»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
00:02 wamba left
gfldex lolibloggedforcale2: gfldex.wordpress.com/2017/02/02/ca...-question/ 00:03
00:05 titsuki left
gfldex WP tried to be clever again :-| 00:08
timotimo "Today cale2 refused to asked an easy question instead" what is this grammar %) 00:11
man, wordpress really is hell 00:12
gfldex i think i got it now
it ate a <pre> again :-/
timotimo yeah, and it ate at least one list between <> 00:13
in the @a.pick(*) example
and in the multi sub whatever case, too
gfldex i'm cleary updating it faster then you can read it :-> 00:14
timotimo could be
i find the last sentence of the post could be interpreted to be rather mean
and personally i'd have put a note about "multiple * in a WhateverCode make multiple arguments" or something 00:15
just a tiny hint
00:15 pyrimidine joined
samcv pretty is good 00:15
timotimo holds a pretty grudge
gfldex the last sentence is the best I could come up with to shoehorn a link to the docs in 00:17
00:17 cpage_ left
gfldex whereby WTFM works better then RTFM 00:18
timotimo "what the fucking manual"? :) 00:19
samcv wtf Manuel! 00:20
hobbs timotimo: I've read a few of those
timotimo oh manuel, what have you done this time
hobbs: pardon me, a few of what? 00:21
00:21 cdg_ left
hobbs what the fucking manuals 00:21
documentation with high WTF content :)
00:21 aindilis` left
samcv where's the fucking manual 00:21
gfldex there are write-only manuals 00:22
00:22 pyrimidine left
timotimo ah, hehehe 00:22
samcv: where's the fucking manual, lebowski?
samcv docs.perl6.org
timotimo hm. you could totally make a gif out of that scene with that wording
gfldex i actually meant Write The Fucking Manual, what really works better then reading them
00:22 rurban left
timotimo "let me look again, i think the docs are somewhere down there" 00:22
hobbs clearly you're not a perl golfer 00:23
s/clearly/obviously/ 00:24
gfldex m: constant term:<ℵ₀> = Inf; constant term:<ℵ₁> = Inf; multi sub infix:«<»(ℵ₀, ℵ₁){ True }; say ℵ₀ < ℵ₁;
camelia rakudo-moar 738493: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Invalid typename 'ℵ' in parameter declaration.␤at <tmp>:1␤------> 3t term:<ℵ₁> = Inf; multi sub infix:«<»(ℵ7⏏5₀, ℵ₁){ True }; say ℵ₀ < ℵ₁;␤»
samcv sub foo($a where * < 10) === sub foo($a where {$_ < 10}) ?? 00:25
gfldex m: multi sub infix:«<»(2, 1){ True }; say 2 < 1;
camelia rakudo-moar 738493: OUTPUT«True␤»
timotimo yup 00:26
00:27 travis-ci joined
samcv sub foo(*@a where {.all ~~ Int}){} 00:27
travis-ci Doc build passed. Will "Coke" Coleda 'update bug report notes for #1077'
travis-ci.org/perl6/doc/builds/197411799 github.com/perl6/doc/compare/2219c...c216bf4ab0
00:27 travis-ci left
samcv seems you can't do (*@a where *.all ~~ Int); doesn't seem to work right 00:27
TimToady m: say *.all ~~ WhateverCode 00:29
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken m: say (*.all ~~ 'meow') ~~ WhateverCode
camelia rakudo-moar 738493: OUTPUT«WhateverCode object coerced to string (please use .gist or .perl to do that)␤ in block <unit> at <tmp> line 1␤False␤»
brokenchicken it ain't currying over the ~~
TimToady nope, it's one of the exeptions 00:30
brokenchicken don't need it in that particular case either, 'cause it thunks; (*@a where .all ~~ Int) should work
TimToady note it doesn't curry on the RHS either, or you'd be surprised
m: say so Int.ACCEPTS((1,2,3).all) 00:31
camelia rakudo-moar 738493: OUTPUT«True␤»
TimToady can always write it like that 00:32
brokenchicken not with a *.all curry
TimToady nope, currying has to stop somewhere, and we never curry from inside argument lists to normal functions/methods 00:33
brokenchicken m: sub (*@a where Int.ACCEPS: .all with *.item) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Malformed parameter␤at <tmp>:1␤------> 3sub (*@a where Int.ACCEPS: .all with 7⏏5*.item) { say @a }(1, 2, 3)␤»
brokenchicken aww
Need moar parentheses
m: sub (*@a where (Int.ACCEPS: .all with *.item)) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«No such method 'ACCEPS' for invocant of type 'Int'␤ in sub at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken m: sub (*@a where (Int.ACCEPTS: .all with *.item)) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«Constraint type check failed for parameter '@a'␤ in sub at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken Oh well
m: dd so 42 ~~ (1, 2, 3).item.all 00:34
camelia rakudo-moar 738493: OUTPUT«Bool::False␤»
brokenchicken I see
Is there an identity method. That just returns self for any object?
TimToady m: say 42.self 00:35
camelia rakudo-moar 738493: OUTPUT«42␤»
brokenchicken doh :) It's .self :)
m: sub (*@a where (Int.ACCEPTS: .all with *.self)) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«Constraint type check failed for parameter '@a'␤ in sub at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken Oh well
gfldex where is .self defined? 00:36
TimToady m: sub (*@a where (Int.ACCEPTS: .all)) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«[1 2 3]␤»
brokenchicken gfldex: in Mu
TimToady m: sub (*@a where @a.all ~~ Real) { say @a }(1, 2, 3) 00:38
camelia rakudo-moar 738493: OUTPUT«[1 2 3]␤»
TimToady m: sub (*@a where @a.all ~~ Int) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«[1 2 3]␤»
brokenchicken m: sub (*@a where *.all.&{Int.ACCEPTS: $_}) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«[1 2 3]␤»
brokenchicken m: sub (*@a where *.all ∈ Int.Range) { say @a }(1, 2, 3) 00:41
TimToady that probably curries
camelia rakudo-moar 738493: OUTPUT«(timeout)»
TimToady however, it probably also flattens Int.Range 00:42
brokenchicken m: dd Int.Range
camelia rakudo-moar 738493: OUTPUT«(-Inf, -Inf, -Inf, -Inf, -Inf, -Inf, -Inf, -Inf, -Inf, -Inf... lazy list)␤»
TimToady we don't really do infinite sets...
brokenchicken m: sub (*@a where .all ~~ Int.Range) { say @a }(1, 2, 3)
camelia rakudo-moar 738493: OUTPUT«[1 2 3]␤»
TimToady that's why it's called "smart" :P 00:43
brokenchicken :)_
Geth doc: 1f1947b763 | (Wenzel P. P. Peppmeyer)++ | doc/Type/Mu.pod6
doc Mu.self
 00:44
TimToady m: say i ~~ Int.Range
camelia rakudo-moar 738493: OUTPUT«True␤»
TimToady hrmmm 00:45
m: say i.Int
camelia rakudo-moar 738493: OUTPUT«Can not convert 0+1i to Int: imaginary part not zero␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
TimToady m: say i < Inf
camelia rakudo-moar 738493: OUTPUT«Can not convert 0+1i to Real: imaginary part not zero␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
00:47 aindilis joined
brokenchicken s: Int.Range, 'ACCEPTS', \(i) 00:47
SourceBaby brokenchicken, Sauce is at github.com/rakudo/rakudo/blob/7384...ge.pm#L419
00:48 aborazmeh joined, aborazmeh left, aborazmeh joined
brokenchicken m: say i after -Inf 00:48
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken That's why. 00:49
m: say i cmp -Inf 00:50
camelia rakudo-moar 738493: OUTPUT«More␤»
brokenchicken hmmmm
m: say i cmp -42
camelia rakudo-moar 738493: OUTPUT«More␤»
brokenchicken m: say i cmp 42
camelia rakudo-moar 738493: OUTPUT«Less␤»
Geth doc: d4d9df69a9 | (Wenzel P. P. Peppmeyer)++ | doc/Language/syntax.pod6
doc term:<> (could do with better example)
 00:51
TimToady it shouldn't really be using cmp for that
cmp tries hard to work across types
brokenchicken s: &infix:<after>, \(i, -Inf)
SourceBaby brokenchicken, Sauce is at github.com/rakudo/rakudo/blob/7384...ny.pm#L479
brokenchicken Yeah, it's just (a cmp b) > 0
TimToady: so before/after should throw on Complex<->Real comparisons unless img. part is zero? 00:52
TimToady a numeric range should be using <=>, I'd think
brokenchicken m: say i <=> 42
camelia rakudo-moar 738493: OUTPUT«Can not convert Complex to Real: Complex is not numerically orderable␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken Aha. Cool. 00:53
TimToady m: say 'bb' ~~ 'a'..'z' 00:54
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken :( 00:55
TimToady that's correct
brokenchicken Huh?
m: say eager 'a'..'z'
camelia rakudo-moar 738493: OUTPUT«(a b c d e f g h i j k l m n o p q r s t u v w x y z)␤»
brokenchicken but it's not one of items?
TimToady nope, ranges are, er, ranges, not sequences
00:56 skids joined
TimToady gotta use any('a'..'z') for that 00:56
brokenchicken Ah ok
TimToady m: say 'zoo' ~~ 'a'..'z' 00:57
camelia rakudo-moar 738493: OUTPUT«False␤»
TimToady you want to carp, carp about that one :)
brokenchicken I don't get how 'zz' fits in 'a'..'z' range tho
I mean 'bb'
and 'zz' doesn't
m: say 'bbbbbbbb' ~~ 'a'..'z' 00:58
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken hehe
m: say 'bbbaaaabbbbb' ~~ 'a'..'z'
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken m: say 'bbbaaaabbbbb' ~~ 'z'..'a' 00:59
camelia rakudo-moar 738493: OUTPUT«False␤»
TimToady m: say 'a!!!!!' ~~ 'a'..'z'
camelia rakudo-moar 738493: OUTPUT«True␤»
TimToady m: say "a\0" ~~ 'a'..'z'
camelia rakudo-moar 738493: OUTPUT«True␤»
TimToady m: say "a\0" ~~ 'a' ^.. 'z'
camelia rakudo-moar 738493: OUTPUT«True␤»
01:04 kurahaupo__ left 01:05 kurahaupo__ joined 01:06 john51 joined 01:09 kurahaupo__ left 01:13 pyrimidine joined 01:18 pyrimidine left 01:20 cpage_ joined
brokenchicken Simply using `(nqp::istype(a, Numeric)||nqp::istype(b, Numeric) ?? a <=> b !! a cmp b) < 0` for before/after doesn't work 01:21
Spectest fails, like with 'a'..*
.oO( add Numeric, Numeric candidate... )
 01:22
But then we get weirdness like `'42' after i` gives something, but `42 after i` says can't make a Real outta i 01:23
m: say '42' after i
camelia rakudo-moar 738493: OUTPUT«True␤»
brokenchicken m: say '-42' after i
camelia rakudo-moar 738493: OUTPUT«False␤»
01:23 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg
brokenchicken hm, maybe it's fine, since cmp uses .Stringy in that case 01:25
01:26 cpage_ left, agentzh joined
brokenchicken You know what'd be a cool bot feature? You give it two ops and it tells you which one got higher precedence 01:27
m: &infix:«>».prec.say
camelia rakudo-moar 738493: OUTPUT«{}␤»
brokenchicken hmm
01:36 cpage_ joined 01:39 astj_ left
brokenchicken Can't get precedence of an op programatically? 01:40
what's .prec?
m: &infix:«+».prec.say 01:41
camelia rakudo-moar 738493: OUTPUT«{assoc => left, prec => t=}␤»
brokenchicken oh, you can. Wonder why there's nothing for `>`
01:45 geekosaur left 01:46 pierrot left, pierrot joined 01:47 geekosaur joined 01:49 astj joined
brokenchicken m: say i <=> Inf 01:49
camelia rakudo-moar 738493: OUTPUT«Less␤»
brokenchicken And the rabbit hole deepens :P 01:50
MasterDuke samcv: did you see my comment about ternaries? i tested on your code, about 5-10% faster 01:51
samcv no i didn't MasterDuke 01:52
which part?
brokenchicken m: say i <=> 42
camelia rakudo-moar 738493: OUTPUT«Can not convert Complex to Real: Complex is not numerically orderable␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken s: &infix:«<=>», \(i, 42)
SourceBaby brokenchicken, Sauce is at github.com/rakudo/rakudo/blob/7384...ex.pm#L477
brokenchicken I don't get how that managed to throw
m: say i <=> 42.Complex 01:53
camelia rakudo-moar 738493: OUTPUT«Can not convert Complex to Real: Complex is not numerically orderable␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken m: say 42.Complex
camelia rakudo-moar 738493: OUTPUT«42+0i␤»
brokenchicken m: say i <=> <42+0i>
camelia rakudo-moar 738493: OUTPUT«Can not convert Complex to Real: Complex is not numerically orderable␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken :S wtf
s: &infix:«<=>», \(i, <42+0i>)
SourceBaby brokenchicken, Sauce is at github.com/rakudo/rakudo/blob/7384...ex.pm#L467
mspo m: say 0i.perl
camelia rakudo-moar 738493: OUTPUT«<0+0i>␤»
brokenchicken # Fail unless imaginary parts are relatively negligible, compared to real parts. 01:54
I see
A bit of a confusing error, since I thought it was trying to convert X to Real
m: say Inf.Complex 01:55
camelia rakudo-moar 738493: OUTPUT«Inf+0i␤»
brokenchicken Seems there's a bit of a glitch either in that thing or in the =~= op involving Infs 01:57
.oO( how did I manage to start debugging this instead of playing my video games :/ )
MasterDuke samcv: replacing gist.github.com/samcv/566950400b06...p6-L57-L62 with `%decomp_spec{$cp}<type> = nqp::eqat(nqp::atpos(@dec, 0), '<', 0) ?? nqp::shift(@dec) !! 'Canonical'` was faster 01:58
samcv ah 01:59
02:01 agentzh left
brokenchicken tis the tolerance computation. Ends up being Inf 02:01
samcv thanks 02:03
MasterDuke np
02:04 noro left, melezhik left, travis-ci joined
travis-ci Doc build passed. Will "Coke" Coleda 'learn new label' 02:04
travis-ci.org/perl6/doc/builds/197420375 github.com/perl6/doc/compare/33c21...1d921bcb40
02:04 travis-ci left, dugword joined 02:08 dugword left
brokenchicken m: say <42+42i> after <42+43i> 02:10
camelia rakudo-moar 738493: OUTPUT«False␤»
brokenchicken m: say <42+42i> after <42+41i>
camelia rakudo-moar 738493: OUTPUT«True␤»
02:11 agentzh joined
brokenchicken TimToady: ^ that now throws. Should it work? The <=> allows Complex comparison only if imaginary parts are negligible compared to Real parts. While with `cmp` just `cmp`s real/imaginary parts separately 02:14
I wish I knew more about Complex numbers :(
SmokeMachine m: role R[::T] {has T:U $.t} #is that right? 02:15
camelia rakudo-moar 738493: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable definition of type T:U requires an initializer␤at <tmp>:1␤------> 3role R[::T] {has T:U $.t7⏏5} #is that right?␤ expecting any of:␤ constraint␤»
brokenchicken Don't know if `cmp` on Complex is wrong or my patch to before/after for Complex is wrong.
SmokeMachine m: role R[::T] {has T $.t} 02:16
camelia ( no output )
SmokeMachine m: role R[::T] {has T:_ $.t}
camelia ( no output )
SmokeMachine m: role R[::T] {has T:U $.t = T}
camelia ( no output )
brokenchicken m: role { has Int:U $.t }
camelia ( no output )
brokenchicken weird
SmokeMachine: I'm no expert, but it looks buggy to me
SmokeMachine brokenchicken: yes... weird... 02:17
brokenchicken: to me to...
too
that would make sense with T:D
m: role R[::T] {has T:D $.t} 02:18
camelia rakudo-moar 738493: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable definition of type T:D requires an initializer␤at <tmp>:1␤------> 3role R[::T] {has T:D $.t7⏏5}␤ expecting any of:␤ constraint␤»
02:18 cpage_ left
brokenchicken .ask TimToady should `<42+42i> after <42+41i>` throw? It used to work, but now that I made `after` use `<=>` for Numeric+Numeric it only works if img part is negligible github.com/rakudo/rakudo/commit/ab3162c127 02:24
yoleaux brokenchicken: I'll pass your message to TimToady.
02:28 cale2 joined 02:29 yqt left
brokenchicken I think I misunderstood TimToady :) I now see in the log he said the numeric *range* should use <=> not before/after 02:35
TimToady no, it shouldn't throw; I was talking about changing Range's smartmatch, not changing cmp
yoleaux 02:24Z <brokenchicken> TimToady: should `<42+42i> after <42+41i>` throw? It used to work, but now that I made `after` use `<=>` for Numeric+Numeric it only works if img part is negligible github.com/rakudo/rakudo/commit/ab3162c127
brokenchicken :D 02:36
OK. Thanks for clarifying :)
02:39 agentzh left
brokenchicken added it to my todo list to fix tomorrow evening 02:39
02:39 pyrimidine joined
TimToady cmp has to stay mellow, so we can always sort a list 02:41
brokenchicken Yeah, makes sense :) 02:44
'night \o
02:44 pyrimidine left 02:45 ilbot3 left 02:47 ilbot3 joined, ChanServ sets mode: +v ilbot3 02:52 melezhik joined 03:08 Actualeyes left 03:10 cpage_ joined 03:11 Actualeyes joined 03:20 dugword joined 03:24 xtreak joined 03:28 travis-ci joined
travis-ci Doc build passed. Will "Coke" Coleda 'link to "how to report a bug" ; closes #1077' 03:28
travis-ci.org/perl6/doc/builds/197421011 github.com/perl6/doc/compare/d91d9...cd18d3143d
03:28 travis-ci left 03:31 agentzh joined 03:35 cdg joined 03:36 cale2 left 03:37 Sgeo joined 03:38 noganex_ joined 03:40 noganex left 03:44 zacts left 03:56 pyrimidine joined 04:04 zacts joined 04:09 agentzh left 04:29 agentzh joined 04:40 hartenfels joined 04:49 geekosaur left 04:50 geekosaur joined 04:51 agentzh left 04:53 sakawaka left 04:58 Cabanoss- joined 04:59 itcharlie left 05:00 agentzh joined, Sgeo left 05:02 Cabanossi left, Cabanoss- is now known as Cabanossi 05:14 khw left, kyan left 05:15 agentzh left 05:23 agentzh joined 05:24 CIAvash joined 05:26 cdg left 05:27 aborazmeh left, Xliff joined 05:28 Xliff left 05:30 cpage_ left 05:37 Dandel_ left 05:41 BenGoldberg left 05:56 Xliff joined 05:57 cpage_ joined, Xliff left 05:59 cpage_ left, agentzh left 06:00 rurban joined, rurban left 06:05 cpage_ joined 06:06 cpage_ left, cpage_ joined 06:09 cog_ joined, noganex joined 06:11 cpage_ left 06:12 lep-delete left, stmuk_ joined, noganex_ left, _28_ria left, cognominal left, captain-adequate left, inokenty left 06:13 wamba joined, avuserow left, breinbaas left, sivoais left, cgfbee left, Celelibi left, _28_ria joined, inokenty joined 06:14 japhb joined, sunnavy joined, sivoais joined 06:15 captain-adequate joined 06:16 lep-delete joined, cpage_ joined 06:18 Celelibi joined 06:21 breinbaas joined 06:22 TEttinger left 06:27 avuserow joined 06:29 TEttinger joined 06:33 cgfbee joined 06:35 hartenfels1 joined, domidumont joined 06:36 hartenfels left, hartenfels1 is now known as hartenfels
masak lol, I blog'd! strangelyconsistent.org/blog/deep-git 06:37
06:41 domidumont left 06:42 domidumont joined 06:51 RabidGravy joined 06:56 xtreak left 07:00 xtreak joined 07:03 dugword left 07:08 domidumont left
tadzik :o 07:09
07:15 darutoko joined 07:26 bjz joined, KDr2 joined 07:31 KDr2 left 07:32 wamba left 07:35 astj_ joined, astj left 07:41 matt_ left 07:42 skids left 07:49 domidumont joined 07:51 matt_ joined, matt_ is now known as Guest15861
moritz masak++ 07:57
08:08 zacts left 08:09 Guest15861 is now known as matt_, tokomer joined 08:21 zakharyas joined 08:30 tokomer left
jast masak: very good post. one minor nitpick: git knows two different meanings for 'merge'. one is commit-level merge, one is file-level three-way merge. the latter is used in rebase, cherry-pick etc., too, so technically those conflicts can still be called merge conflicts. :) 08:32
08:35 tokomer joined 08:41 bjz_ joined, rurban joined 08:42 dakkar joined, bjz left 08:49 xtreak left
samcv nice my UCD-gen.p6 script is now fast enough that it takes only 80 minutes to finish the slowest routine ;) 08:50
08:51 zacts joined
samcv faster on my laptop i think but that 80 minutes was on the bisectable server 08:51
running again to see how my latest speedups have fared
08:51 xtreak joined 08:55 matt_ left 08:59 tokomer left 09:00 astj joined, astj_ left 09:01 matt_ joined, matt_ is now known as Guest17802 09:02 Guest17802 is now known as matt_ 09:06 Vynce joined
samcv i-i think maybe it's slowing down because it has too many keys in the hash? 09:09
toward the end of the file it seems to be very very slow, and i'm not sure if it's caused by collisions in the hash library? 09:10
it starts getting super slow around 1048576 items in it
lizmat hows memory usage going ? 09:13
samcv uh using 6% of my memory, i have 16GB
09:13 xtreak left
samcv not at a million yet on the current run, but at a million it seems like 09:15
i mean it probably slows down before amillion, but it's most perceptible cause that's like almost at the end of the fgile
and it still takes minutes to get to the end
09:18 xtreak joined
lizmat hmmm.... 09:21
09:21 Actualeyes left
samcv could be caused by hash collisions maybe? or something else? 09:21
09:21 andrzejku joined
lizmat I'm not sure how we hash nowadays: I seem to recall it was changed not too long ago by jnthn 09:22
andrzejku hey people
I know that some of you are web developers
tell me please how do you prototype your apps
do you use some prototype tools?
or just a pencil
samcv so the file has 30,000 lines it it. maybe i'll make a timer every 1,000 lines so i can get some data about this 09:23
09:29 ufobat joined
samcv wow so i get to use `gather/take` for the first time now 09:29
always wanted to use this
09:39 pyrimidi_ joined, pyrimidine left 09:40 cpage_ left 09:45 Actualeyes joined
Geth doc: d1788cadb3 | (Samantha McVey)++ | util/trigger-rebuild.txt
Trigger rebuild to pull in atom-language-perl6 v1.14.0
 09:50
samcv andrzejku, i often find thinking about any computer stuff drawing can help.
i really should employ that more often tbh
09:57 rindolf left, agentzh joined
masak jast: interesting point. tempted to include that as an aside in the post ;) 09:58
10:02 agentzh left 10:04 labster left 10:06 riatre joined 10:08 rindolf joined 10:12 riatre left 10:13 riatre joined 10:17 rindolf left
samcv lizmat, here's a CSV of timings taken every 500 lines gist.github.com/samcv/6290f3ac0667...b5347e65c2 10:18
that one doesn't look particularly bad 10:20
but definitely gets higher as it goes on
arnsholt Those are timings *between* each batch, not cumulative, right?
samcv not cummulative
err wait that's not all of them i think 10:21
err maybe it is
yeah that's all that showed up at least
running it again though
i think it might not be measuring a very last number 10:23
which is where it's the slowest but 10:24
Took 1137.77617592 secs to process 30564 and 37.2260 ms/line 10:25
that doesn't sound right 10:26
it can't have taken that little time
10:26 abraxxa joined
samcv i guess 18 sounds close though 10:26
it maybe is just taking a long time to process lines which apply to a range of codepoints hmm 10:27
so maybe it's somewhat that. it does seem to get slower as it goes on but i think most of the slowdown could be those lines which it has apply for thousands of cp's 10:28
i will change it so that it takes timings every 1000 cp's instead of per lines of the file
but it's weird because i timed how long it'd take to apply a pretty big range in a test file and it was pretty fast 10:29
but that was with a previously empty hash, not wit one with almost a million things in it
10:30 astj left, abraxxa left, astj joined 10:32 astj left, astj joined
samcv yeah it looks like that's what's slowing down it. i'm unsure how to make this faster argh 10:33
10:33 bjz joined 10:34 astj left, bjz_ left
samcv i mean it shouldn't take THAT long to bind 65,000 keys? but all the keys are identical 10:35
10:35 melezhik_ joined
melezhik_ Hi! 10:35
samcv i'm using nqp::bindkey(%points, $cp_s, $hashy);
melezhik_ need a little help with regexp
samcv fire away melezhik_
arnsholt, is it possible it's duplicating the object?
melezhik_ I want to match none empty symbols between '[' and ']' 10:36
10:36 astj joined, cpage_ joined
samcv none empty? 10:36
10:36 cpage_ left
samcv what does that mean 10:36
10:36 AlexDaniel joined
melezhik_ in perl5 I would have /\[(\S+)\]/ 10:36
10:37 eiro left
melezhik_ how to say this in perl6? 10:37
samcv / '[' ( \S+ ']' ) /
that should work
melezhik_ s/none empty/none spaces/
samcv you don't need the space
melezhik_ samcv: thanks, will try
samcv that i put in my example. but in perl 6 spaces are not significant
melezhik_ yeah, I know 10:38
jast the exact same expression as in perl5 works for me, too
melezhik_ but I still perl5 in mind as for the regex, so it's hard for me every time I start writting perl6 ones , hehh (:
moritz blatantly advertises the regexes and grammar chapter in leanpub.com/perl6 10:39
10:41 tokomer joined 10:43 MasterDuke left 10:44 eiro joined
CIAvash melezhik_: also /'[' ~ ']' (\S+)/ 10:44
melezhik_ CIAvash: - thanks 10:46
Indeed my initial perl6 regex was fine, but I used $1 instead of $0 (:
I forgot that perl6 starts count numbered captures with $0 10:47
10:49 abraxxa joined 10:54 araujo joined, araujo left, cschwenz joined, araujo joined 10:55 cschwenz left 10:57 espadrine joined 10:58 AlexDaniel left, tokomer left 11:00 AlexDaniel joined 11:03 xtreak left 11:04 xtreak joined, xtreak left, xtreak joined 11:07 cpage_ joined
masak m: multi sub trait_mod:<is>(Routine $r, :$ohhai!) is export(:DEFAULT, :traits) {}; sub foo($x) is ohhai {} 11:10
camelia ( no output )
masak ...but...
m: multi sub trait_mod:<is>(Routine $r, :$ohhai!) is export(:DEFAULT, :traits) {}; sub foo($x) is ohhai($x) {} 11:11
camelia rakudo-moar 18e6f0: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Can't use unknown trait 'is ohhai' in a sub declaration.␤at <tmp>:1␤ expecting any of:␤ rw raw hidden-from-backtrace hidden-from-USAGE␤ pure default DEPRECATED inlinable nodal␤ …»
masak why is the trait suddenly unknown when I pass it a parameter? :)
brokenchicken melezhik_: that can be avoided with named captures: /'[' ~ ']' $<stuff>=\S+/
masak m: multi sub trait_mod:<is>(Routine $r, :$ohhai!) is export(:DEFAULT, :traits) {}; sub foo($x) is ohhai(42) {}
camelia ( no output )
brokenchicken melezhik_: and you use $<stuff> to access it
masak oh, it's only when I give it an $x from the parameter list
arnsholt samcv: Hard to say. But you could very well be running into performance issues in large hashes
masak m: multi sub trait_mod:<is>(Routine $r, :$ohhai!) is export(:DEFAULT, :traits) {}; my $x; sub foo() is ohhai($x) {}; say "alive" 11:12
camelia rakudo-moar 18e6f0: OUTPUT«alive␤»
masak everything works except passing an `$x` from the parameter list. that's what I thought.
samcv arnsholt, it seems it made it hugely faster by moving some stuff from the sub it called into the calling sub
and it was not much code in there anyway so idk
arnsholt samcv: At any rate, there does look to be at least one non-trivial GC-pause in the timings you posted: Most of them are tens of seconds, but one's 10 seconds
samcv so that seems the bulk of the issue
masak problem is, `$x` is in *scope* at that point, but (due to timing issues) it doesn't really make *sense* to pass `$x` to the trait
arnsholt Ah, cool
masak we should probably have a dedicated error message for this, just in case other people are as unreasonable as I am :P 11:13
samcv also doing for %points.keys -> $code; i get Use of uninitialized value <element> of type Any in string context.
Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.
in block at ./UCD-gen.p6 line 366
11:14 ufobat left
samcv and i have no clue what it can't use in a string context and $code is populated with the key inside the body of the for loop 11:14
it functions perfectly so i used to have a quietly on there. but that makes it slower so
i need to figure out how to solve this :\
this is the code github.com/samcv/UCD/blob/master/U...#L362-L366 11:16
11:20 confundus joined
samcv even if i do `my @iter = %points.keys.List;` i still get the problem when iterating over @iter 11:23
oh looks like it's not the iterating but another place. himm. nvm then 11:24
arnsholt masak: Should parameters be in scope for a trait application though? 11:26
That sounds odd, to me
11:29 AlexDaniel left
jnthn To do otherwise would mean a scope "hole" 11:30
It's not really a scoping problem, more a timing problem.
masak what jnthn++ said 11:31