking if one of them is actually wrong 19:16
19:16 ribasushi_ joined 19:18 pupoque left, Util_ joined, huf_ joined 19:19 cooper_ joined 19:20 jercos left, BenGoldberg left, grondilu left, xilo left, Guest61163 left, zby_home left, jtpalmer_ left, gabriel_1 left, Bucciarati left, SamuraiJack_ left, ribasushi_ is now known as ribasushi, Bucciarati joined, pupoque joined 19:22 Vlavv joined, xilo joined
dalek rl6-roast-data: 332e577 | coke++ | / (4 files):
today (automated commit)
 19:23
19:25 lue joined
moritz r: say 26210 - 25988 19:27
timotimo lost quite a few tests on rakudo.parrot
camelia rakudo a1e237: OUTPUT«222␤»
moritz r: say 222 / 26210
camelia rakudo a1e237: OUTPUT«0.008470␤»
timotimo actually, gained
jnthn rebuf should give us another 50ish passing on JVM, but flussence++ tested and saw a regression on the socket tests... 19:29
jnthn doesn't have the regression hee
*here
[Coke] there was a mystery failure a few days ago that cleared up
jnthn If anybody fancies having a poke at it, please do
[Coke] jnthn: want me to test a branch on the smoker?
jnthn [Coke]: Yeah; rebuf branch of Rakudo with rebuf branch of spectest repo. 19:30
[Coke]: On Parrot. 19:31
[Coke] ... good thing you said that. 19:32
what about nqp?
jnthn I updated NQP_REVISION
So no branch, just the matching one or later
(though I think that currently what's in NQP_REVISION is nqp master) 19:33
We don't have sockets working yet on JVM
19:33 kaare__ is now known as kaare_
jnthn But I'd rather not regress them on Parrot 19:33
hm 19:34
jnthn notes that while it does work for him, it seems to take quite a while to do so...
19:36 jfried_ joined 19:37 colomon joined 19:41 pjcj joined 19:42 kaare__ joined 19:43 jtpalmer_ joined, jfried left, jfried_ is now known as jfried, cooper joined 19:44 cooper left, cooper joined, vaelxon left, risou_awy_ joined 19:46 sivoais_ joined 19:47 sorear_ joined 19:48 anocelot_ joined 19:49 Ben_Goldberg left, dalek left, silug joined 19:51 stevan__ joined, wk_ joined 19:52 ribasushi_ joined, yoleaux left, yoleaux joined, ChanServ sets mode: +v yoleaux, pjcj_ left, berekuk left, cooper_ left, kst left, sivoais left, kaare_ left, silug_ left, jtpalmer left 19:53 yoleaux left, colomon left, ribasushi left, gabriel left, Guest71337 left, risou_awy left, orafu left, dagurval left, zacts left, stevan_ left, anocelot left, sorear left, eiro left, Tene left, timotimo left, ribasushi_ is now known as ribasushi
gfldex panda is endangered for me: 19:53
==> Testing File::Tools
t/01-file-find.t ...... Failed 1/10 subtests
19:54 pupoque left 19:56 timotimo joined 19:57 fgomez_ joined, nwc10_ joined, anocelot joined, kaare joined 19:58 sciurius joined, kaare is now known as Guest4583, FOAD left 19:59 pmurias_ joined, simcop2387 left, sergot left, sciurius_ left, pmurias left, breinbaas left, anocelot_ left, genehack_ left, nwc10 left, awwaiid left, fgomez left, kaare__ left, sorear_ left 20:01 nwc10_ is now known as nwc10 20:02 sivoais_ is now known as sivoais, sivoais left, sivoais joined 20:03 dalek joined, ChanServ sets mode: +v dalek, awwaiid joined, Guest4583 left, cognominal joined 20:04 sorear joined, breinbaas joined
lizmat jnthn: will run rebuf spectest now 20:04
will run while I'm getting some fresh, not too hot air whil cycling
back in ~ hour&
20:06 Guest4583 joined, Guest4583 is now known as kaare_, zacts joined
timotimo tadzik: i was thinking of rewriting the pod parsing to be more "modular"; have parameters that specify if whitespace should be preserved, if margins should be removed, if formatting codes (and which) are allowed, if nested blocks should be parsed, ... 20:06
20:07 dagurval joined 20:08 sergot joined 20:10 kaare_ left, gabriel_ joined, genehack joined
timotimo and also: parameters that decide what would terminate the piece ofpod 20:10
i'm not sure it can sensibly be made with "lots and lots of rules", given the sheer amount of combinations 20:11
20:11 eiro joined 20:12 berekuk joined, Tene joined, Tene left, Tene joined 20:13 simcop2387 joined 20:15 orafu joined
pmurias_ jnthn: I found the source of my positional_dispatch problems, I was using a old copy of nqp-mo.pm which didn't pass the info to the compose method :( 20:15
jnthn pmurias_: ah... :( 20:17
tadzik timotimo: seems worth trying. The existing code has a lots of duplicate code 20:28
20:31 vaelxon joined
gfldex tadzik: does panda bootstrap for you right now? 20:47
timotimo doesn't for me :( 20:51
problems with file::find
20:54 FOAD joined
gfldex not ok 10 - found one of two files due to X::IO::Dir 20:57
that's the offending test
timotimo tadzik: do you have an intuition on wether i could know what exact columns the table will have after a specific line was parsed? 21:01
21:01 spider-mario left 21:08 pmurias_ left
jnthn 'night, #perl6 21:16
tadzik gfldex: tring 21:18
timotimo: when you reach the separator you should have a fairly good idea, yes
timotimo what about tables without a separator? 21:19
lizmat gnight jnthn!
timotimo two or more spaces in a row don't guarantee "a cell ends here" 21:20
lizmat .tell jnthn "my Buf $a = "asdf".encode" fails with "Type check failed in assignment to '$a'; expected 'Buf' but got 'utf8'" 21:21
dalek ast/rebuf: a1e23de | (Elizabeth Mattijsen)++ | S02-names-vars/perl.t:
Fudge failing test, although the test itself looks faulty to me
 21:25
timotimo "expected bug" ?:)( 21:30
? :)
lizmat .ping 21:32
seems we lost yoleaux again 21:33
21:33 colomon joined
lizmat jnthn: the faliing IO tests in rebug branch: gist.github.com/lizmat/6050073 21:34
21:35 Guest1337 joined, Guest1337 left 21:36 Guest1337 joined
lizmat jnthn: fwiw, the same numbers happen with every test, so it does seem something systematic rather then depending on load or so 21:38
21:41 PacoAir left 21:44 yoleaux joined, ChanServ sets mode: +v yoleaux
lizmat .tell jnthn "my Buf $a = "asdf".encode" fails with "Type check failed in assignment to '$a'; expected 'Buf' but got 'utf8'" 21:48
yoleaux lizmat: I'll pass your message to jnthn.
lizmat .tell jnthn faliing IO tests in rebug branch: gist.github.com/lizmat/6050073
yoleaux lizmat: I'll pass your message to jnthn.
lizmat gnight #perl6!
colomon o\ 21:57
22:08 colomon left 22:17 zby_home__ left 22:44 colomon joined 22:50 colomon left 23:06 colomon joined 23:15 tokuhirom joined 23:24 grondilu joined
grondilu rn: constant α = 1; say α 23:25
camelia rakudo a1e237, niecza v24-88-g1f87209: OUTPUT«1␤»
grondilu for some reason I've been unable to use unicode variable names, latelly :-( 23:26
yet I know my terminal uses unicode. I mean, I failed running perl6 -e 'constant α = 1;', so I ran echo 'constant α = 1;' |xxd and I got: 23:27
0000000: 636f 6e73 7461 6e74 20ce b120 3d20 313b constant .. = 1;
0000010: 0a
which tells me that α is encoded in ceb1. Isn't that unicode? 23:28
colomon .u α
yoleaux U+03B1 GREEK SMALL LETTER ALPHA [Ll] (α)
colomon That's unicode alpha 23:29
(and I got it by cutting-n-pasting your text there
)
grondilu ok, but according to en.wikipedia.org/wiki/Alpha, CEB1 is unicode too (I confess I don't know the details about utf-8 and unicode) 23:30
why doesn't it work, anyway? 23:32
23:33 Guest1337 left 23:34 colomon left
flussence ~ $ perl6 -e 'constant α = 1; say α' 23:35
1
works for me, and the xxd is identical, so I dunno
grondilu upgrades its CPAN modules and after that will recompile nqp/parrot/rakudo. Maybe this will help. 23:49
23:58 raiph 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!