»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:14 yqt left 00:27 rangerprice left 00:29 laouji joined 00:40 laouji left 00:41 laouji joined 00:47 BenGoldberg joined 00:53 BenGoldberg left 00:55 BenGoldberg joined 01:03 BenGoldberg left 01:15 BenGoldberg joined 01:21 llfourn joined 01:26 llfourn left, Averna joined 01:27 Averna left 01:28 Averna joined 01:39 gfldex left 01:44 mr-foobar left 01:45 ilbot3 left 01:46 ilbot3 joined 02:02 skids left 02:04 rangerprice joined 02:05 xxx left 02:09 dayangkun joined 02:10 xxx joined 02:15 kaare__ joined 02:17 noganex joined 02:20 noganex_ left 02:21 araujo left 02:22 araujo joined 02:23 raiph left 02:31 ab5tract joined 02:32 smls left 02:33 Averna left 02:34 jordman left 02:35 adu joined 02:40 yeahnoob joined 02:42 jordman joined
dalek ast: cf775d3 | ShimmerFairy++ | S02-literals/numeric.t:
Fix S02-literals/numeric.t

Some of the tests for different forms of Rat literals did so in a way that calls will call .Numeric on Str (as well as any other non-Numeric argument). Since a Rat literal that's not parsed correctly will be interpreted as a Str, then as long as Str.Numeric can handle the given Rat form any issues with Rat literals for those forms were hidden. These tests were changed to ensure we don't numify any strings.
Also updated some of the tests to use the new 'does-ok' to check roles, and replaced a smartmatch against Str with an 'isa-ok' check.		 02:44
02:51 araujo_ joined 02:54 araujo left 03:05 CQ_ joined 03:07 CQ left, CQ_ is now known as CQ, AlexDaniel left 03:12 bin_005 joined 03:22 llfourn joined 03:27 ggherdov left, llfourn left, preyalone left 03:28 pnu left 03:29 pnu joined 03:31 ggherdov joined 03:33 preyalone joined, adu left 03:44 prevost joined 03:45 BenGoldberg left 03:46 bin_005 left 03:56 laouji left 03:59 telex left 04:00 telex joined 04:06 llfourn joined, cognominal joined 04:16 rurban joined
colomon m: say (1..5)[1..*] 04:19
camelia rakudo-moar 30d6fe: OUTPUT«2 3 4 5␤»
GLRelia rakudo-moar 087a6c: OUTPUT«2345␤»
04:27 laouji joined
colomon do we have something in GLR for list concatenation? 04:28
04:31 pflanze left 04:35 domm left 04:37 Sqirrel left
ShimmerFairy you mean besides .join() ? 04:45
colomon: or are you talking about the stuff Slip() was meant for? (As I recall, jnthn's exploration gist mentioned that Slip() took care of various calls for some kind of "list concatenation" operator) 04:48
colomon err… join is string concatenation on lists. I’m looking for list concatenation on lists. 04:49
though yeah, slip ought to help a lot. hmm.
ShimmerFairy colomon: I was think of the (probably somewhat outdated) S07 draft in specs: "Pm's Answer: I think that the conjectured Slip type above eliminates much of the need for a list concatenation operator, since given two lists, one can do a join by wrapping one or both in Slip objects:" 04:51
*thinking of
colomon hmmm, that gist was one of the tabs I was keeping open when my browser crashed hard repeatedly at the hackathon yesterday. :)
m: say [1..5].slip 04:52
GLRelia rakudo-moar 087a6c: OUTPUT«Method 'slip' not found for invocant of class 'Array'␤ in block <unit> at /tmp/fXsXQpTv2H:1␤␤»
camelia rakudo-moar 30d6fe: OUTPUT«Method 'slip' not found for invocant of class 'Array'␤ in block <unit> at /tmp/bUEuLeuVRn:1␤␤»
colomon m: say slip([1..5])
GLRelia rakudo-moar 087a6c: OUTPUT«[1 2 3 4 5]␤»
camelia rakudo-moar 30d6fe: OUTPUT«5===SORRY!5=== Error while compiling /tmp/L_0skvjs38␤Undeclared routine:␤ slip used at line 1. Did you mean 'flip'?␤␤»
colomon m: say slip(1..5)
GLRelia rakudo-moar 087a6c: OUTPUT«1..5␤»
camelia rakudo-moar 30d6fe: OUTPUT«5===SORRY!5=== Error while compiling /tmp/5qGVEKkwQq␤Undeclared routine:␤ slip used at line 1. Did you mean 'flip'?␤␤»
ShimmerFairy .Slip, since it's a type :) 04:53
m: say [1..5].Slip
GLRelia rakudo-moar 087a6c: OUTPUT«12345␤»
camelia rakudo-moar 30d6fe: OUTPUT«Method 'Slip' not found for invocant of class 'Array'␤ in block <unit> at /tmp/QNaGJESMN0:1␤␤»
04:59 khw left
colomon m: my @a = 1..3; my @b = 2, 5, 3; say (@a, @b).perl; say (@a.Slip, @b.Slip).perl 05:00
camelia rakudo-moar 30d6fe: OUTPUT«([1, 2, 3]<>, [2, 5, 3]<>)␤Method 'Slip' not found for invocant of class 'Array'␤ in block <unit> at /tmp/JRpvpvTXe4:1␤␤»
GLRelia rakudo-moar 087a6c: OUTPUT«([1, 2, 3], [2, 5, 3])␤(1, 2, 3, 2, 5, 3)␤»
ShimmerFairy colomon: note that prefix:<|> can also be used to Slip-ify something 05:01
m: say (1,2,3).Slip.perl
GLRelia rakudo-moar 087a6c: OUTPUT«(1, 2, 3)␤»
camelia rakudo-moar 30d6fe: OUTPUT«Method 'Slip' not found for invocant of class 'Parcel'␤ in block <unit> at /tmp/WxtRKIbuoF:1␤␤»
ShimmerFairy m: say (|(1,2,3)).perl
GLRelia rakudo-moar 087a6c: OUTPUT«(1, 2, 3)␤»
camelia rakudo-moar 30d6fe: OUTPUT«5===SORRY!5=== Error while compiling /tmp/bPhfVd6072␤Arg-flattening | is only valid in an argument list␤at /tmp/bPhfVd6072:1␤------> 3say (|7⏏5(1,2,3)).perl␤»
colomon ShimmerFairy++
ShimmerFairy m: my @a = 1..3; my @b = 2, 5, 3; say (@a, @b).perl; say (|@a, |@b).perl 05:02
GLRelia rakudo-moar 087a6c: OUTPUT«([1, 2, 3], [2, 5, 3])␤(1, 2, 3, 2, 5, 3)␤»
camelia rakudo-moar 30d6fe: OUTPUT«5===SORRY!5===␤Arg-flattening | is only valid in an argument list␤at /tmp/3LWC_lp1aX:1␤------> 3 @b = 2, 5, 3; say (@a, @b).perl; say (|7⏏5@a, |@b).perl␤Arg-flattening | is only valid in an argument list␤at /tmp/3LWC_lp1aX:1␤------> …»
colomon m: say ([]).map({ 1 }) 05:04
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
camelia rakudo-moar 30d6fe: OUTPUT«␤»
05:06 RabidGravy joined 05:14 prevost left
colomon m: say (Array.new).map({ 1 }) 05:24
camelia rakudo-moar 30d6fe: OUTPUT«␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
05:24 rangerprice left
colomon m: say [[]].map({ 1 }) 05:24
camelia rakudo-moar 30d6fe: OUTPUT«1␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
colomon okay, so under GLR, how does one return a list consisting of exactly one empty array? 05:29
ShimmerFairy m: say [[],].perl 05:31
camelia rakudo-moar 30d6fe: OUTPUT«[[]]␤»
GLRelia rakudo-moar 087a6c: OUTPUT«[[],]␤»
ShimmerFairy m: say ((1,2,3),).perl
camelia rakudo-moar 30d6fe: OUTPUT«((1, 2, 3),)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«((1, 2, 3))␤»
ShimmerFairy m: say ((1,2,3),).map: {1}
GLRelia rakudo-moar 087a6c: OUTPUT«1␤»
camelia rakudo-moar 30d6fe: OUTPUT«1␤»
ShimmerFairy m: say ((1,2,3)).map: {1} 05:32
camelia rakudo-moar 30d6fe: OUTPUT«1 1 1␤»
GLRelia rakudo-moar 087a6c: OUTPUT«111␤»
colomon m: say [[]].perl; say [[],].perl
camelia rakudo-moar 30d6fe: OUTPUT«[[]]␤[[]]␤»
GLRelia rakudo-moar 087a6c: OUTPUT«[]␤[[],]␤»
ShimmerFairy colomon: you have to explicitly add a comma, like how in pre-GLR you have to add a comma to force a Parcel 05:33
m: say (42).perl; say (42,).perl
camelia rakudo-moar 30d6fe: OUTPUT«42␤(42,)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«42␤(42)␤»
ShimmerFairy m: say (42).WHAT; say (42,).WHAT
camelia rakudo-moar 30d6fe: OUTPUT«(Int)␤(Parcel)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of uninitialized value of type Int in string context in block <unit> at /tmp/vHIwTjkdrC:1␤␤Use of uninitialized value of type List in string context in block <unit> at /tmp/vHIwTjkdrC:1␤␤»
colomon m: [[],].map({ 1 }).say
camelia rakudo-moar 30d6fe: OUTPUT«1␤»
GLRelia rakudo-moar 087a6c: OUTPUT«1␤»
colomon ShimmerFairy++ 05:34
ShimmerFairy m: say 42.WHAT
camelia rakudo-moar 30d6fe: OUTPUT«(Int)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of uninitialized value of type Int in string context in block <unit> at /tmp/axPsCAQUP7:1␤␤»
ShimmerFairy m: say Any
camelia rakudo-moar 30d6fe: OUTPUT«(Any)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of uninitialized value of type Any in string context in block <unit> at /tmp/VdN99AQ1rz:1␤␤»
ShimmerFairy looks like GLR doesn't like :U objects on say atm :P 05:35
m: print Mu
camelia rakudo-moar 30d6fe: OUTPUT«Use of uninitialized value <element> of type Mu in string context in block <unit> at /tmp/3OUWV7s8aN:1␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of uninitialized value <element> of type Mu in string context in block <unit> at /tmp/nGH3YXR3uL:1␤»
ShimmerFairy m: print Mu.gist
camelia rakudo-moar 30d6fe: OUTPUT«(Mu)»
GLRelia rakudo-moar 087a6c: OUTPUT«(Mu)»
ShimmerFairy m: say Int.gist
camelia rakudo-moar 30d6fe: OUTPUT«(Int)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«(Int)␤»
colomon m: ([],).map({ 1 }).say 05:36
GLRelia rakudo-moar 087a6c: OUTPUT«1␤»
camelia rakudo-moar 30d6fe: OUTPUT«1␤»
05:45 rurban left 05:58 rurban joined 05:59 colomon left 06:02 colomon joined
colomon O/ from hotel breakfast 06:02
nine m: [$[]].perl.say 06:04
camelia rakudo-moar 30d6fe: OUTPUT«[Any]␤»
GLRelia rakudo-moar 087a6c: OUTPUT«[[],]␤»
nine colomon: ^^^
m: [$[]].map(say $_.perl)
camelia rakudo-moar 30d6fe: OUTPUT«Any␤Cannot call map(Array: Bool); none of these signatures match:␤ ($: Whatever, *%_)␤ ($: &block, :$label, *%_)␤ in block <unit> at /tmp/Y7dK4aPPL8:1␤␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Any␤Cannot call map(Array: Bool); none of these signatures match:␤ ($: &block, :$label, *%_)␤ (HyperIterable:D $: &block, :$label, *%_)␤ in block <unit> at /tmp/S2pegLq2Tf:1␤␤»
nine m: [$[]].map({say $_.perl})
camelia rakudo-moar 30d6fe: OUTPUT«Any␤»
GLRelia rakudo-moar 087a6c: OUTPUT«[]␤»
06:05 colomon left, Woodi left
ShimmerFairy oh yeah 06:06
m: ($()).perl.say
camelia rakudo-moar 30d6fe: OUTPUT«Use of Nil in string context in block <unit> at /tmp/fDruYzXKAY:1␤""␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of Nil in string context in block <unit> at /tmp/yhdkXixgPd:1␤""␤»
06:06 Woodi joined
ShimmerFairy since [] enforces scalar containers, it should suffice to use $ inside a [] :) 06:06
m: [$()].perl.say
camelia rakudo-moar 30d6fe: OUTPUT«Use of Nil in string context in block <unit> at /tmp/0ONHQXvll5:1␤[""]␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of Nil in string context in block <unit> at /tmp/x7jFzMM_CP:1␤[""]␤»
ShimmerFairy derp, $() is something else :P
m: [$( )].perl.say
camelia rakudo-moar 30d6fe: OUTPUT«[]␤»
GLRelia rakudo-moar 087a6c: OUTPUT«[(),]␤»
ShimmerFairy m: ($( )).perl.say 06:07
camelia rakudo-moar 30d6fe: OUTPUT«$( )␤»
GLRelia rakudo-moar 087a6c: OUTPUT«()␤»
JimmyZ bug!
06:07 colomon joined
colomon nine++ 06:08
ShimmerFairy JimmyZ: no, $() is specific to Matches
m: "A" ~~ /A/; say $()
camelia rakudo-moar 30d6fe: OUTPUT«A␤»
GLRelia rakudo-moar 087a6c: OUTPUT«A␤»
ShimmerFairy m: "A" ~~ /A {make "B"}/; say $()
camelia rakudo-moar 30d6fe: OUTPUT«B␤»
GLRelia rakudo-moar 087a6c: OUTPUT«B␤»
ShimmerFairy m: "A" ~~ /A {make "B"}/; say $/.ast // ~$/
camelia rakudo-moar 30d6fe: OUTPUT«B␤»
GLRelia rakudo-moar 087a6c: OUTPUT«B␤»
ShimmerFairy m: "A" ~~ /A/; say $/.ast // ~$/ 06:09
camelia rakudo-moar 30d6fe: OUTPUT«A␤»
GLRelia rakudo-moar 087a6c: OUTPUT«A␤»
nine ShimmerFairy: but $( ).perl.say should print $(), not (). That's the remaining t/spec/S02-types/list.t failure
JimmyZ m: "A" ~~ /A {make "B"}/; say $( )
camelia rakudo-moar 30d6fe: OUTPUT«␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
ShimmerFairy nine: oh, that's the bug :)
nine: so [$()].perl should be '[$(),]' IIUC. 06:10
nine For some reason the nqp::iscont check on SELF in List.Perl returns 0
ShimmerFairy: yes
JimmyZ SELF or self?
nine SELF 06:11
self would be decont'ed in any case
ShimmerFairy ( and for the record, I only learned about $() a couple days ago, and I personally wouldn't mind if it went away. I don't see where it's useful enough to have a short form of {$/.ast // ~$/} )
JimmyZ wish $() and $( ) is the same thing 06:12
*are
ShimmerFairy JimmyZ: I've never seen it used, except for a few spots in nqp and/or rakudo, and it doesn't seem like it has anything to do with regexes. I'm inclined to agree :) 06:13
JimmyZ: and on top of it, $() feels to me like a case of creeping "punctuation scalars" that I know existed in Perl 5, and we worked to get rid of :) (that is, Perl 5's $=, $', and probably everything else that could match /'$' <:Punct>/ :P) 06:15
JimmyZ well, at least, for consistency, they should be same thing 06:16
TimToady agrees that $() can go
JimmyZ $[] vs $[ ] vs $() vs $( )
06:16 colomon_ joined 06:17 colomon left, colomon_ is now known as colomon
ShimmerFairy TimToady: for what it's worth, I kinda like the idea of something that automatically picks the right thing like that, but I think in most cases you're not working off $/ in the first place, and it's not worth a magical-variable-looking thing. 06:18
in most cases where you're interested in C<.ast>s, is what I meant.
TimToady it would make more sense to use $<> for that anyway 06:19
colomon What time are people starting to go to the hackathon? 06:20
dalek ast/glr: fe7a750 | (Stefan Seifert)++ | S03-operators/assign.t:
Add explicit flattening, to fix array assignment tests
 06:21
TimToady registration is at 9, program starts at 9:45, so I have no idea when people will actually go :)
JimmyZ specs git:(master) grep -r '$()' * | wc -l
13
colomon TimToady: think the building is open yet? I’m getting sick of sitting on a footstool thing…. :) 06:22
TimToady yesterday opened at 9
colomon TimToady++ 06:23
06:26 diana_olhovik_ joined, salv0 joined
ShimmerFairy m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $<> 06:27
camelia rakudo-moar 30d6fe: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
ShimmerFairy $<> predictably returns $/ (at least on nom. huh.). In that case, it might not be a bad thing to make $() actually be $<>, since current $<> is redundant (and then also $<xyzzy><> and so on) 06:29
m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/ === $<>
camelia rakudo-moar 30d6fe: OUTPUT«True␤»
GLRelia rakudo-moar 087a6c: OUTPUT«True␤»
ShimmerFairy m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/
camelia rakudo-moar 30d6fe: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
ShimmerFairy that's not good for glr
m: "A" ~~ /$<xyzzy>=(A {make "B"})/; print $/ 06:30
camelia rakudo-moar 30d6fe: OUTPUT«A»
GLRelia rakudo-moar 087a6c: OUTPUT«A»
ShimmerFairy m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/.perl
camelia rakudo-moar 30d6fe: OUTPUT«Match.new(ast => Any, list => (), hash => EnumMap.new(:xyzzy(Match.new(ast => "B", list => (), hash => EnumMap.new(), orig => "A", to => 1, from => 0))), orig => "A", to => 1, from => 0)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Match.new(ast => Any, list => (), hash => EnumMap.new(:xyzzy(Match.new(ast => "B", list => (), hash => EnumMap.new(), orig => "A", to => 1, from => 0))), orig => "A", to => 1, from => 0)␤»
ShimmerFairy m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/.gist
camelia rakudo-moar 30d6fe: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤»
GLRelia rakudo-moar 087a6c: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤»
ShimmerFairy yeah, seems glr's &say fails to call .gist on things now, for some reason
m: say Int
camelia rakudo-moar 30d6fe: OUTPUT«(Int)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«Use of uninitialized value of type Int in string context in block <unit> at /tmp/4qyPWrfzNo:1␤␤»
ShimmerFairy m: class Foo { method Str { "fail" }; method gist { "win" } }; say Foo 06:33
GLRelia rakudo-moar 087a6c: OUTPUT«fail␤»
camelia rakudo-moar 30d6fe: OUTPUT«win␤»
06:33 salva is now known as Guest54260, Guest54260 left, salv0 is now known as salva 06:34 Guest54260 joined
moritz m: say "foo" 06:34
camelia rakudo-moar 30d6fe: OUTPUT«foo␤»
GLRelia rakudo-moar 087a6c: OUTPUT«foo␤»
moritz ShimmerFairy: spectesting a fix
ShimmerFairy moritz++ :) 06:35
m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/.Str
camelia rakudo-moar 30d6fe: OUTPUT«A␤»
GLRelia rakudo-moar 087a6c: OUTPUT«A␤»
ShimmerFairy m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $/
camelia rakudo-moar 30d6fe: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
ShimmerFairy that's a bit weirder though O_o
m: "A" ~~ /$<xyzzy>=(A {make "B"})/; say $_ for $/; say $/ ~~ Iterable
camelia rakudo-moar 30d6fe: OUTPUT«｢A｣␤ xyzzy => ｢A｣␤False␤»
GLRelia rakudo-moar 087a6c: OUTPUT«False␤»
ShimmerFairy Capture may not do Iterable (at least in the nom branch), but it does have AT-POS and friends (as well as AT-KEY and friends) 06:37
moritz m: say Capture ~~ Positional
camelia rakudo-moar 30d6fe: OUTPUT«False␤»
GLRelia rakudo-moar 087a6c: OUTPUT«False␤»
ShimmerFairy and Match is Capture, so I'm guessing glr's &say accidentally doesn't treat $/ as a singular thing. Whether that's say's fault, or Capture not doing Iterable, I don't know.
06:39 domidumont joined
dalek kudo/glr: 64cbc0d | moritz++ | src/core/io_operators.pm:
say must use .gist, not .Str

this way, say(Int) prints "(Int)\n" again		 06:39
ShimmerFairy My guess is that Capture should be doing Positional, Associative, and Iterable anyway. 06:41
JimmyZ moritz: note needs too
one is Str and one is Gist :)
m: note Int 06:42
camelia rakudo-moar 30d6fe: OUTPUT«(Int)␤»
GLRelia rakudo-moar 087a6c: OUTPUT«␤»
ShimmerFairy wishes often that the core roles weren't just empty things that are basically glorified annotations :/
dalek ast/glr: d2c7d71 | (Stefan Seifert)++ | S03-operators/assign.t:
Fix remaining failures in S03-operators/assign.t

Need to take more care when to flatten now. Scalar assignment needs flattening of the whole expression while list assignment needs flattening of the input lists.		 06:43
06:43 domidumont left 06:44 domidumont joined
dalek kudo/glr: 48ecd4d | moritz++ | src/core/io_operators.pm:
note needs to .gist its arguments too, JimmyY++
 06:47
JimmyZ wishes we have is-output($Code, $expectet, $msg) 06:49
06:49 domidumont left
JimmyZ so I can added a bit more test about say/note 06:50
*add
06:53 colomon left
[Tux] test 50000 44.456 44.364 06:55
test-t 50000 45.259 45.167
nom
TravisCi has a panda problem: travis-ci.org/Tux/CSV/builds/77634324 07:03
test 50000 48.296 48.212
test-t 50000 45.212 45.128
glr
masak morning, #perl6
[Tux] m: class C{method foo{gather { take 3..Inf}}};C.new.foo[^9].perl.say 07:06
camelia rakudo-moar 30d6fe: OUTPUT«(3, 4, 5, 6, 7, 8, 9, 10, 11)␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«(3..Inf)␤»
[Tux] I got no answer on that yet
07:09 Sqirrel joined 07:10 domidumont joined 07:12 xinming joined, xinming_ left 07:18 darutoko joined, larion joined 07:19 luiz_lha joined 07:20 luiz_lha is now known as Guest76846 07:28 zakharyas joined 07:30 araujo_ left
dalek ast/glr: 388b0df | (Stefan Seifert)++ | S02-types/hash.t:
Adopt GLRR semantics in S02-types/hash.t
 07:30
07:31 Ven joined, llfourn left
JimmyZ m: my %hash = %('a'..'d' Z 1..4); say %hash.elems; 07:31
camelia rakudo-moar 30d6fe: OUTPUT«4␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«2␤»
dalek kudo/nom: 28850d9 | (Tim Smith)++ | src/core/Exception.pm:
Perl5Var $. maps to .ins method, not .line
kudo/nom: 70f6519 | FROGGS++ | src/core/Exception.pm:
Merge pull request #517 from softmoth/patch-1

Perl5Var $. maps to .ins method, not .line
07:31 FROGGS left
nine m: my %hash = %(flat 'a'..'d' Z 1..4); say %hash; 07:32
camelia rakudo-moar 30d6fe: OUTPUT«a => 1, b => 2, c => 3, d => 4␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«a => 1d => 4c => 3b => 2␤»
07:32 Guest76846 is now known as luiz_lha
nine JimmyZ: ^^^ 07:32
JimmyZ well, I didn't the segfault
nine JimmyZ: Z creates a list of lists 07:33
JimmyZ see
dalek kudo/nom: cd9580f | (Jimmy Zhuo)++ | src/core/Int.pm:
add native int version of pow
kudo/nom: 906ca06 | FROGGS++ | src/core/Int.pm:
Merge pull request #506 from zhuomingliang/patch-4

add native int version of pow
07:33 abraxxa joined
JimmyZ nine: I'd like to see the segfault :P 07:33
nine JimmyZ: I think that's long gone
07:34 FROGGS joined
JimmyZ ok, so no need to fix moarvm 07:34
FROGGS o/ 07:35
07:36 brrt joined 07:39 lizmat joined
FROGGS JimmyZ: note that is-output might want to be more like: is-output($Code, $expectet-out, $expected-err, $msg) 07:40
JimmyZ: unless merging out+err is implicit, which makes it less useful 07:41
JimmyZ we can do $*ERR = $*OUT before testing 07:42
07:42 Ven left
FROGGS JimmyZ: then you do you test that something does (not) warn? 07:43
you need the distinction sometimes
JimmyZ: is_run is good enough for me I think 07:44
JimmyZ oh 07:45
FROGGS look at this for example: t/spec/S06-other/main-usage.t:13
07:46 lizmat left 07:47 lizmat joined, lizmat left
nine Someone needs to rewrite S16-io/say.t to use is_run to check the output. The old, clever way just doesn't work anymore since array .gists contain the brackets now 07:47
FROGGS nine: yeah, I think so too 07:48
07:48 colomon joined
FROGGS but I won't do it now, since I'm running a spectest, reapplying a patch and respectest afterwards 07:49
07:51 lizmat joined
tadzik good morning #perl6 07:55
mfw my battery is at 5% but will still last for a 40-minute talk （´ー｀　） 07:56
07:57 rindolf joined
nine tadzik: you can't be running many spectest then ;) 07:57
tadzik oh no I'm not :)
nine probably needs to move inside within two hours
08:05 xnrand left 08:07 xnrand joined
tadzik m: sub foo { my ($a, $b) = 1, 2; return :$a, :$b }; say foo().perl 08:08
GLRelia rakudo-moar 48ecd4: OUTPUT«(:a(Mu), :b(Mu))␤»
camelia rakudo-moar 906ca0: OUTPUT«(:a(Mu), :b(Mu))␤»
tadzik ...how did it work before :o
nine m: sub foo { my ($a, $b) = 1, 2; return :$a, :$b }; my (:$a, :$b) = foo; say $a; say $b;
GLRelia rakudo-moar 48ecd4: OUTPUT«a => (Mu)␤b => (Mu)␤» 08:09
camelia rakudo-moar 906ca0: OUTPUT«a => (Mu)␤b => (Mu)␤»
nine m: sub foo { my ($a, $b) = 1, 2; return :$a, :$b }; my (:$a, :$b) := foo; say $a; say $b;
camelia rakudo-moar 906ca0: OUTPUT«(Mu)␤(Mu)␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«Too many positionals passed; expected 0 arguments but got 2␤ in block <unit> at /tmp/56MotrnerV:1␤␤»
dalek kudo-star-daily: d390db3 | coke++ | log/ (9 files):
today (automated commit)
08:11 domm_holiday joined
timotimo Timo super tired. Timo sleep a little more 08:15
dalek ast/glr: cca514e | (Jimmy Zhuo)++ | S16-io/print.t:
update S16-io/print.t to use is-run
FROGGS timotimo will miss the Q&A with Larry 08:16
lizmat timotimo will miss Larry Wall Q&A ?
:-)
moritz JimmyZ: I'm working on S16-io/say.t
nine JimmyZ++
moritz++
lizmat moritz++ JimmyZ++ nine++ timotimo++ 08:17
:-)
JimmyZ :)
lizmat FROGGS++ :-)
FROGGS O.o 08:18
08:20 domidumont left
dalek ast/glr: f11010e | moritz++ | S16-io/say.t:
Rewrite say.t by faking $*OUT
 08:21
nine my $foo = "FOO"; my $gorch = "foo bar"; my @q := «a b c "$foo" f g $gorch m n '$bar' x y z»; say @q.perl; 08:22
m: my $foo = "FOO"; my $gorch = "foo bar"; my @q := «a b c "$foo" f g $gorch m n '$bar' x y z»; say @q.perl; 08:23
camelia rakudo-moar 906ca0: OUTPUT«("a", "b", "c", "FOO", "f", "g", ("foo", "bar"), "m", "n", "\$bar", "x", "y", "z")␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«("a", "b", "c", "FOO", "f", "g", ("foo", "bar"), "m", "n", "\$bar", "x", "y", "z")␤»
lizmat ah, so it didn't flatten before either ?
moritz no, but it didn't matter that much before
lizmat yeah
nine m: my $foo = "FOO"; my $bar = "BAR"; my $gorch = "foo bar"; @q1 = «$foo $gorch $bar»;
camelia rakudo-moar 906ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/e7U2MLHwfi␤Variable '@q1' is not declared␤at /tmp/e7U2MLHwfi:1␤------> 3my $bar = "BAR"; my $gorch = "foo bar"; 7⏏5@q1 = «$foo $gorch $bar»;␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/QDfc5hGAYE␤Variable '@q1' is not declared␤at /tmp/QDfc5hGAYE:1␤------> 3my $bar = "BAR"; my $gorch = "foo bar"; 7⏏5@q1 = «$foo $gorch $bar»;␤»
dalek ast/glr: d663fb5 | (Jimmy Zhuo)++ | S16-io/print.t:
add more test to print.t
 08:24
nine m: my $foo = "FOO"; my $bar = "BAR"; my $gorch = "foo bar"; my @q1 = «$foo $gorch $bar»; say @q1.perl;
camelia rakudo-moar 906ca0: OUTPUT«["FOO", "foo", "bar", "BAR"]<>␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«["FOO", ("foo", "bar"), "BAR"]␤»
08:24 domidumont joined
nine But in this case it did flatten ^^^^ 08:24
Which does not make sense to me
moritz if I knew where it was constructed, I'd just stick in a "flat" and be done
nine Test descriptions are "4 elements in unquoted «» list" vs. "13 elements in mixed quoted/unquoted «» list, non-flattened"
moritz: I already have a pach. It just happens to break the cases where it did not flatten before. And I don't understand why it should not flatten in those cases. 08:25
moritz nine: IMHO it should always flatten, and every test assuming otherwise is a vicitim of Stockholm Syndrome 08:26
nine moritz: ok, I think it's really just a bug in the spec test. 08:27
Test description is: "second quote is both parts of $gorch interpolated as sublist in non-flat context"
Point is: there is no second quote.
So I assume $gorch was meant to be quoted there. 08:28
moritz m: my $a = 'a b'; say <<x "$a" y>>.perl
camelia rakudo-moar 906ca0: OUTPUT«("x", "a b", "y")␤»
GLRelia rakudo-moar 48ecd4: OUTPUT«("x", "a b", "y")␤»
moritz yes, then it's a single element, not a sublist
FROGGS with my patch that makes Any.elems == 1 it still does not iterate over Any 08:29
:/
lizmat :-(
FROGGS I guess it is an issue in MapIterCommon or so, but I don't see it yet
dalek kudo/glr: 3e83f87 | FROGGS++ | src/core/Any-iterable-methods.pm:
remove comment about missing label support
 08:31
08:32 domidumont left
dalek ast/glr: 3177163 | (Stefan Seifert)++ | S02-literals/quoting.t:
Fix missing quotes in S02-literals/quoting.t

The test description talks about a "second" quote involving $gorch, but
  $gorch was not quoted. This did not matter before GLR due to implicit
flattening.		 08:32
08:33 llfourn joined
dalek kudo/glr: bf1307e | (Stefan Seifert)++ | src/Perl6/Actions.nqp:
Fix my $g = "a b"; <<c $g d>> not interpolating the sublist

  $g gets turned into ('a', 'b') which does no longer interpolate
implicitly. Have to create explicit code to do that to keep << >> semantics.		 08:33
08:34 Ven joined
lizmat nine++ 08:35
08:35 lizmat left 08:42 lizmat joined
FROGGS hmmm, I could bet I tried the very same thing :o) 08:44
08:45 domidumont joined
ShimmerFairy nine: that's actually not how it's supposed to work in the first place 08:47
domm_holiday Hey, I could use some feedback on a small Perl 6 script I will give a talk on at YAPC::Europe:
github.com/domm/potd/blob/master/potd.p6
ShimmerFairy ah, I misread the section on S02 :)
domm_holiday It's a more or less straightforward port of github.com/domm/potd/blob/master/potd.pl (but the actual Image-resizing etc is still done in Perl5)It's a more or less straightforward port of github.com/domm/potd/blob/master/potd.pl (but the actual Image-resizing etc is still done in Perl5)
I'd appreciate any hints on how to make the code more Perl6ish
As I'm currently traveling through Morocco, my Internet is rather flaky, so feedback via email would work best, although I've got my irssi running in screen :-)
Thanks for any help!
some more context ist linked from here: github.com/domm/potd 08:48
nine ShimmerFairy: it is how it's supposed to work _now_ because the spec tests are the authorative source ;)
ShimmerFairy nine: thinking about it, I kinda want «1 $a» to be equivalent to «1 "$a"»
despite what S02's section on <> and «» says :) 08:49
Ven domm_holiday: our `if` doesn't need parentheses :)
08:49 lizmat left
moritz domm_holiday: in post_template you can indent the heredoct (both contents and end marker) 08:49
Ven domm_holiday: I think instead "shell" it's usually recommended to use the quoting construct (to prevent injection etc) 08:50
moritz domm_holiday: also most people tend to put a space between the subroutine and the signature
ah yes, run() if you don't actually need shell interpretation 08:51
08:52 bin_005 joined
domm_holiday wow, thanks, see the last few commits in the repo :-) 08:52
ok, gotta leave now, but any more feedback is of course appreciated :-)
08:54 colomon left
moritz domm_holiday: you can actually indent the contents of the heredoc, and Perl 6 will strip the leading whitespace for you (at most as much as the delimiter is indented) 08:55
08:55 colomon joined
domm_holiday moritz: ok, so even if I need the file content to start without any spaces, I can indent the heredoc in the code, and perl6 will strip the leading spaces? 08:56
nice, will have to give this a try 08:57
now off for real...
dalek ast/glr: 60232eb | colomon++ | integration/99problems-21-to-30.t:
Move combination sub to the GLR.
Ven I opened a PR for GTK::Simple, maybe somebody can look at it? :)
tadzik Thanks to laben++ and nine++ panda master now works on GLR
colomon \o/
Ven: it works, right?
Ven The tests definitely pass, at least :) 08:58
colomon Merged.
dalek k-simple: 64f1603 | vendethiel++ | / (3 files):
Remove duplicate `is repr`, add a missing `my`

and fix the test deprecation warnings
k-simple: ffe5757 | colomon++ | / (3 files):
Merge pull request #13 from vendethiel/master

Bring GTK::Simple up to date
Ven colomon++
colomon Ven++
hoelzro o/ # 08:59
08:59 adu joined
tadzik Looking for `rakudobrew triple` users 08:59
colomon tadzik: errr, me, sort of. :)
tadzik colomon: will you be willing to maintain it as a rakudobrew plugin if I were to split it?
colomon …. yes? 09:00
tadzik k :)
dalek kudo/glr: 94e6d13 | (Stefan Seifert)++ | src/core/metaops.pm:
Allow DWIM RHS of infix HYPER to be lazy

If the right hand side of an infix hyper like >>+>> is DWIM (the pointy end points outwards), it may actually be lazy (even infinite) because we shorten it to the length of the left hand side anyway.
Fixes (0,0,0,0,0,0) >>+>> (Slip(1,2) xx *)
09:00 AlexDaniel joined
JimmyZ m: note Int 09:01
camelia rakudo-moar 906ca0: OUTPUT«(Int)␤»
GLRelia rakudo-moar bf1307: OUTPUT«␤»
09:01 espadrine joined
dalek ast/glr: fd0153c | (Jimmy Zhuo)++ | S16-io/ (2 files):
add more test to say.t, update note.t to use Fake IO
 09:02
colomon m: say ([1, 2], [1, 4]).map(-> $e { say :$e.perl }) 09:08
camelia rakudo-moar 906ca0: OUTPUT«:e([1, 2])␤:e([1, 4])␤True True␤»
GLRelia rakudo-moar bf1307: OUTPUT«:e($[1, 2])␤:e($[1, 4])␤TrueTrue␤»
colomon m: say ([1, 2], [1, 4]).perl
camelia rakudo-moar 906ca0: OUTPUT«([1, 2], [1, 4])␤»
GLRelia rakudo-moar bf1307: OUTPUT«([1, 2], [1, 4])␤»
09:09 broquaint joined 09:10 |Tux| left 09:13 |Tux| joined 09:14 brrt left, brrt joined, brrt left 09:16 xinming left, xinming joined
colomon m: say ($_ + 1) if $_ > 5 for 1..10 09:17
camelia rakudo-moar 906ca0: OUTPUT«7␤8␤9␤10␤11␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«7␤8␤9␤10␤11␤»
colomon m: say (($_ xx 3).Slip if $_ > 5 for 3, 5, 7, 9) 09:18
camelia rakudo-moar 906ca0: OUTPUT«Method 'Slip' not found for invocant of class 'List'␤ in block <unit> at /tmp/UAUVKjZf_m:1␤␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«(Any)(Any)(Any)(Any)(Any)(Any)␤»
colomon m: say ($_ xx 3 if $_ > 5 for 3, 5, 7, 9)
GLRelia rakudo-moar 94e6d1: OUTPUT«(7 7 7)(9 9 9)␤»
camelia rakudo-moar 906ca0: OUTPUT«7 7 7 9 9 9␤»
colomon m: say (($_ xx 3).list if $_ > 5 for 3, 5, 7, 9)
GLRelia rakudo-moar 94e6d1: OUTPUT«((Any) (Any) (Any))((Any) (Any) (Any))␤»
camelia rakudo-moar 906ca0: OUTPUT«(Any) (Any) (Any) (Any) (Any) (Any)␤»
colomon m: say (($_ xx 3) if $_ > 5 for 3, 5, 7, 9) 09:19
camelia rakudo-moar 906ca0: OUTPUT«(Any) (Any) (Any) (Any) (Any) (Any)␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«((Any) (Any) (Any))((Any) (Any) (Any))␤»
colomon m: say ($_ xx 3 if $_ > 5 for 3, 5, 7, 9)
camelia rakudo-moar 906ca0: OUTPUT«7 7 7 9 9 9␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«(7 7 7)(9 9 9)␤»
dalek ast/glr: ccf062e | colomon++ | integration/advent2011-day23.t:
Add "flat"s needed for GLR.
 09:21
09:22 larion left
colomon m: my @a = (1,2,3).map(*.Int); say @a.perl 09:27
camelia rakudo-moar 906ca0: OUTPUT«[1, 2, 3]<>␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«[1, 2, 3]␤»
colomon m: my ($x, @a) = “dummy”, (1,2,3).map(*.Int); say @a.perl
camelia rakudo-moar 906ca0: OUTPUT«[1, 2, 3]<>␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«[(1, 2, 3).Seq,]␤»
colomon GLRers: why is the second result different from the first?
JimmyZ bug!
colomon okay, switching over to zip.t for the moment. 09:36
09:36 aborazmeh joined, aborazmeh left, aborazmeh joined
colomon m: my @a = 1, 3, 5; my @b = 2, 4, 6; say (@a Z @b).perl 09:36
camelia rakudo-moar 906ca0: OUTPUT«((1, 2), (3, 4), (5, 6))␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«((1, 2), (3, 4), (5, 6)).Seq␤»
colomon m: my @a = 1, 3, 5; my @b = 2, 4, 6; say zip(@a, @b).perl 09:37
camelia rakudo-moar 906ca0: OUTPUT«((1, 2), (3, 4), (5, 6))␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«((1, 2), (3, 4), (5, 6)).Seq␤»
09:37 laben joined
laben hello p6'ers 09:37
colomon huh. That’s not what I’m getting in the test suite.
laben: o/
dalek ast/glr: d13a1bb | moritz++ | S32-list/map_function_return_values.t:
Add a flat
 09:38
hoelzro o/ laben
nine m: my @r; say |@r; 09:39
camelia rakudo-moar 906ca0: OUTPUT«␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/jSKdW6JNtT:1␤␤»
adu are there any Swagger libs yet? 09:40
nine m: my @r; say |@r;
camelia rakudo-moar 906ca0: OUTPUT«␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/Kme20NQhCS:1␤␤»
nine jnthn: I might need a hint for this one ^^^
jnthn: looks to me like some .DEFINITE check is missing, but I don't know where. 09:41
laben what is Swagger?
09:41 yeahnoob left
tadzik Flogging Molly album, for what I know :) 09:41
09:41 rmgk is now known as Guest39027, rmgk_ joined, Guest39027 left, rmgk_ is now known as rmgk
adu laben: swagger.io/ 09:43
colomon what I’m seeing in zip.t is (with slightly different args), @a Z @b goes into infix:<Z> with \([0, 2, 4], [1, 3, 5]), while zip(@a; @b) goes into it with \(([0, 2, 4]), ([1, 3, 5]))
adu laben: It's a JSON format for describing web apis
09:44 larion joined
dalek ast/glr: 9928593 | moritz++ | S32-exceptions/misc.t:
Make a test for empty list-like thingy more liberal
 09:44
adu laben: There's also a codegen thing for Java, but what I was thinking is something like it for Perl6
dalek ast/glr: 7e08156 | moritz++ | S32-exceptions/misc.t:
Fix test plan

a554c9f3aa2429a0ab43325cbd2d70a18239916c removed four tests without adjusting the plan		 09:45
adu I don't think you would even need codegen with Perl6, you could probably do everything at runtime 09:46
laben nine: your recent patch to DWIM hypers is only half, also, in the <<op>> case, it hangs instead of throwing. want me to do the rest?
nine laben: yeah, please, if you know what's missing, go ahead :0 09:47
:)
laben adu: there is a YAML parser/generator module, the rest seems to be missing 09:48
colomon m: say "a" «~« set(<pple bbot rmadillo>)
camelia rakudo-moar 906ca0: OUTPUT«set(rmadillo, bbot, pple)␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«Cannot modify an immutable Set␤ in block <unit> at /tmp/LQ9VZtmkp8:1␤␤»
laben nine: i was thinking i also need to add some tests that we're missing on dwims, like the case i fixed the other day
nine laben: FWIW I ran a full spectest and did not notice any regressions 09:49
laben: so yes, more tests are certainly needed 09:50
laben nine: yeah, it's because those cases were not tested. i'll try to add them so they dont regress anymore :)
adu laben: I guess they are kind of pushing for YAML, but my swagger files are usually JSON 09:52
moritz ++laben
jdv79 is it nomal that t/spec/S32-io/IO-Socket-Async can hang?
looks stuck in a futex:(
moritz jdv79: uhm, I haven't seen it hang, just fail randomly
jdv79 maybe i'll update and try again 09:53
laben adu: there are loads of JSON parsing/generating modules, the recommended ones are JSON::Fast or JSON::Tiny
09:53 esh joined
laben adu: both work already on glr 09:53
adu laben: what's glr? 09:54
masak TimToady++ # As to people's Qs
adu as in the parsing algo?
moritz adu: the Great List Refactoring, and a branch of the same name in rakudo
09:54 colomon left
adu ah, as in a branch name 09:55
laben adu: it means that it will work when the development glr branch will be merged in
adu I should probably test my parser
Ven adu: great list refactor 09:56
09:56 lizmat joined, bin_005 left
moritz m: my @b = map { hash("v" => $_, "d" => $_ * 2)}, (1, 2, 3); say @b.elems 09:57
camelia rakudo-moar 906ca0: OUTPUT«6␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«3␤»
09:57 bin_005 joined
laben btw thanks to nine++ spearheading the work, it seems the majority of the glr troubles were done with. the rest is fixing up the warts and the ecosystem 09:57
moritz laben: uhm, there are still ~60 spectest files failing
nine 59 :) 09:58
FROGGS :P
laben yeah but 60 vs 1000+, i would say quite good
masak what's 59 test failures between friends, eh?
moritz not test failures, test *file* failures
lizmat yeah, what moritz said 09:59
moritz many files fail more than one test
laben work is not complete, but it's 94% of the way there
nine according to Coke++'s spec test reports, we're actually 99.68 % there
09:59 colomon joined
tadzik [ptc]: ...so I'm thinking to maybe just make an aggresive decision and make panda refuse to install things that don't have version specified with a git tag to go along with it 10:00
colomon m: <a b b c c c d d d d>.Set.maxpairs
camelia ( no output )
GLRelia rakudo-moar 94e6d1: OUTPUT«Type check failed for return value; expected 'List' but got 'Seq'␤ in block <unit> at /tmp/spiDCiKGF8:1␤␤»
moritz m: say (1..*).map({}).is-lazy
GLRelia rakudo-moar 94e6d1: OUTPUT«Cannot call map(Range: Hash); none of these signatures match:␤ ($: &block, :$label, *%_)␤ (HyperIterable:D $: &block, :$label, *%_)␤ in block <unit> at /tmp/PjPjvvV98g:1␤␤»
colomon m: <a b b c c c d d d d>.Set.maxpairs.say
GLRelia rakudo-moar 94e6d1: OUTPUT«Type check failed for return value; expected 'List' but got 'Seq'␤ in block <unit> at /tmp/b8EtbDYBlf:1␤␤»
moritz m: say (1..*).map({;}).is-lazy
GLRelia rakudo-moar 94e6d1: OUTPUT«True␤»
camelia rakudo-moar 906ca0: OUTPUT«(timeout)»
rakudo-moar 906ca0: OUTPUT«a => True c => True b => True d => True␤»
rakudo-moar 906ca0: OUTPUT«Method 'is-lazy' not found for invocant of class 'List'␤ in block <unit> at /tmp/jqe4gt54S6:1␤␤»
dalek ast/glr: 1ea8920 | lizmat++ | S02-types/baghash.t:
Fix one BagHash test
 10:01
colomon anyone object to me changing the signature for minpairs and maxpairs in GLR to return Seq instead of List? 10:02
hoelzro I know there has been a lot of discussion about flattening in the GLR; is the flattening of slurpy arguments on the chopping block?
lizmat colomon: I would prefer losing the return sig 10:03
hoelzro this fellow here: blogs.perl.org/users/aaron_baugher/...erl-6.html
lizmat for now
hoelzro used run(< ffmpeg -i >, $k, < -ab 96k >, "new/%v<newfile>", :out )
10:03 laouji left
colomon lizmat: works for me. 10:03
hoelzro and I'm wondering if the <...> will still flatten
(they currently do, and from my understanding of the GLR, they should still, but things seem to be in flux)
10:03 dayangkun left
colomon lizmat: what are those methods meant to do, anyway? 10:03
moritz m: sub f(*@s) { say @s.perl }; f < ffmpeg -i >, $k, < -ab 96k >, "new/%v<newfile>"
camelia rakudo-moar 906ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/UvArV1cTw0␤Variable '$k' is not declared␤at /tmp/UvArV1cTw0:1␤------> 3(*@s) { say @s.perl }; f < ffmpeg -i >, 7⏏5$k, < -ab 96k >, "new/%v<newfile>"␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jhqp50Tfjr␤Variable '$k' is not declared␤at /tmp/jhqp50Tfjr:1␤------> 3(*@s) { say @s.perl }; f < ffmpeg -i >, 7⏏5$k, < -ab 96k >, "new/%v<newfile>"␤»
moritz m: sub f(*@s) { say @s.perl }; f < ffmpeg -i >, (my $k = 2), < -ab 96k >, "new/%v<newfile>" 10:04
camelia rakudo-moar 906ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/mRj1jbJB_o␤Variable '%v' is not declared␤at /tmp/mRj1jbJB_o:1␤------> 3eg -i >, (my $k = 2), < -ab 96k >, "new/7⏏5%v<newfile>"␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«5===SORRY!5=== Error while compiling /tmp/DHmEycVwT5␤Variable '%v' is not declared␤at /tmp/DHmEycVwT5:1␤------> 3eg -i >, (my $k = 2), < -ab 96k >, "new/7⏏5%v<newfile>"␤»
lizmat colomon: not sure... watching jnthn now
moritz m: sub f(*@s) { say @s.perl }; f < ffmpeg -i >, (my $k = 2), < -ab 96k >, "new/bla"
camelia rakudo-moar 906ca0: OUTPUT«["ffmpeg", "-i", 2, "-ab", "96k", "new/bla"]<>␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«["ffmpeg", "-i", 2, "-ab", "96k", "new/bla"]␤»
nine hoelzro: we only flatten implicitly if you pass a container as a single argument to a function. And only if it's not itemized.
10:04 lizmat left
colomon lizmat: he’s not talking that fast. ;) 10:04
nine hoelzro: in all other cases you have to flatten manually.
moritz nine: but here it's the slurpy that flattens
hoelzro indeed 10:05
nine And from all the spec tests I've fixed where things relied on implicit flattening, this is a very good thing. Sometimes things didn't really work for the reasons the programmer thought they did. Often it was very confusing to find out what was really going on.
dalek ast/glr: c10975b | moritz++ | S32-list/map.t:
GLRify S32-list/map.t
 10:06
moritz 58 :-)
hoelzro should slurpies still implicitly flatten, then?
nine hoelzro: really only if they slurp a single container. Think @a.push(@new-elems); 10:07
hoelzro: versus @a.push($@sublist)
moritz why not? There's always ** for when you don't it
laben can someome explain me the differences between make test, make spectest and make localtest etc?
hoelzro laben: make test is a quick compiler sanity check 10:08
10:09 |Tux| left
hoelzro make spectest tests rakudo against the langauge specification tests 10:09
and make localtest I've just heard of, from you =)
colomon m: say Set.new([1, 2], 3)
camelia rakudo-moar 906ca0: OUTPUT«set(1 2, 3)␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«312␤»
moritz laben: 'spectest' uses t/spectest.data as a list of test files 10:10
laben: 'localtest' uses t/localtest.data, which is not under version control, so you can fill it with your own set of files
hoelzro .oO( TIL about localtest! ) 10:11
laben thanks moritz! are there other kind of make *test?
hoelzro laben+ moritz++
laben: make fulltest
tadzik make stresstest
hoelzro laben++
colomon laben: make t/spec/S02-types/set.t
(for instance)
laben and what do they do "fulltest" and "stresstest"? 10:12
tadzik iirc fulltest is spectest + stresstest?
makefile probably knows :)
laben sorry my Makefile fu is not good
oh
moritz laben: some files are marked as "stress" in t/spectests.data, because they take rather long
tadzik oh, no
10:13 |Tux| joined
tadzik fulltest is make test + make stresstest, from what I see 10:13
moritz and "stresstest" includes them, while normal "make spectest" excludes them
10:13 lizmat joined
moritz "fullltest" runs "test" and "stresstest" 10:14
colomon m: say Set.new($[1, 2], 3)
camelia rakudo-moar 906ca0: OUTPUT«set(3, (Any))␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«[1 2]3␤»
colomon huh
moritz m: say Set.new($[1, 2], 3).^name 10:15
camelia rakudo-moar 906ca0: OUTPUT«Set␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«Set␤»
moritz m: say Set.new($[1, 2], 3).gist
GLRelia rakudo-moar 94e6d1: OUTPUT«set(3, [1 2])␤»
camelia rakudo-moar 906ca0: OUTPUT«set(3, (Any))␤»
laben thanks again moritz! now i'm ready to enter the world of test-making ^.^
moritz colomon: it seems the slurpy in sub say(*@) flattens out the set
m: say Set ~~ Iterable
camelia rakudo-moar 906ca0: OUTPUT«False␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«False␤»
moritz m: say Set ~~ Positional 10:16
camelia rakudo-moar 906ca0: OUTPUT«False␤»
GLRelia rakudo-moar 94e6d1: OUTPUT«False␤»
colomon moritz: preivously Set.new([1, 2], 3) does what Set.new($[1, 2], 3) does now.
right, flattening
moritz m: sub f(*@s) { say @s.perl }; f Set.new([1, 2], 3)
GLRelia rakudo-moar 94e6d1: OUTPUT«[set(3,1,2)]␤»
moritz uhm, no
colomon hmm?
moritz the slurpy doesn't flatten it out 10:17
timotimo will there be a recording of the q&a?
lizmat timotimo: yes, there was
timotimo phew.
i'm not 100% screwed
tadzik we will never forget
moritz colomon: I don't understand why &say pretends the set() isn't there
timotimo the only question i could have thought of was "how did you reach the conclusion that you - out of all the people - would get the colon?"
colomon moritz: ah 10:18
lizmat colomon moritz : I'm looking at set/bag/mix behaviour in glr
we probably need a decision from TimToady
colomon lizmat: okay.
lizmat m: say Set.new(<a b c>,<d e>) # is this correct?
GLRelia rakudo-moar 94e6d1: OUTPUT«acbed␤»
colomon lizmat: any objections to my checking in the updated signatures (per your suggestion) for minpairs/maxpairs? 10:19
lizmat no objection from me :-)
moritz m: say Set.new(<a b c>,<d e>).perl
GLRelia rakudo-moar 94e6d1: OUTPUT«set("a","c","b","e","d")␤»
moritz seems rather inconsistent with the rest of GLR
lizmat yes
however, my first try at fixing this, resulted in some strange issues 10:20
gist.github.com/lizmat/72a9b56ea23f4da02363 10:21
tadzik hoelzro: tadzik uses Pull Request! It's super-effective! 10:22
hoelzro Bug has fainted!
moritz is this some cool new technology from github?
tadzik we gamified it :)
yay 10:23
hoelzro tadzik++
tadzik: are you just going through random ecosystem modules?
10:23 |Tux| left
tadzik hoelzro: I started from HTTP::UserAgent 10:24
that got me to File::Temp
and then I wanted a repl :)
hoelzro ah ha =)
thanks for the fix
moritz sounds like Yaks all the way down 10:25
lizmat moritz: yeah 10:26
10:26 |Tux| joined 10:27 pnu left, espadrine left, braintwist left, dalek left, [Tux] left, skarn left, SHODAN left, synbot6 left, hobbs left, krakan left, camelia left, larion left, AlexDaniel left, zakharyas left, Woodi left, Peter_R left, abaugher left, nowan left, Gardner left, jferrero left, JimmyZ left, pink_mist left, colomon left, bin_005 left 10:28 lizmat left, colomon joined 10:29 pnu joined, larion joined, AlexDaniel joined, zakharyas joined, Woodi joined, Peter_R joined, abaugher joined, nowan joined, Gardner joined, jferrero joined, pink_mist joined, JimmyZ joined, Khisanth joined, kst joined, erdic joined, psch joined, apejens joined, mathw joined, El_Che joined 10:30 espadrine joined, braintwist joined, dalek joined, [Tux] joined, skarn joined, SHODAN joined, synbot6 joined, hobbs joined, krakan joined, camelia joined, mls joined, oha joined, ruoso joined, dg joined, timbunce joined, weber.freenode.net sets mode: +vv dalek camelia, colomon left, espadrine left, braintwist left, dalek left, [Tux] left, skarn left, SHODAN left, synbot6 left, hobbs left, krakan left, camelia left, mls left, oha left, ruoso left, dg left, timbunce left 10:33 espadrine joined, braintwist joined, [Tux] joined, skarn joined, SHODAN joined, synbot6 joined, hobbs joined, krakan joined, camelia joined, mls joined, oha joined, ruoso joined, dg joined, timbunce joined, weber.freenode.net sets mode: +v camelia, dalek joined, ChanServ sets mode: +v dalek
nine m: my @a = 1,2; my @b = 1,2,3,4; sub test($a, $b) {say $a.perl; say $b.perl;}; test(@a, @b); 10:34
GLRelia rakudo-moar 64db54: OUTPUT«[1, 2]␤[1, 2, 3, 4]␤»
camelia rakudo-moar 906ca0: OUTPUT«[1, 2]␤[1, 2, 3, 4]␤»
tadzik ha, looks like HTTP::UA works on GLR
nine m: my @a = 1,2; my @b = 1,2,3,4; sub test($a is copy, $b) {say $a.perl; say $b.perl;}; test(@a, @b);
camelia rakudo-moar 906ca0: OUTPUT«[1, 2]␤[1, 2, 3, 4]␤»
GLRelia rakudo-moar 64db54: OUTPUT«[1, 2]␤[1, 2, 3, 4]␤»
dalek ast: 5e66cf2 | hoelzro++ | S32-io/copy.t:
Test copy() when destination exists
 10:35
ast: 5a81fde | hoelzro++ | S32-io/copy.t:
Add tests for copy(:createonly)
nine m: my @a = 1,2; my @b = 1,2,3,4,5; use nqp; say nqp::getattr(Range.new(@a, @b), Range, <$!max>).perl; 10:36
camelia rakudo-moar 906ca0: OUTPUT«[1, 2, 3, 4, 5]␤»
GLRelia rakudo-moar 64db54: OUTPUT«[1, 2, 3, 4, 5]␤»
10:38 Fleurety joined
dalek ast: c4385c4 | hoelzro++ | S32-io/copy.t:
Change copy(:createonly) test to check for X::IO::Copy
 10:40
10:44 Averna joined, Averna left
tadzik huh, we have 'authors' defined in S22, but most of the ecosystem uses 'author' 10:44
dalek kudo/nom: 17a7747 | hoelzro++ | src/core/io_operators.pm:
Add :createonly to sub copy()
tadzik arglebargle
10:45 telex left
tadzik also Please note that for identification purposes, only the owner (who uploaded the distribution to the content storage) should be used 10:45
so why even have a list there?
10:45 TEttinger left
tadzik oh, the owner *field* 10:45
...which nobody uses :)
oh, nvm
okay, I got de-confused 10:46
10:46 telex joined 10:47 Averna joined, adu left 10:48 diana_olhovik_ left
dalek kudo/glr: 30cc6a3 | (Stefan Seifert)++ | src/core/Range.pm:
Handle lists as right end point of Ranges

  @a..@b should mean @.elems .. @b.elems.
We already had code to explicitly handle a list on the left hand side. That it worked on the right hand side was probably due to some side effect of the implementation that we lost when refactoring. Add code to explicitly handle it once and for all instead.		 10:49
10:51 diana_olhovik_ joined, Ven left
nine 54! 10:54
timotimo after much delay, i'll finally actually head out 10:55
will i get a bit of noms at the venue to replace my missed breakfast?
nine timotimo: noms just started 10:56
timotimo great! 10:59
dalek ast/glr: 0edc6f8 | (Stefan Seifert)++ | S03-operators/repeat.t:
Remove obsolete todo marker. Works on GLR now.
nine 52! 11:00
laben nine: can you check on your machine if S03-metaops/hypers.t has regressed? 11:03
dalek ast/glr: bb702e3 | (Stefan Seifert)++ | S03-operators/assign.t:
Remove another obsolete todo marker. Passes on GLR!
 11:04
laben m: use Test; eval-dies-ok 'my @a >>[=]>> (1,2,3)', "hypering assignment dies correctly";
camelia rakudo-moar 906ca0: OUTPUT«not ok 1 - hypering assignment dies correctly␤␤# Failed test 'hypering assignment dies correctly'␤# at /tmp/FNmySmOsSu line 1␤»
GLRelia rakudo-moar 30cc6a: OUTPUT«not ok 1 - hypering assignment dies correctly␤␤# Failed test 'hypering assignment dies correctly'␤# at /tmp/dww0KnenjM line 1␤»
laben m: use Test; is True «+« (), (), "left-dwim hyper against empty RHS doesn't hang"; 11:05
camelia rakudo-moar 906ca0: OUTPUT«ok 1 - left-dwim hyper against empty RHS doesn't hang␤»
GLRelia rakudo-moar 30cc6a: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/L4WrYbKr5Y:1␤␤»
11:05 FROGGS left
laben m: True <<+<< () 11:06
camelia ( no output )
GLRelia rakudo-moar 30cc6a: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/PcziSd9P7K:1␤␤»
laben m: ().elems
camelia ( no output )
GLRelia ( no output )
laben m: True.elems
camelia ( no output )
GLRelia ( no output )
laben m: my @a >>[=]>> (1,2,3)
GLRelia ( no output )
camelia ( no output )
laben m: my @a >>[=]>> (1,2,3); dd @a 11:07
camelia rakudo-moar 906ca0: OUTPUT«@a = []<>␤»
GLRelia rakudo-moar 30cc6a: OUTPUT«@a = []␤»
11:13 rurban left
nine laben: the elems error is due to my @r; say |@r; 11:17
m: my @r; say |@r;
camelia rakudo-moar 906ca0: OUTPUT«␤»
GLRelia rakudo-moar 30cc6a: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/5It1QKwFCs:1␤␤»
laben due to "say" or to flattening? it errors out even without "say", see "True <<+<< ()" 11:18
m: True <<+<< ()
GLRelia rakudo-moar 30cc6a: OUTPUT«This type does not support elems␤ in block <unit> at /tmp/AL3b2r0YCp:1␤␤»
camelia ( no output )
laben the eval one evals without dieing anymore, someone (stub-)implemented hyper assign? 11:19
m: my @a >>[=]>> (1,2,3); dd @a
camelia rakudo-moar 906ca0: OUTPUT«@a = []<>␤»
GLRelia rakudo-moar 30cc6a: OUTPUT«@a = []␤»
nine laben: due to the |@a when @a is empty 11:21
11:21 lizmat joined
laben i guess i'll wait on that one. what to do on the eval one? change to eval-ok and test that it does not assign anything? 11:24
11:29 Ven joined 11:31 FROGGS joined
FROGGS yay, I think I got 'for Any {' and friends nailed down 11:31
11:33 gfldex joined
tadzik wow 11:33
dalek kudo/glr: 348886d | FROGGS++ | src/core/ (2 files):
make Any.elems==1 and Any.list==(Any)

  ... as discussed with jnhtn and others. This fixes the issue that you were
not able to iterate over a single type object.		 11:34
tadzik panda's META.info now has 41 lines of "authors"
11:34 lizmat left
tadzik github.com/tadzik/panda/blob/maste...TA.info#L7 11:36
that's pure happiness
11:36 domidumont left, larion left
timotimo \o/ 11:36
JimmyZ FROGGS: multi method list(Any:D:) and multi method list(Any:U:) can be just multi method list(Any:)? since the code is same
FROGGS JimmyZ: aye, forgot to change that... I left it in case I needed it, but it can be simplified 11:37
11:38 aborazmeh left
dalek ast/glr: fb33c96 | FROGGS++ | S (3 files):
follow Any.list beeing 1
 11:38
nine FROGGS: btw. I don't see any reason to use infix:<,> to create a List anymore 11:39
11:39 domidumont joined
FROGGS nine: ohh, I just was lazy I guess 11:39
didn't thought about it much
timotimo d'oh, the big room isn't filled very much, but the small one is packed to the brim 11:43
ab5tract++ is filling up the room quite well
11:49 lizmat joined
FROGGS my patch potentially causes regressions in t/spec/S17-supply/classify.t 11:50
but I'm not sure
dalek ast: 5e7d42f | vendethiel++ | S03-metaops/cross.t:
fix cross metaop with nested/itemized lists
nine FROGGS: classify.t has been failing before, too. 11:51
[Coke] The Q&A with larry should eventually end up on youtube, I'll let you know when we have a URL.
(they're double checking audio, etc.)
[Coke] is reminded how empty github.com/coke/mojo6 is. :( 11:52
tony-o why doesn't panda precompile anymore? 11:53
11:54 larion joined
moritz m: my Int %h{Str} = a => 1, b => 2; say %h.pairs.sort.map({.perl(:arglist)}).join(', '); 11:54
camelia rakudo-moar 17a774: OUTPUT«"a" => 1, "b" => 2␤»
GLRelia rakudo-moar 348886: OUTPUT«:a(1), :b(2)␤»
moritz m: my Int %h{Str} = a => 1, b => 2; say %h.perl
camelia rakudo-moar 17a774: OUTPUT«Hash[Int,Str].new("a" => 1, "b" => 2)␤»
GLRelia rakudo-moar 348886: OUTPUT«Hash[Int,Str].new(:a(1), :b(2))␤»
moritz m: my Int %h{Str} = a => 1, b => 2; say EVAL(%h.perl).perl 11:55
camelia rakudo-moar 17a774: OUTPUT«Hash[Int,Str].new("a" => 1, "b" => 2)␤»
GLRelia rakudo-moar 348886: OUTPUT«Hash[Int,Str].new()␤»
Ven github.com/rakudo/rakudo/pull/518
moritz m: my Int %h{Str} = a => 1, b => 2; say EVAL(%h.perl).elems
camelia rakudo-moar 17a774: OUTPUT«2␤»
GLRelia rakudo-moar 348886: OUTPUT«0␤»
moritz m: say Hash.new(:a(1)).perl
camelia rakudo-moar 17a774: OUTPUT«{}<>␤»
GLRelia rakudo-moar 348886: OUTPUT«{}␤»
11:56 abraxxa left
moritz ah, that's the problem; Hash.perl produces named args, but Hash.new doesn't accept named args 11:56
dalek kudo/glr: eda95ad | vendethiel++ | src/core/metaops.pm:
make cross metaop respect itemization
 11:58
kudo/glr: c68db93 | niner++ | src/core/metaops.pm:
Merge pull request #518 from vendethiel/glr

make cross metaop respect itemization
moritz TimToady: ^^ the Hash.perl regression seems to be your doing 11:59
tony-o yea precomp only sped up parsing by a second, damn.
dalek osystem: e570cb3 | (Tadeusz “tadzik” Sośnierz)++ | META.list:
Retire IO::Select
 12:01
tadzik oh now dalek doesn't know who I am
where do we fix this?
timotimo parsing what exactly? 12:02
tony-o timotimo: wrong channel, seeing what performance i can get on a from-json method 12:03
timotimo ah 12:04
this is about json::fast, too? 12:05
or are you trying a different approach?
tony-o just trying some different approaches
not sure i can beat json::fast without nativecall, though
timotimo there's still one thing that's bad about json::fast which is that the loop in nom_ws allocates a BOOTCode every time it starts 12:07
which means it doesn't get inlined
12:08 Ven left
JimmyZ timotimo: Moarvm has a roadmap about it 12:10
FROGGS what is the current test file count?
12:11 Loren joined 12:14 rindolf left
nine I think 50 or 51 12:15
12:15 Loren_ joined, Averna left
nine Fixing List.ACCEPTS may raise that number though :( 12:15
12:17 lizmat left 12:18 Loren left 12:20 blackcat_ joined
FROGGS huh: 12:21
t/spec/integration/precompiled.rakudo.moar (Wstat: 0 Tests: 7 Failed: 0)
TODO passed: 6
I even think that was my patch... as it turned up before I pulled other patches 12:22
12:23 Loren joined, Loren_ left
nine FROGGS: I see nothing wrong with fixing TODOs :) 12:24
FROGGS aye :o)
12:25 blackcat_ left 12:27 tinyblak joined 12:31 tinyblak left 12:32 coffee` left
nine If I give List a real ACCEPTS method, S05-modifier and S05-substitution tests start to fail. 12:34
12:34 lizmat joined
moritz FROGGS: fwiw the precompiled.t test is a flapper here 12:34
12:34 lizmat left
moritz (the passing TODO) 12:34
FROGGS ohh, sad 12:35
also this?
t/spec/S09-typed-arrays/arrays.rakudo.moar (Wstat: 0 Tests: 77 Failed: 0)
TODO passed: 47, 63
ohh, and: 12:36
t/spec/S04-phasers/begin.rakudo.moar (Wstat: 0 Tests: 13 Failed: 0)
TODO passed: 13
12:36 brrt joined, Ven joined
timotimo JimmyZ: that's not a moarvm problem, really; this is about having an intlexref preventing the block from being inlined 12:37
which i've been trying to improve, but i've been failing, too :|
JimmyZ you mean lex => local optimization? 12:38
timotimo yes
12:39 Ven left
JimmyZ that looks like will reduce the core setting size 12:42
timotimo it is likely to, yeah, but it's also a big performance win for tight loops 12:43
JimmyZ agrees
12:44 Ven joined
dalek ast/glr: f5c11d0 | (Stefan Seifert)++ | S09-typed-arrays/arrays.t:
Remove obsolete TODO markers in S09-typed-arrays/arrays.t

Thanks to FROGGS++ for pointing this out!		 12:47
nine m: use Test; my $bound; BEGIN { $bound := 'foo'; } is $bound, 'foo'; 12:48
camelia rakudo-moar 17a774: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Ud_rvyt5Yw␤Strange text after block (missing semicolon or comma?)␤at /tmp/Ud_rvyt5Yw:1␤------> 3t; my $bound; BEGIN { $bound := 'foo'; }7⏏5 is $bound, 'foo';␤ expecting any of:␤ infix…»
GLRelia rakudo-moar c68db9: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PZTttnvq96␤Strange text after block (missing semicolon or comma?)␤at /tmp/PZTttnvq96:1␤------> 3t; my $bound; BEGIN { $bound := 'foo'; }7⏏5 is $bound, 'foo';␤ expecting any of:␤ infix…»
nine m: use Test; my $bound; BEGIN { $bound := 'foo'; }; is $bound, 'foo';
camelia rakudo-moar 17a774: OUTPUT«not ok 1 - ␤␤# Failed test at /tmp/5YhQoM8lCY line 1␤# expected: 'foo'␤# got: (Any)␤»
GLRelia rakudo-moar c68db9: OUTPUT«not ok 1 - ␤␤# Failed test at /tmp/KfG5bMQIdu line 1␤# expected: 'foo'␤# got: (Any)␤»
12:48 Loren_ joined
nine I don't know why this passes in the spec test file, but it obviously is not fixed yet. 12:49
12:51 xfix joined, xfix left, xfix joined, Loren left 12:56 larion left
nine m: my @a = any(3, 4); say @a.perl; 12:56
camelia rakudo-moar 17a774: OUTPUT«[any(3, 4)]<>␤»
GLRelia rakudo-moar c68db9: OUTPUT«[4]␤»
[Coke] O_o 12:57
12:57 larion joined
nine m: my @a = (any(3, 4),); say @a.perl; 12:57
camelia rakudo-moar 17a774: OUTPUT«[any(3, 4)]<>␤»
GLRelia rakudo-moar c68db9: OUTPUT«[any(3, 4)]␤»
nine m: say any(3, 4).item; 12:58
camelia rakudo-moar 17a774: OUTPUT«any(3, 4)␤»
GLRelia rakudo-moar c68db9: OUTPUT«(timeout)»
nine That's even more worrying
12:58 adu joined
nine m: say any(3, 4).item.perl; 12:58
camelia rakudo-moar 17a774: OUTPUT«any(3, 4)␤»
GLRelia rakudo-moar c68db9: OUTPUT«any(3, 4)␤»
12:59 adu left, skids joined, rurban joined
ShimmerFairy m: my @a = any(3, 4); say @a.perl; 13:03
camelia rakudo-moar 17a774: OUTPUT«[any(3, 4)]<>␤»
GLRelia rakudo-moar c68db9: OUTPUT«[4]␤»
dalek pan style="color: #395be5">perl6-examples: 8a9c60d | paultcochrane++ | / (3 files):
Add --no-highlight option to htmlify

Which passes information about whether or not to run syntax highlighting on to the website object which builds the website.		 13:04
pan style="color: #395be5">perl6-examples: 19b56c7 | paultcochrane++ | / (2 files):
Turn off syntax highlighting if --no-highlighting is set
pan style="color: #395be5">perl6-examples: 7c832ab | paultcochrane++ | Makefile:
Add html-nohighlight make target

This generates the html without syntax highlighting to speed things up for development work.
pan style="color: #395be5">perl6-examples: 133f103 | paultcochrane++ | lib/Pod/Htmlify.pm6:
Swap back to previous category sorting code
pan style="color: #395be5">perl6-examples: 6ffeb8a | paultcochrane++ | lib/Pod/Htmlify.pm6:
Sort lists of subcategories by key when generating the index page
[Coke] hack is MUCH faster today. Probably because everyone is in the talks. 13:06
[ptc] _itz: sorting in example subcategories should work again now. The change probably happened after hashes lost their implicit ordering (that was a while ago now)
timotimo heh.
13:08 muraiki joined
[Coke] m: my ($foo, $bar = "baz") = <a>; say "$foo $bar"; 13:09
camelia rakudo-moar 17a774: OUTPUT«Use of uninitialized value $bar of type Any in string context in block <unit> at /tmp/zhNVztssuh:1␤a ␤»
GLRelia rakudo-moar c68db9: OUTPUT«Use of uninitialized value $bar of type Any in string context in block <unit> at /tmp/Wrqq6JMeKx:1␤a ␤»
13:10 Loren_ left
[Coke] m: enum A (b => 42, <c d e>); say A.enums.perl 13:12
camelia rakudo-moar 17a774: OUTPUT«{:b(42), "c d e" => 43}<>␤»
GLRelia rakudo-moar c68db9: OUTPUT«{:b(42), "c d e" => 43}␤»
[Coke] ^^ does that look acceptable?
RT #115052
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=115052
13:13 smls joined 13:17 plicease left, plicease joined, Ven left
moritz [Coke]: I agree with RT #123191 that it should flatten the list, i.e. produces b => 42, c => 43, d => 43, e => 44 13:18
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=123191
[Coke] moritz: can I merge those tickets? 13:19
(I see they are referring to each other) 13:20
moritz [Coke]: no, I'd simply close RT #115052
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=115052
moritz [Coke]: the error doesn't appear anymore
brrt javascript is becoming more.. .perly, apparantly 13:21
v8project.blogspot.nl/2015/08/v8-release-46.html 13:22
skids Reminder for hackathon folks, when you get to unflattenning the sub forms in Any-iterable, see gist.github.com/skids/4f6ff2e1c1e3b6855773 it fixes some stuff that can be tricky to find.
I never sent a PR for the brach I had where I was doing that, because it was pre-GLRR.
FROGGS m: my $x = do if 0 { 1 } elsif 0 { 2 }; say $x 13:23
camelia rakudo-moar 17a774: OUTPUT«Empty␤»
GLRelia rakudo-moar c68db9: OUTPUT«␤»
FROGGS m: my $x = do if 0 { 1 } elsif 0 { 2 }; say $x.defined
camelia rakudo-moar 17a774: OUTPUT«False␤»
GLRelia rakudo-moar c68db9: OUTPUT«True␤»
FROGGS m: my $x = do if 0 { 1 } elsif 0 { 2 }; say $x.WHAT
camelia rakudo-moar 17a774: OUTPUT«Empty␤»
GLRelia rakudo-moar c68db9: OUTPUT«(Slip)␤»
FROGGS should this be Nil by any chance?
13:24 larion left
moritz how do I iterate a Perl 6 Iterable from NQP? 13:26
FROGGS uff 13:28
moritz this is for rt.perl.org/Public/Bug/Display.html?id=123191 where we should flatten a sublist 13:32
nine skids! 13:33
skids o/
13:34 bloonix joined
FROGGS m: for 1..2 { FIRST { say "hello $_" } } 13:34
camelia rakudo-moar 17a774: OUTPUT«hello 1␤»
GLRelia ( no output )
13:34 virtualsue joined
nine skids: I found out that our List .ACCEPTS just about everything. If I fix that by giving it an implementation the S05-modifier tests start failing. Can you have a look? 13:34
13:35 Loren_ joined
skids nine: I'm at $DAYJOB, unfortunately. 13:35
13:37 lizmat joined
nine skids: any idea where I could look at? 13:42
13:44 Ven joined
skids nine: Well there are a couple smartmatches in there but they should only happen to Ranges. 13:45
dalek c: 2e91532 | hoelzro++ | lib/Language/unicode_entry.pod:
Mention Vim's unicode digraphs
 13:46
skids nine: which tests fail? 13:47
13:51 lizmat left
timotimo phew, it's really rather warm 13:52
skids nine: I'd look at '$multi = Positional.ACCEPTS($nth);' which is supposed to determine which code to run at the bottom. (e.g. detect :nth(3) vs :nth(2,3) or :nth(2,4...*)) 13:53
13:56 virtualsue left 14:07 coffee` joined
dalek kudo/nom: 97ae81f | hoelzro++ | src/core/Exception.pm:
Add X::IO::Move
 14:09
kudo/nom: eb6de6d | hoelzro++ | src/core/ (2 files):
Implement IO.move() and move()
ast: 66de147 | hoelzro++ | S32-io/move.t:
Add test for move() and IO.move()
14:10 andreoss joined
nine skids: ok('abab' ~~ m:2x/ab/, ':2x (repetition) modifier (+)'); 14:10
andreoss m: my @x = do for 1..10 { start { } }; map await, @x
camelia rakudo-moar 17a774: OUTPUT«Must specify a Promise or Channel to await on␤ in block <unit> at /tmp/GXm1XcYyU6:1␤␤»
GLRelia rakudo-moar c68db9: OUTPUT«Must specify a Promise or Channel to await on (got an empty list)␤ in block <unit> at /tmp/OV08Wr31k_:1␤␤»
andreoss where does this empty list come from? 14:11
masak phew, talks are over.
correction: I don't have to give more talk :) 14:12
14:12 Sqirrel left
nine skids: Positional.ACCEPTS?? But Positional is just a Role and an almost empty at that. 14:13
skids nine: ok then not that ACCEPTS line it does not get hit without :nth. The line (0..Inf).Array would be hit only when :x but no :nth. 14:14
nine Ah, timotimo++ cleared it up. Positional.ACCEPTS($nth) is just a different way to write nqp::istype($nth, Positional) 14:15
andreoss m: my @z = 1..10; my @x = do while my @z.shift { start { $_ } }; say await @x 14:16
camelia rakudo-moar 17a774: OUTPUT«Potential difficulties:␤ Redeclaration of symbol @z␤ at /tmp/TWd0nXQ198:1␤ ------> 3my @z = 1..10; my @x = do while my @z7⏏5.shift { start { $_ } }; say await @x␤␤»
GLRelia rakudo-moar c68db9: OUTPUT«Potential difficulties:␤ Redeclaration of symbol @z␤ at /tmp/xncxmTdYzt:1␤ ------> 3my @z = 1..10; my @x = do while my @z7⏏5.shift { start { $_ } }; say await @x␤␤»
nine I think, I start to understand what's happening. match returns a list in those nth-cases. That's why implementing ACCEPTS on List changes the outcome.
andreoss m: my @z = 1..10; my @x = do while @z.shift { start { $_ } }; say await @x
camelia rakudo-moar 17a774: OUTPUT«␤»
GLRelia rakudo-moar c68db9: OUTPUT«␤»
14:17 andreoss left
skids nine: Yes and it should hit the listgen code for anything with an :x as well. The above test does not have an :nth so if it is the only one failing I'd look at (0..Inf).Array. 14:18
14:18 andreoss joined
andreoss m: my @z = 1..10; my @x = do while my $u = @z.shift { start { $u } }; await @x 14:19
camelia ( no output )
GLRelia ( no output )
hoelzro m: ''.substr(0, *-16)
camelia ( no output )
GLRelia ( no output )
hoelzro m: say(''.substr(0, *-16))
camelia rakudo-moar 17a774: OUTPUT«Number of characters argument to substr out of range. Is: -16, should be in 0..Inf; use *-16 if you want to index relative to the end␤ in block <unit> at /tmp/0VjtwxJfe7:1␤␤»
GLRelia rakudo-moar c68db9: OUTPUT«Number of characters argument to substr out of range. Is: -16, should be in 0..Inf; use *-16 if you want to index relative to the end␤ in block <unit> at /tmp/NksYpL0ufB:1␤␤»
nine Soo...is it possible, that nth does not actually work all that well?
hoelzro ^ is that a bug?
14:19 lizmat joined
skids nine: (not that I understand the bottom part of that method too well :-) 14:19
nine m: say "foo" ~~ m:2x/foo/;
camelia rakudo-moar 17a774: OUTPUT«␤»
GLRelia rakudo-moar c68db9: OUTPUT«␤»
masak m: class C { has $.foo is rw where { $++ < 2 } }; given C.new { .foo = 1; say .foo; .foo = 2; say .foo; .foo = 3 }
camelia rakudo-moar 17a774: OUTPUT«1␤2␤Type check failed in assignment to '$!foo'; expected '<anon>' but got 'Int'␤ in block <unit> at /tmp/U0mQkWTVfZ:1␤␤» 14:20
GLRelia rakudo-moar c68db9: OUTPUT«1␤2␤Type check failed in assignment to '$!foo'; expected '<anon>' but got 'Int'␤ in block <unit> at /tmp/_93IAw7zwk:1␤␤»
masak hee hee
state variables in where blocks... :)
nine Aaaaaah...now it makes sense!
List had the following ACCEPTS method: multi method ACCEPTS(List:D:, $topic) { self }
timotimo ... how does it make sense? 14:21
does having that method make sense? or does behavior caused by this make sense now?
nine Str.match will return a list of matches for the nth-stuff. If this list was empty, List.ACCEPTS would boolify to False. If it contained Matches (because the match was successfull) it boolifies to True.
TimToady List is one of the distinguished "result" types that prevent a redispatch to .ACCEPTS
timotimo oooooh
masak wow, I didn't know about that 14:22
TimToady like True, False
timotimo i remember something about that
JimmyZ m: class C { has $.foo is rw where { $++ < 2 } }; given C.new { .foo = 3; say .foo; .foo = 5; say .foo; .foo = 3 }
GLRelia rakudo-moar c68db9: OUTPUT«3␤5␤Type check failed in assignment to '$!foo'; expected '<anon>' but got 'Int'␤ in block <unit> at /tmp/BciYcQ8pDD:1␤␤»
camelia rakudo-moar 17a774: OUTPUT«3␤5␤Type check failed in assignment to '$!foo'; expected '<anon>' but got 'Int'␤ in block <unit> at /tmp/QpoUT_f16t:1␤␤»
nine TimToady: this messes up (1, 2, 3) ~~ (1, 2, 3) now
timotimo the number of times we typecheck isn't guaranteed, though
14:23 Sqirrel joined
smls Is there a list versio of join? 14:23
TimToady m: say (1,2,3) ~~ [1,2,3]
camelia rakudo-moar 17a774: OUTPUT«True␤»
GLRelia rakudo-moar c68db9: OUTPUT«True␤»
TimToady list matching wants its own pattern language anyway
tadzik smls: list version of join?
timotimo what does "list version of join" mean?
yeah
nine TimToady: or rather (1, 2, 3) ~~ (4, 5) 14:24
m: say (1, 2, 3) ~~ (4, 5)
camelia rakudo-moar 17a774: OUTPUT«False␤»
GLRelia rakudo-moar c68db9: OUTPUT«45␤»
TimToady m: say (1,2,3) ~~ [1,2]
camelia rakudo-moar 17a774: OUTPUT«False␤»
GLRelia rakudo-moar c68db9: OUTPUT«False␤»
smls i.e. joining (2, 4, 5) with 0 --> 2, 0, 4, 0, 5
timotimo looks a bit like a zip + flatten
well, and you'd hack off the last bit
smls except zip adds an extra one at the start or end
TimToady you can match using an array currently, but we want something like (1,2,3) ~~ pattern 1,2,3,*
timotimo m: my @in = 2, 4, 6; my @our = flat (@in X 0); say @our.perl 14:25
camelia rakudo-moar 17a774: OUTPUT«[2, 0, 4, 0, 6, 0]<>␤»
GLRelia rakudo-moar c68db9: OUTPUT«[2, 0, 4, 0, 6, 0]␤»
timotimo i meant X rather than Z btw
smls trying to find the best way to generate an argument list of the form \( -name "*.txt" -o -name "*.htm" \) for pasing to the Linux `find` command
(but kept as a list so it can be called without shell quoting) 14:26
14:26 szabgab joined
masak m: say (1, 2, 3) ~~ (1, 2, 3) 14:26
camelia rakudo-moar 17a774: OUTPUT«True␤»
GLRelia rakudo-moar c68db9: OUTPUT«123␤»
14:26 khw joined
nine TimToady: then I will change the S03-smartmatch/array-array.t tests to match against Array instead of List 14:26
szabgab good afternoon
timotimo heyo szabgab 14:27
masak szabgab! \o/
timotimo i remember having results from multi-match was rather hairy
14:27 lizmat left
tadzik hey szabgab! 14:27
geekosaur if you're avoiding the shell then it's just '(', '-name', '*.txt', '-o', '-name', '*.htm', ')'
TimToady well, we really do want to steal the wildcarding and give it to a special list pattern type
szabgab is there going to be a new rakudo * release soon? I'd like to build environment for my YAPC::EU talk
and I wonder if I should use 2015.07 or wait for 08 ? 14:28
masak was there a compiler release yesterday?
szabgab masak: are you asking me? 14:29
I have no idea
or was that a rhetorical question?
geekosaur general channel question
masak it was a question to the channel
but it was connected to your question :)
skids m: (gather for 1,2,3 { once { take $_; next }; take 0; take $_; }).perl.say 14:30
camelia rakudo-moar 17a774: OUTPUT«(1, 0, 2, 0, 3)␤»
GLRelia rakudo-moar c68db9: OUTPUT«(1, 0, 2, 0, 3).Seq␤»
nine TimToady: if I use Array.ACCEPTS' logic in List for Iterable $topics and the old ?self for everything else, we pass S03-smartmatch/array-array.t without any regressions. 14:32
timotimo i know moritz worked towards a release
masak # Announce: Rakudo Perl 6 compiler, Development Release #91 ("TBD") 14:33
so not fully released.
dalek ast: 87d1e68 | alexghacker++ | S15-nfg/cgj.t:
Create cgj.t
ast: 79585f0 | alexghacker++ | S15-nfg/cgj.t:
Update cgj.t

Add comments with references for the behavior being tested for
ast: 8bb05dc | (Zoffix Znet)++ | S15-nfg/cgj.t:
Merge pull request #66 from alexghacker/master

Create tests for the COMBINING GRAPHEME JOINER special case
masak szabgab: coming back to your question, I don't see why we wouldn't have a star release after the compiler release. but we haven't had the compiler release just yet. 14:34
szabgab ok
dalek c: 7678c0f | hoelzro++ | lib/Type/IO/Path.pod:
Remove not implemented warning for move()

Because now it *is* implemented!		 14:35
c: 12172fc | hoelzro++ | lib/Type/IO/Path.pod:
Document copy()
masak (I don't usually make the star release, so this is a guess, not a promise) :)
skids m: (flat (slip(),0,0...*) Z, (1,2,3)).perl.say # smls 14:36
GLRelia rakudo-moar c68db9: OUTPUT«(0, 1, 0, 2, 0, 3)␤»
camelia rakudo-moar 17a774: OUTPUT«5===SORRY!5=== Error while compiling /tmp/XlZL0oeqHq␤Undeclared routine:␤ slip used at line 1. Did you mean 'flip'?␤␤»
timotimo if you made a promise, szabgab could non-blockingly wait, rather than having to poll :)
szabgab and this will be a star release using panda with breaks
:)
masak 'panda with breaks'?
tadzik panda's broken?
timotimo "brakes"
i bet :P
masak 'brakes'?
szabgab brakes
right
tadzik oh:) yeah
masak like, on a bike?
timotimo yes 14:37
szabgab tadzik knows :)
tadzik a slowdown so shocking you'll hit the pavement with your face
timotimo poor pavement
tadzik yeah, take that!
szabgab: thought I hope that it'll make updating panda on Star much more reliable 14:38
which I see is a pain for a while
smls skids: How come replacing 0,0...* with (0 xx *) does not work in that example?
(hangs indefinitely)
skids smls: the whole thing is one sequence and gets flattenned due to first arg rule. 14:39
(the slip is part of the ...)
smls but isn't flattening lazy?
14:40 Ven left
hoelzro m: say(''.substr(0, *-16)) 14:41
GLRelia rakudo-moar c68db9: OUTPUT«Number of characters argument to substr out of range. Is: -16, should be in 0..Inf; use *-16 if you want to index relative to the end␤ in block <unit> at /tmp/k8j5ZO1v7c:1␤␤»
camelia rakudo-moar 17a774: OUTPUT«Number of characters argument to substr out of range. Is: -16, should be in 0..Inf; use *-16 if you want to index relative to the end␤ in block <unit> at /tmp/0a7smABWrP:1␤␤»
hoelzro ^ is that a bug?
szabgab I think I move myself to the background again and check the web site for announcement 14:42
see some of you in Granada!
14:43 szabgab left 14:44 lizmat joined
lizmat gist.github.com/lizmat/72a9b56ea23f4da02363 14:44
TimToady hoelzro: looks LTA to me 14:45
hoelzro TimToady: I'm thinking it should result in an empty string, does that same ok?
14:46 tinyblak joined
TimToady we've been treating off-the-beginning as an error, but it shouldn't recommend what you just said... 14:47
14:48 colomon joined
skids smls: 1,0..* Z, (1,2,3) would produce (1,1,0..*,2) which is then fed to flat. Which should be lazy but maybe it is eagerized after that somehow. 14:48
smls skids: xx not .. 14:49
(though I guess the same applies?)
skids right, sorry. 14:50
colomon lizmat: have you managed to talk with TimToady about set / bag construction?
colomon is just now back on the Internet. 14:51
14:53 itz joined, yqt joined
hoelzro TimToady: what about arrays? should my @a; @a[0..*-16] doesn't result in an error 14:53
_itz [ptc]: yay! that looks great again 14:54
nine m: say (1,2,3) ~~ (4,5)
GLRelia rakudo-moar c68db9: OUTPUT«45␤»
camelia rakudo-moar eb6de6: OUTPUT«False␤»
TimToady hoelzro: that's just a null range
nine m: say (1,2,3) ~~ (1,2,3)
camelia rakudo-moar eb6de6: OUTPUT«True␤»
GLRelia rakudo-moar c68db9: OUTPUT«123␤»
hoelzro ahhhh
14:54 tinyblak left
hoelzro that makes sense 14:54
TimToady m: say(''.substr(0 .. *-16)) 14:55
GLRelia rakudo-moar c68db9: OUTPUT«Type check failed in assignment to '$from'; expected 'Int' but got 'Range'␤ in block <unit> at /tmp/Ybkh9OSlHY:1␤␤»
camelia rakudo-moar eb6de6: OUTPUT«Type check failed in assignment to '$from'; expected 'Int' but got 'Range'␤ in block <unit> at /tmp/0Yh8hsXKLB:1␤␤»
andreoss is there something soft close on channels? like saying to it that this element is the last one?
14:55 virtualsue joined
TimToady hmm, I thought we had a range form of substr 14:55
14:55 ab6tract joined
FROGGS m: say substr("foobar", 1..4) 14:56
camelia rakudo-moar eb6de6: OUTPUT«ooba␤»
GLRelia rakudo-moar c68db9: OUTPUT«ooba␤»
14:56 tinyblak joined
FROGGS m: say "foobar".substr(1..4) 14:56
camelia rakudo-moar eb6de6: OUTPUT«ooba␤»
GLRelia rakudo-moar c68db9: OUTPUT«ooba␤»
ab6tract m: sub arr-ret { [<for many things>] }; my ($one,$two,$three) = arr-ret
camelia ( no output )
GLRelia ( no output )
_itz how late are the talks in the main room now running roughly?
ab6tract m: sub arr-ret { [<for many things>] }; say my ($one,$two,$three) = arr-ret
camelia rakudo-moar eb6de6: OUTPUT«for many things (Any) (Any)␤»
GLRelia rakudo-moar c68db9: OUTPUT«formanythings␤»
ab6tract so that question is answered :) 14:57
TimToady m: say substr("foobar", 1..-9) 14:58
camelia rakudo-moar eb6de6: OUTPUT«Number of characters argument to substr out of range. Is: -9, should be in 0..Inf; use *-9 if you want to index relative to the end␤ in block <unit> at /tmp/XDEI7GXCJF:1␤␤»
GLRelia rakudo-moar c68db9: OUTPUT«Number of characters argument to substr out of range. Is: -9, should be in 0..Inf; use *-9 if you want to index relative to the end␤ in block <unit> at /tmp/PllVFxG9cQ:1␤␤»
dalek kudo/glr: 51b313b | (Stefan Seifert)++ | src/core/List.pm:
Make smartmatching Iterables against Lists use same semantics as Arrays

S03-smartmatch/array-array.t tests smartmatching against Lists.
  TimToady++ suggests adding some pattern type for smart matching against
lists. However, we're trying to get to the same functional level as nom now. We can always create new semantics later.
Makes array-array.t pass.
TimToady m: say substr("foobar", 2..2) 14:59
camelia rakudo-moar eb6de6: OUTPUT«o␤»
GLRelia rakudo-moar c68db9: OUTPUT«o␤»
TimToady m: say substr("foobar", 2..1)
camelia rakudo-moar eb6de6: OUTPUT«␤»
GLRelia rakudo-moar c68db9: OUTPUT«␤»
colomon TimToady: I think we talked about having a range form of substr, but there were objections? But I may have missed something, I’ve been not faithfully backlooging the last few months.
TimToady m: say substr("foobar", 2..-1)
camelia rakudo-moar eb6de6: OUTPUT«Number of characters argument to substr out of range. Is: -2, should be in 0..Inf; use *-2 if you want to index relative to the end␤ in block <unit> at /tmp/mWgvVrORU9:1␤␤»
GLRelia rakudo-moar c68db9: OUTPUT«Number of characters argument to substr out of range. Is: -2, should be in 0..Inf; use *-2 if you want to index relative to the end␤ in block <unit> at /tmp/avzl8U6xIq:1␤␤»
TimToady that should really treat any null range the same
and it's just a wrong message anyway 15:00
15:05 virtualsue_ joined, virtualsue left
tony-o got json grammar down to 3.8s but that's still a 2 full seconds slower than built-in..gonna try streaming 15:05
15:05 virtualsue_ is now known as virtualsue
tony-o and coffee. 15:05
15:06 diana_olhovik_ left
timotimo mhhh, streaming coffee 15:06
tony-o so gooood 15:08
_itz my $coffee=Supply.new() 15:09
$coffee.tap
dalek ast/glr: 971583a | (Stefan Seifert)++ | S12-introspection/WHAT.t:
Skip TODO tests that now fail fatally
 15:10
hoelzro I missed that while moving rooms - we have sub substr(Range), but not method substr(Str:D: Range)? 15:11
FROGGS hoelzro: we have both
hoelzro oh, ok, I must've mistread
*misread
FROGGS just negative and watevercode endpoints are handled wrongish
hoelzro ahhhhh yes 15:12
dalek pan style="color: #395be5">perl6-examples: 45f1eb6 | paultcochrane++ | t/categories/euler.t:
Skip long running examples when running tests
 15:13
pan style="color: #395be5">perl6-examples: 13177d5 | paultcochrane++ | t/categories/euler.t:
Improve skipping code
pan style="color: #395be5">perl6-examples: f418282 | paultcochrane++ | t/categories/euler.t:
Correct number of planned Euler tests
pan style="color: #395be5">perl6-examples: 9f9f3d0 | paultcochrane++ | t/categories/euler.t:
Skip prob034 as it takes too long in tests
pan style="color: #395be5">perl6-examples: 4122609 | paultcochrane++ | t/categories/euler.t:
Fix spacing between Euler subtests
pan style="color: #395be5">perl6-examples: 55ec5cb | paultcochrane++ | t/categories/euler.t:
Fix Euler test ordering
15:14 Ven joined 15:15 andreoss left 15:16 itz left, larion joined 15:20 virtualsue left
masak by the way -- I just read that p6u thread -- my opinion is that we should have 'use strict' on in one-liners. 15:23
I've never felt helped by it being off, and I've sometimes been burned by it.
15:23 tinyblak left 15:24 tinyblak joined 15:25 larion left
Ven masak: isn't it on these days? 15:25
no, I'm wrong. Anyway, I agree. `alias 6=perl6 -M-no-strict` or w/e is fine by me
15:26 Ven left
TimToady so mot it be 15:26
tadzik do we have no-strict?
FROGGS we have
tadzik oh, neat :)
masak TimToady: 'mot'? does that mean you agree we can change it to strict-on-by-default for one-liners? :) 15:29
TimToady nodnod
masak \o/\o/
kudos to whoever makes it so 15:30
15:30 tinyblak left
masak I'm gonna scurry back to the venue 15:30
if it's not fixed before I get there, I'll make it so :P
&
15:32 zakharyas left 15:35 spebern joined, spebern left
dalek kudo/glr: 901d7c4 | (Stefan Seifert)++ | src/core/array_slice.pm:
Implement multi-dim array assignment
 15:36
FROGGS I think I fixed the FIRST issue
15:38 brrt left 15:39 ab6tract left, tinyblak joined 15:43 cognominal left 15:44 tinyblak left
nine m: my @a = [0]; @a[0][0]++; say @a.perl; 15:45
camelia rakudo-moar eb6de6: OUTPUT«[[1]]<>␤»
GLRelia rakudo-moar 901d7c: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/XYlDK5Q7GT:1␤␤»
nine Why is this even allowed to work much less expected?
dalek kudo/nom: 5fb81ff | (Carl Masak)++ | src/Perl6/World.nqp:
strict is now on by default, even for -e

This should make (most of) p6u happy.
laben m: my @a = [0]; dd @a 15:46
camelia rakudo-moar eb6de6: OUTPUT«@a = [[0]]<>␤»
GLRelia rakudo-moar 901d7c: OUTPUT«@a = [0]␤»
laben m: my @a = [0]; dd @a[0]
camelia rakudo-moar eb6de6: OUTPUT«@a = [0]␤»
GLRelia rakudo-moar 901d7c: OUTPUT«@a = 0␤»
laben m: my @a = [0]; dd @a[0][0] 15:47
camelia rakudo-moar eb6de6: OUTPUT«<element> = 0␤»
GLRelia rakudo-moar 901d7c: OUTPUT«0␤»
laben m: my @a = [0]; dd @a[0][0]++
camelia rakudo-moar eb6de6: OUTPUT«0␤»
GLRelia rakudo-moar 901d7c: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/DQShfIsIeL:1␤␤»
nine Oh, so this test really means my @a = (0)?
15:47 salva left
nine err... my @a = $[0]; really 15:49
laben m: my @a = [0],; dd @a
camelia rakudo-moar eb6de6: OUTPUT«@a = [[0]]<>␤»
GLRelia rakudo-moar 901d7c: OUTPUT«@a = [[0],]␤»
laben m: my @a = [0],; dd @a[0][0]++
camelia rakudo-moar eb6de6: OUTPUT«0␤»
GLRelia rakudo-moar 901d7c: OUTPUT«0␤»
timotimo - QAST::Stmt my int $foo
- QAST::Var(lexicalref $foo :decl()) $foo
why the F... do we declare this to be lexicalref?
dalek ast/glr: 4e91c07 | (Stefan Seifert)++ | S09-autovivification/autoincrement.t:
Fix creation of nested arrays in S09-autovivification/autoincrement.t
 15:50
kudo/glr: 0b76e79 | lizmat++ | src/core/ (2 files):
Give Set/Bag/Mix single arg semantics, nine++
colomon m: say rindex( 'xxy','y', $pos )
camelia rakudo-moar eb6de6: OUTPUT«5===SORRY!5=== Error while compiling /tmp/QdgMvkrHBC␤Variable '$pos' is not declared␤at /tmp/QdgMvkrHBC:1␤------> 3say rindex( 'xxy','y', 7⏏5$pos )␤»
GLRelia rakudo-moar 901d7c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ZuxG_JNOxK␤Variable '$pos' is not declared␤at /tmp/ZuxG_JNOxK:1␤------> 3say rindex( 'xxy','y', 7⏏5$pos )␤»
tadzik timotimo: ¯\_(ツ)_/¯ 15:51
dalek kudo/glr: 15b06d0 | FROGGS++ | src/core/Any-iterable-methods.pm:
fire FIRST phaser
 15:52
FROGGS another test file down!
15:53 andreoss joined
colomon lizmat++ nine++ 15:56
m: say rindex( 'xxy','y', -1e34 )
camelia rakudo-moar eb6de6: OUTPUT«Nil␤»
GLRelia rakudo-moar 901d7c: OUTPUT«Nil␤»
nine 48 current official count 15:58
jdv79 is Pod::To::HTML broken for anyone else on nom?
colomon lizmat: are you working on getting the bag/set/etc test files working, or shall I?
15:58 FROGGS left
tadzik I think panda just broke on glr 15:58
15:59 Ven joined, Ven left
dalek kudo/glr: 186223e | (Stefan Seifert)++ | src/core/List.pm:
Fix List.ACCEPTS regression

List.ACCEPTS really has to return the list and not boolify too early. Otherwise my @matches = $foo ~~ m:g/bar/; breaks		 16:01
colomon m: say (set [ foo => 10, bar => 17, baz => 42 ]).perl 16:02
GLRelia rakudo-moar 15b06d: OUTPUT«set(:baz(42),:foo(10),:bar(17))␤»
camelia rakudo-moar eb6de6: OUTPUT«set([:foo(10), :bar(17), :baz(42)])␤»
16:03 domidumont left, smls left
nine 47! 16:04
Oh, 46 actually! The 1 was lock.t 16:05
JimmyZ Is it ok to remove Array.ACCEPTS ?
looks Array is List
16:05 adu joined
nine JimmyZ: no, there are slight differences 16:05
16:09 kanl left, kanl joined 16:11 lizmat_ joined 16:14 lizmat left
dalek ast/glr: b8a19c9 | (Stefan Seifert)++ | S04-phasers/begin.t:
Remove obsolete TODO marker
 16:14
nine 45
colomon still has no clue what set / bag / etc constructors should be doing — once that’s figured out cleaning up the tests will be easy… 16:16
16:16 virtualsue joined 16:17 kanl left, Sufrostico joined
nine m: my @a = 1,2,3; my @b = 3,4,5; say zip(@a; @b).perl; 16:18
camelia rakudo-moar eb6de6: OUTPUT«((1, 3), (2, 4), (3, 5))␤»
GLRelia rakudo-moar 186223: OUTPUT«(([1, 2, 3], [3, 4, 5])).Seq␤»
16:18 dwarring left
Sufrostico sorry the silly question... where can i get the status of the implemetation of threads in perl6 .. rakudo 16:19
tadzik what status do you mean?
they are implemented and work on both supported backends
16:19 kanl joined
Sufrostico nice 16:20
lightweight?
btyler_ Sufrostico: here's a very recent talk showing off features in that area: jnthn.net/papers/2015-yapcasia-concurrency.pdf
16:20 rarara left
Sufrostico nice 16:20
my python co worker was complaining about perl threads just yesterday 16:21
tadzik he might have meant perl 5 threads
colomon nine: I stuck a say inside infix:<Z>, and it gets different arguments when you call it via zip 16:22
tadzik "0b76e797c245892b9a5ddbc4fe7b8d3829d116f0 is the first bad commit"
lizmat_: ^
colomon nine: extra list wrapping it, I think.
16:22 kaare__ left
sjn Sufrostico: that talk btyler_ linked to, it's already on youtube 16:22
sjn just watched it, and liked it a lot. Very instructive presentation 16:23
hahainternet wg 28
:(
too many irc windows
tadzik lizmat_: I reverted it and it fixed panda, but I have no idea what can it possibly be impacting 16:24
16:25 itz joined
dalek pan style="color: #395be5">perl6-examples: 115ad18 | (Ben Tyler)++ | categories/rosalind/trie-grondilu.pl:
trie-grondilu: fix error with postfix condition.

Not yet sure how this broke, investigating what might be more or less tightly bound under GLR.		 16:26
16:26 rurban left 16:28 AlexDaniel left 16:29 araujo joined
_itz pamda is broken on 'brew I think 16:29
err panda
tadzik yes
you mean on glr, right? 16:30
_itz no on nom
tadzik oh
that's straneg
_itz I'll try from scratch
colomon just worked fine for me a couple of minutes ago?
(on nom)
tadzik seems to work for me
is broken on glr though 16:31
16:33 TimToady left 16:34 TimToady joined
masak m: constant k = 1, map { k[$_] }, 0..*; .say for k[^10] 16:37
camelia rakudo-moar 5fb81f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/h2ttbus2EC␤Variable '&k' is not declared␤at /tmp/h2ttbus2EC:1␤------> 3constant k = 1, map { 7⏏5k[$_] }, 0..*; .say for k[^10]␤»
GLRelia rakudo-moar 186223: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1TJghQX9ug␤Variable '&k' is not declared␤at /tmp/1TJghQX9ug:1␤------> 3constant k = 1, map { 7⏏5k[$_] }, 0..*; .say for k[^10]␤»
tony-o is it the ',' instead of ';' ?.. 16:38
oh nvm
masak tony-o: it's rt.perl.org/Ticket/Display.html?id=111674
not having the constant just defined available in the rhs of the constant assignment currently prevents a bunch of cool self-referential constant lazy-list definitions :) 16:39
I'm almost tempted to debug this one and see what causes it.
16:43 Sufrostico left 16:47 FROGGS joined, yqt left
moritz idly wonders how many of those cool self-referential definitions are readably by mere mortals 16:51
laben nine: github.com/rakudo/rakudo/pull/519 16:54
FROGGS m: sub foo { say &?ROUTINE.name; for 1 { say &?ROUTINE.name; } }; foo 16:55
camelia rakudo-moar 5fb81f: OUTPUT«foo␤foo␤»
GLRelia rakudo-moar 186223: OUTPUT«foo␤foo␤»
FROGGS m: sub foo { say &?ROUTINE.name; gather for 1 { say &?ROUTINE.name; } }; foo 16:56
camelia rakudo-moar 5fb81f: OUTPUT«foo␤␤»
GLRelia rakudo-moar 186223: OUTPUT«foo␤␤»
dalek kudo/glr: 6e8b97b | laben++ | src/core/metaops.pm:
Fix dwim hypers to manage lazy iterables correctly
kudo/glr: fc8ab7e | moritz++ | src/core/metaops.pm:
Merge pull request #519 from laben/glr

Fix dwim hypers to manage lazy iterables correctly
FROGGS m: sub foo { say &?ROUTINE.perl; gather for 1 { say &?ROUTINE.perl; }; gather for 1 { say &?ROUTINE.perl; } }; foo
camelia rakudo-moar 5fb81f: OUTPUT«sub foo () { #`(Sub|59748232) ... }␤Code.new␤Code.new␤»
GLRelia rakudo-moar 186223: OUTPUT«sub foo () { #`(Sub|53058688) ... }␤Code.new␤Code.new␤»
FROGGS m: sub foo { say &?ROUTINE.perl; for 1 { say &?ROUTINE.perl; }; gather for 1 { say &?ROUTINE.perl; } }; foo
camelia rakudo-moar 5fb81f: OUTPUT«sub foo () { #`(Sub|42434376) ... }␤sub foo () { #`(Sub|42434376) ... }␤Code.new␤»
GLRelia rakudo-moar 186223: OUTPUT«sub foo () { #`(Sub|54308288) ... }␤sub foo () { #`(Sub|54308288) ... }␤Code.new␤»
FROGGS m: sub foo($,$,$) { say &?ROUTINE.arity; for 1 { say &?ROUTINE.arity; }; gather for 1 { say &?ROUTINE.arity; } }; foo 1, 2, 3 16:57
camelia rakudo-moar 5fb81f: OUTPUT«3␤3␤0␤»
GLRelia rakudo-moar 186223: OUTPUT«3␤3␤0␤»
laben moritz: about the roast tests, do i push to my repo and create a PR or push directly (dont know if i can)
FROGGS m: sub foo($,$,$) { say &?ROUTINE.arity; for 1 { say &?ROUTINE.arity; }; gather for 1 { say &?ROUTINE(); } }; foo 1, 2, 3 16:58
camelia rakudo-moar 5fb81f: OUTPUT«(timeout)3␤3␤»
GLRelia rakudo-moar 186223: OUTPUT«(timeout)3␤3␤»
moritz laben: you push directly
laben: and if you can't, you tell me
FROGGS m: sub foo($,$,$) { say &?ROUTINE.arity; for 1 { say &?ROUTINE.arity; }; gather for 1 { if 1 { &?ROUTINE.arity } } }; foo 1, 2, 3
camelia rakudo-moar 5fb81f: OUTPUT«3␤3␤»
GLRelia rakudo-moar 186223: OUTPUT«3␤3␤»
FROGGS m: sub foo($,$,$) { say &?ROUTINE.arity; for 1 { say &?ROUTINE.arity; }; gather for 1 { if 1 { say &?ROUTINE.arity } } }; foo 1, 2, 3
camelia rakudo-moar 5fb81f: OUTPUT«3␤3␤3␤»
GLRelia rakudo-moar 186223: OUTPUT«3␤3␤3␤»
FROGGS awesome bug
laben moritz: i'm part of the perl6 org but it seems i cant push directly, maybe some git stuff i dont know? 16:59
moritz laben: what does 'git remove -v' say?
laben moritz: if you meant "git remote -v" it's "" 17:00
$ git remote -v
origin git://github.com/perl6/roast.git (fetch)
origin [email@hidden.address] (push)
darn
moritz laben: hurm, looks correct; what error message do you get if you try to push? 17:01
laben moritz: permission denied (publickey)
_itz ssh-add? 17:02
moritz laben: does a 'git push' work to any other repo from the same shell?
laben: might be that you have multiple SSH keys, and the one that's used isn't added on github 17:03
laben moritz: yea, i just pushed to my gh repo of rakudo for the last pr
moritz huh
laben maybe it's better i push to gh fork and create a pr 17:04
moritz laben: if that works for you, sure 17:05
laben: you can even merge yourself
(I hope :-)
colomon tadzik: emmentaler is eating my projects.json file?!
moritz nom nom nom 17:06
masak m: role R[::N, $result = N] { say $result }; my $r = R[Int].new
camelia rakudo-moar 5fb81f: OUTPUT«Cannot find method 'collisions'␤ in any apply at src/gen/m-Metamodel.nqp:2633␤ in any compose at src/gen/m-Metamodel.nqp:2810␤ in any make_pun at src/gen/m-Metamodel.nqp:1642␤ in any find_method at src/gen/m-Metamodel.nqp:1670␤ in block <uni…»
GLRelia rakudo-moar fc8ab7: OUTPUT«Cannot find method 'collisions'␤ in any apply at src/gen/m-Metamodel.nqp:2635␤ in any compose at src/gen/m-Metamodel.nqp:2812␤ in any make_pun at src/gen/m-Metamodel.nqp:1644␤ in any find_method at src/gen/m-Metamodel.nqp:1672␤ in block <uni…»
FROGGS reported an awesome bug right now as RT #125934
masak submits rakudobug
17:06 virtualsue left
moritz m: role R[::N, $result = N] {} 17:07
GLRelia ( no output )
camelia ( no output )
laben moritz: can you check it before merging? github.com/perl6/roast/pull/67 (btw it seems i can indeed merge it myself...)
17:07 itz left
moritz laben: in line 204, it might be nicer to declare new lexicals @r and @e 17:08
laben: in fact, I don't think you use those values at all 17:10
because in line 218, you re-assign to them
laben: otherwise, looks fine
17:10 Loren_ left
laben moritz: i do use them in line 207 17:11
should i add more blank lines?
moritz laben: oh, count me blind :-)
laben: no, it's fine, but declare them with 'my' please, so that global defintions in that file (which aren't relevant for the tests) aren't overridden
(and the block of tests can easier be moved around, for exapmle) 17:12
laben moritz: i updated the commit, now it should be ok 17:16
17:17 FROGGS left
tadzik colomon: I don't know, is it? :) 17:20
17:20 lizmat joined
tadzik colomon: I haven't touched emmentaler in ages 17:20
17:20 lizmat_ left
tadzik I learned that panda now has a really strong competition :) 17:20
in form of github.com/ugexe/zef
in many ways it feels like Mojolicious of module management, with its own bundled HTTP library and all 17:21
muraiki interesting 17:23
nine laben: is it really ok to have one side being lazy without being DWIM?
tadzik I'll have to evaluate if I should see it as competition or a successor :)
laben nine: what do you mean?
nine laben: the DWIM makes it auto truncate to the non-DWIM side's length, so infinitiy is not a problem. But without DWIM we have to iterate infinitely. 17:24
tadzik ugexe++ and tony-o++ in any case :)
laben nine: can you illustrate an example? i thought i covered all sides... 17:25
17:25 cognominal joined
dalek ast/glr: a88ba52 | laben++ | S03-metaops/hyper.t:
Add tests for binary op (non-)dwim cases and fudge broken ones
 17:30
ast/glr: 1846e7f | laben++ | S03-metaops/hyper.t:
Merge pull request #67 from laben/glr

Add tests for binary op (non-)dwim cases and fudge broken ones
17:31 FROGGS joined
laben nine: with this ^^^ one test file failing less 17:32
oh cmon now panda fails a test again on newest glr 17:34
17:34 llfourn left
tadzik it doesn't work at all 17:34
FROGGS :S
tadzik since 0b76e797c245892b9a5ddbc4fe7b8d3829d116f0
suddenly arrays are Seqs
not sure who to blame :P 17:35
nine tadzik: Seq is usually the correct answer. .list it if you need to iterate it more than once
tadzik it was a bit more complicated than that :)
there'd be .flat and .list and .grep(*.defined) and dragons and shit 17:36
and suddenly, instead of ["File::Find", "Shell::Command", "JSON::Fast"] there'd be ((stuff).Seq, (Any).Seq, (Any.Seq)).Seq, in random order
dalek kudo/glr: 18261c9 | lizmat++ | src/core/set_operators.pm:
Give set()/bag()/mix() single arg semantics
 17:45
nine laben: ah sorry for the noise. I misread your patch. Too tired to process boolean expressions it seems :) 17:49
laben nine: dont worry, you were tired for perl6 goodness, i forgive you ^.^ 17:50
17:50 domidumont joined
dalek rl6-roast-data: a689b9e | coke++ | / (10 files):
today (automated commit)
 17:52
colomon wants to do module with .dragons and .shit now….
dalek kudo/glr: 1a05c67 | lizmat++ | src/core/set_operators.pm:
Be more smart about singel arging bag/set/mix()

  FROGGS++ for inspiration
colomon lizmat: I keep seeing your patches, celebrating, then discovering they don’t quite fix set.t… 17:53
tadzik :D
timotimo ;(
lizmat colomon: that's because set.t / bag.t etc. need to be fixed
working on that now
colomon okay, if you’ve got that, I’ll go back to being utterly confused by emmentaler 17:54
timotimo zef has a "smoke" command, too 17:55
maybe emmentaler could be ported over to zef? :P
dalek pan style="color: #395be5">perl6-examples: 4e3e0eb | paultcochrane++ | categories/wsg/advanced-2008/event001-eric256.pl:
Fix failure in event001-eric256

Using `split('')` to get the individual elements of a number or string is a Perl 5-like way of doing this. Unfortunately, `split` matches the empty string at the *ends* of the string as well, hence one gets 9 elements from a 7 digit string. The solution is to use `comb`, which is the correct way to do this in Perl 6.
colomon timotimo, tadzik: I’m deadly serous that I’ve started emmentaler with “./smoker projects.json”, get a weird error message, and then projects.json is just gone. 17:56
17:56 cognominal left
tadzik colomon: hmm 17:56
colomon: oh...
colomon oh?
tadzik colomon: so panda recently started updating $!projectsfile automatically
I guess that's not always desirable 17:57
especially if it fails
colomon it’s completely 100% undesirable, because that file isn’t necessarily projects.json
tadzik right
colomon I use subsets for testing pretty regularly 17:58
timotimo argh!
tadzik colomon: yeah, I even had to override that for panda tests :/ 17:59
stupid thing
colomon: should be fixed now
18:00 andreoss left
tadzik thanks for letting me know 18:00
colomon tadzik: hmmm
==> Installing panda from a local directory '/home/colomon/.rakudobrew/moar-nom/panda'
Could not download module metadata: No such method 'field' for invocant of type 'Array'
tadzik wat 18:01
glr?
colomon nope, nom
laben m: say defined my @c
camelia rakudo-moar 5fb81f: OUTPUT«True␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«True␤»
tadzik . o O ( field? )
colomon (I’m running on hack here)
tadzik there's no 'field' in panda
laben m: say defined (my @c=Any)
camelia rakudo-moar 5fb81f: OUTPUT«True␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«True␤»
colomon tadzik: I know, that’s what’s been driving me nuts when I was trying to use emmentaler
tadzik: let me try blowing away my rakudo and rebuilding from scratch
tony-o what is emmentaler?
tadzik colomon: oh! 18:02
colomon tony-o: module smoke tester (via panda)
tadzik do you have HTTP::UserAgent installed?
colomon tadzik: maybe?
tadzik colomon: panda will use it if it's installed
maybe it's broken
colomon just kind of blew away everything
tadzik oh well :)
laben tadzik: is there a reason you want to use sets instead of list.unique? 18:03
tadzik laben: I never wanted to use sets 18:04
they sort of appeared
as long as you mean seqs
laben tadzik: no i mean sets that you create with (|) 18:05
tadzik I now think I should've waited a bit longer with the glr merge
laben: oh
laben: I don't think I wrote that :P
laben: no reason 18:06
colomon laben: what’s the context?
tadzik laben: and frankly, I don't think that's a right way of doing this anyway
I mean, I'd rather have code that takes META.info and gives me something that actually conforms to specs that I can process
dalek ast/glr: 0da7828 | lizmat++ | S02-types/bag.t:
Make S02-types/bag.t pass all tests
ast/glr: a7dafef | lizmat++ | S03-metaops/hyper.t:
Merge branch 'glr' of github.com/perl6/roast into glr
tadzik rather than taking whatever mess META.info contains and trying to make the best of it in the latest possible moment 18:07
honestly I really feel like giving panda a good, solid shakeup from the ground up 18:08
it's got 4 years of legacy all over its codebase, and some parts of it are either stupid or plain wron g:)
[Coke] wonders if the p6u is like the mcu. 18:13
[ptc] _itz: the perl6-examples are testing green on travis again! Yay!
colomon \o/ 18:14
FROGGS tadzik: go for it
tadzik FROGGS: part of my head says "nooo, rewrite is always stupider than fixing things carefully" 18:15
but I guess that's not always true 18:16
FROGGS tadzik: call it refactoring and you are fine :o)
tadzik haha
18:19 dre_ joined 18:20 dre_ left
dalek ast/glr: 869e1a9 | lizmat++ | S02-types/mix.t:
Make S02-types/mix.t pass all tests
 18:21
[Coke] notes that he and colomon are on the 2nd floor. 18:22
colomon notes that it is cool and fairly quiet up here.
tadzik oh!
[Coke] what did we miss? 18:23
*golf clap* *golf clap*
ugexe a Distribution object is meant to be the API to META.info 18:24
tadzik hm?
oh
jdv79 [Coke]: is it cool up there?
b2gills
.oO( I just noticed that I have exactly 14,000 rep on StackOverflow, most of it from Perl 5/6 answers )
 18:25
lizmat m: my %h = set < a a b b c c d d >; say %h.perl
camelia rakudo-moar 5fb81f: OUTPUT«Odd number of elements found where hash initializer expected␤ in block <unit> at /tmp/I8hgYvvthd:1␤␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«{:a("c"), :b("d")}␤»
jdv79 oh, colomon already said that. duh.
[Coke] jdv79: coolish. 18:26
b2gills m: my %h := set < a a b b c c d d >; say %h.perl
camelia rakudo-moar 5fb81f: OUTPUT«set("a","c","b","d")␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«set("a","c","b","d")␤»
18:26 diana_olhovik joined 18:33 telex left
jdv79 what is the diff between META6.json and META.info? 18:33
tadzik history 18:34
jdv79 ugexe: what does that mean?
tadzik :P
18:34 telex joined
jdv79 ok, so same; effectively. 18:34
tadzik FROGGS: rebecca.blackfriday/
laben b2gills: about our recent discussion about >>op<< not throwing and shortening the longer side, it seems there is no reason to go about doing this given Z already has this behaviour 18:35
tony-o tadzik: FROGGS: www.youtube.com/watch?v=qe6ZbL6KbEE 18:36
moritz TIL that blackfriday is now a TLD
laben colomon: the context is method init-projects in Ecosystem.pm where it creates the projects from the parsed json
colomon: specifically where it sets dependencies
FROGGS tony-o: that's not his real hair :o) 18:37
dalek rl6-roast-data: 8b52920 | coke++ | / (2 files):
another glr run for 'today'
tony-o oh my
18:37 khw left 18:40 diana_olhovik left
tadzik tony-o: I think I'll spare the rest of the hackathoners and watch it later at the hotel :o 18:41
moritz: endless.horse/
18:42 diana_olhovik__ joined
[Coke] !!! everyone working at the hackathon - please privmsg me or email me a small writeup of what you did, even if you were working remotely. I'm writing up a blog post so our hosts know what we got up to. 18:42
18:42 kaare__ joined
ugexe jdv79: i mean a distributions meta information should be consumed from a Distribution object, which can provide normalized values (ver vs version) as well as the .content method for accessing files in the `provides`. 18:42
tadzik Distribution being part of Zef 18:44
or is it specced?
ugexe its specced in S22
18:44 Loren joined
tony-o it's specced 18:44
tadzik oh!
tony-o we have questions about it though, it's kind of vague
tadzik well, it's a spec written before an implementation :) 18:45
same with hooks
they just say "there should be hooks"
[Coke] down to 261 failing tests. 18:46
arnsholt And TimToady said: "Let there be hooks!" =)
[Coke] we also have a few more skips than nom.
ugexe there is an implementation. called Distribution.pm in rakudo core. its not up to spec 18:47
tony-o ugexe: wrote up my json::faster benchmark crap here: ugexe.com/speed-quest-1-0-or-0-1-or-something/ 18:48
tadzik: the question more revolves around the way these things were imagined, ugexe is imagining centralized CURs and i'm imagining decentralized CURs and both would work but the implementation is different. that's kind of a quick way of saying it for lack of time (i'm hungry) 18:49
it's a theory question, i wrote a CUR that works over HTTP and not sure how to handle .install yet 18:50
tadzik hm
ugexe im writing content storage for Distribution objects now. Distribution::Local and then the Distribution::HTTP::Github. Distribution.content can then pull from the hard disk or from github.com via http requests by pulling from whatever content storage is attached to it 18:53
18:55 virtualsue joined, domidumont left
tadzik sounds good 18:55
18:56 Ven joined
colomon tony-o++ 18:56
moritz m: my Enum @a = :a; say @a.perl 19:01
camelia rakudo-moar 5fb81f: OUTPUT«Array[Enum].new("a" => Bool::True)␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«Array[Enum].new(:a)␤» 19:02
moritz m: my Enum @a = :a; say EVAL(@a.perl).perl
camelia rakudo-moar 5fb81f: OUTPUT«Array[Enum].new("a" => Bool::True)␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«Array[Enum].new()␤»
moritz m: say :a.perl, :a.perl(:arglist) 19:05
camelia rakudo-moar 5fb81f: OUTPUT«:a"a" => Bool::True␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«:a:a␤»
laben tadzik: i got it, what has changed that broke panda (other than glr): json::fast returns Any instead of () and that got containerized, so it didnt get flattened since the array got an element
tadzik: i got a change that fix this
tadzik: unless you want to rewrite it all
dalek ast/glr: 7e4a1c0 | (Stefan Seifert)++ | S32-array/push.t:
GLRify S32-array/push.t
tadzik laben: ooh 19:07
laben: I definitely won't blindly rewrite it while throwing all the smart thing it's doing away, so if you have the fix please go ahead :) 19:09
19:10 bin_005 joined
laben tadzik: well, you could simply want to rewrite those lines that calculate deps, anyway this is it gist.github.com/46655b37a8f21834b737 or do you prefer a PR? 19:10
colomon m: say rindex( 'xxy','y', -1e34 )
camelia rakudo-moar 5fb81f: OUTPUT«Nil␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«Nil␤»
dalek kudo/glr: 2178f78 | moritz++ | src/core/Enum.pm:
Restore Enum.perl(:arglist) behavior

fixes S32-hash/perl.t		 19:11
19:12 cognominal joined
FROGGS m: for 1..5 { LAST { .say } } 19:12
camelia rakudo-moar 5fb81f: OUTPUT«5␤»
GLRelia ( no output )
colomon lizmat: you still working on the set/bag test files? 19:13
lizmat colomon: indeed I am
colomon okay, I’ll work on something else. lizmat++ 19:14
lizmat okidoki
tadzik laben: just applied it :)
thanks!
19:14 brummett joined, virtualsue left 19:15 Loren left
colomon m: say ([] ... [])[0] 19:15
camelia rakudo-moar 5fb81f: OUTPUT«␤»
GLRelia rakudo-moar 1a05c6: OUTPUT«Cannot get sequence endpoint from an empty list (use * or :!elems instead?)␤ in block <unit> at /tmp/IC4uUeKFzM:1␤␤»
colomon Am I correct in thinking that shouldn’t work under GLR? 19:16
laben m: dd ([] ... [])
camelia rakudo-moar 5fb81f: OUTPUT«([],)␤»
GLRelia rakudo-moar 2178f7: OUTPUT«Cannot get sequence endpoint from an empty list (use * or :!elems instead?)␤ in block <unit> at /tmp/4HLHlsTkas:1␤␤»
laben m: dd ([] .. []) 19:17
camelia rakudo-moar 5fb81f: OUTPUT«0..[]␤»
GLRelia rakudo-moar 2178f7: OUTPUT«0..0␤»
19:18 rangerprice joined 19:21 yqt joined 19:24 espadrine left 19:26 xinming left, xinming joined
tadzik laben: hmm, it still seems to have list issues 19:28
now I get resolve stage failed for panda: Dependency File::Find JSON::Fast Shell::Command is not present in the module ecosystem
19:28 virtualsue joined
laben tadzik: yeah i just noticed it too trying rebootstrap 19:28
i'll check it out in an hour or so, cant now 19:29
tadzik (╯°□°）╯︵ ɹןb
19:30 llfourn joined
moritz tadzik: looks like a missing flat 19:32
tadzik yeah 19:33
ugexe say from-json("projects.json".IO.slurp) <-- output is correct. my @a = from-json("projects.json".IO.slurp); say @a; <-- output: [] 19:34
19:34 llfourn left 19:35 brummett left
tadzik strange 19:36
19:36 itz joined
colomon ugexe: did you have a module or list of modules that was borking panda when you tried to install it? 19:37
FROGGS m: say from-json '{ "foo" : [ "bar", "baz" ] }'
camelia rakudo-moar 5fb81f: OUTPUT«foo => bar baz␤»
GLRelia rakudo-moar 2178f7: OUTPUT«foo => [bar baz]␤»
FROGGS m: say my @a = from-json '{ "foo" : [ "bar", "baz" ] }'
camelia rakudo-moar 5fb81f: OUTPUT«foo => bar baz␤»
GLRelia rakudo-moar 2178f7: OUTPUT«foo => [bar baz]␤»
moritz m: say $*IN.slurp-rest.^name 19:38
camelia rakudo-moar 5fb81f: OUTPUT«Str␤»
GLRelia rakudo-moar 2178f7: OUTPUT«(timeout)»
ugexe colomon: no, i think it was generally nativecall modules or things with Build.pm
colomon ugexe: thanks
moritz m: say $*IN.slurp-rest.^name # checking if that timeout is reproducible
camelia rakudo-moar 5fb81f: OUTPUT«Str␤»
GLRelia rakudo-moar 2178f7: OUTPUT«(timeout)» 19:39
colomon m: say $*IN.slurp-rest
camelia rakudo-moar 5fb81f: OUTPUT«Céad slán ag sléibhte maorga Chontae Dhún na nGall␤Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;␤Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí␤I mo dhiaidh bhí gleanntáin ghlas’ G…»
GLRelia rakudo-moar 2178f7: OUTPUT«(timeout)» 19:40
moritz cannot reproduce on my machine
moritz@hack:~/p6/rakudo$ cat ../evalbot/stdin | ./perl6-m --setting=RESTRICTED -e 'say $*IN.slurp-rest.chars'
1134
colomon moritz: I keep on finding consistently reproducable differences between GLRelia and my local builds. 19:43
m: say rindex( 'xxy','y', -1e34 )
GLRelia rakudo-moar 2178f7: OUTPUT«Nil␤»
camelia rakudo-moar 5fb81f: OUTPUT«Nil␤»
colomon > rindex( 'xxy','y', -1e34 ) 19:44
index start offset out of range
hold on
moritz gets Nil too 19:45
19:46 Loren joined
colomon FROGGS++ just explained I needed to update MoarVM 19:47
and now it works properly
19:48 Ven left
dalek rl6-roast-data: 301b56e | coke++ | / (2 files):
another glr run for 'today'
 19:48
FROGGS jnthn / nine / moritz / lizmat: I've got a totally ugly patch for next.t and last.t: gist.github.com/FROGGS/6ad7ea03a39cc5453c7a 19:49
colomon m: roundrobin((); 1; 2..4; (5..7); <a b>).join(' ‘).say
camelia rakudo-moar 5fb81f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/M6TnNj13n1␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/M6TnNj13n1:1␤------> 3); 1; 2..4; (5..7); <a b>).join(' ‘).say7⏏5<EOL>␤ expecting any of:␤ ar…»
GLRelia rakudo-moar 2178f7: OUTPUT«5===SORRY!5=== Error while compiling /tmp/wm62xe6e3F␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/wm62xe6e3F:1␤------> 3); 1; 2..4; (5..7); <a b>).join(' ‘).say7⏏5<EOL>␤ expecting any of:␤ ar…»
FROGGS jnthn / nine / moritz / lizmat: it works but I think it is not commitable this way :o)
colomon m: roundrobin((); 1; 2..4; (5..7); <a b>).join(“ “).say
camelia rakudo-moar 5fb81f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/U9otqS7iGO␤Unable to parse expression in smart double quotes; couldn't find final '”' ␤at /tmp/U9otqS7iGO:1␤------> 3); 1; 2..4; (5..7); <a b>).join(“ “).say7⏏5<EOL>␤ expecting any of:␤…»
GLRelia rakudo-moar 2178f7: OUTPUT«5===SORRY!5=== Error while compiling /tmp/EcbhlfBafV␤Unable to parse expression in smart double quotes; couldn't find final '”' ␤at /tmp/EcbhlfBafV:1␤------> 3); 1; 2..4; (5..7); <a b>).join(“ “).say7⏏5<EOL>␤ expecting any of:␤…»
colomon m: roundrobin((); 1; 2..4; (5..7); <a b>).perl.say
camelia rakudo-moar 5fb81f: OUTPUT«((1, 2, 5, "a"), (3, 6, "b"), (4, 7))␤»
GLRelia rakudo-moar 2178f7: OUTPUT«([1, 2, 3, 4, 5, 6, 7, "a", "b"]).Seq␤»
colomon seeing this is several other tests as well — seems like GLR is not honoring semi-colons 19:50
semi-colon is not LoL composer anymore (FROGGS++ says)?
function signature is roundrobin(**@lol) 19:51
FROGGS m: roundrobin([], [1], [2..4], [5..7], <a b>).perl.say
camelia rakudo-moar 5fb81f: OUTPUT«((1, 2, 5, "a"), (3, 6, "b"), (4, 7))␤»
GLRelia rakudo-moar 2178f7: OUTPUT«([1, 2, 5, "a"], [3, 6, "b"], [4, 7]).Seq␤»
19:54 ab6tract joined
maettu that was a fine day at the spw215 :-) 19:58
19:58 larion joined
_itz jnthn++ # particularly good talks 20:00
20:00 muraiki left 20:02 FROGGS left 20:06 virtualsue_ joined, virtualsue left
moritz is yary (from p6u) on IRC? 20:06
20:07 virtualsue_ is now known as virtualsue
moritz and if yes, under which nick? 20:07
dalek ast/glr: 14b065f | colomon++ | S32-container/roundrobin.t:
Update to conform to GLR. FROGGS++

roundrobin((); 1; 2..4; (5..7); <a b>) becomes roundrobin([], [1], [2..4], [5..7], <a b>)
and
is(roundrobin([1..3]).elems, 1, 'roundrobin does not flatten array items');
just goes away, as that's not how these sorts of arguments work anymore.
ugexe something seems particularly messed up here 20:09
Use of uninitialized value $v of type Any in string context in block at /home/nickl/perl6/zef/lib/Zef/Utils/Any
there is no /Any file/folder
20:10 Loren left
colomon uninit’ed last part of the file path? 20:10
ab6tract what's p6u? 20:11
tadzik glr is fun :)
20:11 kaare__ left
tadzik ab6tract: perl6-users mailing list 20:11
ab6tract also: great SPW so far :D
dalek ast/glr: 2b5bf32 | colomon++ | S32-container/zip.t:
Also switch to current "LoL" standard.
ab6tract tadzik: thanks :)
tadzik [email@hidden.address]
SPW is awesome :) 20:12
maettu++ # and the rest of the crew
colomon maettu++ indeed!
20:15 adu left 20:16 FROGGS joined 20:18 colomon left
dalek ast/glr: 96b12b1 | (Stefan Seifert)++ | S04-statements/for.t:
GLRify S04-statements/for.t
 20:18
20:19 darutoko left
dalek kudo/glr: 441d54d | (Stefan Seifert)++ | src/core/List.pm:
Fix "This type does not support elems" on my @a; say |@a;
 20:20
FROGGS m: use Test; my $a = ''; for $[1..3, 4..6] { $a ~= $_.WHAT.gist }; is($a, Array.gist, 'List context'); 20:21
camelia rakudo-moar 5fb81f: OUTPUT«not ok 1 - List context␤␤# Failed test 'List context'␤# at /tmp/cfAEH_1jF0 line 1␤# expected: '(Array)'␤# got: '(Any)(Any)(Any)(Any)(Any)(Any)'␤»
GLRelia rakudo-moar 2178f7: OUTPUT«not ok 1 - List context␤␤# Failed test 'List context'␤# at /tmp/vMz8rI5vqa line 1␤# expected: '(Array)'␤# got: '(Range)(Range)'␤»
FROGGS m: use Test; my $a = ''; for ([1..3, 4..6],) { $a ~= $_.WHAT.gist }; is($a, Array.gist, 'List context');
camelia rakudo-moar 5fb81f: OUTPUT«ok 1 - List context␤»
GLRelia rakudo-moar 2178f7: OUTPUT«ok 1 - List context␤»
FROGGS aha 20:22
but yeah, makes sense
20:22 itz left 20:25 virtualsue left
nine Woah, we're at 37 latest official count 20:25
ab6tract nine: amazing!! :D 20:26
20:26 colomon joined, xfix left
FROGGS much awesome :o) 20:27
moritz such GLR
dalek ast/glr: 382ffeb | lizmat++ | S02-types/set.t:
Make S02-types/set.t pass all tests
kudo/glr: 7591fc4 | lizmat++ | src/core/ (3 files):
Simplify grep and friends on Set/Bag/Mix

Since Set is now also returning pairs for .list		 20:28
20:30 llfourn joined
dalek ast/glr: 0e04a3a | laben++ | S03-metaops/hyper.t:
Defudge hyper tests that incorrectly die no more

Fixed by 441d54d42e1601d555e601c3068a935dbb4360d4		 20:32
ast/glr: edcf987 | laben++ | S03-metaops/hyper.t:
Merge pull request #68 from laben/glr

Defudge hyper tests that incorrectly die no more
nine laben++ # good catch! 20:33
20:35 llfourn left
laben time to see what's up with panda, again :/ 20:35
ugexe panda installs with zef, so it should just be the bootstraper that is not working 20:37
moritz does zef precompile?
colomon does zef install with panda?
laben lizmat: did you try panda rebootstrap recently? it dies with "Use of uninitialized value %!elems of type Any in string context" that seem to reference a field from rakudo code
ugexe yes zef precompiles, and it will install with panda
lizmat laben: it has been ages since I did a panda rebootstrap
could you give me a stack trace somehow? 20:38
nine m: my $pair = a => 1; for $pair.value -> $v is rw { $v += 1 }; say $pair.value; 20:39
camelia rakudo-moar 5fb81f: OUTPUT«2␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/TZ11hQ2jzw:1␤␤»
nine m: class Foo { has $.value; }; my $foo = Foo.new(value => 1); for $foo.value -> $v is rw { $v += 1 }; say $foo.value; 20:40
camelia rakudo-moar 5fb81f: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/C0atOOJKsZ:1␤␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/Q3IFHhrAmb:1␤␤»
nine m: class Foo { has $.value is rw; }; my $foo = Foo.new(value => 1); for $foo.value -> $v is rw { $v += 1 }; say $foo.value;
camelia rakudo-moar 5fb81f: OUTPUT«2␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/ygrqpiCGUp:1␤␤»
laben lizmat: sorry, stacktrace seems unavailable, it complains about panda funcs but there's no %!elems there, i thought you would know better since it uses sets to calculate deps 20:41
nine m: class Foo { has $.value is rw; }; my $foo = Foo.new(value => 1); $foo.value += 1; say $foo.value;
camelia rakudo-moar 5fb81f: OUTPUT«2␤»
GLRelia rakudo-moar 7591fc: OUTPUT«2␤»
20:41 Loren joined
nine my $a = 1; for $a -> $v is rw { $v += 1; }; say $a; 20:42
m: my $a = 1; for $a -> $v is rw { $v += 1; }; say $a;
camelia rakudo-moar 5fb81f: OUTPUT«2␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/wm1cSXjnLp:1␤␤»
20:47 diana_olhovik__ left
nine m: my $a = 1; $a.list += 1; 20:47
camelia rakudo-moar 5fb81f: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/ClaHKbb4QA:1␤␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/BJd52Bh6ii:1␤␤»
nine m: my $a = 1; $a.list.[0] += 1; 20:48
camelia rakudo-moar 5fb81f: OUTPUT«Cannot modify an immutable Int␤ in block <unit> at /tmp/PHFHYbWCfL:1␤␤»
GLRelia rakudo-moar 7591fc: OUTPUT«Cannot modify an immutable Int␤ in block <unit> at /tmp/isVT4ES9jn:1␤␤»
timotimo o/ 20:50
lizmat m: mix( (foo => 10.1, bar => 17.2)).perl.say 20:51
camelia rakudo-moar 5fb81f: OUTPUT«(:foo(10.1)=>1,:bar(17.2)=>1).Mix␤»
GLRelia rakudo-moar 7591fc: OUTPUT«(:foo(10.1)=>1,:bar(17.2)=>1).Mix␤»
timotimo hoelzro: were we agreeing to sign on to irc before bed or were we agreeing to not do that? %)
hoelzro timotimo: umm...
timotimo ven's gone :( 20:52
nine
.oO( what bed? )
hoelzro I don't know if we agreed or not, but here we are =)
timotimo oh snap
moritz you guys *sleep* when you should be hacking? :-)
lizmat ok, sethash.t and mixhash.t have some issues I'm too tired too grep now
timotimo moritz: psch!
lizmat so I'll take care of them tomorrow
timotimo moritz: btw, you started the release progress; what can i do to help? 20:53
fill out the moarvm changelog perhaps?
moritz timotimo: yes (and maybe even release MoarVM) 20:54
timotimo right
dalek ast/glr: 0d1b8bf | lizmat++ | S02-types/mixhash.t:
Fix at least one mixhash.t test

More tomorrow
20:54 espadrine joined
dalek rl6-roast-data: b82045e | coke++ | / (2 files):
another glr run for 'today'
 20:57
20:58 skids left 20:59 espadrine left 21:01 Loren left 21:04 lucasb_ joined 21:05 bin_005 left 21:06 FROGGS left
lizmat m: Bag.new( (a => 42, b=>666)).perl.say 21:07
camelia rakudo-moar 5fb81f: OUTPUT«(:b(666)=>1,:a(42)=>1).Bag␤»
GLRelia rakudo-moar 7591fc: OUTPUT«(:b(666)=>1,:a(42)=>1).Bag␤»
lizmat m: Hash.new( (a => 42, b=>666)).perl.say
GLRelia rakudo-moar 7591fc: OUTPUT«{:a(42), :b(666)}␤»
camelia rakudo-moar 5fb81f: OUTPUT«{:a(42), :b(666)}<>␤»
lizmat m: Bag.new-from-pairs( (a => 42, b=>666)).perl.say 21:08
camelia rakudo-moar 5fb81f: OUTPUT«("a"=>42,"b"=>666).Bag␤»
GLRelia rakudo-moar 7591fc: OUTPUT«("a"=>42,"b"=>666).Bag␤»
lizmat fades away
21:08 lizmat left
lucasb_ Is there some rationale for print() using .Str and say() using .gist? Shouldn't they just differ with regard to "\n"? 21:10
timotimo say's supposed to be "for humans" and print "for computers" 21:11
[Coke] lucasb_: there was an argument made years ago for that, it was lost, it's probably not getting changed at this point. (I wasn't a fan either) 21:12
21:12 Sqirrel left 21:13 Sqirrel joined
jdv79 [Coke]: what was the short reason? 21:13
lucasb_ timotimo: I thought that was the difference between .perl and .gist. But if it applies to say/print, ok then
jdv79 seems nice there's a differentiation otherwise there's no need for 2 methods
lucasb_ timotimo, [Coke]: thanks for answering 21:14
[Coke] what timotimo said.
(for jdv79 )
PerlJam greetings! How's SPW going? 21:15
jdv79 i thought you meant for the other side
21:16 ab6tract left
hoelzro PerlJam: greetings! it's going quite well, thanks! 21:17
21:17 larion left, bin_005 joined, bin_005 left
PerlJam Did someone cut a nom release? 21:18
21:18 kanl left
PerlJam Or ... is someone planning on doing it? 21:18
timotimo we're working towards it, a little bit 21:19
but not really the people at the spw
21:21 bin_005 joined
PerlJam Well, I'd volunteered for this month and I'll have some free time later tonight (in about 4 hours or so). Should I just make a nom release? 21:21
timotimo moritz has started working on that today (or was it yesterday?); i'm going to bed real soon now, so i can't really be of help :( 21:23
japhb Anybody have links to SPW slide decks? (If there have been any, I've missed them in the voluminous backlog. Which is sortof like luminiferous aether, and likewise problematic in that you can't detect anything in it.) 21:24
laben good night|work #perl6 21:25
laben heads off to bed
PerlJam g'night laben
japhb o/ laben
21:25 laben left
PerlJam patiently waits for someone to answer japhb too ... 21:27
21:27 [Sno] left
japhb Yeah, here I am hoping to get a Friday afternoon Perl 6 talk fix .... 21:27
timotimo ohai japhb
how are you today? 21:28
lucasb_ has anybody mentioned that yapcasia videos are online? I watched timtoady's and rjbs's talk and liked it :)
timotimo oh!
where can they be found?
ah
lucasb_ youtube channel: www.youtube.com/user/yapcasia/videos
timotimo www.youtube.com/user/yapcasia/videos
look, it's jnthn! 21:29
japhb ohai timotimo 21:32
Not doing too bad; recovering from latest iteration of oral surgery, wheee
timotimo japhb: i've been meaning to run benchmarks against a rakudo/glr, but somehow i never got around to it 21:33
japhb Seems like enough bugs have been fixed that it might even work. :-) 21:34
timotimo oh yes
japhb Although, since last year, it should at least record the failure modes it finds ...
timotimo some of the tests may want changes so that they do the same thing, basically
huh 21:35
japhb True. Any obvious performance differences need to be investigated to make sure they still do the same amount of work.
timotimo someone put JSON::Fast into panda
but not a sufficiently new version of it
bootstrap now fails, because one of the tests requires a single trailing , to be added to the expected value
japhb Though TBH, that's why I added the ability to specify expected output -- so that we could go back and make the tests output an expected value that should only be reached if all the work was done. 21:36
timotimo: That seems like a wrong test ...
timotimo yes, glr made it required
ugh, bench list-checkouts is very unhappy with GLR :) 21:37
japhb lucasb_: Thank you for the YAPC::Asia video pointer. :-)
Oh, ouch, yeah, I bet.
There's a fair amount of sugar in bench that's probably flattening/itemization-sensitive.
21:38 itz joined
timotimo very much so 21:38
japhb Double pluses if you figure out patches that will work correctly under both, FWIW.
timotimo i'm just now rebuilding rakudo on nom on my desktop at home
so that i can bench on ahead :)
japhb Probably the best immediate course of action 21:39
timotimo immediate means "right before bedtime" ;)
japhb wonders what percentage of his collected Perl 6 code will have to be rewritten after the merge. :-/
heh
.oO( "Why didn't you do your chores already?" "Because you told me to do them immediately, and it's not bedtime yet!" )
 21:40
timotimo okay, building a very recent nom and a very recent glr now 21:42
i'm at a crossroads, and a decision is needed
do i buy new ram for my desktop because the 8gb just don't cut it any more 21:43
or do i just wait for my next big update and get DDR4 along with that?
21:46 virtualsue joined
timotimo now how am i supposed to go to bed when there's a benchmark going on and i could be watching for the exact moment it finishes so that i can see the performance difference immediately :| 21:46
japhb timotimo: Do you need to swap out all of your current memory sticks in order to add more memory to the existing desktop?
timotimo yes 21:47
but i have someone who's very interested in getting at least one of the sticks
japhb timotimo: Have it alarm/SMS/page/otherwise wake you when it's done?
timotimo ugh, that's even worse! :)
just started the "./bench time" call :) 21:48
japhb timotimo: Well, that makes it easier ... go to 16 GB then. :-)
timotimo Run command wrote to STDERR: install/bin/perl6 --optimize=3 -e 0
Compiler rakudo-moar is failing for test zero, continuing to next compiler/test.
well, that's not so good
hahahahaha
Useless use of constant integer 0 in sink context (line 1)
:D
japhb There aren't many times that I've regretted a memory upgrade. (The only time was when populating more slots in a single channel dropped the memory clock enough to steal much of the expected improvement in overall throughput) 21:49
timotimo oh! yikes
i have only two slots to begin with
at the moment they have two 4gb sticks in them
in theory, it'd already be a nice upgrade to replace one of them with an 8gb stick
japhb Yeah, I'm wondering if it ought to be possible to turn off the STDERR detection, because overzealous warnings/alerts are causing a problem. Or maybe I should just -Mno-strict the -e tester. 21:50
timotimo we can just have the "zero" program do something that does something that doesn't warn :) 21:51
like perl6 -e ''
or is that "empty" already?
japhb timotimo: ... as long as those two slots can be safely mismatched. Not all of them can. (Or at least, back the last time I had a desktop at home. Which was ... a while ago, actually.)
timotimo mhh 21:52
i expect it'll be fine
for_empty also warns about the final "1"
japhb empty is '', and it's intentionally there to not even do any parsing. It should be the fastest the compiler can be started in non-help/REPL mode.
timotimo we should implement a special case of "empty program" so we don't even have to deserialize the grammar :P
japhb :-P
timotimo in the repl, we bring up the first input prompt before the compiler causes the grammar to be loaded 21:53
hoelzro ok, bed time for realz
good night #perl6
PerlJam sleep well hoelzro!
timotimo gnite hoelzro!
japhb o/ hoelzro
timotimo still watching jnthn's concurrency talk 21:54
(which i missed via oversleeping tonight)
japhb Did he do the same one at SPW as at YAPC::Asia?
timotimo slightly changed, i believe
japhb Dang it, now I want that slide deck even more. :-)
timotimo you already have the previous one, yeah? 21:55
japhb Yeah. But he does so much CDD, that often the first place I find out about his latest cool stuff is in those slide decks. :-)
timotimo ah 21:56
21:58 rurban joined
timotimo first component is at 56/73 already 21:58
21:59 lucasb_ left
japhb thinks that 'my $z = 0;' might be a decent replacement for '0' that would be identical across all compilers (delta := for nqp) 21:59
timotimo ah
sounds good
that'll not warn any more 22:00
japhb Yeah, and still maintains its use as a near-minimal use of the parser (probably a much better one, too, come to think of it -- more likely to light up more of the grammar.)
timotimo yeah 22:01
japhb Care to patch, or shall I?
timotimo please go ahead :)
i'd be interested to see if rc-forest-fire is going to ever work properly in the future >_>
oh! 22:02
it works!
cool
dalek rl6-bench: 8b8d049 | (Geoffrey Broadwell)++ | microbenchmarks.pl:
Fix 'zero' microbenchmark not to sink-warn
 22:03
japhb Oh cool!
wheee, the joy of freehand patching 22:04
timotimo are fires supposed to randomly go out?
japhb timotimo: If they can't find more to burn nearby, yes. It's been a long time since I looked, but some implementations of that CA can put a new tree on top of a burning one. I don't *think* we do that. 22:06
22:06 rurban left
timotimo nah, i see burning trees with lots of neighbours and these neighbours are spared from burning up 22:06
22:06 rangerprice left
japhb If it goes out when there's adjacent trees, that's probably a bug 22:06
22:07 RabidGravy left
japhb Do the fires ever spread in the first place? 22:07
timotimo yes
japhb Oh, hmmm, that's more subtle territory
timotimo yeah
second component is now running
japhb There was a long time ago a bug in which neighbor information started pointing to old versions of the forest; IIRC TimToady++ fixed that in the nom days. 22:08
I wonder if it has reared it's ugly head again. 22:09
japhb imagines a t-shirt with the logo "UGLY BUG"
So many possible images to go with that logo/tagline. Or maybe none at all, and implicitly that role falls to the wearer. :-) 22:10
timotimo BUGLY 22:11
22:11 byterock joined
timotimo ;; AUTHORITY SECTION: 22:11
ly. 600 IN SOA dns.lttnet.NET. khaleds.lttnet.NET. 2015082889 3600 900 2419200 600
^- for bug.ly :)
japhb wishes there were a decent way to guarantee an equivalent random number stream between compilers, without limiting their freedom to write a fast PRNG
heh
timotimo yeah, that's a hard problem™
on GLR, the forest doesn't change any more at all 22:12
japhb I thought about just reading from the same file of random bits, but then you end up testing binary I/O performance ... 22:13
timotimo so we should take the benchmark results with more spoonsful of salt than we usually do
yeah :|
you could implement a dumb prng yourself
japhb true.
My all time favorite is one from TAOCP. Lemme see if I can find it. 22:14
timotimo github.com/Microsoft/WinObjC/blob/...isc.mm#L28 22:16
japhb Oh. Oh dear.
That's ... wow.
timotimo twitter.com/FioraAeterna/status/63...3636004865 - "This is a very creative RC4 implementation, huh" 22:17
who decided the monitor concurrency pattern should be called that? 22:18
22:19 itz left, virtualsue left
timotimo Type check failed in binding; expected 'Positional' but got 'Seq' 22:20
in sub MAIN at perl6/rc-forest-fire-stringify:8
quite GLR-ish
that's "my @grid := [ ... ] xx $h"
japhb Yeah, sounds like it. 22:22
timotimo yup. probably really wants to be = instead of :=
10 minutes of jnth's yapc::asia talk remaining, but only a single benchmark
time to remember the ./bench compare incantation again 22:24
japhb Not finding the algorithm online, I went to my bookshelf. It's algorithm 3.2.2.A in my second edition copy: Basically fill 55 elements of a 56-element circular queue with arbitrary numbers (not all even), and then just keep doing Xn = X(n-24) + X(n-55) mod machine word size. Works whether the bit sequences are interpreted as integers or as floats, even. Deterministic but long-period and REALLY fast. 22:25
timotimo neat 22:28
we'd really want a mersenne twister in our ecosystem at some time
there's quite fast implementations of those using simd instructions even
hm 22:30
now which one was glr and which one was nom?
t.h8.lv/p6bench/2015-08-29-nom_vs_glr.html - blue is nom, orange is glr 22:31
22:31 llfourn joined
timotimo looks pretty bad 22:31
m: my @a; my $i = 0; @a[$i] = $i; while (++$i <= SCALE) { @a[ $i ] = $i }; say @a[*-1];' 22:33
camelia rakudo-moar 5fb81f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/M447ppGz9O␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/M447ppGz9O:1␤------> 3 SCALE) { @a[ $i ] = $i }; say @a[*-1];'7⏏5<EOL>␤ expecting any of:␤ sing…»
GLRelia rakudo-moar 7591fc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/hOGsBrnyXJ␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/hOGsBrnyXJ:1␤------> 3 SCALE) { @a[ $i ] = $i }; say @a[*-1];'7⏏5<EOL>␤ expecting any of:␤ sing…»
timotimo m: constant SCALE = 1024; my @a; my $i = 0; @a[$i] = $i; while (++$i <= SCALE) { @a[ $i ] = $i }; say @a[*-1];'
camelia rakudo-moar 5fb81f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Wj6P7eW9Qa␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/Wj6P7eW9Qa:1␤------> 3 SCALE) { @a[ $i ] = $i }; say @a[*-1];'7⏏5<EOL>␤ expecting any of:␤ sing…»
GLRelia rakudo-moar 7591fc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ojhqffgMU1␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/ojhqffgMU1:1␤------> 3 SCALE) { @a[ $i ] = $i }; say @a[*-1];'7⏏5<EOL>␤ expecting any of:␤ sing…»
timotimo m: constant SCALE = 1024; my @a; my $i = 0; @a[$i] = $i; while (++$i <= SCALE) { @a[ $i ] = $i }; say @a[*-1];
camelia rakudo-moar 5fb81f: OUTPUT«1024␤»
GLRelia rakudo-moar 7591fc: OUTPUT«1024␤»
timotimo what is going on here ... 22:34
that benchmark i pasted above has a 4x slowdown :(
well, we haven't really optimized things at all yet on the glr side 22:35
japhb Actually, it looks like mixed results to me. Some of the tests, like while_push*, show a good benefit from GLR, but while_array_set is much worse.
timotimo i still think it's bad ;)
japhb Ooh, for loops took a huge hit.
22:35 llfourn left
japhb I'm still scrolling through it 22:35
reduce is mixed. any_equals is way worse 22:36
timotimo i wonder if the for-to-while optimization had been in effect in nom, but isn't any more in glr
that optimization is super fiddly
japhb As are split_string*
timotimo breaks every few revision
japhb I bet you are correct.
Hmmm, charrange (and a couple others) seem more unstable
timotimo how do you feel about building a nom and a glr locally and comparing their --target=optimize while i rest? :) 22:37
japhb timotimo: You should definitely rest. Unfortunately, I should definitely get back to $day-job. 22:38
22:40 rangerprice joined 22:41 rarara joined 22:42 TEttinger joined
timotimo good luck at $day-job! 22:45
ttyl :)
22:56 skids joined 22:58 yqt left 23:00 rarara left 23:03 Timbus_ joined, Timbus left 23:04 byterock left 23:05 rangerprice left 23:07 larion joined 23:13 larion left
jdv79 wonder what a p6 "brush" for github.com/syntaxhighlighter would look like 23:20
where are the existing p6 syntax highlighers?

Powered by Cro and the Raku® Programming Language.

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

Thank you!