»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.
00:00 lizmat left 00:16 llfourn joined 00:20 llfourn left 00:21 tinyblak left 00:26 laouji joined 00:31 tinyblak joined 00:39 laouji left 00:40 laouji joined, laouji left 00:48 araujo_ joined 00:50 laouji joined, araujo left 00:55 khisanth_ is now known as Khisanth
pmichaud GLR-related challenge (a problem that once I know the solution will help with GLR design): 01:03
Write a function that given an array @a, lazily produces a list in the form (0, Write a Perl 6 function that given an array @a, lazily produces a list in the form (0, @a[0,0], 1, @a[1,1], 2, @a[2,2], 3, @a[3,3], ... ).
(in standard Perl 6)
01:03 lizmat joined
pmichaud you can assume that while/until/etc. loops are able to lazily return lists of elements 01:04
grondilu you sure you wrote that right?
pmichaud which part are you unsure of? ;-)
grondilu it's not some recursive thing is it? 01:05
pmichaud no, it should be iterative
(at least I'm thinking it should be iterative)
it's a question of syntax and composition more than anything else.
grondilu but I'm confused about the part "(0, Write"
pmichaud oh 01:06
yeah, copy/paste error
fixing
Write a Perl 6 function that given an array @a, lazily produces a list in the form (0, @a[0,0], 1, @a[1,1], 2, @a[2,2], 3, @a[3,3], ... ).
*there*
grondilu does not look so hard 01:07
pmichaud it has a couple of gotchas :)
grondilu oh it must flatten, right?
pmichaud no, no flattening
the equivalent output would be (0, (@a[0], @a[0]), 1, (@a[1], @a[1]), 2, (@a[2], @a[2]), ... ) 01:08
grondilu sub (@a) { gather for ^Inf -> $n { take $n, @a[$n, $n].item } } # is that too naice? 01:09
*naive
01:09 telex left
pmichaud grondilu: #1. It's itemizing the slices 01:09
(which my output doesn't have) 01:10
01:10 telex joined
pmichaud the list that sub is producing is actually ((0, [@a[0], @a[0]]), (1, [@a[1], @a[1]]), ... ) 01:11
grondilu fine
well then just double the takes
sub (@a) { gather for ^Inf -> $n { take $n; take @a[$n, $n].item } }
or 01:12
sub (@a) { gather for ^Inf -> $n { take $n; take @a[$n, $n] } }
01:12 araujo_ left
grondilu no? 01:12
pmichaud thinking
that seems plausible, yes.
grondilu m: (sub (@a) { gather for ^Inf -> $n { take $n; take @a[$n, $n] } })( my @ = "a".."z" )
camelia rakudo-moar 80f8ef: OUTPUT«(timeout)» 01:13
grondilu m: say (sub (@a) { gather for 0..* -> $n { take $n; take @a[$n, $n] } })( my @ = "a".."z" )
camelia rakudo-moar 80f8ef: OUTPUT«0 a a 1 b b 2 c c 3 d d 4 e e 5 f f 6 g g 7 h h 8 i i 9 j j 10 k k 11 l l 12 m m 13 n n 14 o o 15 p p 16 q q 17 r r 18 s s 19 t t 20 u u 21 v v 22 w w 23 x x 24 y y 25 z z 26 (Any) (Any) 27 (Any) (Any) 28 (Any) (Any) 29 (Any) (Any) 30 (Any) (Any) 31 (Any) …»
grondilu m: say (sub (@a) { gather for ^@a -> $n { take $n; take @a[$n, $n] } })( my @ = "a".."z" ) 01:14
camelia rakudo-moar 80f8ef: OUTPUT«0 a a 1 b b 2 c c 3 d d 4 e e 5 f f 6 g g 7 h h 8 i i 9 j j 10 k k 11 l l 12 m m 13 n n 14 o o 15 p p 16 q q 17 r r 18 s s 19 t t 20 u u 21 v v 22 w w 23 x x 24 y y 25 z z␤»
pmichaud I'd need to see the .perl
grondilu m: say (sub (@a) { gather for ^@a -> $n { take $n; take @a[$n, $n] } })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, ("a", "a"), 1, ("b", "b"), 2, ("c", "c"), 3, ("d", "d"), 4, ("e", "e"), 5, ("f", "f"), 6, ("g", "g"), 7, ("h", "h"), 8, ("i", "i"), 9, ("j", "j"), 10, ("k", "k"), 11, ("l", "l"), 12, ("m", "m"), 13, ("n", "n"), 14, ("o", "o"), 15, ("p", "p"), 16, ("q",…»
01:14 AlexDaniel joined
pmichaud and it would *really* be better if this is done without gather/take. Either that or we'll have to really fix gather/take. 01:14
grondilu ah that's the hard part then. Doing lazy lists without gather/take is tricky.
pmichaud I did say you can assume that while/until is able to lazily return lists of elements 01:15
(as designed in S04)
grondilu m: say (sub (@a) { for ^@a -> $n { $n, @a[$n, $n] } })( my @ = "a".."z" ).perl 01:16
camelia rakudo-moar 80f8ef: OUTPUT«Nil␤»
pmichaud actually, let me relax the problem a bit, then. Take laziness out as a requirement.
grondilu m: say (sub (@a) { do for ^@a -> $n { $n, @a[$n, $n] } })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«((0, ("a", "a")), (1, ("b", "b")), (2, ("c", "c")), (3, ("d", "d")), (4, ("e", "e")), (5, ("f", "f")), (6, ("g", "g")), (7, ("h", "h")), (8, ("i", "i")), (9, ("j", "j")), (10, ("k", "k")), (11, ("l", "l")), (12, ("m", "m")), (13, ("n", "n")), (14, ("o", "o…»
pmichaud i.e., it's okay to do it eagerly.
so, I'll reformulate the problem a bit.
grondilu m: say (sub (@a) { ( $++, @a[$++,$++] for ^@a ) })( my @ = "a".."z" ).perl 01:17
camelia rakudo-moar 80f8ef: OUTPUT«((0, ("a", "a")), (1, ("b", "b")), (2, ("c", "c")), (3, ("d", "d")), (4, ("e", "e")), (5, ("f", "f")), (6, ("g", "g")), (7, ("h", "h")), (8, ("i", "i")), (9, ("j", "j")), (10, ("k", "k")), (11, ("l", "l")), (12, ("m", "m")), (13, ("n", "n")), (14, ("o", "o…»
pmichaud Q2: Write a Perl 6 function that takes an array as a parameter @a and produces a list in the form (0, @a[0,0], 1, @a[1,1], 2, @a[2,2], 3, @a[3,3], ... ). Avoid us 01:18
ing gather/take.
grondilu m: say (sub (@a) { ^Inf Z (@a Z @a) })( my @ = "a".."z" ).perl 01:19
01:19 silug left
camelia rakudo-moar 80f8ef: OUTPUT«((0, "a"), (1, "a"), (2, "b"), (3, "b"), (4, "c"), (5, "c"), (6, "d"), (7, "d"), (8, "e"), (9, "e"), (10, "f"), (11, "f"), (12, "g"), (13, "g"), (14, "h"), (15, "h"), (16, "i"), (17, "i"), (18, "j"), (19, "j"), (20, "k"), (21, "k"), (22, "l"), (23, "l"), (…» 01:19
pmichaud You can assume that the list-producing qualities of while and until (as described in S04) are available.
Util m: sub pm (@a) { @a.keys.map({ $_, item(@a[$_, $_]) }) }; say [my @ = pm(["A".."D"])].perl; 01:20
camelia rakudo-moar 80f8ef: OUTPUT«[0; "A", "A"; 1; "B", "B"; 2; "C", "C"; 3; "D", "D"]␤»
Util What is the significance of the semi-colons in the output?
pmichaud it's showing that you have lists-of-lists (i.e., a LoL object). 01:21
grondilu m: say (sub (@a) { ( flat $++, @a[$++,$++] for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e", 5, "f", "f", 6, "g", "g", 7, "h", "h", 8, "i", "i", 9, "j", "j", 10, "k", "k", 11, "l", "l", 12, "m", "m", 13, "n", "n", 14, "o", "o", 15, "p", "p", 16, "q", "q", 17, "r", "r", 18, "s", "s",…»
grondilu m: say (sub (@a) { ( flat $++, @a[$++,$++].item for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, $("a", "a"), 1, $("b", "b"), 2, $("c", "c"), 3, $("d", "d"), 4, $("e", "e"), 5, $("f", "f"), 6, $("g", "g"), 7, $("h", "h"), 8, $("i", "i"), 9, $("j", "j"), 10, $("k", "k"), 11, $("l", "l"), 12, $("m", "m"), 13, $("n", "n"), 14, $("o", "o"), 15, $("p",…»
grondilu what about that? 01:22
m: say (sub (@a) { ( flat $++, (@a[$++,$++]) for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e", 5, "f", "f", 6, "g", "g", 7, "h", "h", 8, "i", "i", 9, "j", "j", 10, "k", "k", 11, "l", "l", 12, "m", "m", 13, "n", "n", 14, "o", "o", 15, "p", "p", 16, "q", "q", 17, "r", "r", 18, "s", "s",…»
pmichaud the inner parcels have been itemized, so no
grondilu m: say (sub (@a) { ( flat $++, (@a[$++,$++],) for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e", 5, "f", "f", 6, "g", "g", 7, "h", "h", 8, "i", "i", 9, "j", "j", 10, "k", "k", 11, "l", "l", 12, "m", "m", 13, "n", "n", 14, "o", "o", 15, "p", "p", 16, "q", "q", 17, "r", "r", 18, "s", "s",…»
grondilu m: say (sub (@a) { ( flat $++, (@a[$++,$++],Nil) for ^@a ) })( my @ = "a".."z" ).perl 01:23
camelia rakudo-moar 80f8ef: OUTPUT«(0, "a", "a", Nil, 1, "b", "b", Nil, 2, "c", "c", Nil, 3, "d", "d", Nil, 4, "e", "e", Nil, 5, "f", "f", Nil, 6, "g", "g", Nil, 7, "h", "h", Nil, 8, "i", "i", Nil, 9, "j", "j", Nil, 10, "k", "k", Nil, 11, "l", "l", Nil, 12, "m", "m", Nil, 13, "n", "n", Nil,…»
grondilu m: say (my @ = ^10)[4,5] 01:24
camelia rakudo-moar 80f8ef: OUTPUT«4 5␤»
grondilu m: say (my @ = ^10)[4,5].perl
camelia rakudo-moar 80f8ef: OUTPUT«(4, 5)␤»
grondilu (just checking what a splice looks like with perl)
m: say ("foo", (my @ = ^10)[4,5]).perl
camelia rakudo-moar 80f8ef: OUTPUT«("foo", (4, 5))␤»
grondilu i c
lizmat ( pmichaud has temporarily left the room ) 01:25
01:25 araujo_ joined
pmichaud I'm in a different room :) 01:25
grondilu m: say (sub (@a) { ( flat $++, @a[$++,$++][] for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e", 5, "f", "f", 6, "g", "g", 7, "h", "h", 8, "i", "i", 9, "j", "j", 10, "k", "k", 11, "l", "l", 12, "m", "m", 13, "n", "n", 14, "o", "o", 15, "p", "p", 16, "q", "q", 17, "r", "r", 18, "s", "s",…»
grondilu m: say (sub (@a) { ( @a[$++,$++] for ^@a ) })( my @ = "a".."z" ).perl 01:26
camelia rakudo-moar 80f8ef: OUTPUT«(("a", "a"), ("b", "b"), ("c", "c"), ("d", "d"), ("e", "e"), ("f", "f"), ("g", "g"), ("h", "h"), ("i", "i"), ("j", "j"), ("k", "k"), ("l", "l"), ("m", "m"), ("n", "n"), ("o", "o"), ("p", "p"), ("q", "q"), ("r", "r"), ("s", "s"), ("t", "t"), ("u", "u"), ("v…»
grondilu m: say (sub (@a) { ^Inf Z ( @a[$++,$++] for ^@a ) })( my @ = "a".."z" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«((0, "a"), (1, "a"), (2, "b"), (3, "b"), (4, "c"), (5, "c"), (6, "d"), (7, "d"), (8, "e"), (9, "e"), (10, "f"), (11, "f"), (12, "g"), (13, "g"), (14, "h"), (15, "h"), (16, "i"), (17, "i"), (18, "j"), (19, "j"), (20, "k"), (21, "k"), (22, "l"), (23, "l"), (…»
grondilu m: say (sub (@a) { ^Inf Z, ( @a[$++,$++] for ^@a ) })( my @ = "a".."e" ).perl 01:27
camelia rakudo-moar 80f8ef: OUTPUT«((0, "a"), (1, "a"), (2, "b"), (3, "b"), (4, "c"), (5, "c"), (6, "d"), (7, "d"), (8, "e"), (9, "e"))␤»
01:27 araujo_ left
grondilu Z flattens the RHS apparently 01:27
01:28 araujo_ joined, cognominal joined
grondilu wait 01:28
m: say List.^methods
camelia rakudo-moar 80f8ef: OUTPUT«new to from fmt flat list lol flattens Capture Parcel Supply eager elems gimme iterator munch pick pop shift plan roll reverse rotate splice sort uniq unique squish rotor REIFY FLATTENABLE_LIST FLATTENABLE_HASH reduce sink STORE_AT_POS combinations permuta…»
01:29 cognominal left
grondilu m: say (my @ = ^10).^methods 01:29
camelia rakudo-moar 80f8ef: OUTPUT«new BIND-POS DELETE-POS flattens name of default dynamic REIFY STORE Method+{<anon>}.new Method+{<anon>}.new perl ACCEPTS elems iterator pick plan sort uniq unique Method+{<anon>}.new infinite fmt list flattens gimme sink STORE_AT_POS Method+{<anon>}.new M…»
pmichaud also, Z uses gather/take internally I suspect, so it's not legal. :)
grondilu m: say (my @ = ^10).kv
camelia rakudo-moar 80f8ef: OUTPUT«0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9␤»
grondilu m: say (sub (@a) { ( @a[$++,$++] for ^@a ).kv })( my @ = "a".."e" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(0, ("a", "a"), 1, ("b", "b"), 2, ("c", "c"), 3, ("d", "d"), 4, ("e", "e"))␤»
grondilu what about that? 01:30
01:30 araujo_ left
grondilu maybe using .kv is cheating? 01:30
pmichaud kv is using gather/take also. 01:31
01:31 araujo_ joined
grondilu well if the list is not lazy, we can push in a variable, right? 01:33
01:33 silug joined
grondilu m: say (sub (@a) { my @b; push @b, $++, @a[$++, $++] for ^@a })( my @ = "a".."e" ).perl 01:33
camelia rakudo-moar 80f8ef: OUTPUT«Nil␤»
grondilu m: say (sub (@a) { my @b; push @b, $++, @a[$++, $++] for ^@a; return @b })( my @ = "a".."e" ).perl 01:34
camelia rakudo-moar 80f8ef: OUTPUT«[0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e"]<>␤»
grondilu m: say (sub (@a) { my @b; push @b, $++, (@a[$++, $++]) for ^@a; return @b })( my @ = "a".."e" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«[0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e"]<>␤»
grondilu oh
m: say (sub (@a) { my @b; for ^@a { push @b, $_; push @b, @a[$_, $_] }; return @b })( my @ = "a".."e" ).perl 01:35
camelia rakudo-moar 80f8ef: OUTPUT«[0, "a", "a", 1, "b", "b", 2, "c", "c", 3, "d", "d", 4, "e", "e"]<>␤»
grondilu push flattens ?!
lizmat push takes a slurpy, which flattens, no? 01:36
grondilu m: say (sub (@a) { my @b; for ^@a { @b[2*$_] = $_; @b[2*$_+1] = (@a[$_, $_]) }; return @b })( my @ = "a".."e" ).perl 01:37
camelia rakudo-moar 80f8ef: OUTPUT«[0; "a", "a"; 1; "b", "b"; 2; "c", "c"; 3; "d", "d"; 4; "e", "e"]<>␤»
grondilu semicolons again :/ 01:38
m: say (sub (@a) { my $b; for ^@a { $b[2*$_] = $_; $b[2*$_+1] = (@a[$_, $_]) }; return $b })( my @ = "a".."e" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«[0; "a", "a"; 1; "b", "b"; 2; "c", "c"; 3; "d", "d"; 4; "e", "e"]␤»
pmichaud I have to get ready for my session -- bbl 01:40
will backlog later
01:42 BenGoldberg joined
grondilu m: say (sub (@a) { ( $++, @a[$++, $++]) xx * }; return @b })( my @ = "a".."e" ).perl 01:42
camelia rakudo-moar 80f8ef: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ceoJACQ167␤Variable '@b' is not declared␤at /tmp/ceoJACQ167:1␤------> 3) { ( $++, @a[$++, $++]) xx * }; return 7⏏5@b })( my @ = "a".."e" ).perl␤»
grondilu m: say (sub (@a) { ( $++, @a[$++, $++]) xx * })( my @ = "a".."e" ).perl
camelia rakudo-moar 80f8ef: OUTPUT«(ListIter.new,)␤»
grondilu m: say $++ xx 5
camelia rakudo-moar 80f8ef: OUTPUT«0 1 2 3 4␤»
grondilu m: say (sub (@a) { ( $++, @a[$++, $++]) xx * })( my @ = "a".."e" )[^10].perl
camelia rakudo-moar 80f8ef: OUTPUT«((0, ("a", "a")), (1, ("b", "b")), (2, ("c", "c")), (3, ("d", "d")), (4, ("e", "e")), (5, (Any, Any)), (6, (Any, Any)), (7, (Any, Any)), (8, (Any, Any)), (9, (Any, Any)))␤»
grondilu m: say (sub (@a) { |( $++, @a[$++, $++]) xx * })( my @ = "a".."e" )[^10].perl 01:43
camelia rakudo-moar 80f8ef: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ilV2BgSHoT␤Arg-flattening | is only valid in an argument list␤at /tmp/ilV2BgSHoT:1␤------> 3say (sub (@a) { |7⏏5( $++, @a[$++, $++]) xx * })( my @ = "a"␤»
grondilu well that's hard
01:50 araujo__ joined 01:53 ssqq joined, araujo_ left 01:55 SevenWolf joined 02:07 llfourn joined 02:08 dayangkun joined 02:09 laouji left, laouji joined, dayangkun left 02:10 dayangkun joined
AlexDaniel method antipair() returns Pair:D # - what does this smiley means? 02:11
02:13 ssqq left 02:15 yqt left 02:23 Petit_Dejeuner left 02:29 noganex joined 02:32 noganex_ left
b2gills It means it returns a defined Pair object 02:34
02:35 rmgk left, rmgk_ joined, rmgk_ is now known as rmgk
AlexDaniel b2gills: thanks 02:41
lizmat m: my $a := $a; say $a # jackpot! 02:53
camelia rakudo-moar 80f8ef: OUTPUT«(signal SEGV)»
lizmat can someone rakudobug this ?
awwaiid Hello! 02:54
lizmat ok, I lost my pair programmer, I'll do it myself :-) 02:55
awwaiid o/
awwaiid I'm having great fun using matplotlib.pyplot Inline::Python; I created a simple wrapper, gist.github.com/awwaiid/ef3f0abcfa96e34977b4 02:56
... and now I don't want to keep adding methods. What witches brew can we stir up today?! 02:57
a nice evil autoload type thing would do
if this was a Class, I could inherit from it and be done; alas this is a collection of defs in a module 02:58
lizmat there's 'method FALLBACK' 02:59
awwaiid googles
03:00 thezip joined
lizmat design.perl6.org/S12.html#FALLBACK_methods 03:00
awwaiid lizmat: OMG that worked 03:01
lizmat++ # Rock!
lizmat :)
03:03 kurahaupo1 joined
awwaiid I updated the gist with the new version 03:03
I'm slurping data out of git which has github PRs; using that to estimate PR lifetime (time delta from first commit to PR merge). I was previously dumping that as a CSV and plotting in libreoffice -- but no longer! 03:08
so awesome
lizmat :-) 03:10
m: my $a := $a; say $a.Str
camelia rakudo-moar 80f8ef: OUTPUT«Cannot call method 'Str' on a null object␤ in block <unit> at /tmp/bu5hgP8_WK:1␤␤»
03:11 bin_005 joined 03:13 tinyblak left, tinyblak joined
awwaiid update gist.github.com/awwaiid/ef3f0abcfa96e34977b4 to have the full example -- simple wrapper lib, script for slurping git, and an example result chart png 03:14
Great way to end the day. Bedtime! 03:15
03:19 silug left
lizmat good night, awwaiid 03:26
03:26 kaare_ joined
thezip lizmat, thanks for your help in pmichaud's presentation tonight 03:30
lizmat yw ! 03:31
03:32 silug joined 03:33 skids joined 03:39 jack_rabbit left 03:41 laouji left 03:46 thezip left 03:50 bin_005 left, BenGoldberg left
dalek kudo/nom: 1372f77 | lizmat++ | src/core/Pair.pm:
Make sure that Pair<doesnotexist> returns Any
 03:51
lizmat rather than Mu spotted at the YAPC::NA Perl 6 intro course 03:52
zostay is there a better way to say: subtype IntOrPositional of Mu where any(Int, Positional) ? 03:54
03:57 vendethiel joined 03:58 AlexDaniel left
lizmat zostay: don't know offhand 04:12
zostay or using valid syntax (:-p): subset IntOrPositional of Mu where Int|Positional; 04:15
was just thinking returns Int|Positional would be handy
04:17 davido_ left 04:18 davido_ joined
lizmat yes, but that will probably not happen before this year's Christmas 04:21
m: sub a returns Bool { fail }; a
camelia rakudo-moar 1372f7: OUTPUT«Earlier failure:␤ ␤ in block <unit> at /tmp/H579fGZYP9:1␤␤Final error:␤ Type check failed for return value; expected 'Bool' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:639␤ in sub a at /tmp/H579fGZYP9:1␤ in block…»
lizmat m: sub a returns Bool|Failure { fail }; a 04:22
camelia rakudo-moar 1372f7: OUTPUT«5===SORRY!5=== Error while compiling /tmp/6FKCwO2RKZ␤Missing block␤at /tmp/6FKCwO2RKZ:1␤------> 3sub a returns Bool7⏏5|Failure { fail }; a␤ expecting any of:␤ new name to be defined␤»
zostay np ^_^ 04:23
04:28 laouji joined 04:39 kurahaupo1 is now known as kurahaupo, jack_rabbit joined 04:43 TimToady left 04:45 TimToady joined 04:46 quester joined 04:50 dagurval joined 04:56 diana_olhovik_ joined 05:34 diana_olhovik_ left 05:38 jack_rabbit left 05:42 vendethiel left 05:44 lizmat left 05:46 Psyche^ joined 05:48 xinming left 05:50 Patterner left 06:00 xinming joined 06:05 llfourn left 06:08 vendethiel joined 06:10 diana_olhovik joined 06:15 kurahaupo left 06:18 quester left 06:21 FROGGS joined 06:25 Sqirrel left, Sqirrel joined 06:26 tinyblak left, tinyblak joined 06:27 yeahnoob joined, RabidGravy joined 06:31 tinyblak left 06:33 rindolf joined 06:34 lizmat joined 06:35 domidumont joined 06:37 vendethiel left, domidumont left 06:38 domidumont joined, bjz joined 06:39 bjz left
lizmat m: say say 2**4294967296 # seems we got a 64bit int power problem 06:39
camelia rakudo-moar 1372f7: OUTPUT«0␤True␤»
06:40 vendethiel joined
lizmat m: say 2**4294967295 # this will probably timeout 06:40
camelia rakudo-moar 1372f7: OUTPUT«(timeout)»
06:42 bbkr_ joined, bbkr left
lizmat m: use nqp; say nqp::pow_I(2,4294967296,Num,Int) # basically wrong at NQP level 06:44
camelia rakudo-moar 1372f7: OUTPUT«0␤»
lizmat m: use nqp; say nqp::pow_I(2,4294967295,Num,Int) # basically wrong at NQP level
camelia rakudo-moar 1372f7: OUTPUT«(timeout)»
06:45 tinyblak joined 06:50 yeahnoob left 06:57 bjz joined, gfldex joined 07:05 _mg_ joined 07:10 bjz left 07:17 cognominal joined 07:25 ely-se joined
ely-se hi 07:25
07:26 FROGGS left
ely-se I had this idea of adding "alias Larry=clang" to my zshrc so I could say "Larry -Wall" 07:29
probably a bad and old joke :(
07:31 ely-se left, llfourn joined 07:33 Ven joined 07:36 ely-se joined 07:41 g4 joined, g4 left, g4 joined 07:46 SevenWolf left 07:53 ely-se left, [Sno] left, [Sno] joined 07:55 Ven left, Ven_ joined
Mouq m: say sub (@a){my $i = 0; do while $i <= @a { say ($++ %% 2) ?? $i++ !! @a[$i - 1, $i - 1] }}(my @ = "a".."e" )[^10].perl # not sure what's wrong here 07:56
camelia rakudo-moar 1372f7: OUTPUT«0␤a a␤1␤b b␤2␤c c␤3␤d d␤4␤e e␤5␤(Bool::False,)␤»
Mouq (while always seems to return (Bool::False,)
(also I know that's not a perfect solution, but it's close 07:57
)
Oh hey 08:04
RabidGravy is there anyway I can tell whether the perl is 32 or 64 bit?
nothing stands out in $*VM
Mouq m: gist.github.com/Mouq/50660c17d7cd617f8537
camelia rakudo-moar 1372f7: OUTPUT«((Bool::True, 0), (Bool::False, 0), (Bool::True, 1), (Bool::False, 1), (Bool::True, 2), (Bool::False, 2), (Bool::True, 3), (Bool::False, 3), (Bool::True, 4), (Bool::False, 4))␤(0, ("a", "a"), 1, ("b", "b"), 2, ("c", "c"), 3, ("d", "d"), 4, ("e", "e"))␤»
Mouq Err
08:04 amurf left
RabidGravy cool 08:05
Mouq m: gist.github.com/Mouq/50660c17d7cd617f8537
camelia rakudo-moar 1372f7: OUTPUT«(0, ("a", "a"), 1, ("b", "b"), 2, ("c", "c"), 3, ("d", "d"), 4, ("e", "e"))␤»
Mouq Yayy
08:05 abraxxa joined
Mouq RabidGravy: Wish I could answer your question, but I have no idea 08:08
[ptc] tadzik: ping 08:11
08:11 spider-mario left 08:12 darutoko joined
pmichaud Mouq++ # that's helpful, thanks 08:13
Mouq Cleaned up the gist a little as well, but no p 08:14
(It's lazy! :D)
pmichaud that helps me formulate my next Q. :) 08:15
Mouq Bring it on ;)
jnthn morning, #perl6
pmichaud jnthn: o/
Mouq o/
pmichaud jnthn: the Perl 6 tutorial was a huge hit tonight at yapc::na
nwc10 we're hoping for a neverending September, but in a good way? :-) 08:16
pmichaud definitely over 30 attendees, most of whom stayed 3.5 hours or more
Ven_ wow :-) 08:17
pmichaud without a break.
jnthn pmichaud: Wow o.O 08:18
pmichaud: Did you fix some of the bugs? ;)
pmichaud jnthn: some, but missed some
08:18 bjz joined
pmichaud and we ran into some other oddities that we'll likely investigate at hackathon on thu 08:18
jnthn OK; Swiss Perl Workshop are interested in having it, so please share your fixes ;)
pmichaud Will do.
jnthn (At a point that's convenient.)
pmichaud I'll send back the src file
jnthn Cool, thanks. 08:19
pmichaud I also have a few suggestions for content reordering, etc.
jnthn OK. I was fairly careful to avoid forward dependencies in the material, but everything is vulnerable to forward dependencies in curiosity... :-)
Would be good to make it less so, though. :) 08:20
pmichaud the use of the whatever-star in the given/when statement was a major one
since it's not introduced prior to that
jnthn Hm 08:21
pmichaud getting from given topicalizer to whatever star seemed to be way too big a leap
most people leapt to "oh, * is an alias for $_" 08:22
anyway, I'll come up with a list of places we got hung up
jnthn Hm, I don't remember that being problematic, though I suspect I did a "lies to children" :) 08:23
But yeah, if people actually want to dig into *why* it works, it's a big leap :)
pmichaud yeah, we had a lot of *why*
and "how"
I quickly did a whatever segue at that point, though, and people were very pleased to learn about whatever. 08:24
jnthn *nod*
pmichaud it might help to cover @a[*-1] earlier in subscripting
jnthn I guess you could re-write it as $_ in the given/when example
pmichaud that at least introduces whatever, without the obvious $_ reference
then when we get to * < 10 they have the notion of "oh, whatever"
er * < 0... whatever. 08:25
jnthn That could also work, yes.
If you'll also be at Swiss Perl Workshop I'm also open to co-hosting it. :-) 08:26
pmichaud I'm still uncertain about swiss pw, but the odds are going up drastically with each passing day. :) 08:27
okay, time for my next GLR-related coding challenge....
08:27 abraxxa left
pmichaud Q3: Each of the standard looping constructs (for, while, until) has the ability to create a list. At present I only know how to get them to return a list with the same number or values as the number of iterations. Is there a way to get a loop that produces more than one value for some or all of its iterations? 08:27
diegok wants 3.5h of Perl 6 tutorial at YAPC::EU
pmichaud one of the things that worked really well here is that the tutorial was scheduled as an evening activity... 08:28
i.e., it was at the same time as YAPC::NA's traditional "game night"
08:28 abraxxa joined
pmichaud so, it didn't clobber any other talks or events, it could be open-ended as to ending time 08:28
i.e., we ran it more like an extended BoF than a tutorial or scheduled talk 08:29
08:29 tinyblak left
jnthn diegok: Should be do-able 08:29
[Tux] RT#124191 is fixed \o/ 08:30
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124191
jnthn pmichaud: "redo"?
diegok jnthn: good!, should I start asking for space?
jnthn diegok: You're one of the organizers, iiuc?
08:30 FROGGS joined
diegok jnthn: yup, I'm helping the Granada folks 08:30
pmichaud jnthn: well, I'm really asking if there's a way to get more than one value out of each block invocation 08:31
and have them concatenate into a list
jnthn diegok: Cool. Anyway, I'm fine with doing it.
pmichaud: Not as far as I'm aware; all kinds of "multiple return vlaue" in Perl 6 really mean "return a Parcel"
pmichaud: And we decided we want to do away with flattening of those in most contexts 08:32
pmichaud: So it'd take an explicit .flat somewhere on the resulting list
pmichaud right. and I'm thinking we need _something_ that allows us to iteratively build a list but where some iterations add more elements to the list than one
jnthn gather/take... :)
pmichaud using an explicit .flat on the resulting list over flattens
diegok jnthn: ok, I'll send it to the orgas list and I'll come back to you!, thank you!. 08:33
jnthn pmichaud: True. and I guess you want a more efficient thing than gather/take
pmichaud: Trouble is we want to avoid introducing checks all over the place
pmichaud We need either a more efficient thing than gather/take, or we need a far faster gather/take. 08:34
Mouq m: my $s = 0; say do for ^10 -> $i { LEAVE { say "redo" and redo if $s++ < $i+2 }; $i } # alas
camelia rakudo-moar 1372f7: OUTPUT«redo␤redo␤0 1 2 3 4 5 6 7 8 9␤»
skids pmichaud: maybe something with catching/resuming an exception that occurs at just the right time?
pmichaud skids: catching/resuming exception is basically gather-take kind of slowness
jnthn skids: The catch/resume of the exception is, at least once we support batching, the main slow in gather/take.
Today it shares that cost with "take a continuation"
But if we get to batch we can avoid that.
But there'll still be the catch/resume cost. We can cheapen that but it's never going to be free. 08:35
pmichaud essentially I feel like we're missing a list primitive
skids Yeah I wasn't suggesting it as a "way to" but the only way I could see something like that happenning.
(with existing constructs)
jnthn pmichaud: I guess "how do we write kv" is an immediate motivator? 08:36
pmichaud jnthn: you'd be right.
[Tux] if I re-enable the test I excluded for RT#124191, I now get a core dump after 11834 tests :(
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124191
[Tux] /pro/3gl/CPAN/rakudobrew/bin/perl6: line 2: 6958 Segmentation fault PATH=/pro/3gl/CPAN/rakudobrew/moar-nom/install/bin:$PATH perl6 "$@" 08:37
08:37 Ven_ left
pmichaud basically, is .map(&block) constrained to producing lists that have the same number of elements as the number of invocations of &block ? 08:38
Mouq pmichaud: No, it can have less :P 08:40
skids Another way to look at it is is there a way to invoke block more than once per element (and then the block could alternate what it actually does)
FROGGS m: say (^10).map({$_ == 3 ?? Empty !! $_}) # not constraint
camelia rakudo-moar 1372f7: OUTPUT«0 1 2 4 5 6 7 8 9␤» 08:41
pmichaud having the block alternate what it does is definitely not the approach I'm looking for
FROGGS but it cannot have more in the way it is implemented right now
08:41 tinyblak joined
pmichaud I'm more interested in the case of more-than-one-element per iteration 08:41
I know about the zero case
skids (or invoke multiple blocks)
FROGGS what if 'return 1, 2' would return two things, and 'return (1, 2)' would return one? 08:42
I mean, would that bite us somewhere?
m: say (^10).map({$_ == 3 ?? 1, 2 !! $_}) 08:43
camelia rakudo-moar 1372f7: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xxFam3ekAp␤Precedence of , is too loose to use inside ?? !!; please parenthesize␤at /tmp/xxFam3ekAp:1␤------> 3say (^10).map({$_ == 3 ?? 17⏏5, 2 !! $_})␤ expecting any of:␤ infix␤ …»
Mouq m: say sub mapcombin(@a, $f, $g) { (^(2*@a)).map({$_ %% 2 , ($_/2).floor}).map(->[$s,$i]{$s ?? @a[$i] ~~ $f !! @a[$i] ~~ $g})}(^10, *+1, */2)
camelia rakudo-moar 1372f7: OUTPUT«1 0 2 0.5 3 1 4 1.5 5 2 6 2.5 7 3 8 3.5 9 4 10 4.5␤»
Mouq m: say sub mapcombin(@a, $f, $g) { (^(2*@a)).map({$_ %% 2 , ($_/2).floor}).map(->[$s,$i]{$s ?? @a[$i] ~~ $f !! @a[$i] ~~ $g})}(^10, *+1, */2).perl
camelia rakudo-moar 1372f7: OUTPUT«(1, 0.0, 2, 0.5, 3, 1.0, 4, 1.5, 5, 2.0, 6, 2.5, 7, 3.0, 8, 3.5, 9, 4.0, 10, 4.5)␤»
FROGGS m: say (^10).map({$_ == 3 ?? leave 1, 2 !! $_}) # would that be the way to produce more output values? 08:44
camelia rakudo-moar 1372f7: OUTPUT«5===SORRY!5=== Error while compiling /tmp/0Um8Fy_r3m␤Undeclared routine:␤ leave used at line 1␤␤»
jnthn FROGGS: Trouble is that you'd then have to say "well, what is the type that is returned", and you'd get into "is a block-final (1, 2) different from 1, 2" and so on...which wil be...tricky
pmichaud FROGGS: we don't have any notion of "return two things" in Perl 6.
FROGGS jnthn: we already have List, Array and Parcel... just make one of them flatten :P 08:45
pmichaud I would not be opposed to having a list-y type that always immediately flattens. 08:46
FROGGS jnthn: though, in '@a = 1, 2' the LHS decides about the flattening, right?
jnthn FROGGS: yes
FROGGS :/
jnthn pmichaud: We were trying to get rid of List/Parcel, but type would be the right slot to hang things off...
pmichaud At the moment I'm still expecting Parcel to go away. &infix:<,> becomes a List constructor 08:47
but if we had a type that immediately flattened when placed in a list, that would be... useful.
FROGGS (1, 2) == List which does not flatten on its own; 1, 2 == List that will 08:48
pmichaud FROGGS: parens are grouping only
FROGGS would be ncie if we had a way oto see the results of these thoughts within minutes :o)
pmichaud if I have a type that immediately flattens, then I could have something like { ... ; Flat.new($a, @b) } 08:49
FROGGS pmichaud: yes, but somehow two literal lists have to produce two different types
:{ } creates object hashes... maybe |( ) creates flattened lists? 08:50
pmichaud then (1..10).map( { Flat.new($a, @b) } would be able to give me the ($a, @b, $a, @b, $a, @b, ...) sort of List I'm looking for
FROGGS though, the syntax bit might not be so important right now
08:50 bin_005 joined
pmichaud FROGGS: the syntax is actually hugely important to me right now 08:50
FROGGS ohh, okay
pmichaud part of my struggle is that p6 has never had a feature to do this sort of thing, and I think we need one.
FROGGS I tend to think about syntax just being sugar for the semantics 08:51
pmichaud every time I start to work on fixing Lists I run into "we don't have a way to do this seemingly common thing"
08:51 laouji left
pmichaud (this problem predates GLR) 08:51
[ptc] .tell tadzik could you review github.com/tadzik/rakudobrew/pull/42 when you get time please? It allows matching Rakudo/panda versions to be installed on Travis
yoleaux [ptc]: I'll pass your message to tadzik.
skids Extend "proceed" to take a value, maybe? 08:52
jnthn |(...) already has a meaning, so we can't really steal that 08:53
s/so/and/
jnthn agrees with Pm's analysis that we could do with this 08:54
08:54 zengargoyle left
FROGGS jnthn: sad, it would be the obvious syntax :lo) 08:55
:o)*
jnthn I don't have a good syntax suggestion short of trying to pick a word like interpolate(1, 2)
pmichaud many of the problems we've had with Z and xx and Z, and ... come down to "we don't have a way to concatenate lists.
08:56 zengargoyle joined 08:57 yeahnoob joined
pmichaud at this point I'd even consider something like a infix:<~,> or infix:<,~> operator :-) 08:58
08:59 espadrine joined
[ptc] hoelzro, ugexe: I've added docs about building Perl 6 on Travis: github.com/paultcochrane/docs-trav...9104b0dce0 08:59
08:59 laouji joined 09:00 skids left
moritz pmichaud: fwiw I've also suggested to have a list concatenation operator 09:00
[ptc] hoelzro, ugexe: could you please review before I submit the pull request? Thanks in advance!
09:00 xinming left
pmichaud alternatively, I'd be happy with the notion that 09:01
09:01 xinming joined
pmichaud (1..5).map( { 1,2 } ) produces (1, 2, 1, 2, 1, 2, 1, 2, 1, 2) 09:01
and if you want the groups, you do
jnthn [Tux]: Where do I find the code to reproduce that?
FROGGS what about (,) ? 09:02
1, 2 (,) 3, 4
like the set ops
pmichaud (1..5).map( { ((1,2),) } producing ((1,2), (1,2), (1,2), (1,2), (1,2))
jnthn pmichaud: What type will 1,2 produce?
pmichaud jnthn: List.
(1,2) produces a List 09:03
FROGGS the flattenable List?
pmichaud ((1,2),) produces a List containing a single List element
jnthn pmichaud: (1..5).map({ my @ = 1, 2 }) # result?
pmichaud jnthn: (1, 2, 1, 2, 1, 2, 1, 2, 1, 2)
FROGGS aye, that's what the @-sigil should do
that's why we have sigils after all, no? 09:04
pmichaud I'm just speculatin' here to try to move towards an answer.
jnthn But I thought a lot of this is what we were getting rid of in the GLR...
FROGGS ohh
09:04 laouji left
pmichaud jnthn: getting rid of what, exactly ? 09:04
jnthn pmichaud: A lot of the places that things magically flatten. 09:05
So there's a small number of them (slurpies, array assignment, [...]) to remember.
pmichaud this flattening isn't "magical", though.
FROGGS jnthn: this would return five Arrays, no? (1..5).map({ my @ = 1, 2 })
so the result would be ([1,2], [1,2], [1,2], [1,2], [1,2])
pmichaud or, it's "flattening" only because we don't have a way of saying "a map iteration that adds multiple elements to the resulting List" 09:06
jnthn FROGGS: No, mor elike ((1, 2), (1, 2), ...)
FROGGS err, [1, 2]<>
yeah
jnthn Or what you said :)
pmichaud I publicly said tonight that I totally hate the [...]<> notation. :)
jnthn pmichaud: I don't think you're alone in that :P 09:07
pmichaud To me it's a huge design smell.
jnthn *nod*
moritz "box with a beak"
FROGGS *nod*
pmichaud I'm wondering if [...] and {...} should produce flattenable array/hash
the fact that they're itemized is a little bit of a holdover from p5 09:08
and I'm not sure that meaning has a lot of purpose in a post-GLR world.
moritz I'd like to be able to create nested arrays with my @a = [1, 2], [3, 4]
FROGGS or with my @a = (1, 2), (3, 4)
pmichaud moritz: would it be.... what FROGGS said
moritz what would ( , ) construct? 09:09
jnthn That...surprises me. I'd thought Array assignment would flatten non-itemized things...
moritz a List? Array? Parcel?
pmichaud I asked TimToady yesterday what flattens, and he said "only slurpies and .flat" 09:10
(or 'flat()')
he didn't mention list assignment as flattening. Perhaps that was an oversight.
jnthn Perhaps.
pmichaud moritz: there's no ( , )
jnthn I'd had his list plus list assignment and [...]
moritz because if (1, 2) doesn't construct an Array, @a = (1, 2), (3, 4); @a[0][0] = 5 will fail
jnthn (the inside of [...])
pmichaud moritz: I only see &infix:<,> syntactically 09:11
moritz: thinking, but you may be correct.
moritz pmichaud: ok, what would my @a = (1, 2), (3, 4); say @a[0].^name output?
pmichaud Still, my @a = [1,2], [3,4]; can work to construct arrays even if [...] doesn't itemize.
moritz if that works, +1 09:12
pmichaud basically @a gets two values, each of which are Arrays
09:12 bbkr joined
pmichaud itemization isn't needed for that 09:12
because the act of putting the array into @a[0] and @a[1] produces itemization
(since @a[0] and @a[1] are scalar containers) 09:13
then @a.perl could just be "[ ... ]" without the ugly <> at the edn.
09:13 laouji joined
pmichaud *end 09:13
jnthn pmichaud: That could well work 09:14
09:14 bbkr_ left
pmichaud since flattening is no longer default-ish, I suspect we no longer need circumfix [ ] and { } to itemize. 09:14
jnthn means flat [1,2,3], [4,5,6] would have to become flat $[1,2,3], $[4,5,6] ?
pmichaud jnthn: yes.
jnthn OK.
pmichaud although 09:15
flat [[1,2,3], [4,5,6]] wouldn't :-)
indeed, one could expensively itemize anything by just putting it in square brackets. :) 09:16
(and then I guess index to get it back out... but you get the point)
I think I'd be _really_ happy if [1,2,3] was exactly the same as Array.new(1,2,3) and my @ = 1,2,3 09:17
and same for the hash constructor
jnthn It's certainly attractive. :-)
09:17 laouji left
pmichaud I can't think of a post-GLR instance where we need to have [1,2,3] be the same as Array.new(1.2.3).item 09:18
nwc10 and also "Quicker, easier, more seductive"?
(ie, are there downsides of this?)
pmichaud nwc10: it makes things simpler and much faster on the implementation side.
nwc10 which is a down side, if the goal is to torture implementors 09:19
are there down sides for users? :-)
pmichaud nwc10: I'm trying to think of some... but I've not commonly used [...] or { ... } in my p6 programs so I can't be sure.
surprisingly, having this behavior also means that 09:20
my @a = [1,2,3] # works
my @a = [1,2,3] # @a has three elements 09:21
my @a = [1,2,3],[4,5,6] # @a has two Array elements
my @a = [1,2,3], # @a has one Array element
similarly, my %h = { foo => 'bar' }; # hash assignment 09:22
moritz my @a = [1,2,3] vs my @a = [1,2,3],[4,5,6] looks like a design smell to me 09:23
(one of them flattening, the other not)
pmichaud it has nothing to do with flattening
moritz whatever it has to do with, I'd to be the one to document that
pmichaud it has to do with the first being an Array, and the second being a List of Array
it's the same principle we have now, where (1) is an Int but (1,) is a Parcel. 09:24
moritz "to write nested data structures, you can write things like my @a = [1,2,3],[4,5,6]; beware that it only works if you at least two pairs of braces"
pmichaud it's the comma that does it
not the pairs of braces.
my @a = [1,2,3],4 # @a has two elements 09:25
moritz still a design smell
pmichaud we have that design smell no matter what
arnsholt Python does more or less exactly the same thing, FWIW
pmichaud my @a = (1,2,3) # how many elements does @a get?
arnsholt "1," is a single-element tuple, "1" is just the number
moritz (not saying that the Parcel case right now is any better, but at least the parcel gets flattened away in many cases)
pmichaud does it? 09:26
not any more. We're getting rid of flattening.
moritz assignment to array variables still flattens it away
m: my @a = (1, ); say @a[0].^name
camelia rakudo-moar 1372f7: OUTPUT«Int␤»
pmichaud okay, TimToady didn't mention that (as I noted, perhaps an oversight)
m: my @a = ((1,),1); say @a[0].WHAT 09:27
camelia rakudo-moar 1372f7: OUTPUT«(Int)␤»
pmichaud m: say ((1,),1).[0].WHAT
camelia rakudo-moar 1372f7: OUTPUT«(Parcel)␤»
pmichaud m: say ((1,2,3),4).elems 09:28
camelia rakudo-moar 1372f7: OUTPUT«2␤»
09:29 laouji joined
pmichaud m: say (my @ = ((1,2,3),4)).elems 09:29
camelia rakudo-moar 1372f7: OUTPUT«4␤»
pmichaud the whole bit of list assignment flattens but nowhere else does feels more smelly to me
and more smelly still if array assignment flattens and list assignment doesn't 09:30
09:30 laouji left
pmichaud m: say [(1,2),3].elems; say ((1,2),3).elems 09:31
camelia rakudo-moar 1372f7: OUTPUT«3␤2␤»
jnthn I'd hope [...] and my @ = ... would have the same flattening semantics. 09:32
(for the ...)
pmichaud exactly
and I'm thinking I want "no flattening" 09:33
but even here, what are the flattening semantics of @a = ... versus (@a, @b) = ... ?
or ($a, $b) = ... ?
09:34 laouji joined
pmichaud I mean, do we ahve the case that @a = ... flattens the rhs while ($a, $b) = ... doesn't? 09:35
if we say that both flatten, then list assignment flattens but constructing a list doesn't ?
09:35 laouji left
moritz is at the point where he wants to get rid of all implict flattening, and have a separate list concatenation operator for when you need flatten 09:36
pmichaud at the moment I really like the notion that [...] doesn't itemize. 09:37
I can't see where it poses a particular problem.
moritz and maybe a flattening meta operator, so that you can say @lista FZ @listb instead of flat(@lista Z @listb)
09:38 laouji joined
Mouq What's the biggest difference between […] and (…) then? Mutability? 09:38
pmichaud one constructs an Array
OTOH!!!
Mouq Right, my question was more like, what's the difference between Array and List 09:39
or differences
pmichaud Array elements are all Scalars
moritz so you can always assign to Array elements
pmichaud i.e., everything put into an array is itemized
moritz m: my @a = List.new(1, 2, 3); @a[0] = 5;
camelia ( no output )
moritz m: my @a := List.new(1, 2, 3); @a[0] = 5;
camelia rakudo-moar 1372f7: OUTPUT«Cannot modify an immutable Int␤ in block <unit> at /tmp/vZFZm2V3tn:1␤␤»
Mouq heh
pmichaud OTOH!!! (again).... if we don't flatten list/array assignment, then my @a = 1..10; doesn't work.
moritz m: my @a := Array.new(1, 2, 3); @a[0] = 5; say @a
camelia rakudo-moar 1372f7: OUTPUT«5 2 3␤»
Mouq k, thanks :) 09:40
09:40 yeahnoob left
pmichaud still, I think that flattening is separate from [...] being itemized. 09:41
moritz then what's the point of itemizing at all?
FROGGS can't we allow this? my @a = |1..10; 09:42
pmichaud sorry, I should say "flattening of list assignment" is separate from "[...] being itemized"
Mouq FROGGS: Precedence
09:42 laouji left
pmichaud FROGGS: we can allow that, but it's got the totally wrong default. 09:42
my @a = 1..10; ought to dwim 09:43
09:43 brrt joined
FROGGS k 09:43
FROGGS agrees
09:44 zakharyas joined
pmichaud moritz: note that when needed, we can still itemize with prefix-$ 09:44
i.e., if we needed a non-flattening array, then $[...] is always at hand.
and that seems much better to me than having [...]<> which means "deitemize the itemized array"
also I'd like to note that "flattening" is a little overloaded here... because traditionally we've used "flattening" to do nesting, when in most cases we just need/want one level of flattening. 09:45
09:45 laouji joined
moritz FWIW I'd be fine with 'my @a = 1..10' not DWIMing, if that's the price we pay for a very simple model where it's trivial to explain how to flatten and how not, and that works 100% of all times 09:46
09:46 laouji left, laouji joined
pmichaud i.e., sometimes we want to be able to "flatten" (1, (2, 3..7)) to become (1, 2, 3..7) and not (1, 2, 3, 4, 5, 6 7) 09:48
09:48 FROGGS left
pmichaud as an example of where we already have the "flattening" smell now: 09:49
m: for (1,2,3) { .say }
camelia rakudo-moar 1372f7: OUTPUT«1␤2␤3␤»
pmichaud m: for (1,2,3), { .say }
camelia rakudo-moar 1372f7: OUTPUT«5===SORRY!5===␤Expression needs parens to avoid gobbling block␤at /tmp/yqAtdgcsfh:1␤------> 3for (1,2,3), { .say }7⏏5<EOL>␤Missing block (apparently taken by expression)␤at /tmp/yqAtdgcsfh:1␤------> 3for (1,2,3), { .say }7⏏5<EOL>…»
pmichaud m: for ((1,2,3),) { .say }
camelia rakudo-moar 1372f7: OUTPUT«1 2 3␤»
pmichaud m: for (1,2,3),(4,5,6) { .say }
camelia rakudo-moar 1372f7: OUTPUT«1 2 3␤4 5 6␤»
pmichaud afaic, once the decision was made to not flatten parcels by default, that completely changes the implications for itemization. 09:51
El_Che is there a test smoke like tooling/infra for perl 6 implementations (not modules). I only found perl6advent.wordpress.com/2013/12/...kudo-star/ and github.com/coke/perl6-roast-data. I am not sure how to set up a useful smoking setup similar like the ones I run for perl5 (or if it's needed/useful) 09:52
pmichaud anyway, it's almost 04h00 here and I had better get some sleep
jnthn :) 09:53
Sleep well, pmichaud++
09:53 amurf joined
pmichaud before I go, I think I should summarize a bit of what I'm toying with / looking for 09:57
I'd like to separate our notion of "flattening" into "shallow flat" and "deep flat"
09:57 amurf left
pmichaud so, with my @odd = 1, 3, 5; my @even = 2, 4, 6; 09:58
a shallow flat of (1, (2, (@odd, @even))) would result in (1, 2, ((1, 3, 5), (2, 4, 6))) 09:59
while a deep flat of (1, (2, (@odd, @even))) would result in (1, 2, 1, 3, 5, 2, 4, 6)
i.e., that most of the places we need automatic flattening really only need to be doing it at one level deep 10:00
10:00 coffee` left
pmichaud (slurpy params perhaps being the exception to that.) 10:00
Mouq my @a = 1...10 would still DWIM, yes? 10:01
pmichaud yes
because a shallow flat or deep flat of a range expands it 10:02
10:03 coffee` joined
pmichaud anyway, that's what I'm toying with and the sort of answer that keeps popping up. 10:03
but the whole notion of itemizing, scalar containers, and even LoL came from a need to suppress the automatic flattening behavior of Parcels. 10:04
Now that Parcels no longer flatten by default, many of the consequences that led to itemization and LoL no longer apply.
more to the point, now that infix:<,> will be creating a List instead of a Parcel, many of the consequences.... 10:05
anyway, sleep time for sure now
bbl
Mouq sleep well pmichaud++ 10:06
10:07 laouji left, brrt left, coffee` left
Mouq m: say (1, 2, ((1, 3, 5), (2, 4, 6))).deepmap({$_}).perl 10:08
camelia rakudo-moar 1372f7: OUTPUT«(1, 2, 1, 3, 5, 2, 4, 6)␤»
10:09 laouji joined
|Tux| jnthn, github.com/Tux/Text-CSV_XS - edit t/55_combi.t and remove the '#' in line 46 10:10
that '#' was put there when 124191 came into focus
github.com/Tux/CSV of course. the other is perl5 10:11
Mouq A thought, if flattening is split: flat could just be shallow flat and deep flat could be the default case of a bikeshedded version of deepmap
10:11 coffee` joined, Ven joined
|Tux| I'll try to create a shortened version 10:12
Mouq m: say (1, 2, ((1, 3, 5), $(2, 4, 6))).deepmap({$_}).perl
camelia rakudo-moar 1372f7: OUTPUT«(1, 2, 1, 3, 5, $(2, 4, 6))␤»
|Tux| 124191 is definitely fixed 10:17
sergot wut, what is deepmap? 10:18
moritz it descends into arrays/hashes
m: say (1, 2, ((1, 3, 5), $(2, 4, 6))).deepmap(* * 3).perl 10:20
camelia rakudo-moar 1372f7: OUTPUT«(3, 6, 3, 9, 15, 9)␤»
jnthn |Tux|: Good; it took some doing. :) 10:21
jnthn is currently working on RT #124318, which is also mysterious
Mouq I'm a little surprised it doesn't retain structure, honestly. I think there's a function that does 10:22
|Tux| tries to make reproducing the segfault a bit smaller
Mouq m: say ((1, 2, ((1, 3, 5), $(2, 4, 6))) »*» 3).perl
camelia rakudo-moar 1372f7: OUTPUT«(3, 6, ((3, 9, 15), $(6, 12, 18)))␤»
10:23 araujo__ left, araujo joined, araujo left, araujo joined 10:27 _mg_ left 10:35 bowtie joined, telex left, bowtie is now known as Guest85966 10:36 telex joined
|Tux| jnthn, gist.github.com/Tux/08b2083d9eaff31d8bc6 10:38
that makes the segfault go away. It must be a hint :)
FWIW it doesn't matter if the match is correct or not 10:41
10:41 dayangkun left 10:43 bin_005 left
dalek p: bf329da | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for new frame return API.
 10:46
10:46 AlexDaniel joined
dalek kudo/nom: 24c39ea | jnthn++ | / (2 files):
Don't run exit handlers on fake frames.

When we are seeing if a multi can pass the bind check, we create a frame to hold lexicals used in the binding checking. This frame is never actually entered and not fully set up, however. If the block in question had exit handlers, we would end up looking at uninitialized memory in trying to run them, which led to all kinds of oddness when a frame with exit handlers (LEAVE, temp, etc.) had a signature with where constraints. Fixes RT #124318.		 10:50
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124318
jnthn Turns out it wasn't an optimization bug at all like the ticket suggested, and Valgrind whined if you got rid of the loop 10:51
*even if
arnsholt jnthn: A QAST/HLL::Grammar question: Is there a mechanism for an infix:sym<foo> that wants to create a QAST::Op.new(:op<not>, QAST::Op.new(:op<foo>, $lhs, $rhs))? 10:54
If the infix makes a negated foo as its AST, the arguments end up as arguments to the not, not the foo =( 10:55
jnthn arnsholt: Not really; we tend to handle such things in Rakudo using a custom EXPR
arnsholt: EXPR action method, I mean
10:56 tinyblak left
arnsholt Yeah, that's what I thought. Cheers! 10:56
dalek ast: aceeb7c | jnthn++ | integration/weird-errors.t:
Untodo test for RT #123686 & RT #124318.
 10:57
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=123686
10:59 llfourn left 11:01 llfourn joined
|Tux| jnthn, no more segfault, but the tests start failing halfway when using ~~ 11:10
11:11 Sqirrel left, zakharyas left
|Tux| is (%state{1001}, "INI - separator is equal to quote- or escape sequence", "Illegal combo 1001"); 11:15
%state{1001} ~~ m{"separator is equal to"} or die %state{1001};
if I comment out the second line, all is well. If I have that ~~ in there, it dies
INI - separator is equal to quote- or escape sequence
in sub combi at t/55_combi.t:48
and the content is correct! so the ~~ fails 11:16
The 258'th ~~ fails 11:17
timotimo o/ 11:23
11:25 smls joined
smls o/ 11:25
In case someone else finds it useful:
addons.mozilla.org/en-US/firefox/a...-synopses/
addons.mozilla.org/en-US/firefox/a...erl-6-doc/
Only does a simple google site search 11:26
but the advantage is that you can assign a shortcut to it in firefox
So if you addign the "p6" search shortcut to the p6 doc addon, you can simply type "p6 substr" [ENTER] in the address bar to search it. 11:27
*assign
sergot thanks moritz++ 11:30
11:30 llfourn left 11:31 mr-foobar left 11:32 mr-foobar joined
|Tux| jnthn, or whoever liker to dig into *weird* failures: 11:32
gist.github.com/Tux/e85a53243c7bab63858d
adding that warn line makes the die line pass 11:33
remove the warn line and it dies the 258'th time that match is done
11:38 FROGGS joined
masak did someone say weird failures? :P 11:38
jnthn masak: golf plz :P 11:41
11:42 amurf joined 11:43 llfourn joined, _mg_ joined 11:47 amurf left 11:50 Ven left 11:56 tinyblak joined 12:01 Ven joined 12:02 Sqirrel joined, Ven left, domidumont left 12:18 Sqirrel left 12:23 domidumont joined
jnthn |Tux|: Got the SEGV reproduced locally 12:30
smls Inside a grammar, can I get at the filename of the file passed to Grammar.parsefile ?
12:30 silug left
jnthn No, we don't keep it around anywhere 12:30
smls so I can include it in an error message
jnthn You could easily method parsefile($*FILENAME) { nextsame() } though 12:31
And then it's available in $*FILENAME
(or whatever name you prefer) 12:32
smls interesting
I guess I could also simply have { die "invalid line '$/'" } inside the grammar, and then have a CATCH {} in the scope where I call MyGrammar.parsefile to add the filename anddie for real
right?
timotimo jnthn: and this is where we're actually happy that keyword arguments get silently accepted by methods! :) 12:33
jnthn smls: Yes, that's also possible
|Tux|: Seems it's something weird this time though...even if I disable dynamic opt it fails 392 tests. 12:34
12:36 Ven joined
smls m: (grammar {token TOP { {fail "ooops"} }}).parse("aaa") // say "Failed" 12:36
camelia rakudo-moar 24c39e: OUTPUT«Failed␤»
smls looks like it even workswith fail :)
although then I can't get at the failure message anymore, right? 12:37
timotimo in that case you can use "orelse"
m: (grammar {token TOP { {fail "ooops"} }}).parse("aaa") // say "Failed: $_" 12:38
camelia rakudo-moar 24c39e: OUTPUT«Use of uninitialized value $_ of type Any in string context in block <unit> at /tmp/7Xa09UnFCM:1␤Failed: ␤»
timotimo m: (grammar {token TOP { {fail "ooops"} }}).parse("aaa") orelse say "Failed: $_"
camelia rakudo-moar 24c39e: OUTPUT«Use of uninitialized value $_ of type Any in string context in block <unit> at /tmp/Uh4G_hHE56:1␤Failed: ␤»
timotimo m: (grammar {token TOP { {fail "ooops"} }}).parse("aaa") orelse say "Failed: $!"
camelia rakudo-moar 24c39e: OUTPUT«Use of Nil in string context in block <unit> at /tmp/x87a9Bbqdd:1␤Failed: ␤»
timotimo except not
jnthn uh, I don't think fail will work out too well deep in the grammar though
smls ok
jnthn |Tux|: When I say the string in question, it comes out as "INI - separator is equal to quote- or escape sequence", which doesn't match the regex m{"sep_char is equal to"} 12:42
12:43 Ven left
jnthn |Tux|: Making the line 'ok (%state{1001} ~~ m{"separator is equal to"}, "Illegal combo 1001");' passes all the tests 12:44
12:45 silug joined
dalek kudo-star-daily: 989da10 | coke++ | log/ (2 files):
today (automated commit)
 12:46
jnthn |Tux|: Ah, but then when I swtich dynopt back on again I see issues. 12:47
RabidGravy does NativeCall have a way of checking whether a shared library exists without creating a binding and trying to use it?
timotimo i'd also like something to exist that sets up native subs "up front" 12:48
jnthn RabidGravy: Not that I'm aware of 12:49
timotimo but there's still more or less a rewrite of the "internals" of NativeCall on the horizon
smls sigspace is driving me crazy: i.imgur.com/SVb2A3q.png 12:50
jnthn Well, I suspect a lot will stay the same, it's mostly that we need to do something to let us JIT native calls in a good way 12:52
arnsholt smls: What does the comment rule look like?
smls token comment { '//' .*? $$ }
token ws { \h* } 12:53
RabidGravy jnthn, would it possible then for it to throw different exceptions for "Cannot locate symbol" and "Cannot locate native library" ?
jnthn smls: I see no "rule" there? 12:54
arnsholt smls: That shouldn't trigger ws at all, though. token doesn't give sigspace
jnthn RabidGravy: Hm, they don't already?
smls right, the comment was a rule
TOP too
change them to tokens currently to regain sanity :P
arnsholt But The rule you want is probably along the lines of "token comment { '//' \N* }"
RabidGravy well in the sense they have a different message, but both are X::AdHoc
jnthn I was gonna say, why does comment even want to be a rule there 12:55
RabidGravy: ah, you wnat different types...
RabidGravy: Reasonable request; should be do-able
FROGGS El_Che: I'd like to see that we submit spectest reports to testers.perl6.org fwiw 12:56
RabidGravy jnthn, yeah so I can do something like nicely 13:00
m: use NativeCall; sub foo() is native('libxndfile') { * }; try { foo(); CATCH { default { say $_.perl; } } }
camelia rakudo-moar 24c39e: OUTPUT«X::AdHoc.new(payload => "Cannot locate native library 'libxndfile.so': libxndfile.so: cannot open shared object file: No such file or directory")␤»
13:02 domidumont left 13:03 domidumont joined 13:11 brrt joined
hoelzro [ptc]: so far looks good, but I'm concerned about the testing script. Shouldn't we use panda-test instead? 13:12
since some things have a Build.pm that does some custom things?
13:15 yqt joined 13:16 g4 left 13:18 skids joined, dayangkun joined 13:19 dayangkun left, xinming left, dayangkun joined
hoelzro is anyone here going to the Swiss Perl Workshop in August? 13:19
jnthn hoelzro: yes 13:20
hoelzro \o/
FROGGS hoelzro: yes
hoelzro \o/
FROGGS :D
13:20 xinming joined
hoelzro with the hackathon happening, I figured as much 13:20
jnthn |Tux|: So far I've managed to golf it to something not involving Text::CSV :) 13:21
13:21 ZoffixWork joined
hoelzro I'm buying my flight right now 13:21
13:22 laouji left
ZoffixWork FYI: modules.perl6.org/ doesn't have anything listed on it. Failing list generation? I don't see any commits made from the time when I know it was definitely behaving fine, so I'm assuming it's not the code that's the problem. 13:22
masak hoelzro: yes 13:23
ZoffixWork Is it possible for it to choke on invalid JSON in META.info? 'cause I spotted an extra comma in perl6-ANTLR4, which was recently added to the ecosystem.
hoelzro more \o/
itz modules.perl6.org/ is fairly minimal JSON :) 13:24
oops
masak ZoffixWork: if it's possible for it to choke, then we should probably put a `try` statement somewhere...
itz modules.perl6.org/proto.json
ZoffixWork checks the code
OK, false alarm. It works now :) 13:25
I guess I came to the site right in the process of it generating the updated list or something :P
13:25 Ven joined
itz maybe it should generate a new json file and mv it atomically when complete 13:26
El_Che FROGGS: I was preparing an sparc x86 vm for perl5 smoke testing and I was curious how perl6 implementation would run on "exotic" setups. However, I haven't find much to setup such environment
FROGGS El_Che: yeah, we have to do that soonish I think 13:27
El_Che If I may step on some toes while giving kudos for the results, I found test smoke for perl5 a real pita to setup 13:28
(writing the config file felt like assembler)
ZoffixWork It's only 158KB. I doubt my initial guess about generating new files is correct. Oh well, works now. 13:29
ZoffixWork makes motions of sweeping stuff under the rug 13:30
El_Che (for the record: once setup the smoke works fine on autopilot) 13:31
llfourn How do you do the use MyDynamicExporter 'some_method'; some_method(); -- patern that is used a lot in perl5? 13:38
where the arguments to import() end up being created via *$dyanimc_name = sub { } 13:39
or *{"${caller}::dynamic_name}"} = sub {} rather 13:40
I guess I have to get caller in same way come to think of it..
13:42 cdc is now known as cdc_le_co-corse 13:44 cdc_le_co-corse is now known as cdc
jnthn you write an EXPORT sub in the module that receives the arguments and returns a hash of symbols to install 13:46
llfourn ahh is that what this is about: &EXPORT sub did not return an EnumMap 13:47
awesome thanks
13:47 ab6tract joined
ab6tract o/ #perl6 13:48
I've been backlogging and felt compelled to say that anything like having a trailing comma change the semantics of list creation would be a horrible thing to do to anyone 13:49
brrt y
why
a trailing comma may easily change the semantics of an english sentence, 13:50
ab6tract because it is the classic 'change a tiny detail in a Perl script and have everything change'
scenario
brrt well, we operate in a field where tiny details change everything
did not know the argument for changing the semantics, just arguing that it can be reasonable 13:51
13:51 tinyblak left, tinyblak joined
ab6tract brrt: trailing commas are a classic perl-ism 13:54
to have @a = [1,2,3] be different than @a = [1,2,3],
... what can I say, it's the kind of choice that would drive many away from considering the language
including me, depending on how the rest of the GLR settles
13:57 _mg_ left
smls ooh, interesting backlog 13:57
"no longer need circumfix [ ] and { } to itemize" does indeed sound interesting 13:58
but I don;t see how it would have those consequences that pmichaud lists (and ab6tract complains about)
"my @a = [1,2,3],[4,5,6] # @a has two Array elements"
13:59 JimmyZ_ joined
smls ^^since list assignment has flattening *@slurpy semantics for the RHS, that would actually be 6 elements, no? 13:59
also, "my @a = [1,2,3]" and "my @a = [1,2,3]," would still be the same, as both would get flattened by the slurpy
ab6tract smls: IIUC, pmichaud++ was basing that off of TimToady omitting/forgetting 'list assignment' in his list of flatteners 14:00
RabidGravy m: my $f = Sub.new(); say $f
camelia rakudo-moar 24c39e: OUTPUT«(signal SEGV)»
RabidGravy m: my $f = Sub.new(); # is fine
camelia ( no output )
smls I hope it's just forgetting :P
ab6tract me too 14:01
14:02 Ven left
ab6tract brrt: wrt tiny changes big impact, that is not a very good argument against making sure that those tiny changes are explicit, loud, unignorable 14:04
With great expressitivity comes greater responsibility
itz y $f 14:05
oops
github.com/perl6/modules.perl6.org/pull/9
ab6tract And ignoring our own languages' reputation for obscure corner cases will surely doom us 14:06
itz ^^ that should fix the modules.perl6.org issue just seen 14:07
dalek href="https://modules.perl6.org:">modules.perl6.org: ab616e1 | (Steve Mynott)++ | web/lib/P6Project.pm:
make write_file atomic so partial or empty files are never in place
 14:08
href="https://modules.perl6.org:">modules.perl6.org: 542e334 | ab5tract++ | web/lib/P6Project.pm:
Merge pull request #9 from stmuk/stmuk

make write_file atomic so partial or empty files are never in place as probably happened irclog.perlgeek.de/perl6/2015-06-10#i_10729972
RabidGravy just rt'd that segv as #125376 14:11
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=125376
smls is also confused about pmichaud statement that "flat [[1,2,3], [4,5,6]] wouldn't" have to be rewritten if [] were non-itemizing. Will flat no longer be recursive? 14:14
[ptc] hoelzro: haven't heard of panda-test before, will have to look into it 14:15
hoelzro: I'm thinking that things are going to change a bit after people start using Travis for Perl 6 projects 14:16
hoelzro: I had `make test` in the build script beforehand and was thinking while writing the docs: hmm, don't think everyone's likely to do that :-) 14:17
hoelzro: was also wondering about using `ufo` to generate a Makefile like EUMM does, but haven't yet been able to get ufo to work on my system... 14:18
RabidGravy is there another way of creating a named sub with a name chosen at run-time that doing an EVAL? 14:19
hoelzro [ptc]: yeah, I think that using the tools that panda ships is the way to go (for now) 14:20
14:21 dayangkun left 14:23 Ven joined
llfourn RabidGravy: You can make a sub like my $foo = sub {...}; $foo.set_name('foo'); 14:25
that won't declare the symbol in the current lexical scope though 14:26
methods you can do with $obj.add_method 14:27
I'm still in the process of figuring these things out
m: my &Foo = sub { say "win" }; Foo(); 14:29
camelia rakudo-moar 24c39e: OUTPUT«win␤»
llfourn oh that works ^^ :D
14:30 Ven left 14:32 andreoss joined
moritz m: my $name = 'thing'; my $sub = anon sub ::($name) () { say 'win' }; $sub(); say $sub.name 14:35
camelia rakudo-moar 24c39e: OUTPUT«===SORRY!===␤Name ::($name) is not compile-time known, and can not serve as a sub declaration␤»
psch m: sub f($name, &code) { &OUTER::('&' ~ $name) = &code }; f("foo", sub { say "bar" }); foo # similarly forbidden
camelia rakudo-moar 24c39e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/evaTssJ8Wc␤Undeclared routine:␤ foo used at line 1␤␤»
moritz too bad :/
psch ISTR jnthn was against something like that because of opts..?
moritz psch: also if you use &OUTER::, it already has one &, so you don't need '&' ~ $name
psch oh, right 14:36
well, doesn't make it work :)
moritz yes, you can't add to lexpads at run time
which is great, because it means we can use numbers instead of names for lexpad entries in compiled code
RabidGravy I'm going with the EVAL for the time being 14:37
moritz also it allows us to catch undeclared names at compile time (which is a much bigger reason)
14:39 Ven joined
lizmat m: say $*KERNEL.bits # irclog.perlgeek.de/perl6/2015-06-10#i_10728514 RabidGravy 14:40
camelia rakudo-moar 24c39e: OUTPUT«64␤»
RabidGravy lizmat++ # perfect 14:41
I had actually seen that but instantly forgotten clearly 14:42
psch i remember the opt reason was actually about binding 14:43
e.g. «sub f { ... }; &(
[Coke] hoelzro: (SPW) possibly.
psch e.g. «sub f { ... }; &('f') := sub { #`[ do things ] }
»
it might be workable for stubs, but sets bad expectations
and for actual subs it breaks at least inlining, i suppose 14:44
hoelzro [Coke]: cool
psch (plus the missing ::, too)
14:45 Ven left 14:47 msouth joined, cognominal left 14:48 Ven joined
psch "java.lang.RuntimeException: Unknown container config rakudo_scalar" 14:48
humm
that points me at lack of understanding of CUs 14:49
pmichaud (trailing comma) we already have the case that a trailing comma changes the meaning of things 14:50
(1,) versus (1) has been a part of Perl 6 for a long time.
14:52 MilkmanDan left 14:53 Ven left
msouth Is it a known thing that this: my $a=11; my $b:=$a; $b=12; say $b.WHAT gives (Mu) in the REPL in version 2015.03 ? 14:53
jnthn I know I fixed something related to binding and the REPL and it was probably more recent than that. 14:54
msouth ok
psch #122914 was close on april 20th, jnthn++ 14:55
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=122914
msouth I have some pastebin'd reproductions: paste.scsys.co.uk/487896 paste.scsys.co.uk/487897 paste.scsys.co.uk/487898
if that's useful for a regression test or whatever
pmichaud unfortunately, 2015-03 is the latest .msi we have available and several people were using that at last night's tutorial
jnthn pmichaud: Ah
msouth yeah I got mine from homebrew on OS X
14:56 MilkmanDan joined
pmichaud so that particular part of the tutorial failed for quite a few people 14:56
14:57 tinyblak left, tinyblak joined
ab6tract pmichaud: still not very convincing. that's bothered me before, and I just assumed it would be cleaned up in GLR 14:57
I guess I've just been piling "don't like this about lists, must be going away" into a specific bucket
pmichaud ab6tract: sadly it often doesn't work that way. 14:58
that said, trailing commas are currently quite rare, and I don't think my trailing comma example from last night would end up being a common thing either.
we've seen over and over that most people's natural expectation of 14:59
my @a = [1,2,3];
is that @a ends up with three elements. I mean, it looks like an array assignment
tadzik computer, messages
yoleaux 08:51Z <[ptc]> tadzik: could you review github.com/tadzik/rakudobrew/pull/42 when you get time please? It allows matching Rakudo/panda versions to be installed on Travis
ab6tract true
however, it seems like "[ ]" is always an item is a much easier rule to remember 15:01
than "[ ]" is only an item in list context
that sounds like the howling of zombie camels 15:02
pmichaud the whole notion of "[ ] is always an item" came out of the need for a way to suppress flattening in the majority of contexts
because flattening was so much a default part of perl 6 behavior
tadzik [ptc]: done, thanks a lot!
pmichaud that's no longer true. flattening is now the exception, not the rule.
15:02 lizmat left 15:03 brrt left
pmichaud Prior to 2010, we didn't have a notion of "item context" 15:03
|Tux| jnthn++; # I was not paying attention due to $work activities
sorry to keep throwing these more complicated bugs at you
pmichaud at least not one that manifested itself the way it does now
jnthn |Tux|: Well, I'd rather have them now than later...
pmichaud perhaps an example is good here:
for [1,2,3], (4,5,6) { .say } # how many iterations? 15:04
jnthn |Tux|: I've got it down to a script that doesn't involve any modules now, and to a specific optimization; unfortunately it's a very widely applied one that's probably going on wrong info, so it's still going to take some hunting.
|Tux| good luck hunting 15:05
I found it *very* strange that adding the same match made all tests pass 15:06
15:06 lizmat joined
jnthn |Tux|: Yeah; in the end I got it down to gist.github.com/jnthn/cff4c17fcc6d7bd2ece3 but even trying to remove one of the remaining conditionals makes the bug go away. 15:07
15:07 Ven joined
|Tux| O, wow, that is golfed down indeed! 15:08
15:08 Ven left
grondilu m: for False, True -> $b { .say } 15:09
camelia rakudo-moar 24c39e: OUTPUT«(Any)␤(Any)␤»
grondilu m: for False, True -> $b { say $b }
camelia rakudo-moar 24c39e: OUTPUT«False␤True␤»
grondilu was wondering if the parens were necessary in gist.github.com/jnthn/cff4c17fcc6d7bd2ece3
[ptc] tadzik: sweet! Thank you! 15:12
15:19 msouth left
ab6tract pmichaud: well, if you keep [] stable as item-always 15:19
then the answer is quite clear
pmichaud is it?
15:19 tinyblak left
ab6tract yes, it is a single item 15:19
pmichaud for [1,2,3], (4,5,6) { .say } # how many iterations? 15:20
ab6tract it's only when you start switching behaviors based on context that things get terrible
15:20 tinyblak joined
TimToady well, for doesn't flatten, so that would stay 2 iterations in any case 15:20
pmichaud right... so what's the difference between the [] and the () here ?
15:20 amurf joined
TimToady as I understand the proposal without actually backlogging, [] just makes an Array, instead of a List 15:20
ab6tract apparently nothing, except some future mugs hitting future walls
TimToady much like :{} makes Hash[Any,Any] rather than Hash[Any] 15:21
pmichaud TimToady: the proposal is that [] produces an Array, not an itemized array
TimToady yes, I got that
I'm saying that would be all the difference there is
O(Array/List)
s/O//
it's still meaningful, but only in creating assignable elems 15:22
pmichaud somehow I'm not really following this answer. :-/ 15:23
TimToady my $a1 := [1,2,3]; $a1[2] = 'works'; my $a2 := (1,2,3); $a2[2] = 'fails' 15:24
jnthn Did TimToady just mean that [...] creates something with assignable elems (Array) and ...,... creates a List which doesn't?
pmichaud jnthn: okay, that helps clarify part of it :)
15:25 amurf left
TimToady but your proposal as I understand it would be to strip the Scalar from around Array 15:25
when you make a [] one
pmichaud "not strip the scalar", but rather "don't add one"
er
not "strip the scalar", rather "don't add one"
TimToady well, sure, I meant "strip" metaphorically 15:26
pmichaud just making sure :)
15:26 diana_olhovik left
pmichaud yes, I'm proposing that [] simply be an Array constructor, not an itemized-Array constructor. 15:26
15:26 msouth joined
pmichaud s/proposing/exploring/ may be better 15:26
and same for {...} as a Hash constructor 15:27
dalek osystem: 7b56c97 | RabidGravy++ | META.list:
Add LibraryCheck to ecosystem
RabidGravy filthy hack but someone had to do it 15:28
TimToady but we'd probably have to make my @a = [],[]...oh, that's where you're proposing a shallow flatten?
pmichaud right
TimToady hafta think about that
pmichaud it's just now awkward to talk about "item context" when the only place that does flattening is array assignment and slurpies
I noticed this very quickly in last night's tutorial
TimToady nodnod 15:29
pmichaud since parcels no longer flatten
hoelzro is open('-', :r) re-opening standard input spec'd behavior, or a Rakudo-specific detail? the only mention I see of it behaving this way in the spec is in command line processing
15:29 spintronic joined 15:30 dwarring left
TimToady I think I wouldn't call it "shallow flattening" 15:30
pmichaud hoelzro: using '-' to represent standard input is pretty standard , so I'm guessing it's designed behavior
I agree, I don't like the name "shallow flattening"
hoelzro pmichaud: for a command tool, that makes sense, but for the open() function?
pmichaud I just needed a placeholder to talk about interpolation that isn't a deep interpolation
TimToady in the proposed "for $array {}" I'd rather think of it as "for takes one argument which is coerced to give an iterator"
hoelzro granted, not a lot of people have files named '-', but those reasons
TimToady so "my @array = $array" could be the same story 15:31
pmichaud TimToady: hmmm.
TimToady it's one argument, which might have commas in it, but if not, the single item is iterted 15:32
*rat
so "for $a,$b" iterates twice, guaranteed 15:33
pmichaud even under that way of thinking about it, it doesn't seem that [] needs to be an itemized array
TimToady but "for $something {}" depends on the elems of $something
pmichaud a bare array works just the same
TimToady I wasn't proposing that [] be itemized 15:34
pmichaud okay
TimToady I'm exploring :)
pmichaud here in Perl 6 land we've gotten so used to thinking about "item context" that now it's odd to go back to perhaps not needing it :)
s/item context/itemized values/ 15:35
TimToady so "for 42 {}" still listifies to a single elem
pmichaud I still tend to think of that as being the same as (42).map({}) 15:36
but it amounts to the same thing... the automatic treating of scalars like lists in response to list-y methods
skids As long as we can say "look for a top-level comma, if it ain't there, behavior is different" the single-item case is at least explainable IMO.
pmichaud I don't like the "behavior is different" part of that phrasing. I'm not seeing a difference in behavior. 15:37
15:37 tinyblak left
TimToady it's consistently "ask the top-level thing for its list" 15:37
15:37 tinyblak joined
TimToady just when there's commas, the top-level thing is already a list 15:37
skids Well, you don't see it because you aren't looking at [],[] as "two []s" and "[]" as "one []", naively. 15:38
pmichaud [],[] is a list of two arrays
[] is a single array
skids Yes you know that, newbies it is "two []s"
TimToady just like @foo,@br
pmichaud if I iterate a list of two arrays, I get two iterations
if i iterate a single array, I get iterations for each element of the array
or, phrased another way 15:39
if I iterate a list of two arrays, I get two iterations. if I iterate an array of $n elements, I get $n iterations
it's not "difference in behavior" at all.
TimToady I do suspect there's more gotchas from teh viewpoint of a P5 programmer though 15:40
pmichaud maybe. I'm not yet convinced of that. 15:41
If we treat prefix-$ as being the "don't interpolate" thingy, p5'ers might be okay with that
15:41 tinyblak_ joined
pmichaud because it looks a lot like an arrayref 15:42
TimToady it does fit on a deep level with the referential simplification that the context determines the deref
pmichaud my @a = $[1,2,3]; # @a has one element
TimToady so 'for' is just another kind of .[] thingy
pmichaud my @a = [1,2,3]; # @a has three elements
or put another way, perhaps at this point we're going to have gotchas from P5 viewpoint no matter what we do 15:43
TimToady well, but if you do that to [], you have to do it to $ either, so going back to using $ for pseudo-itemization is a problem
s/either/too/
seems to me
pmichaud well, mainly because you're thinking of my @a = $s; iterates $s 15:44
TimToady maybe $ could still imply Scalar, I dunno
pmichaud I hadn't looked at it that way.
I was still thinking that my @a = $s; would give @a a single element, even if $s is iterable
15:45 tinyblak left
TimToady but you were asking for 'for $s' to iterate the other day, which is the same thing 15:45
pmichaud not really
for me, 'for $s' is like $s.map
and in that case, the item/scalarness doesn't apply. 15:46
TimToady well, if you dig deep enough, list assignment is also mapping input values into individual assignments
though not perhaps with an explicit .map 15:47
pmichaud yeah, the idea of list assignment being map-like or for-like hadn't crossed my mind 15:48
there is historical basis for list assignment being iteration-y, though... I just hadn't ever made a connection.
not sure a connection should be made :)
now I'm curious about the meaning of something like (@a, @b) = @b, @a 15:51
15:52 daxim joined, tinybla__ joined 15:53 tinyblak_ left
pmichaud and we do need to deal with / consider the flattening behavior of .[ ] as well 15:54
15:54 lizmat left
pmichaud it's analogue-ish to the [ ] array constructor, I suspect 15:54
daxim what's the equivalent of naked readline/implicit open on stdin? encrypted.google.com/search?q=site....org+stdin gives me no results 15:55
hoelzro daxim: get or lines
smls Would it be so bad to have to use $ to construct an AoA? @a = $[1, 3], $[2, 4] 15:56
i.e. make [] non-itemizing, but keep list assignment as it is
and keep flattening recursive
pmichaud smls: My first thought was "yes, that'd be so bad", but my subsequent thoughts are "well... maybe not so bad" :) 15:57
smls s/construct/assign to a @ variable/
pmichaud my brane is hurting a bit, I'll make my way down to the conference venue now 15:58
daxim there's a difference
both perl -E'print scalar readline' < foo and perl -E'print scalar readline' foo work
smls oh, I guess @a = [[1, 3], [2, 4]] would also still work because the outer [] constructs an array (thereby itemizing the inner ones) before assignment 15:59
daxim but perl6 -e'say $*IN.get' foo hangs
only the redirect version works
pmichaud smls: correct 16:00
smls: as opposed to ([1,3], [2,4]) which wouldn't itemize the inner arrays
smls yeah
ab6tract pmichaud: I can commiserate with the hurt in your brain :)
ZoffixWork I checked the code and it just logs the error and proceeds to the next module. There's no chocking :) RE: [09:24:33] <masak> ZoffixWork: if it's possible for it to choke, then we should probably put a `try` statement somewhere... 16:01
ab6tract fwiw, I still haven't seem something that doesn't look like some form of sharp object to the eye
pmichaud $Pm.torment(<GLR>) 16:02
ab6tract but then again, we are talking about contexts.... I guess it's an immutable property of the topic :(
anyway, out for now 16:03
jnthn daxim: It's waiting for input; the foo there won't magically go to STDIN. If I typed something and hit enter it will say it. Maybe you're looking for $*ARGFILES
ab6tract have fun, #perl6!
ZoffixWork \o
16:03 ab6tract left
daxim why is the magic gone? :( 16:04
16:05 FROGGS left
TimToady $*IN.get is <STDIN>, not <> 16:06
that's why the "IN" 16:07
16:07 diana_olhovik joined, yqt left
TimToady daxim: so if you want the magic, just do this: perl6 -e 'say get' foo 16:11
get will behave like <>
or use lines for all of them 16:12
moritz (but hopefully without the security implications of <>)
daxim ah, so get is a function, too, not just a method
16:12 domidumont left
daxim I'm glad the magic is not gone after all, that makes porting much easier 16:12
16:13 uncleyear left, uncleyear joined
TimToady m: say get 16:13
camelia rakudo-moar 24c39e: OUTPUT«Céad slán ag sléibhte maorga Chontae Dhún na nGall␤»
moritz m: say $*ARGS.get 16:14
camelia rakudo-moar 24c39e: OUTPUT«Dynamic variable $*ARGS not found␤ in block <unit> at /tmp/baPSkM5wxe:1␤␤»
moritz uhm, what's that thing called...
daxim jnthn said $*ARGFILES
moritz ah, right
pmichaud as to why "@a = $scalar" might not iterate $scalar whereas "for $scalar" does, part of the difference to me may come down to the former being list prefix precedence, whereas the latter doesn't seem to have that precedence. 16:15
dalek osystem: f39c42c | (Jeffrey Goff)++ | META.list:
Add ANTLR4, Marpa
 16:16
daxim how do I get a complete list of those built-in functions, then? docs.perl6.org/routine-sub.html does not list `get` 16:17
DrForr The pull request didn't seem to do the trick yesterday. And Marpa is very much raw.
16:18 uncleyear left
ZoffixWork DrForr, ANTLR got a JSON error in its meta so it's not indexing. I sent you a pull to fix it 16:18
16:19 uncleyear joined, lizmat joined 16:23 spider-mario joined, tinybla__ left 16:24 yqt joined
DrForr Yeah, I just saw that this morning. Merged, I also added Marpa in the update. 16:24
RabidGravy daxim, right now it's look in the source then add the missing functions to the docs ;-)
16:25 tinyblak joined
PerlJam DrForr: the README for Marpa is a little confusing ;) 16:27
DrForr Sigh, fixing.
JimmyZ_ somebody wrote a parser for core setting to search subs and signature 16:28
it is in a gists, iirc
16:28 silug left
PerlJam no worries. I just tuned in and saw Marpa in Perl6 context and wondered what, exactly, that meant. I can read the source in lieu of the README :) 16:28
DrForr I haven't even managed to get the library compiled, but I'll get it work. 16:29
16:29 tinyblak_ joined 16:30 mj41 joined 16:31 ZoffixWork left 16:32 tinyblak left
skids WRT the "how do we implement .kv-ish things question" I think that whatever happens to flattening, it would also be good to have a procedural "take"-like (but strictly lexical and efficient) construct so you can to "loop { calc; takeish-thing; calc; takeish-thing; }" rather than "loop { calc; store-calc; calc; stored-calc, calc; }" 16:33
16:33 tinyblak joined
skids Being lexical it could do label stuff. 16:33
TimToady what's the matter with just using gather/take? 16:34
dalek kudo/nom: 14458a9 | lizmat++ | src/core/Str.pm:
Add some camels and beer mugs
 16:35
16:35 tinyblak left 16:36 tinyblak_ left, tinyblak joined
skids TimToady: "It will never be efficient" is my understanding. 16:36
Plus, nesting and breaking the nesting.
TimToady that's horse pucky
(the efficiency) 16:37
16:37 mj41 left
TimToady part of the GLR is making gather/take fast 16:37
DrForr README updated, though it's still very much a work in progress.
I haven't even gotten the library to compile locally, and the API will probably change when I get back from YAPC. 16:39
skids TimToady: I'm mostly referring to irclog.perlgeek.de/perl6/2015-06-10#i_10728696
16:40 tinyblak left
TimToady a gather/take in eager context will be able to negotiate an implementation with no context switching 16:41
PerlJam (Travis Chase)++ combining Perl + Testing + Legos 16:44
DrForr I missed a talk with LEGO? :( 16:46
PerlJam yeah, he should have put that in the title IMHO 16:47
(I just stumbled upon his talk looking through the live streams)
DrForr Yes, an improved draw. I'll have to check it out on stream.
16:49 tinyblak joined
RabidGravy If anyone is interested, Test::is doesn't work properly with enums 16:50
timotimo probably because enums stringify to EnumName::ValueName? 16:51
and "is" somewhat explicitly does string comparison
RabidGravy yeah that's basically it 16:52
timotimo though if you want to compare to an enum, why doesn't the thing you're comparing return an enum instance, too?
RabidGravy m: enum Foo(:Bar(2)); say Bar eq 2;
camelia rakudo-moar 14458a: OUTPUT«False␤»
RabidGravy timotimo, because I don't know how to make the corresponding thing from an int 16:54
timotimo m: enum Foo <Bar Baz Quux>; say Foo(2);
camelia rakudo-moar 14458a: OUTPUT«Quux␤»
RabidGravy ah-ha! 16:55
I tried just about every other thing
timotimo wow :)
PerlJam RabidGravy: next time, just ask timotimo ;)
timotimo Type($value) is "coercion syntax" in general
colomon hmmm, port to p6? www.retro11.de/ouxr/211bsd/usr/src/...arp.c.html 16:57
andreoss i've just tried git clone this link 16:59
>You may copy the warp kit in whole or in part as long as you don't try to 17:02
make money off it, or pretend that you wrote it.
it's not free software
timotimo can this game be built easily on a modern linux system? 17:05
ab5tract TimToady: it is great to hear some optimism about gather/take 17:06
colomon andreoss: with enough work I think we can contact the author. ;)
timotimo it would be quite surprising if the game could compile without any digging for older library versions 17:07
ab5tract It's an awesome construct 17:08
timotimo it very much is
i like it a whole lot more than python's "yield"
colomon woah, like half of the game is command line processing. 17:09
timotimo i don't think warp.c is all there is to that game :D
colomon oh, right.
timotimo :D
i made the same mistake as you
colomon indeed, www.retro11.de/ouxr/211bsd/usr/src/games/warp/
itz www.retro11.de/ouxr/211bsd/usr/src/...arp.c.html 17:10
oops
github.com/RetroBSD/retrobsd/tree/...games/warp
colomon itz++ 17:11
ab5tract timotimo: does 'yield' have control flow implications?
in Python, I mean 17:12
pmichaud stevan's talk: Q: "How many of you know what a mop is?" 17:13
timotimo what do you mean?
yield does the typical coroutine thing
pmichaud . o O ( If you understand what a mop is, you don't understand what a mop is ) # :-)
timotimo if we do modernize warp, we should get it included in the typical bsdgames packages 17:14
17:15 Possum left
skids lua's "yield" has a slight advantage over "take" in that it allows bidirectional coroutine communications (but you can hack that together in Perl6 with an extra lazy list). 17:17
andreoss itz: it's not the original code i guess, some kind of mips port 17:18
ab5tract timotimo: nevermind, I was misremembering the behavior of 'yield' in Ruby 17:21
timotimo skids: we have take-rw which allows the "receiving" end to change values previously take'd before returning control to the gathering code
arnsholt I'm getting an error when building NQP head using MoarVM HEAD: "MoarVM op 'neverrepossess' is unknown as a core or extension op"
Anyone seen this before?
ab5tract for some reason I had the impression that it behaved like a return statement, which would only allow one
timotimo arnsholt: could be you've got an older version of MoarVM in your path that Configure is picking up? 17:22
ab5tract timotimo: that's eff-ing brilliant
timotimo i think it's hacky and sort of weird, but that's just me :)
skids timotimo: that does the trick, too. Good point. 17:23
17:23 FROGGS joined
andreoss it even compiles 17:24
ab5tract timotimo: a :with(Code &a) adverb might be cool
timotimo i'd rather people just use threads that block on each other
where would you put that adverb? 17:25
17:25 JimmyZ_ left
ab5tract gather :with($transform) 17:26
timotimo i think you'll have to give me a bit more than that :)
ab5tract Well, I might just be wildly interpreting 'bidirectional coroutine communications'
and am still learning the adverbial syntax 17:27
timotimo i just don't understand when you expect that transform to be called and with what
dalek kudo/nom: c6925b2 | (Steve Mynott)++ | tools/build/create-moar-runner.pl:
correct minor typo and grammar
kudo/nom: 4f36870 | FROGGS++ | tools/build/create-moar-runner.pl:
Merge pull request #437 from stmuk/stmuk

correct minor typo and grammar
17:28 Possum joined
FROGGS rjbs++ and TimToady++ # chatter 17:29
rjbs :) 17:30
smls TimToady: The reason why space is forbidden before postfixes, is to avoid ambiguity with infixes, right?
In that case, wouldn't it be feasible to make an exception for method calls, and solve the ambiguity by instead fobidding people from defining infix operators that look like ".foo" ?
17:30 captain-adequate left
smls I thought I would get used to the .foo(...)\ unspace syntax for multi-line method chains, but it still bothers me a little every time I write it. 17:31
ab5tract timotimo: I was imagining that $transform would be called with the taken values as args at the end of each iteration
lizmat FWIW, smls++
ab5tract within the scope of wherever gathering is happening
timotimo ab5tract: that looks like you just want to use .map on the lazy list and give that to whoever is interested
ab5tract indeed 17:32
dalek rl6-roast-data: 24af6fd | coke++ | / (8 files):
today (automated commit)
pmichaud I really want the "unspace" question listed as a FAQ on faq.perl6.org 17:33
also, oddly there's no link to faq.perl6.org from the perl6.org home page 17:34
ab5tract Not very useful after all :)
smls Ideally, faq.perl6.org would be merged into doc.perl6.org IMO
ab5tract is used to being lazy, less used to the code being as lazy as he is
smls separated into multiple pages grouped by topic, like perldoc's FAQs
ab5tract timotimo: thinking in adverbs has me asking the question though, why we use the *-rw idiom rather than a :rw one 17:35
timotimo too many adverbs are a design smell 17:36
andreoss smls: how about p6doc faq?
17:37 tinyblak left 17:38 abraxxa left
ab5tract timotimo: I get the feeling you speak from deep, perhaps painful, experience 17:38
17:38 espadrine left
ab5tract It sounds reasonable, regardless 17:38
timotimo haha 17:39
no, not really
17:40 SevenWolf joined
dalek q: d6e9dbf | pmichaud++ | answers.md:
Stub in a FAQ for unspace and method calls.
 17:43
pmichaud note that's just the Q without the A.
colomon my recollection is there was a general feeling that we were too gung ho about adverbs in the early days of p6, and have come to the conclusion they’re best used sparingly.
ab5tract pmichaud: Note that there is likely useful A's for some things in doc.perl6.org/language/traps 17:45
pmichaud my recollection differs a bit. For a very long time, we didn't have a good way to parse adverbs, so we tended to avoid them. 17:46
ab5tract which I think has fallen into an unfortunate naming scheme. 'gotchas' might be less forboding
pmichaud I'm totally fine with migrating faq.perl6.org into doc.perl6.org... it's just not my focus at the moment :)
17:47 lizmat left
pmichaud basically, I'm thinking we might want to come up with the standard answers for the frequently bikeshedded topics :) 17:47
ab5tract pmichaud: On the contrary, I was more considering the live scenario you are in right now, offering a resource which fields some known areas of confusion
pmichaud ab5tract: yeah, I'm reading it now
(plus I'm half listening to stevan++'s talk )
actually, perhaps I should create a "bikeshedded answers" section of faq.perl6.org :) 17:48
17:52 telex left
PerlJam presumes that TimToady told Stevan an easier/better way to get at the bits of data tacked onto a package. 17:54
17:54 telex joined
smls pmichaud: Will your quest to make @( ) / .list unnecessary in many places, also extend to making things like .tree less necessary? 17:55
e.g. this does not DWIM currently: @a X (@b Z @c) 17:56
to make it do the "natural" thing, we have to write: @a X (@b Z @c).tree.list
...which is quite ugly.
17:56 lizmat joined
smls if the GLR would make that go away too, I'd be happy :) 17:57
FROGGS btw, I'd also like to see an implicit unspace before the method call dot
PerlJam smls: @a X [@b Z @c] is quite close to what you want :) 17:59
lizmat m: my $a = 42; say $a .Str # only when we would get a TTIAR
camelia rakudo-moar 4f3687: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YnBCRQfLrk␤Two terms in a row␤at /tmp/YnBCRQfLrk:1␤------> 3my $a = 42; say $a7⏏5 .Str # only when we would get a TTIAR␤ expecting any of:␤ infix␤ infix stopper␤ po…»
smls it gets me a very different output
lizmat so basically turn a TTIAR in which the 2nd term starts with "." into something useful 18:00
arnsholt timotimo: Looks like you were sort of right. I had --with-moar but no --prefix, which made it try to use nqp/install/bin/moar even though I told it to use some other moar
18:01 rindolf left
FROGGS lizmat: Slang::Tuxic does it and it seems there are no false positives there 18:01
timotimo arnsholt: ah, right 18:02
lizmat FROGGS: but that's not the only thing that Slang::Tuxic does, is it ? 18:03
18:03 andreoss left
timotimo one problem with .foo with space in front is that ".foo" is already shorthand for "$_.foo" 18:03
FROGGS lizmat: correct, but it is one of the sane things it does :o)
timotimo: it is still nice to line-up long method.method.method calls 18:04
timotimo yes
lizmat timotimo: yes, and it can stay that way
timotimo i mean if we allowed .foo without a \ on the line before that 18:05
lizmat because then it is the first term
m: .say for ^5 # would continue to work
camelia rakudo-moar 4f3687: OUTPUT«0␤1␤2␤3␤4␤»
timotimo i was thinking of the potential version of .foo where we're allowed to put whitespace and newline between $foo and .bar 18:06
RabidGravy had gone name blind when Test.pm was kebabed github.com/rakudo/rakudo/pull/438
arnsholt FROGGS: From my understanding, it's precisely that kind of helpful parsing rules that makes parsing Perl 5 the horror that it is
FROGGS timotimo: right, .say will continue to work
timotimo of course 18:07
raydiak if I did "say sub {}␤.defined" would that be "say(sub {}); $_.defined" or "say(sub {}.defined)"?
timotimo i just mean it can confuse people
FROGGS arnsholt: yes, but you want to have nicely readable code, right?
timotimo the way we currently have it is good, IMO
smls raydiak: the former 18:08
}␤ is a statement terminator
allowing space befor .method would not change that
raydiak got it 18:09
18:09 rindolf joined
RabidGravy boo! is_approx isn't approximate enough 18:11
;-) 18:12
lizmat RabidGravy: please fix! soon! :-) 18:13
18:14 domidumont joined
lizmat that one was specifically *not* kebab-cased to allow for a better version 18:14
named "is-approx"
RabidGravy ah okay
I only noticed because ffmpeg can't make a WAV file of exactly 1 second - for some reason it insists on some multiple of 1024 frames 18:16
ab5tract suddenly imagines libav bindings 18:17
RabidGravy I finally got around to libsoundfile
ab5tract RabidGravy++ :D 18:18
RabidGravy er, libsndfile - libsoundfile is different
ab5tract I think I parsed it properly anyway 18:19
awwaiid Is there a "binding.pry" type thing in rakudo, so I can do like "start a REPL here"? 18:21
(I might have asked this before, but can't remember) 18:22
18:25 yqt left 18:30 bin_005 joined
daxim p6: my $i = -1; #`{ computed last field } say <q w e r t>[$i..2] 18:31
camelia rakudo-moar 4f3687: OUTPUT«Index out of range. Is: -1, should be in 0..Inf␤ in block <unit> at /tmp/tmpfile:1␤␤»
daxim why does accessing indexes from the right end not work anymore? 18:32
FROGGS daxim: you have to indicate that you really want to do it 18:33
masak are you sure that ever worked?
arnsholt It's never been supposed to work
AFAIK
masak right.
FROGGS p6: my $i = -1; #`{ computed last field } say <q w e r t>[*$i..2]
camelia rakudo-moar 4f3687: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> 3 computed last field } say <q w e r t>[*7⏏5$i..2]␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ …»
FROGGS err 18:34
p6: my $i = -1; #`{ computed last field } say <q w e r t>[(* + $i)..2]
camelia rakudo-moar 4f3687: OUTPUT«␤»
FROGGS p6: my $i = -1; #`{ computed last field } say <q w e r t>[0..*]
camelia rakudo-moar 4f3687: OUTPUT«q w e r t␤»
FROGGS p6: my $i = -1; #`{ computed last field } say <q w e r t>[0..(*-1)] 18:35
camelia rakudo-moar 4f3687: OUTPUT«q w e r t␤»
FROGGS p6: my $i = -1; #`{ computed last field } say <q w e r t>[(*-1)]
camelia rakudo-moar 4f3687: OUTPUT«t␤»
FROGGS hmmm
I thought it would mean .elems here
[Coke] gist.github.com/coke/f852b3728e459867f622 # of lines per author in rakudo blame.
m: say 706/27668 # surprised i'm this high. 18:36
camelia rakudo-moar 4f3687: OUTPUT«0.025517␤»
FROGGS I am twice as much to blame than TimToady? O.o
[Coke] I alwso see some people are in there under 2 or more addresses.
FROGGS yeah 18:37
[Coke]: is that for rakudo?
ahh yes
you mentioned it :D 18:38
daxim FROGGS, masak: perl -e'my $i = -1; print for (qw"q w e r t")[$i..2]' # expected result, want to "wrap around" 18:39
[Coke] maybe I can edge past perljam. :) 18:43
tony-o_ jnthn: is there a mechanism in nativecall to call perl6 functions? 18:46
18:47 lizmat left
FROGGS tony-o_: as a callback? 18:48
tony-o_ yea
similar to a callback
FROGGS doc.perl6.org/language/nativecall#F..._arguments
tony-o_ FROGGS: tyvm 18:49
FROGGS tony-o_: yw
raydiak m: print qw"q w e r t".rotate(-1)[^4] # for lack of a concise way to wrap the indices, you could wrap the list around instead...
camelia rakudo-moar 4f3687: OUTPUT«tqwe»
tony-o_ FROGGS: on the C side, do i end up calling 'callback(<whatever>)' to run the perl6 side sub/method ? 18:51
daxim thanks, raydiak 18:52
FROGGS tony-o_: I think so
raydiak daxim: you're welcome :) 18:54
daxim when I dump something with $foo.perl.say, it shows up as $( ………something……… ) 18:55
what's that?
gives me hard jquery flashbacks :/ 18:56
vendethiel daxim: itemization
m: for $(1, 2, 3) { say .perl; }
camelia rakudo-moar 4f3687: OUTPUT«$(1, 2, 3)␤»
vendethiel m: for $(1, 2, 3), 2, 3 { say .perl; }
camelia rakudo-moar 4f3687: OUTPUT«$(1, 2, 3)␤2␤3␤» 18:57
daxim .WHAT says the type is List, what does itemization mean then? 18:58
18:58 amurf joined
timotimo [Coke]: it seems like i'll have to be doing some core setting contributions :3 18:58
19:03 uncleyear left, uncleyear joined, amurf left 19:07 nys joined 19:10 yqt joined
timotimo [Coke]: can you give me a list of what file has the least timo-lines in it? :) 19:14
[Coke] I leave that as an exercise to the reader. 19:16
vendethiel Util++ #nice talk 19:19
hoelzro what's the policy for accessing globaly things (eg. $*ARGFILES) from separate threads? "play with fire and you're going to get burned", or do we try to protect people from themselves? 19:22
19:22 domidumont left 19:25 smls left
DrForr OpenCV actually looks like a fun library to play with. 19:25
hoelzro DrForr: OpenCV is one of the libraries I have on my "write a Perl 6 binding for this" list 19:27
19:27 pullphinger joined, pullphinger left 19:28 pullphinger joined
DrForr The API is almost all C++, it looks like. 19:28
hoelzro yes, but I think that they at least have extern "C" symbols you can link to for C support 19:29
19:29 Peter_R joined
DrForr waltman's talk got me thinking :) 19:30
Util vendethiel: Thanks! 19:31
19:33 avalenn left
DrForr Building opencv takes quite a while. 19:36
19:36 avalenn joined
vendethiel "Michael Michaud - "How (not) to create a language specification for Perl 6‎" uh. 19:37
jnthn ...who? :) 19:38
Util www.yapcna.org/yn2015/talk/6274 19:40
By Patrick Michaud (‎Pm‎) from DFW.pm
How (not) to create a language specification for Perl 6
vendethiel (my titme was from youtube)
jnthn
.oO( How (not) to type "Patrick" )
 19:41
vendethiel s/me/le/
Util Ah
colomon Is pmichaud talking now? 19:42
Have him streaming on my iPad, yes! 19:44
seems like they’re doing a really nice job of it, too. (knock on wood) 19:46
masak if we ever clone pmichaud, I move that we name the clone "Michael Michaud" :D 19:47
PerlJam watches pmichaud's talk too
masak with the nick "mmichaud", of course
flussence heh, xhtml2 would've been a better example there - the browsers went off and invented html5 outside the w3c instead :) 19:48
PerlJam masak: Patrick A Michaud (who spoke at YAPC::NA 2013) should speak at more conferences concurrently with Patrick R. Michaud too 19:49
RabidGravy I while back someone suggested of taking a CArray and putting its contents into a Buf but buggered if I can remember how
19:50 molaf joined
masak PerlJam: I wonder if there's both a Michael A and a Michael R. too? 19:50
RabidGravy a way of taking
PerlJam I think pmichaud's brother is named Michael ... I don't remember. 19:51
19:51 lizmat joined 19:53 diana_olhovik left 19:56 lizmat_ joined, lizmat left 19:58 FROGGS left
ab5tract pmichaud: FWIW, It only took a few hours for me to (mostly) drop my issues with `$a` vs `$a,` 19:58
erm `[]` and `[],`
though it essentially boils down to the same thing
19:59 mohij joined
ab5tract `,` is an infix constructor (or is it a coercion?) to List 19:59
So it's not about "list context" in any way like my day-to-day work with Perl 5 would lead me to understand it 20:00
Put another way, the 'rule' of `,` is sufficiently simple to offset that essential frustration that elicited my earlier reaction 20:02
20:13 darutoko left 20:14 dwarring joined 20:16 colomon left
dalek kudo/nom: 88e1464 | lizmat++ | src/core/Perl.pm:
Make this version Perl 6 A
 20:18
[Coke] Do we have any buyin on that version numbering scheme? last I heard it was undecided. 20:19
masak the "Patrick A Michaud" suddenly makes a lot more sense... :P
[Coke] want to make sure we have buyin before it goes out in a monthly, is all. 20:20
masak agreed.
last I heard it was undecided, but there's also no alternative that will please everybody.
PerlJam masak: except that this one is Patrick R Michaud
jnthn is guessing lizmat's commit came after some discussion at YAPC::NA :) 20:21
lizmat_ I guess the buyin is that pmichaud just mentioned the lettering scheme for the version of the *language*
20:21 lizmat_ is now known as lizmat
jnthn ah :) 20:21
lizmat and I hated the vunknown pretty badly :-) 20:22
20:22 msouth left
nwc10 .u   20:22
yoleaux U+2006 SIX-PER-EM SPACE [Zs] ( )
lizmat m: say $*PERL.version
camelia rakudo-moar 4f3687: OUTPUT«vunknown␤»
nwc10 there, that's got a "SIX" in it. That will do nicely.
everyone can fill in their own preference. No nasty prescribed design :- 20:23
)
masak no alternative will please everybody, but "vunknown" is likely to satisfy no-one :P
PerlJam lizmat: so ... how does one specify that version in code? use v6.A ? 20:24
lizmat in code, one would specify a version of the *compiler* I would think ?
PerlJam lizmat: no, I want to say my code assumes this version of the language.
jnthn m: v6a
camelia rakudo-moar 4f3687: OUTPUT«5===SORRY!5=== Error while compiling /tmp/IHetXE0eAq␤Undeclared routine:␤ v6a used at line 1␤␤»
jnthn m: v6.a 20:25
camelia rakudo-moar 4f3687: OUTPUT«Method 'a' not found for invocant of class 'Version'␤ in block <unit> at /tmp/6ioYvjpcsY:1␤␤»
jnthn aw :)
itz can't we have String::Koremutake strings based on git sha hashes for versions? :)
20:27 ShimmerFairy left, bin_005_j joined 20:28 bin_005 left
dalek kudo/nom: 2b2a1a1 | lizmat++ | src/core/Perl.pm:
Got corrected by TimToady++, it's Advent, not A
 20:28
PerlJam nice!
lizmat and please forgive me if consensus turns out this is wrong :-) 20:29
$ 6 'say $*PERL.version'
v6.Advent
flussence I'm trying to get a silly idea to work: making "is cached" use a typed %cache if possible, instead of calling .gist every time
lizmat flussence: that would call .WHICH everytime
flussence still in the "trying to get it to build at all" phase, though :)
lizmat which *could* be better, I agree
[Coke] cached can't use gist, can it, since that loses info? 20:30
lizmat current is cached implementation is pretty naive in that respect
more like a proof of concept and nice for demonstrations :-)
flussence my current line of thinking is: if there's only one param and it ~~ Cool, it's probably safe to use directly as a hash key 20:31
20:31 sjn joined
masak std: say v6.Advent 20:32
camelia std 28329a7: OUTPUT«ok 00:00 135m␤»
20:33 colomon joined 20:34 mohij left
masak std: say v6.Ekke.Ekke.Ekke.Ptangya.Zoooooooom.Boing.Ni 20:34
camelia std 28329a7: OUTPUT«ok 00:00 135m␤»
20:34 mohij joined 20:37 pecastro left 20:39 espadrine joined, ShimmerFairy joined, rindolf left 20:42 mohij left 20:43 rurban joined, rurban left
flussence ooh, 5.22 has a grapheme thingy 20:43
lizmat sorta, yes, for .combing :-) 20:46
wonder how that would work for the cases that jnthn had in his lightning talk
flussence I'm liking these other things going on :)
20:48 rurban joined
jnthn Time for some rest...and will be offline in the morning traveling to London. o/ 20:48
lizmat good night jnthn and safe travelling! 20:49
PerlJam flussence: yeah ... too bad this perl (5) didn't exist years ago already
spintronic will PDL traits be implemented in Perl6? 20:50
it's from Synopsis 9
PerlJam spintronic: eventually (probably)
spintronic ^_^ 20:51
lizmat hopes danaj will get with something soon :-)
20:53 pullphinger left 21:05 sivoais left 21:06 sivoais joined 21:07 skids left 21:15 Ven joined 21:16 colomon left
[Coke] maybe not by christmas. 21:16
(pdl traits, that is) 21:17
21:21 lizmat left
timotimo shaped arrays are on the roadmap 21:25
PerlJam yeah, but pdls are more holy-grail-ish than just shaped arrays 21:26
21:28 rurban left 21:31 rurban joined 21:36 molaf left 21:37 hoelzro_trying_w joined
timotimo well, to be fair 21:37
when we say "we have routines", we also mean something more "holy-grail-ish" than just routines
21:37 colomon joined
RabidGravy *** Error in `/home/jonathan/.rakudobrew/moar-nom/install/bin/moar': corrupted double-linked list: 0x00000000040ae1a0 *** 21:39
that's not good is it
DrForr I've gotten quite a few of those when playing with readline. 21:40
zostay i've been doing some deep thinking about PSGI for Perl 6 and i've come up with a proposal for a Perl 6 version of the PSGI standard 21:43
if you're interested: github.com/zostay/P6SGI
21:46 lizmat joined
ab5tract zostay++ 21:47
21:49 yqt left 21:51 rurban left
timotimo zostay: i think you meant "where { when ... {" instead of "where { where ... {" in the Application section 21:53
21:54 rurban joined 21:55 SevenWolf left 21:57 kaare_ left
lizmat m: gather { take fail } 21:57
camelia rakudo-moar 2b2a1a: OUTPUT«Unhandled exception: Could not find symbol '&Return'␤ at <unknown>:1 (/home/camelia/rakudo-inst-1/share/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from src/gen/m-CORE.setting:16775 (/home/camelia/rakudo-inst-1/share/perl6/runtime/CORE.set…»
21:59 Ven left 22:01 SevenWolf joined 22:02 yqt joined
dalek kudo/nom: 975bcc3 | lizmat++ | src/core/control.pm:
Make gather { take fail } a proper failure
 22:03
lizmat $ 6 'gather { take fail }'
Unhandled exception: Attempt to return outside of any Routine
22:04 Foxcool left 22:07 Foxcool joined
pmichaud good afternoon, #perl6 22:07
22:07 Foxcool left 22:08 Foxcool joined
lizmat pmichaud o/ 22:08
pmichaud some current GLR thinking, from lunchtime:
[ ... ] becomes an array composer without itemization 22:09
[ ... ] imposes a flattening context on the values inside it
22:10 frew joined
pmichaud gather/take will need to become very efficient 22:10
frew I think --prefix is broken in Configure.pl when you use --gen-moar 22:11
is there a better place to say this?
pmichaud we still don't know of a way to have a loop (e.g. 'map') that returns a list with more elements than there were iterations in the loop
frew: this is good, also [email@hidden.address]
frew ok; I'll leave it at that for now then
pmichaud (map returning more elements -- not being able to do that feels like an arbitrary limitation) 22:12
TimToady quietfanatic++ suggested that Empty was a degenerate case of a more general list inserter 22:13
frew details: gist.github.com/frioux/7922db83735d1a5984f8 22:15
dalek kudo/nom: 9b25d80 | lizmat++ | src/core/Compiler.pm:
Remove deprecations that are 1 year old now
frew I don't think there's an obvious solution other than a better error
or .deb's :)
timotimo frew: do you have write access to /opt as the user that executes that?
frew no
that's the problem
timotimo that explains it 22:16
frew but I don't like to compile as root?
timotimo right
this is not about "--prefix is broken", this is "our --prefix behaves very differently from how you'd expect it to"
frew sure
timotimo which essentially boils down to "our --prefix is broken"
frew lol
I mean, it's arguably the subtle --gen-moar (and I bet --gen-nqp) behaviour 22:17
timotimo right; those also contain the "make install" step
frew I figured that
timotimo we've been told before that it's very not cool to be installing during Configure.pl 22:18
frew ok
timotimo even compiling during Configure.pl isn't very neat
frew right
I mean
I get it
and really, the solution is to have nqp and moar installed by a package I think
timotimo yeah
frew but bootstrapping; I get it.
frew googles ppa rakudo
slavik is it possible and is there any documentation on meta-object programming in perl6? I am wondering if it's possible to create a class at run time. 22:19
timotimo it is very possible
lizmat ClassHOW.new ? 22:20
m: Int.^HOW.say
camelia rakudo-moar 975bcc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/vgde32XoO2␤Cannot use .^ on a non-identifier method call␤at /tmp/vgde32XoO2:1␤------> 3Int.^HOW7⏏5.say␤ expecting any of:␤ method arguments␤»
lizmat m: Int.HOW.say
camelia rakudo-moar 975bcc: OUTPUT«Perl6::Metamodel::ClassHOW.new␤»
22:20 bin_005_j left
timotimo github.com/timo/ADT/blob/master/lib/ADT.pm6#L55 here's one example library that does it 22:20
zostay timotimo: thx... corrected 22:21
slavik timotimo: ty
pmichaud quietfanatic++ # general list inserter
timotimo yw
pmichaud a general list inserter feels.... elegant. 22:22
timotimo perhaps this is what we've always wanted parcel to be? 22:23
or maybe at one point
lizmat but isn't that what [ ... ] is ?
timotimo we'll call it Syringe :P
pmichaud [ ... ] is (to be) an array composer
RabidGravy is there any way of making a CArray of a certain length? other than looping and assigning each element?
timotimo i think if you just assign the 100th element it'll become 100 elements big 22:24
RabidGravy ooh let's try that
lizmat .map: { |[ ... ] } ? # did not backlog yet 22:25
RabidGravy timotimo, yep that appears to work 22:26
pmichaud I kind of like the idea of |[...] and |(...) being used for this, yes. 22:29
dalek ast: 43ad905 | lizmat++ | S02-types/deprecations.t:
Remove tests for $*PERL deprecations
timotimo the Weekly is in quite some peril tonight; the cat is very adorable and cuddly at the moment 22:31
lizmat timotimo: are you sure it's not a squirrel ? :-) 22:32
timotimo hehe
i'm the squirrel, if anything
lizmat ah. Confuse-A-Cat Ltd. :-)
timotimo hm?
lizmat www.youtube.com/watch?v=B2Je1CEPkUM 22:34
pmichaud "Perl 6 is geyserware"
labster++
22:35 amurf joined
frew how do I do this in perl6? s/[^[:ascii:]]//g; 22:36
frew is alrady porting stuff
or where can I find out?
lizmat s:g to start with
frew s:g/foo/bar/ ? 22:37
RabidGravy boo! this Audio::Sndfile is taking 15s to load a 1 second file
pmichaud m: my $_ = 'hello'; s:g/l/L/; .say
camelia rakudo-moar 975bcc: OUTPUT«Potential difficulties:␤ Redeclaration of symbol $_␤ at /tmp/SwuYea3EDX:1␤ ------> 3my $_7⏏5 = 'hello'; s:g/l/L/; .say␤heLLo␤»
pmichaud m: $_ = 'hello'; s:g/l/L/; .say
camelia rakudo-moar 975bcc: OUTPUT«heLLo␤»
lizmat RabidGravy: is it precomped ? 22:38
RabidGravy no, not yet
frew hm 22:39
lizmat I guess the source is 10K lines + ?
RabidGravy I've only just made it not crash
lizmat ah, actually loading a file, not just loading the module 22:40
frew hmm
lizmat forget what I said
22:40 amurf left
frew Method 'subst-mutate' not found for invocant of class 'Any' 22:40
TimToady m: say "föo".subst(/ <-ascii> /, '', :g)
camelia rakudo-moar 975bcc: OUTPUT«Method 'ascii' not found for invocant of class 'Cursor'␤ in block <unit> at /tmp/EB0ld3Rn6W:1␤␤»
RabidGravy no, load a 53200 byte WAV file into a buffer 22:41
TimToady m: say "föo".subst(/ <-[\0..\x7f]> /, '', :g)
camelia rakudo-moar 975bcc: OUTPUT«fo␤»
TimToady m: say "föo".encode('ASCII')
camelia rakudo-moar 975bcc: OUTPUT«Blob[uint8]:0x<66 3f 6f>␤»
22:42 rurban left
TimToady m: say "föo".encode('ASCII').decode 22:42
camelia rakudo-moar 975bcc: OUTPUT«f?o␤»
frew trying to port this: github.com/frioux/dotfiles/blob/ma.../ascii-ify
RabidGravy say $*IN.slurp.subst(/ <-ascii> /, '', :g); 22:47
would do it
22:48 jack_rabbit joined 22:49 rurban joined
timotimo lizmat: monty python is nice :) 22:49
lizmat: i only recently watched Fawlty Towers with friends 22:50
frew RabidGravy: well I'd rather continue to stream the IO
huh, well when I do $_ = $_.subst(/ <-ascii> /, '', :g) I get Bogus postfix? 22:51
RabidGravy say $_.subst(/ <-ascii> /, '', :g) for $*IN.lines;
frew I'm doing more than the subst 22:52
22:52 rurban left
frew but yeah I'll do that 22:52
Method 'ascii' not found for invocant of class 'Cursor'
TimToady m: say "föo".subst(/ <-[\0..\x7f]> /, '', :g) 22:53
camelia rakudo-moar 9b25d8: OUTPUT«fo␤»
frew TimToady: thanks
so is that a missing feature?
TimToady I dunno, not sure I want to encourage people to asciify :) 22:54
frew it's just a little toy script for the most part 22:55
seemed like it should be easy to convert
timotimo hm, doesn't trans also support this kind of thing
m: my $r = r/foo/; say $r 22:56
camelia rakudo-moar 9b25d8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/yHcp5G8L7e␤Missing required term after infix␤at /tmp/yHcp5G8L7e:1␤------> 3my $r = r/foo/7⏏5; say $r␤ expecting any of:␤ prefix␤ term␤»
timotimo m: my $r = rx/foo/; say $r
camelia rakudo-moar 9b25d8: OUTPUT«rx/foo/␤»
timotimo m: my $r = regex { foo }; say $r
camelia rakudo-moar 9b25d8: OUTPUT«{ foo }␤»
timotimo hoelzro: should we also be saving the declarator into the regex source in this case
22:57 skids joined
hoelzro timotimo: ah, good point 22:58
22:59 RabidGravy left
timotimo the weekly is useful for this kind of thing :) 22:59
23:00 BooK left, rurban joined 23:01 BooK joined, amurf joined, quietfanatic joined
quietfanatic For the record, a general list inserter class was not what I was thinking of 23:02
but I guess I'm glad to have accidentally inspired a good idea.
if it is good, we don't know yet. :) 23:03
hoelzro timotimo: could you ticket that up?
dalek kudo/nom: a6d2edc | lizmat++ | src/core/Code.pm:
Make sure we don't create Code objects willy nilly
 23:04
timotimo rakudo ticket or rt? 23:13
lizmat it's preventing #125376 23:14
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=125376
23:24 cognominal joined
hoelzro timotimo: RT, please 23:29
23:34 bjz left
timotimo it's #125383 23:34
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=125383
23:37 bjz joined 23:38 bjz left 23:41 gfldex left
lizmat yapc::na shutting down& 23:48
23:48 lizmat left 23:51 quietfanatic left 23:58 espadrine left

Powered by Cro and the Raku® Programming Language.

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

Thank you!