»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
timotimo i should head to bed 00:29
dalek kudo-star-daily: b11617e | coke++ | log/ (9 files):
today (automated commit)
 01:36
Herby_ Evening, everyone! 01:55
\o 01:56
o/ 02:01
grondilu I was kind of hoping that ($a..$b).max was $b even if $a < $b. Turns out it's the case. Not sure if it will always be. 02:38
m: say (2..1).max;
camelia rakudo-moar 845a95: OUTPUT«1␤»
grondilu that matches the semantics of what I was trying to do, but I'm not sure I can rely on it. 02:39
grondilu FYI I was coding a range of altitude values taken by my KSP spaceship during a descent. For instance 1000..100 meant a fall from 1000m to 100m. I meant max not as in "max value" but as "last value". 02:41
sorry I meant $a > $b earlier.
TEttinger perlbal space program 02:47
Herby_ ksp a pretty fun game? 02:51
grondilu I've discovered the KOS addon lately and I've been obsessed 02:54
with writing a script to land in a single burn at full throttle.
I'm making a draft in Perl 6. 02:55
Herby_ hmmm 02:57
i'll have to check that out
grondilu if you're curious here's what I got so far: gist.github.com/grondilu/011df8d34834a804ba18 02:59
grondilu although frankly I had already written a draft in kerboscript but launching KSP to test and debug it was quite a hassle so I translated into something equivalent in Perl 6 to do the tests. 03:05
timotimo o/ 06:42
timotimo m: say (2...1).max 06:43
camelia rakudo-moar 845a95: OUTPUT«2␤»
timotimo m: say (1...2).max
camelia rakudo-moar 845a95: OUTPUT«2␤»
timotimo grondilu: ^ ... does what you want
FROGGS[mobile] o/ 07:57
timotimo ho-hum. it's already 9am, and i still have those 2.5h of driving in front of me
FROGGS[mobile] uhh 08:00
are you still comming? 08:01
timotimo yeah
is that bad?
FROGGS[mobile] hehe, no :o)
FROGGS[mobile] but better hurry up :o) 08:04
timotimo i intend to 08:05
FROGGS[mobile] okay, I think I am a little bit early at the hackathon 08:24
moritz: you are not there yet, right? 08:25
timotimo starts at 10 officially 08:45
in the car now
nadim morning, what's the adverb to return a result from a regex substitution rather than doing the substitution in place? 08:49
Timbus nadim, can't use $str.subst ? 08:54
nadim I can! I just confused subst with in-place subst. thank you 08:56
nadim although I'd prefer an adverb to s/// to do the same thing 08:58
moritz \o from the Hackathon at GPW2016 09:02
lizmat moritz o/
lizmat is looking forward to seeing a lot of commits when she's online again 09:03
happy hacking!
commute&
moritz m: my %h = 'foo' => [1,2,3], 'bar' => [4,5,6]; say %h{*}».[1] 09:13
camelia rakudo-moar 845a95: OUTPUT«(5 2)␤»
moritz m: my %h = 'foo' => [1,2,3], 'bar' => [4,5,6]; say %h{*}»[1]
camelia rakudo-moar 845a95: OUTPUT«(5 2)␤»
moritz that's RT#64768 09:13
dalek ast: 5c4dafa | moritz++ | S32-hash/slice.t:
RT #64768: Tests

Whatever-Slicing plus hyperop array indexing		 09:21
timotimo I'm about half way to the city, will take a little break from driving though 09:53
tadzik don't irc and drive! :P 09:56
nadim tadzik++ 09:57
mst otherwise you might get into a bash.org 09:58
tadzik hah
mst best line from a debugging article ever: "Up to this point, I am on the right track. That will not last long." 10:06
ianthehenry.com/2016/3/9/lazy-io/
moritz lol 10:13
arnsholt I quite liked "If you have any significant Haskell experience, the answer is probably obvious. Stay with me! It’s going to get a lot worse before it gets any better." as well 10:16
dalek c: a0fd237 | paultcochrane++ | doc/Language/testing.pod:
Document the does-ok() test function
 10:23
FROGGS hmmm, can I create a new subroutine from a given Signature object? 10:24
RabidGravy I don't think so, but I've wanted that several times 10:27
nine You people have weird needs 10:28
llfourn you need to call stuff on $*W but you can do it :P
$*W.create_code_object($qast,'Sub',nqp::decont($sig)); # I have some code like this 10:29
FROGGS ohh, but this is QAST :/ 10:30
llfourn yeah you need to write QAST =(
I've been in situations where you want to specify a signature for some code but in the body all you want to do is take a capture and pass it on somewhere 10:32
so it would be nice to take a singuture and code block and merge them
FROGGS m: use nqp; my &r = nqp::create(Routine); nqp::bindattr(nqp::decont(&r), Code, '$!signature', :(--> int32)); say &r.signature 10:34
camelia rakudo-moar 845a95: OUTPUT«( --> int32)␤»
FROGGS I think that's good enough for me
RabidGravy yeah, I think I wanted it to be able to create a multi based on some configuration for "dispatch purposes"
moritz but how do you write a code block that references parameters without having the parameters? 10:35
llfourn moritz: you have a magic capture variable
FROGGS I just need something for NativeCall, so the routine does not even have to have a code object
moritz FROGGS: makes sense in your use case
and I guess all the other crazy use cases that need it also make sense on some level :-)
FROGGS hehe
kalkin- I have a grammar, which hangs up rakudo: gist.github.com/kalkin/f0a64ff21cfbf7310d3d 10:38
the part which hangsup is line 8 the asterisk (*) neojdomg [<element>] 10:39
s/neojdomg/behind
Or am I doing something wrong? 10:41
moritz kalkin-: you shouldn't quantify something with * or + that might match zero characters 10:41
kalkin- 1) I always thought * is zero or more. 10:42
2) + means 1 or more?
moritz: Did I misunderstood you? 10:43
RabidGravy no, that's right
llfourn kalkin-: I think what he's saying is like doing <token>* where token can match 0 characters means that it will be an infinite loop 10:44
RabidGravy but if you quantify as "zero or more" something that itself may match zero items then bad happens
moritz kalkin-: but <element> can match zero characters
kalkin-: so it does, and then it tries, at the same position, to match again 10:45
kalkin- moritz: Ahh I think I understand now. I must fix the rule element
moritz kalkin-: and it matches zero characters again
kalkin-: correct
kalkin- Thank you for the hint moritz & RabidGravy 10:46
moritz kalkin-: you're very welcome 10:47
RabidGravy it's all good fun
dalek kudo/nom: 3a050fb | moritz++ | src/core/Buf.pm:
#127642: Fix blob eqv Blob
 10:50
ast: 524d930 | moritz++ | S03-operators/buf.t:
RT#127642: Fix for Blob eqv Blob
 10:54
timotimo what button do i press to ring the right doorbell? 10:57
noris network?
moritz timotimo: you're at the wrong entry 10:58
timotimo oh, ok
which way to the right one?
moritz timotimo: you should go to the left entry
timotimo left on the inside?
moritz wait a sec, I'm coming out 10:59
timotimo i has a seat \o/ 11:06
timotimo it's annoying that we can get NaN values in the profiler :\ 11:32
azawawi hi 11:34
timotimo hello azawawi!
azawawi Got the following error message "Type check failed in assignment to $archive; expected NativeCall::Types::Pointer[LibZip::NativeCall::zip] but got NativeCall::Types::Po..."
why are we truncating error messages?
timotimo huh?
i wonder what is responsible for cutting that off
i'm not aware that's happening anywhere
moritz azawawi: from plain rakudo? 11:35
azawawi at first i thought it was atom that was truncating output and then i checked command line
2016.02
timotimo huh! 11:36
huh, indeed 11:37
X::TypeCheck does that
dalek kudo/nom: 5221023 | FROGGS++ | / (2 files):
let nativecast take signatures instead of routine targets

  timotimo++		 11:38
timotimo github.com/rakudo/rakudo/blob/nom/...n.pm#L1827
moritz ah, that's because it's .perl output, and that can be, like, very long 11:39
several pages long, in fact
dalek c: 92d34a1 | FROGGS++ | doc/Language/nativecall.pod:
state that nativecast will take signatures now
 11:40
sjn_phone moritz: hey, I'm downstairs and the doorbell doesn't seem to work
azawawi so i caught a semi-bug
:)
while being annoyed
moritz sjn_phone: you probably took the left entry; please use the right one instead
azawawi so what do i get the got ... part :) 11:40
moritz sjn_phone: or simply walk through to the data center; I'm coming out 11:41
mst moritz: /alias entry say $0: you probably took the left entry; please use the right one instead 11:42
moritz: /entry sjn_phone
since you've already had to say that twice, clearly it deserves an irssi alias
timotimo i'm surprised that was wrong. i went to the left, but not left enough
so if sjn went too far to the left ... :| 11:43
mst /alias entry say $0: YOU WENT TOO FAR TO THE LEEEEEEFT
moritz I already put a sign on the right entry pointing to the left one 11:45
and I wasn't allowed to take and publish photos of the entries, for "security" reasons
mst LOL 11:46
what about a three quarters perspective sketch done in pencil?
makes it easier to eat if you're worried they're going to take you alive
sjn is slowly turning right 11:47
timotimo :) 11:48
sjn why not just call it "the wrong entry" since the other one obviously is the right one?
timotimo what do i hack on now? :S 11:49
azawawi Expected: NativeCall::Types::Pointer[LibZip::NativeCall::zip], Got: NativeCall::Types::Pointer[LibZip::NativeCall::zip].new(55436656) # huh? 11:52
Shouldnt be the same? 11:53
Shouldnt they be the same?
timotimo can you inspect the $expected.^of (or what it's called) and the $got.^of? 11:54
it's just .of 11:55
and see if they =:= or eqv or whatever?
azawawi github.com/azawawi/perl6-libzip/bl...ve.pl6#L17
arnsholt Parametric types are a bit special. IIRC there's some extra legwork that needs to be done to get that to work properly, and it might be that NativeCall doesn't do it
RabidGravy azawawi, yeah, I took a lot of those out, it did work somewhat at some point before Christmas then stopped working again 11:56
timotimo ho-hum
azawawi :) 11:57
so basically forget about error checking
when nativecall types are involved
RabidGravy it works in some places, I think parameter type constraints is one place it doesn't 11:58
azawawi now here is the fun thing
timotimo i wonder if we ought to use parametric 6model extensions thingies 11:58
azawawi github.com/azawawi/perl6-libzip/bl...ve.pl6#L15 # working here
but failing in OO sugar github.com/azawawi/perl6-libzip/bl...ip.pm6#L12 11:59
timotimo well, a Pointer is just as good as any Pointer[foo]
arnsholt timotimo: Yeah, that's the extra legwork that needs to be done 12:00
azawawi meanwhile could we increase this to a reasonable value github.com/rakudo/rakudo/blob/nom/...n.pm#L1827
24 character length is too short 12:01
timotimo up it to ~4kb :)
azawawi 24 is a magic value also :)
timotimo: why not 8k? :) 12:02
RabidGravy: current streak (aka steak) is 45 days :) 12:03
RabidGravy I'm at 1 12:03
I've been working on a branch of a fork 12:04
azawawi i wanted to stop today but then i said to myself i need to break the 60 day barrier :)
RabidGravy but I did have a revelation when I was in the shower and we may be able to input audio data from jack after all 12:05
timotimo you are "at 1"?
azawawi RabidGravy: checkout this www.pjrc.com/store/teensy32.html :) 12:06
RabidGravy yeah they're good
funny firmware license though if you care about such things
timotimo, actually apparently "2" github.com/jonathanstowe 12:07
kalkin- say i have tokens A & B. My language grammar specifies that a valid word is a word consisting of a string of Bs, which might have up to one A on any position.
timotimo ah, streaks
i see
kalkin- how can i express it in a new token
azawawi off to enjoy a Spring sunny walk. Have fun :)
moritz m: my \a = gather { for 1..3 { take $_; say a.cache.perl } }
camelia ( no output )
timotimo kalkin-: you can match Bs until the end or an A followed only by Bs
moritz m: my \a = gather { for 1..3 { take $_; say a.cache.perl } }; 1 for a
camelia rakudo-moar 522102: OUTPUT«WARNINGS for /tmp/TlNpGFipsI:␤Useless use of constant integer 1 in sink context (use Nil instead to suppress this warning) (line 1)␤===SORRY!===␤This Seq has already been iterated, and its values consumed␤(you might solve this by adding .cache on u…»
moritz m: my \a = gather { for 1..3 { take $_; say a.cache.perl } }; say a.perl 12:08
camelia rakudo-moar 522102: OUTPUT«List_54742512␤List_54742512␤List_54742512␤(my \List_54742512 = (1, 2, 3)).Seq␤»
timotimo so it's like B+ A? B*
kalkin- no A could also be on the first position
timotimo OK, then it'd just be B* A? B*
kalkin- valid words: ABBB, BBA BBBAB, BBBA, BBB... 12:09
RabidGravy right off out to get vegetables
kalkin- timotimo: hmm it's simpler than i thought, thanks
kalkin- timotimo: no, it would be B*A*B*, but this also would match empty tokens 12:10
[B*A*B*]{1,} ? 12:11
timotimo wait, i think i misunderstood your idea 12:12
timotimo it's either only Bs or it begins with an A and then only Bs? 12:12
i thought you meant Bs until the end or until an A and then Bs until the end
kalkin- I'm trying to parse haml. In haml an html element might have one id (A) and as many classes as you like (B) which are concated in form like: #id.foo .foo#id.bar and so on 12:13
i have to match .foo, #bar, #bar.foo, .buz#bar.foo, .buz.foo#bar and so on.. 12:14
and of course .foo.buz
timotimo ah 12:16
well, in that case it's like i described. any amount of classes is fine. any amount of classes followed by an id is fine. any amount of classes followed by an id and then any amount of classes is also fine
you can write it as an alteration of those cases if you like, or collapse two fo those together 12:17
kalkin- you mean like [B*A?B*] | B+ ? 12:18
no more like [B*A?B*] | B+ | A
timotimo in the firstmost version you'll want to have A instead of A? 12:22
and then the last alternative is superfluous
kalkin- i got it 12:23
my token css-id-combination { <css-class>* <id>+ <css-class>* | <id> | <css-class>+ } 12:24
timotimo no need for the one in the middle, and the + after <id> is wrong
kalkin- oh you are right 12:25
timotimo: thank you
timotimo you're welcome :)
dalek c: be1f0cc | (Siavash Askari Nasr)++ | doc/Language/testing.pod:
Add missing commas for can-ok
 12:33
moritz m: grammar A { token TOP { <so> }; token so { 'foo' | 'bar' } }; say A.parse('foo') 12:40
camelia rakudo-moar 522102: OUTPUT«Too many positionals passed; expected 1 argument but got 2␤ in regex so at /tmp/xV8XBdhiFd line 1␤ in regex TOP at /tmp/xV8XBdhiFd line 1␤ in block <unit> at /tmp/xV8XBdhiFd line 1␤␤»
AlexDaniel m: my $x = ‘abc’; say S/b/z/ with $x; say $x 12:45
camelia rakudo-moar 522102: OUTPUT«azc␤abc␤»
AlexDaniel nadim: ↑
perhaps it is time to make a list of underrated perl 6 features 12:48
timotimo :) 12:50
AlexDaniel m: say exp 2, 8 12:53
camelia rakudo-moar 522102: OUTPUT«64␤»
AlexDaniel m: say exp 2: 8 # ;)
camelia rakudo-moar 522102: OUTPUT«64␤»
moritz m: my $c; my $name; BEGIN { $c = { say "OH HAI $name" } }; $name = "masak"; $c() 12:56
camelia rakudo-moar 522102: OUTPUT«Use of uninitialized value $name of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block at /tmp/qxYfuY_Rl0 line 1␤OH HAI ␤»
moritz m: my uint8 $x = 255; print $x, " -> "; $x = $x + 1; say $x; 13:20
camelia rakudo-moar 522102: OUTPUT«255 -> 0␤»
moritz m: my uint8 $x = 255; print $x, " -> "; $x++; say $x; 13:21
camelia rakudo-moar 522102: OUTPUT«255 -> 256␤»
moritz (RT#127144)
moritz www.users.globalnet.co.uk/~loxias/h...potter.htm # for the hackathon participants 13:26
m: say &die ~~ Callable 13:29
camelia rakudo-moar 522102: OUTPUT«True␤»
moritz m: say so 'xyzabcde' !~~ /[ ab | abc ]: de/ 13:32
camelia rakudo-moar 522102: OUTPUT«False␤»
moritz m: say so 'xyzabcde' ~~ /[ ab | abc ]: de/
camelia rakudo-moar 522102: OUTPUT«True␤»
moritz this looks correct to me 13:33
rt.perl.org/Ticket/Display.html?id=121307 complains about it
m: say so 'xyzabcde' ~~ /[ ab || abc ]: de/
camelia rakudo-moar 522102: OUTPUT«True␤»
moritz dunno if that is correct though
moritz m: say uniname("\x[80]") 13:38
camelia rakudo-moar 522102: OUTPUT«␤» 13:39
dalek kudo/distribution: df9bf6b | FROGGS++ | src/ (4 files):
implement and expose $?DISTRIBUTION
 13:41
[Tux] test 22.221 13:49
test-t 14.360
csv-parser 51.160
OOPS?
jnthn [Tux]: Hm, that's a good bit slower than yesterday? 13:50
jnthn wonders if it'd not be worth setting something up to run the thing that gets those numbers on every commit so we can easily figure out what caused such slowdowns :) 13:54
timotimo it's quite a bit slower :\ 13:55
llfourn that would be pretty cool. Each commit on the X axis
(of the charts tux.nl/Talks/CSV6/speed4.html ) 13:56
[Tux] all timing is available in the git repo
I just run make tt
llfourn [Tux]: so you do it manually? are the charts done manually too? 13:57
[Tux] no, that is done semi-automatically
the moment I run 'make tt' it writes the results into a log file after which the graphs are generated 13:58
llfourn ah ok. using the github API to subscribe to commits would be neat.
where would one put a server to collect build stats if one were so inclined? 13:59
I guess making it part of travis wouldn't work out because the performance of the machine may vary 14:00
jnthn If you want a machine-independent-er measure (but it'll be slow) then recording instructions retired using callgrind or something would be one way 14:01
Though it doesn't account for all slowdowns
llfourn having travis always report some callgrind stats to irc might be worth it. But might get spammy. 14:04
jnthn Yeah 14:05
llfourn wasn't moritz++ saying something about he has infinite VMs from work or something :)? 14:06
RabidGravy if the general trend is faster then finding what made that 25% slower might be a big win
moritz llfourn: not infinite, but elastic :-) 14:07
llfourn: ... and only for the next half year
AlexDaniel moritz: that's good enough
llfourn moritz: ^ we need infinite and forever 14:08
I agree. What [Tux]++ is already doing is awesome. But having a bunch of speed tests that are auto-run for every moar,nqp,rakudo commmit would be pretty valuable. And a nice way to display them. 14:11
RabidGravy I really should do "nuke rakudo and test all my modules" more often, already found one unstated dependency 14:12
moritz m: sub a($x is rw) { $x; }; { a 42 }() 14:21
camelia rakudo-moar 522102: OUTPUT«Parameter '$x' expected a writable container, but got Int value␤ in sub a at /tmp/vL_bQs_DZb line 1␤ in block <unit> at /tmp/vL_bQs_DZb line 1␤␤»
BenGoldberg m: sub a($x is raw) { $x; } { a 42 }() 14:23
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/nXFKZOCSsG␤Unexpected block in infix position (missing statement control word before the expression?)␤at /tmp/nXFKZOCSsG:1␤------> 3sub a($x is raw) { $x; }7⏏5 { a 42 }()␤ expecting any of:␤ …»
BenGoldberg m: sub a($x is raw) { $x; } { a 42 }
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1kKgw30044␤Unexpected block in infix position (missing statement control word before the expression?)␤at /tmp/1kKgw30044:1␤------> 3sub a($x is raw) { $x; }7⏏5 { a 42 }␤ expecting any of:␤ …»
BenGoldberg m: sub a($x is raw) { $x; }; { a 42 }
camelia ( no output )
BenGoldberg m: sub a($x is raw) { $x; }; { a 42 }()
camelia ( no output )
rindolf Hi all. With latest rakudobrew, this self-contained program gives 594452426380733382 instead of 605857431263981935 (which is what perl 5 gives and is the correct solution ). This is Euler #288 - paste.debian.net/414557/
can anyone verify? 14:27
llfourn Solution == 594452426380733382 # :) 14:28
moritz rindolf: might be easier if you supplied the p5 version as well 14:30
BenGoldberg m: gist.github.com/BenGoldberg1/41fdc...aab9ea8847
camelia rakudo-moar 522102: OUTPUT«Solution == 594452426380733382␤»
rindolf moritz: see github.com/shlomif/project-euler/b...-288-v1.pl 14:31
moritz rindolf: Can't locate Euler288.pm in @INC 14:34
rindolf moritz: it's in the same directory.
moritz: wait a sec - i can generate a perl 5 version as well. 14:35
moritz rindolf: don't bother, I have it now 14:36
dalek p: 3106f89 | jnthn++ | src/ (4 files):
Prepare for supporting different kinds of profile.
rindolf moritz: here you go - paste.debian.net/414563/ 14:39
moritz: what do you get when you run the p6 program? 14:41
moritz rindolf: I'm pretty confident that the problem is in your use of [+] (somelist), $another_value 14:42
rindolf: which coerces (somelist) into numerical context, using the number of elements
rindolf: when I fix that, I get 605857431263981935 14:44
[+] flat(...) instead of [+] (...)
rindolf moritz: ah.
jnthn Hmm...Rakudo HEAD is busted for me? 14:52
C:\consulting\MoarVM\install\bin\moar --libpath="C:\consulting\MoarVM\install\share\nqp\lib" perl6.moarvm --target=mbc --ll-exception --output=RESTRICTED.setting.moarvm src/RESTRICTED.setting 14:53
Cannot invoke this object
at gen/moar/m-CORE.setting:26470 (./CORE.setting.moarvm:val:150)
moritz hm, all fine for me on last recompile 14:54
RabidGravy I rebuilt at around 11ish and have re-tested all my modules
jnthn The val call in question is: %ENV{$key} = val(nqp::p6box_s(nqp::iterval($envelem)));
FROGGS[mobile] m: class Foo { has &.bar:(--> int32); } # jnthn: that would be the desired syntax for function pointer members, right?
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/BubQqaAJRt␤You can't adverb has &.bar␤at /tmp/BubQqaAJRt:1␤------> 3class Foo { has &.bar:(--> int32)7⏏5; } # jnthn: that would be the desired s␤»
moritz (though not on MOAR HEAD, just on the recommended revision)
FROGGS[mobile] jnthn: ups, I added that 14:55
jnthn FROGGS[mobile]: I'm guessing val is fragile in some way that something in my environment hits
FROGGS[mobile] hmmm, let's fix that
FROGGS[mobile] sadly I've not enough battery power to power up my windows vm 14:56
jnthn Well, if I can get the value, it may be reproducable platform-independent 14:57
moritz just print $key before this line? 14:59
or nqp::p6box_s(nqp::iterval($envelem)) 15:00
jnthn m: say val('10.00.30319.01')
camelia rakudo-moar 522102: OUTPUT«10.00.30319.01␤»
jnthn hm, that was the last one 15:01
jnthn m: say val('') 15:02
camelia rakudo-moar 522102: OUTPUT«␤»
jnthn m: say val(Str)
camelia rakudo-moar 522102: OUTPUT«Value of type Str uselessly passed to val() in block <unit> at /tmp/33TESslLwY line 1␤(Str)␤»
jnthn hmm
jnthn OK, got the key, it's whatever is in MAKEFLAGS that seems to upset it 15:06
jnthn Which is only defined when under make. Hm 15:08
FROGGS[mobile] maybe it tries to use a type that is not available that early? 15:09
jnthn But it's not in compilation, it's at setting loading time 15:10
It happens when compiling restricted setting
FROGGS[mobile] but yeah, dumping the value as moritz said might give us a better hint 15:11
jnthn m: val(' ') 15:11
camelia rakudo-moar 522102: OUTPUT«Cannot invoke this object␤ in block <unit> at /tmp/f_yrJyKsrU line 1␤␤»
jnthn There you go :)
m: val(' ') 15:12
camelia rakudo-moar 522102: OUTPUT«Cannot invoke this object␤ in block <unit> at /tmp/1rKKnLb4h1 line 1␤␤»
jnthn It's actually 18 empty spaces
FROGGS[mobile] space?
jnthn Yes!
FROGGS[mobile] -.-
jnthn I was darn confused, though it was the empty string...but spaces
RabidGravy computers are evil like that 15:13
nadim For those interested in the Dumper I have been posting about (and asking a lot for help, with great answers), I added the ability to diff two structures. it's just a little beta so far imgur.com/xQZkgJx 15:17
FROGGS[mobile] that looks nice! 15:19
nadim++
nadim the diff-glyph column will probably be removed or optional
llfourn nadim: nice :)
FROGGS[mobile] bbl & 15:20
timotimo cool 15:20
nadim I am open to suggestions. right now everything that is in the dumper is still there, filtering, custom type handles, ... so, normally, it should be easy to take a large dump, filter it out and diff with another output 15:21
option I am thinking about are: custom diff glyph, mine are helpful but make it harder to read, display diff onlly, custom matcher for containers (allow second entry in one to match fifth in the other), and since DDT is also a Data::Visitor I think it will be easy to say "here are two structs, callme back for the elements that are different". 15:24
moritz m: my $promise = start { fail 42 }; await $promise; say $promise.status 15:25
camelia rakudo-moar 522102: OUTPUT«42␤ in block <unit> at /tmp/CggkgQIwd7 line 1␤␤»
moritz m: my $promise = start { fail 42 }; await $promise; say 'alive' 15:25
camelia rakudo-moar 522102: OUTPUT«42␤ in block <unit> at /tmp/ArNDKZ38hL line 1␤␤»
moritz m: my $promise = start { die 42 }; await $promise; say 'alive'
camelia rakudo-moar 522102: OUTPUT«42␤ in block <unit> at /tmp/12MUZwUahS line 1␤␤»
jnthn await rethrows exceptions 15:26
moritz ok, so that's "worked as designed"
jnthn aye
moritz m: my $promise = start { fail 42 }; say (await $promise).^name
camelia rakudo-moar 522102: OUTPUT«42␤ in block <unit> at /tmp/4g6jFVZp_t line 1␤␤»
moritz that makes rt.perl.org/Ticket/Display.html?id=123204 closable 15:26
jnthn The fail turns into a die
Because there's no handler available
moritz m: my $promise = start { fail 42 }; say ($promise.result).^name 15:27
camelia rakudo-moar 522102: OUTPUT«42␤ in block <unit> at /tmp/cvlmPMwZ5o line 1␤␤»
moritz that OTOH surprises me
moritz ah well 15:27
jnthn Arguably we could make start blocks handle failures, but then we'd also be making them ahndle return
moritz maybe it's for the best
jnthn That may or may not be clever :)
moritz m: my $p = start { return 42 }; say $p.result 15:27
camelia rakudo-moar 522102: OUTPUT«Attempt to return outside of any Routine␤ in block <unit> at /tmp/h_oYypC6Y_ line 1␤␤»
jnthn It's one of those ones where I'd want to know what TimToady thinks 'cus I can't do a good enough simulation of what he'd think :) 15:28
moritz m: sub f { my $p = start { return 42 }; $p.result }; say f() 15:28
camelia rakudo-moar 522102: OUTPUT«Internal error: Unwound entire stack and missed handler␤»
moritz jnthn: uhm, better to make it handle returns than throwing internal errors :-) 15:28
jnthn moritz: Well, or to give a good "can't use return here" but yeah, it's tempting to make it DWIM 15:29
jnthn I'm fine with us saying that a start block's block is routine-y 15:29
jnthn Though it brings up all the fun questions like "should be define &?ROUTINE in it", "so then should it being of type Routine", etc. :) 15:30
arnsholt jnthn: For a custom parametric type (like Pointer and friends in NativeCall), is it enough to do a custom nqp::setparameterizer() on them, or are there a few more hoops at the Perl 6 level? 15:33
jnthn arnsholt: See things like CoercionHOW and shaped arrays for examples, but it's basically that, yeah 15:34
arnsholt: Though I'd prefer it if you were to add to src/core/Metamodel/Primitives.pm support for parametric stuff 15:35
And then use that from NativeCall
dalek ast: f87e7b4 | moritz++ | S17-promise/start.t:
Tests for RT #123204
 15:36
skids m: (0..^4000000).first(:end).say; say now - BEGIN now; # mini-optimization LHF. 15:36
camelia rakudo-moar 522102: OUTPUT«3999999␤1.37532110␤»
moritz m: say 09 15:39
camelia rakudo-moar 522102: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6.␤ Please use 0o9 if you mean that.␤ at /tmp/e8qzgM9ho6:1␤ ------> 3say 097⏏5<EOL>␤9␤»
skids That one was a funny ticket. 15:40
moritz and it's still open, because we don't have tests for it yet 15:40
moritz m: my ($x0, \x1) = (1, 2); say $x0; say x1 15:41
camelia rakudo-moar 522102: OUTPUT«1␤2␤»
moritz m: my (\x1) = 1; say x1 15:43
camelia rakudo-moar 522102: OUTPUT«1␤»
moritz m: sub postfix:<^^^>($a) is looser(&infix:<+>) { $a; 77 }; say +6^^^ +8^^^ 15:46
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/XxDlILWbKU␤Variable '&prefix:<^^^>' is not declared␤at /tmp/XxDlILWbKU:1␤------> 3is looser(&infix:<+>) { $a; 77 }; say +67⏏5^^^ +8^^^␤»
jnthn o.O 15:48
moritz that's already in RT
kid51 I have a question re use of ':exists' adverb on hashes
suppose I have %capitals
moritz with a patch by FROGGS++, and a comment that he doesn't know if it's the right fix
kid51 Missouri => Jefferson City, New York => Albany, Slovakia => Bratislava, UK => London
I tried: say %capitals{'New York'}:exists 15:50
Got: (False False)
What am I doing wrong? (Am working thru www.jnthn.net/papers/2015-spw-perl6-course.pdf)
skids m: my %c = Missouri => Jefferson City, New York => Albany, Slovakia => Bratislava, UK => London; say %c{'New York'}:exists
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/E76vpblgxd␤Undeclared names:␤ Albany used at line 1␤ Bratislava used at line 1␤ City used at line 1␤ Jefferson used at line 1␤ London used at line 1␤ New used at line 1␤␤»
moritz need quotes 15:51
skids m: my %c = Missouri => Jefferson City, "New York" => Albany, Slovakia => Bratislava, UK => London; say %c{'New York'}:exists
camelia rakudo-moar 522102: OUTPUT«5===SORRY!5=== Error while compiling /tmp/FRYNsrcNPS␤Undeclared names:␤ Albany used at line 1␤ Bratislava used at line 1␤ City used at line 1␤ Jefferson used at line 1␤ London used at line 1␤␤»
jnthn And on the values :)
llfourn mo quotes
skids oh right. all of the values
m: my %c = Missouri => "Jefferson City", "New York" => "Albany", Slovakia => "Bratislava", UK => "London"; say %c{'New York'}:exists 15:52
camelia rakudo-moar 522102: OUTPUT«True␤»
jnthn kid51: Did you do %capitals<New York>:exists by any chance, instead of %capitals{'New York'}:exists?
llfourn my %c = Missouri => "Jefferson City", "New York" => "Albany", Slovakia => "Bratislava", UK => "London"; say %c{'New York'}:exists 15:53
jnthn kid51: If so, <...> is quote words
llfourn m: my %c = Missouri => "Jefferson City", "New York" => "Albany", Slovakia => "Bratislava", UK => "London"; say %c<New York>:exists # oops
camelia rakudo-moar 522102: OUTPUT«(False False)␤»
jnthn m: say <New York>.perl
camelia rakudo-moar 522102: OUTPUT«("New", "York")␤»
kid51 say %capitals{'New York'}:exists works (True) 15:53
say %capitals{<New York>}:exists does not work (False False) 15:54
jnthn Right, 'cus that's actually doing a slice
arnsholt That's because <New York> is a list of two elements
Juerd <New York> is 'New', 'York'
skids It works, it just does not mean what you think it does.
jnthn It's like...is it qw(...) in Perl 5?
arnsholt Yeah, qw// 15:55
kid51 So, is there any circumstance where I would use '<' and '>' in %capitals...:exists ? 15:56
skids When you are working with a has where you know the keys don't have spaces in them.
llfourn kid51: when your data doesn't have spaces in the keys :)
kid51 k. thx 15:57
moritz well, <...> is *only* for literals
and you always know if a literal contains a space or not. You see it right there
%hash<$key> is like %hash{'$key'}, not like %hash{$key} 15:58
dalek p: 0ccafe5 | timotimo++ | / (2 files):
bump MOAR and map "setdebugtypename"
skids moritz: right.
dalek kudo/nom: e00046b | timotimo++ | / (2 files):
bump nqp and use setdebugtypename in ClassHOW.
skids moritz: as long as you haven't found some contorted reason to EVAL. 15:59
arcetera I'm trying to build Rakudo 2016.01 but it errors out 16:01
sprunge.us/QVTX
^ error log
rakudo pkgfile: sprunge.us/QBdi
nqp pkgfile: sprunge.us/cbfb 16:02
moarvm pkgfile: sprunge.us/AhDg
timotimo that looks really strange
what's causing the build to think it failed? 16:03
moritz cd rakudo && make install
== Installing modules for MoarVM
is that a star installation?
arcetera yes
moritz ah
timotimo is this running in parallel or something?
moritz and are you running that with a make that parallelizes build steps?
arcetera should I try building with `make -j1` 16:04
as my MAKEFLAGS are set to -j5
moritz unset that variable please
arcetera yeah lemme try that
moritz: just using -j1 in the pkgfile should do it 16:05
moritz R* is known not to work well with parallel builds
arcetera ah
moritz arcetera: even for recursively called makes?
arcetera yes
timotimo we're leaving the venue now-ish
arcetera yeppers looks like that did it 16:07
or not
hold on
as MAKEFLAGS is just a var set in /etc/pkgmk.conf, and pkgmk is a shell script 16:09
tryna do a hacky workaround to do MAKEFLAGS="" in the Pkgfile
arcetera wtf 16:10
lemme sprunge this
arcetera oh dear 16:12
arcetera sprunge.us/dZQE error log 16:13
arcetera new pkgfile sprunge.us/QeJK 16:13
BenGoldberg .seen zoffix 16:17
yoleaux I saw Zoffix 26 Jan 2016 10:53 EST in #perl6: <Zoffix> m: class Bar does Associative { has $.foo = 42; method AT-KEY($key) { self."$key"() } }; Bar.new<foo>.say
FROGGS jnthn: from reading the logs I dont see if you attempt to fix val(' ')... do you? 16:18
llfourn .seen ZoffixWin 16:19
yoleaux I saw ZoffixWin 20 Feb 2016 00:42Z in #perl6: <ZoffixWin> \o
jnthn FROGGS: No, didn't 16:20
FROGGS: Working on some other stuff
FROGGS jnthn: k, that's what I wanted to hear :o) 16:20
arcetera any ideas?
llfourn arcetera: where does this line come from: cd modules/panda && /home/arc/src/crux/rakudo/work/pkg/usr/bin/perl6-m bootstrap.pl 16:26
BenGoldberg .tell Zoffix In your design notes for perl6-IRC-Client-Plugin-Factoid, you say that the plugin must handle "addressed, notice, and private message requests and respond using the same method" ... if one follows the IRC RFCs, all bots should *never* respond to a "notice" message with another message, and bots *should* use "notice" to reply to messages they got via "privmsg". 16:27
yoleaux BenGoldberg: I'll pass your message to Zoffix.
FROGGS jnthn: got a patch... was a stupid error :o) 16:30
arcetera llfourn: no ide
a
geekosaur BenGoldberg, I've seen exactly one irc bot that followed that rule re privmsg, and people complained until it was "fixed".
(most bots I'm aware of just ignore notice entirely...)
rudi_s Are there any code coverage tools available for perl6? 16:31
jnthn rudi_s: Haven't seen any...in theory someone could build one using the instrumented profiling data 16:32
rudi_s: Though it'd only have block accuracy 16:33
llfourn arcetera: It's a little perplexing but it seems home/arc/src/crux/rakudo/work/pkg/usr/bin/perl6-m thinks it's .moarvm is in /usr/share where as it's meant to be in the build dir. 16:34
rudi_s jnthn: Ok, thanks. 16:35
arcetera i do have moar installed
that's a seperate port
llfourn arcetera: right but perl6.moarvm is perl6 compiled as moarvm bytecode which is what it's getting confused about :S 16:36
dalek kudo/nom: 7358690 | FROGGS++ | src/core/allomorphs.pm:
fix val(" ")
arcetera so what do i do
is there anything i can do
for that matter
llfourn I'm not that familiar with the build process unfortunetly 16:37
BenGoldberg geekosaur, Don't nickserv and chanserv use notice messages? 17:02
Skarsnik They are server bots 17:06
awwaiid in the operators doc it has "X Sequencer <==, ==>, <<==, ==>>" . Elsewhere these are called "Feed Operators". Is there a preference? I'm tempted to remove the reference to "Sequencer" in the docs and consistently use "Feed" 17:16
jnthn awwaiid: Those operators are called feed operators, the precedence level they are at is called sequencer precedence, which is probably where that naming came from. 17:18
awwaiid: So unless the context is precedence levels, I'd stick with feed
awwaiid It is both contexts, so that helps, thanks 17:30
ugexe is there a concise way to do `my @arr = 1,2,3; while @arr.pop { once { @arr.push(1) }; say $_ } # 1 2 3` where only unique values are pop'd from @arr? 18:00
BenGoldberg What do you mean unique? 18:01
llfourn ugexe: make it a .Set? 18:02
ugexe as if the first line was `state %again; next unless %again{$_}++`
llfourn m: my $a = <1 2 3 1 1>.Set; say $a.perl
camelia rakudo-moar 735869: OUTPUT«set(IntStr.new(1, "1"),IntStr.new(2, "2"),IntStr.new(3, "3"))␤»
ugexe im not sure how to use that in a loop that will add values back into the set 18:03
BenGoldberg m: my %a := set <1 2 3 1 1>; say %a.keys 18:04
camelia rakudo-moar 735869: OUTPUT«(1 2 3)␤»
llfourn hrm
BenGoldberg m: my %a := set <1 2 3 1 1>; say %a.kv
camelia rakudo-moar 735869: OUTPUT«(1 True 2 True 3 True)␤»
llfourn there's .unique 18:05
maybe it's lazy?
m: my @a = <1 2 3>; for @a.unique { @a.push(4) when 2; .say } 18:06
camelia rakudo-moar 735869: OUTPUT«1␤2␤3␤4␤»
llfourn wins the game
lazy-iterators++
ugexe but i need a while and a mutable array
i think
BenGoldberg What exactly are you doing, anyway? 18:07
Making a web spider?
ugexe trying to avoid recursion
llfourn m: my @a = <1 2 3>; my $iter = @a.unique while $iter.pop { @a.push(4) when 2; .say }
camelia rakudo-moar 735869: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MfUUZ91Dgd␤Unexpected block in infix position (missing statement control word before the expression?)␤at /tmp/MfUUZ91Dgd:1␤------> 033>; my $iter = @a.unique while $iter.pop7⏏5 { @a.push(4) when 2; .…»
BenGoldberg What would be the recursive solution, then we'll help you with a non-recursive one.
ugexe building dependency chains
llfourn m: my @a = <1 2 3>; my $iter = @a.unique; while $iter.pop -> $_ { @a.push(4) when 2; .say } 18:08
camelia rakudo-moar 735869: OUTPUT«Method 'pop' not found for invocant of class 'Seq'␤ in block <unit> at /tmp/DZ5kga5Htl line 1␤␤»
llfourn how do you get the next thing from a Seq?
m: my @a = <1 2 3>; my $iter = @a.unique; while $iter.pull-one -> $_ { @a.push(4) when 2; .say }
camelia rakudo-moar 735869: OUTPUT«Method 'pull-one' not found for invocant of class 'Seq'␤ in block <unit> at /tmp/uQkPlpk9hn line 1␤␤»
ugexe github.com/ugexe/zef/blob/master/l...t.pm6#L112 this is what im trying to replace 18:09
llfourn m: my @a = <1 2 3>; my $iter = @a.unique.iterator; while $iter.pull-one -> $_ { @a.push(4) when 2; .say } 18:10
camelia rakudo-moar 735869: OUTPUT«1␤2␤3␤4␤X::Multi::NoMatch exception produced no message␤ in block <unit> at /tmp/yurYQXBvdq line 1␤␤»
llfourn I don't really get the while loop requirement. You can add more stuff on with for loops as shown above :S 18:11
I'm not sure how 18:12
to iterate a Seq in a while loop
ugexe im not sure if i can use a for loop or not, i know i played with the idea at one time
llfourn the @wants.append thing will work with a for loop 18:13
llfourn it's one of my favourite p6 features 18:13
maybe splice is not lazy?
what is splice with no args anyway 18:14
ugexe .splice is not lazy no
llfourn m: <one two three>.splice.say
camelia rakudo-moar 735869: OUTPUT«Method 'splice' not found for invocant of class 'List'␤ in block <unit> at /tmp/IY1OEbPs7M line 1␤␤»
ugexe it used to be .splice(0) which *is* lazy
but thats also broken
llfourn m: my @a = <one two three>; @a.splice.say
camelia rakudo-moar 735869: OUTPUT«[one two three]␤»
ugexe m: my @a = 1,2,3,4; @a[2]:delete; say @a.splice
camelia rakudo-moar 735869: OUTPUT«[1 2 (Any) 4]␤»
ugexe m: my @a = 1,2,3,4; @a[2]:delete; say @a.splice(0)
camelia rakudo-moar 735869: OUTPUT«[1 2]␤»
BenGoldberg splice with no arguments causes an array to become empty, and returns what previously had been in it. 18:15
llfourn ah I see
ugexe broken further down to
ugexe m: my @a = 1,2,3,4; @a[2]:delete; say @a[0..*] 18:15
camelia rakudo-moar 735869: OUTPUT«(1 2)␤»
ugexe m: my @a = 1,2,3,4; @a[2]:delete; say @a[lazy 0..3]
camelia rakudo-moar 735869: OUTPUT«(1 2)␤»
llfourn yeah I rkn try changing it to a for loop should work imo :)
ugexe m: my @a = 1,2,3,4; @a[2]:delete; say @a[3] 18:16
camelia rakudo-moar 735869: OUTPUT«4␤»
BenGoldberg It could have been written as: while( @wants ) { @wanted = @wants; @wants = (); @wanted .= grep(*.defined) or last; ...
ugexe oh i remember why i used a while @wants.splice. i want all the elements in @wants that i can get 18:17
so i can search for them all at once, instead of searching for each one individually
batching i guess
llfourn oic
BenGoldberg Well, there's that @wants.append inside, too.
ugexe yeah. all of those can be fetched at once in the next iteration 18:18
er, s/fetched/processed/
BenGoldberg So inside that loop, I see, my @todo = @wanted.grep(...).grep(...).unique; Wouldn't that be more efficient as @todo = @wanted.unique.grep(...).grep(...) ? 18:22
ugexe yep. the entire method is a mess 18:25
BenGoldberg It 18:25
It's too long ... it makes my brain hurt ;)
ugexe agree. it handles finding all the dependencies recursively itself, so im trying to split that functionality out into a different area 18:27
it ended up that way because searching p6c ecosystem is different than searching metacpan (one has the entire package index, the other needs to do a web search each time a module needs to be looked up and each time for each dependency it must find) 18:28
ugexe (the web requests is why i want to make sure the values are unique *before* processing. otherwise it would make web requests to search for identities it already processed) 18:31
ugexe i suppose i could use a Supply to emit only unique values 18:33
sortiz \o #oerl6 18:53
RabidGravy erp 18:55
sjn o/ 19:02
sortiz For Perl6 objects associated to native allocated resources, I've thinking in the Disposable pattern, a la C#. There is any established/planned nomenclature about a perl6 'Disposable' role? 19:07
sortiz Now I'm using variants of 'with Foo.new { LEAVE { $_.free } ... }' to guaranty the proper deallocation of the resourses, but can be nice some syntax sugar when Foo does Disposable, for example a 'using Foo.new { ... } ' that desugared to the previous. 19:21
Woodi sortiz: I think for now such thing are internal/hidden becouse Perl do this automatically... about C# you should ask jnthn++ and what means "native" ? NC ? 19:25
sortiz And that needs, at least, a consensus about the name of the method: 'free', 'finalize', 'dispose'.
Woodi sortiz: generally discusion is about topic not yet (definitive) resolved in industry :) 19:26
sortiz Woodi, Yes, native in this context is a NC allocated resource, for example a DB connection in DBDish.
Woodi sortiz: so, for now, I didn't saw better thing then LEAVE for things allocated on Perl6 side 19:28
Woodi sortiz: btw. I just two days ago learned about "Lease" pattern and minutes ago that MS call it Dispose :) 19:30
sortiz Woodi, Yes, Perl internally does for some, via its REPRs, but there are lots of other cases when naming conventions can be useful for the future. 19:30
Woodi sortiz: i think problem isn't about syntax, many more layers of things adds problems to such nice feature :) 19:31
sortiz I'm asking about previous art in Perl6, to avoid reinventing the hot water. :) 19:32
RabidGravy I'd go with "dispose", it is more "I'm done with this now, do what you will with it". 19:37
Woodi sortiz: doc.perl6.org/type/Metamodel::Finalization but probably it calls .free at GC time and sometimes not... 19:39
arcetera i still haven't gotten rakudo to compile
Woodi likes lexical scope LEAVE
arcetera it appears to be looking for a file which isn't existent 19:40
because i'm writing a port
Woodi arcetera: that's strange... from where you got sources ?
arcetera the official rakudo site
rakudo.org
rakudo star
i'm trying to write a crux port and that appears to be the issue
here's the Pkgfile sprunge.us/OCjb 19:41
sortiz RabidGravy, that's the idea. I'll go for "dispose" then.
Woodi arcetera: crux ?
arcetera and currently it's compiling with stderr piped to sprunge so let me upload that
Woodi: yes that's my distro
Woodi arcetera: Linux or other os ? :)
arcetera linux 19:42
it compiles fine outside the port itself
but i get this when i try to compile through the Pkgfile and pkgmk sprunge.us/WgCR
the actual issue appears to be around "Unhandled exception: While looking for '/usr/share/perl6/runtime/perl6.moarvm': no such file or directory"
which obviously doesn't exist because it's not installed yet 19:43
Woodi you tried: cd rakudo && make && make install maybe ? 19:44
I know it should be automatic...
El_Che it's weird 19:45
arcetera Woodi: the thing is 19:46
this is not me compiling manually
El_Che I had something similar while trying (and failing) to build it on Solaris. libuv built fine?
arcetera this is in the package
and it doesn't like that
it expects that i'm compiling directly
RabidGravy it is "installing" stuff into /home/arc/src/crux/rakudo/work/pkg/usr/share/ but is adjusting the wrappers for /usr/share 19:47
arcetera exactly
how do I make it stop that
RabidGravy er which one don't you want?
arcetera adjusting the wrappers for /usr/share/ 19:48
Woodi arcetera: maybe that: you compiling rakudo which gets Moar from sources ? better first create and install moar as package
arcetera i did
otherwise it'd error out earlier
i also packaged nqp
RabidGravy there is probably some option to configure
arcetera there's --sdkroot and --sysroot 19:49
ergh
the rakudo port in crux's repos at the moment is from 2014 and still uses parrotvm 19:50
Woodi right. but at the top of log there is problem with configure...
arcetera that doesn't seem to be the problem
arcetera hmm 19:50
Woodi "wrappers" works via links ? 19:51
RabidGravy no it makes "perl6" as a script that does /foo/bar/baz/moar /blah/blag/perl6.moarvm .,... 19:52
arcetera oh hold on
lemme try this
tryna do --sysroot="$PKG"
hopefully that doesn't cause problems later 19:53
RabidGravy you might also want to check the #perl6-toolchain channel 19:53
arcetera the (serverely outdated) existing rakudo port doesn't work either so
Woodi arcetera: btw. Rakudo build system is mark for rewrite if I remember corectly...
arcetera ┐('~'; )┌
Woodi: thank GOD 19:54
nope that didn't do it
Woodi arcetera: I install things into /opt/lib/graft/star-201601/... then link everything into /opt and it works... 19:54
arcetera see that isn't the problem 19:55
that's a hacky workaround
atweiden m: my FatRat $f = FatRat(989898988898989898980909090939838302233848473393040383234234509484.3489523478234723847238432423); say $f eqv FatRat(989898988898989898980909090939838302233848473393040383234234509484.3489523478234723847238432423); say $f eqv FatRat(989898988898989940247844191134641988224436627880004840480824357399.83933482381524942767485346932); 19:56
camelia rakudo-moar 735869: OUTPUT«True␤True␤»
RabidGravy arcetera, I think you need to find the part in the Configure.pl where it makes the lines for "tools/build/create-moar-runner.pl ..." which generates the wrappers and determine how to influence the bahviour 19:57
arcetera oh boy i gotta write a diff bois
RabidGravy but if I understand you are trying to package this so "install" to a staging location, you will need to rewrite the wrappers for perl6 etc before the final installation 19:58
arcetera yeah i don't really feel like this right now 19:59
especially because i've never touched perl5 and i'm only trying to learn perl6 for fun
El_Che arcetera: domidumont ins #perl6-toolschain is the Debian perl packager 20:00
RabidGravy but the #perl6-toolchain (or whatever it's called) is probably the best place to ask
arcetera lel
i've already asked in #perl6-toolchain
crux ports are a *lot* easier to write than debian packages
so hopefully i'll find something there 20:01
RabidGravy also PERL6_LANG_DIR 20:04
El_Che looking at the the homepage of crux. Any relation with slackware (maybe philosophical)?
RabidGravy appears to influence where the create-moar-runner.pl expects the stuff to be 20:05
arcetera it's similar to slackware in a few aspects
atweiden arcetera: are you looking to package p6 modules for crux? 20:17
arnsholt .tell azawawi Is there some code I could look at for your NativeCall type mismatch? I didn't manage to reproduce it initially, so I suspect it interacts with something a bit deeper 20:28
yoleaux arnsholt: I'll pass your message to azawawi.
arcetera atweiden: i'm looking to package the compiler 20:30
rakudo
atweiden is there any issue with crux?
arcetera the issue is that it's trying to look for files in a directory that doesn't exist
namely in /usr/share rather than $PKG/usr/share
atweiden are you packaging moarvm and nqp separate from rakudo? 20:31
arcetera yes.
here's the log sprunge.us/WgCR 20:32
Woodi arcetera: but eg. moar -h works ?
arcetera and here's the Pkgfile sprunge.us/JiUA
Woodi: yes.
moar and nqp are packaged properly
atweiden rakudo-star then 20:33
Woodi atweiden: sprunge.us/JiUA 20:34
arcetera rakudo-star is the package
Woodi ... is strange
arcetera Woodi: thing is 20:34
that's the same way you'd compile it
in the build() function
Woodi arcetera: prefix without $PKG and no --backends=moar 20:35
atweiden arcetera: it looks to be a problem with panda bootstrap.pl ?
arcetera if I don't use make DESTDIR="$PKG" then it installs to / not the package
atweiden also, you're building rakudo-star 2016.01 20:36
arcetera yes i am
that's the latest
atweiden i wonder if anything has changed since 01 and 02
with precompilation
arcetera ah
as i have nqp/moar 02
atweiden does rakudo (non-star) build?
arcetera atweiden: havne't tried 20:38
arcetera agh, shitty ssh connection to the vps i use irc on 20:38
atweiden arcetera: you should be able to build rakudo and install p6 modules through your crux's pkg manager 20:41
not star but it will be the same result 20:42
arcetera got it
i'll focus on rakudo itself then
where can i get a release tarball for non-star
atweiden github.com/rakudo/rakudo/releases
arcetera thanx
atweiden see also: aur.archlinux.org/packages/?K=perl6
arcetera atweiden: next time please use http/s:// 20:43
otherwise my url handler doesn't pick it up :P
atweiden sry
arcetera it's fine 20:46
now i'm tryna compile just rakudo
aight that worked 20:47
now what other shit do i need to actually get up and running with perl 6
do we have a repl? we have a repl!!!! 20:48
rather than packaging modules i'm just gonna package panda 20:51
Woodi arcetera: did perl ./Configure.pl finished without errors in last build ? 20:52
arcetera yes
we have a repl
Woodi ah, ok :)
arcetera packaging panda was much less painless 20:58
awwaiid I see that rakudo is in debian unstable! sweet!
arcetera aight
awwaiid (er, updated 2016.02 version)
arcetera to the five people who use crux 20:59
you can now use my package for an updated ver of rakudo
awwaiid arcetera: nice
arcetera well shit 21:09
panda reports this error post-install sprunge.us/gFLS 21:11
here's the Pkgfile sprunge.us/DHVE
the problem is it installs the required libs in /usr/lib not /usr/share/perl6
arcetera better yet 21:13
it installs a WHOLE LOTTA JUNK
here's the footprint sprunge.us/VRjj
lemme look at the AUR PKGBUILD 21:16
Woodi arcetera: actually this is our newest design ;) 21:18
arcetera oh god i've had to do something oh so very unholy 21:19
arcetera i give up on panda 21:27
RabidGravy psch, I can get a reasonable facsimile of a sine out of Audio::PortAudio now, considerable cheating by precomputing cycle frames * buffer length and creating the CArrays up front 21:48
couldn't find anything that was fast enough to fill the buffer in the required time 21:49
arcetera aight, now to learn perl6 21:53
I've never used Perl 5 so this'll be interesting
geekosaur perl6 is sufficiently different from perl5 that not knowing perl5 should not matter 21:56
the concepts are the same, but perl5 was an agglomeration of decades of special cases and cruft. perl 6 chucked it and started over. 21:58
nine arcetera: you still here? 22:19
arcetera: you posted something on #perl6-toolchain but then left
arcetera nine: i fixed it 22:27

Powered by Cro and the Raku® Programming Language.

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

Thank you!