»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_log/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 July 2018.
00:00 reportable6 left 00:03 reportable6 joined
timotimo nice. 00:04
not quite as effective when + and ~ are different operators though
00:20 Cabanoss- left 00:30 Cabanossi joined 00:34 aborazmeh joined, aborazmeh left, aborazmeh joined
pilne would a rough/quick/dirty way to get solid gtk3 bindings just to use nativecall on the c libs? 00:36
or will that induce too many problems with input lag?
Xliff m: say (! !)
camelia 5===SORRY!5=== Error while compiling <tmp>
Prefix ! requires an argument, but no valid term found
at <tmp>:1
------> 3say (! !7⏏5)
expecting any of:
prefix		 00:37
Xliff m: say !()
camelia True
timotimo Xliff: you want to throw some gtk bindings at pilne?
pilne i mean.... at the moment i'm just a very enthusiastic rookie at doing anything with perl6 :D 00:38
Xliff pilne: There's GTK::Simple, which just gets you to the GUI goodness.
Or my project, which attemps to port the whole thing.
github.com/Xliff/p6-GtkPlus
00:38 wildtrees left
pilne tyvm 00:40
i really do love how just skimming idiomatic perl6 just reads like it works (at least mostly from what i've encountered)
Xliff pilne: I know. I love it! 00:42
pilne that's one thing that i always liked about perl5 being more concered with data structure than type, i know p6 approaches sig/twigles differently, but it it still awesome 00:44
Xliff pilne: If you are looking for something specific from GTK in Perl6 and it's not there, please let me know.
Also, p6-GtkPlus does take a bit to build. That's one of the reasons it's only available via git.
I've updated the installation instructions on the bottom of the project page. 00:45
00:47 aborazmeh left
pilne nice 00:47
00:48 kybr joined
Xliff m: say +!() 00:49
camelia 1
Xliff m: say ! +!()
camelia False 00:50
Xliff m: say ! ! +!()
camelia True
Xliff m: say ! ! !() + ! !()
camelia 1
Xliff m: say ! !() + ! !()
camelia 0
Xliff m: say ! !() ~ ! !() 00:51
camelia FalseFalse
Xliff m: say !() + !() + !() + !()
camelia 4
Xliff m: say !() + !() + !() + !() ~ !() + !()
camelia 42
Xliff :P
say () 00:52
m: say ()
camelia ()
Xliff m: say (!() + !() + !() + !() ~ !() + !()) / 2 00:54
camelia 21
timotimo is / 2 allowed?
Xliff say "42" / 2
m: say "42" / 2
camelia 21
timotimo you can * ! !()
Xliff m: say (!() + !() + !() + !() ~ !() + !()) / (!() + !()) 00:55
camelia 21
Xliff m: say (!() + !() + !() + !() ~ !() + !()) / !() + !()
camelia 43
Xliff m: say !() + !() + !() + !() ~ !() + !() / !() + !()
camelia 43
Xliff m: say !() + !() + !() + !() ~ !() + !() / (!() + !())
camelia 41.5
Xliff m: say (!() + !() + !() + !() ~ !() + !()) / (!() + !()) 00:56
camelia 21
Xliff This is 'urtin' me brayne.
timotimo m: say +(()+()) 01:00
camelia 0
timotimo ah of course
m: say +(!())
camelia 1 01:01
timotimo oh, no need for that
01:03 Xliff left
irced spontaneously combusts. 01:09
pilne if a different low-level language than c, exposes a c abi, would nativecall theoretically work with it? 01:10
timotimo sure 01:11
Elronnd pilne: it even works with high-level languages other than c that expose c abis 01:16
irced says "far out" as he waves his hand in front of his face. 01:17
pilne i think perl6 is the best example of an "anti bondage-and-discipline" language i've ever seen
Elronnd nah that's just how c abi works
at least in this case 01:18
irced leaves a trail of flames.
Elronnd eats the flames
irced must be hallucinating. 01:19
pilne i mean, perl6 encourages writing code in a way that is completely natural to how the programmer is thinking, and provides incredible tools for interfacing with other langauges.
Elronnd I mean, I agree with you in general, but nativecall is nothing special 01:20
in fact, it's kind of mediocre when you get to more complicated c libs
pilne fair enough 01:21
irced Elronnd: in other words, it can become a lot of work to implement a c library with a gazillion functions? 01:22
pilne i haven't dove deep at all... i'm still coming back to things after realizing that since i *am* just a hobbyist doing this for fun, there's absolutely no reason for me to learn a language i'm not having fun with (:
Elronnd irced: gazillion functions are fine, but if you want to pass complex buffer or pointer types then it gets hairy
irced Elronnd: oh i see, like type casting a pointer maybe? 01:23
Elronnd: or not, that can all be done in the library as needed. 01:24
Elronnd well, as an example, one time I had a c library that had a bunch of nested structs and unions
irced scratches his chin.
Elronnd and I gave up trying to implement it in perl6, because you have to make a separate named class for each one whereas it can all be inline in c
pilne that does sound a bit icky 01:25
irced Elronnd: oh, i see. i only perused the native call docs as of yet, and i recall some details that would lead me to agree with you.
Elronnd: so i guess the solution would be to wrap the library with your own library that maybe defines macros for such ... or simply pointers. 01:28
Elronnd: seems plausible 01:29
Elronnd: assuming macro calls are possible. just thinking out loud here.
Elronnd: so like a struct that flattens the nested structs. 01:30
Elronnd: might reduce the "coding footprint". alternatively, you might be able to automate the native call definitions. 01:31
Elronnd yes. All of those are possibilities
automated binding generators are a thing. I don't think they're a thing in perl6 01:32
Macro calls are definitely not possible
in my own project (which is hybrid p6/c), I gave up on actually passing data back and forth, and just use a pointer which is an opaque pointer from the p6 side
irced automated binding generators can be a custom thing, and not fully automated.
like code generation based on inputs such as a self written list and maybe some rules 01:34
Elronnd that is true
irced Elronnd: tell us more about native call. what do you mean opaque pointer? 01:35
Elronnd on C side I have typedef struct { whatever; } Blah; 01:36
irced Ok, starting with struct Blah
Elronnd Blah *make_bla(); and then bunch of functions that do things with a Blah* instance
on Perl6 side, straight up Pointer bla = make_bla();
err
my Pointer $bla
irced ah, so basically you write c in a purist OOP style, no direct attribute access? 01:37
Elronnd well
the c code doesn't work like that
only the perl6 code
irced and you're using pointers to functions in the struct?
Elronnd no
struct is only for data
irced as typical. 01:38
Elronnd yes
although I do do the OOP func pointer thing in another pure-c project
irced cool. so the functions that do things with the Blah* instance are in the p6 code?
aside: yeah the OOP function pointer thing is not all bad. 01:40
aside: just another way to do things.
aside: though i think it's been frowned upon because it obfuscates debuggers (though this i have not confirmed) 01:43
Elronnd it's been frowned upon?
01:43 noisegul_ joined
irced aside: conceivably could affect optimization 01:43
Elronnd: just web chatter i've come across.
Elronnd: anyway, thanks for sharing your nativecall experiences! 01:44
01:46 noisegul left 01:53 khisanth_ left 01:54 harmil left 02:05 Manifest0 left, Manifest0 joined 02:06 khisanth_ joined 02:11 pilne left 02:14 uid9164 joined, irced left
uid9164 What's the difference between Num, Real, Rat? I want to define $a = 6/7, and $a can be val(str $str). 02:16
Which type is the most common choice? In c# I only use "single", so I am a bit not used to this. 02:17
02:20 epony joined 02:27 crella_ joined, Cabanossi left 02:31 uid9164 left 02:33 Cabanossi joined 03:19 aindilis left 03:28 aindilis joined
crella_ How should I define a multiline string in perl6 like <xml>? In my country we can't access google legally, so I can hardly get useful result... 03:36
Elronnd crella_: can you use startpage.com or duckduckgo.com or bing.com? 03:49
crella_: anyway, you can just do it.
my $x = "foo
bar
baz"; say $x
^ works fine 03:50
crella_: or can you use tor, or a vpn or proxy?
crella_ bing.com is always ok to me, but bing usually return results about perl5, not 6.. 03:51
thanks 03:53
Elronnd try startpage, it uses google results
also, see if you can use tor, because it sucks not to be able to use parts of the internet
crella_ OK, I'm not interested in these.Actually I don't live on coding. 03:59
Elronnd plenty of internetage that doesn't have to do with coding though 04:00
04:22 vike left 04:34 jaldhar joined 04:35 jaldhar_ joined, jaldhar_ left 04:36 jaldhar_ joined, jaldhar left 04:39 curan joined 04:41 Itaipu left 04:42 Itaipu joined 04:59 crella_ left, jaldhar_ left 05:00 jaldhar_ joined 05:06 molaf joined 05:09 jaldhar_ left 05:10 jaldhar_ joined 05:11 jaldhar_ left 05:12 jaldhar_ joined 05:16 vike joined 05:39 lizmat_ left 05:46 |oLa| joined
discord6 <Aearnus> hey, the docs for FatRat say the denominator can be 0, but it doesn't work in practice docs.perl6.org/type/FatRat#(Ration...method_new 05:50
05:51 Aearnus joined
Aearnus p6: FatRat.new(10, 0) 05:51
camelia ( no output )
Aearnus huh. on mine it says `Attempt to divide by zero when coercing Rational to Str` 05:52
timotimo that's when you try to "use" the value 05:54
m: FatRat.new(10,0).gist.say
camelia Attempt to divide by zero when coercing Rational to Str
in block <unit> at <tmp> line 1
timotimo m: FatRat.new(10,0).perl.say
camelia FatRat.new(1, 0)
timotimo creating it is fine, but when you try to stringify, it blows up, much like a Failure 05:55
Aearnus ahh
so
timotimo which the REPL does, of course 05:56
Aearnus p6: FatRat(10,0).numerator.say
camelia 2
Aearnus oop
p6: FatRat.new(10,0).numerator.say
timotimo huh
camelia 1
timotimo yeah
Aearnus ah okay, there's the functionality mentioned by the docs
timotimo m: FatRat(10,0).perl.say 05:57
camelia FatRat.new(2, 1)
timotimo well, that's weird?
Aearnus yeah, it, uh, does that
it doesn't matter what params you pass in
i made that mistake earlier
timotimo haha
Aearnus spent a solid 10 minutes debugging it
06:00 reportable6 left
timotimo m: say 10.FatRat(0).perl 06:02
camelia FatRat.new(10, 1)
timotimo oh
m: say (10,0).FatRat.perl
camelia FatRat.new(2, 1)
timotimo m: say (1, 2, 3, 4, 5).FatRat.perl
camelia FatRat.new(5, 1) 06:03
timotimo just the number of elements in the list
Aearnus ahha that's odd behavior but i'll take it
timotimo yeah, when you call Type(1, 2, 3) and Type doesn't implement a CALL-ME, it will turn 1, 2, 3 into a list and call the "Type" method on it
06:04 reportable6 joined, ChanServ sets mode: +v reportable6
Kaiepi wait what? 06:04
ohh wait
timotimo github.com/rakudo/rakudo/blob/mast....nqp#L3517
Kaiepi yeah i forgot about that behaviour
Aearnus seems like something that should be documented 06:05
timotimo coercers is a really hairy topic
Aearnus considering someone coming from python for example would definitely make that mistake
timotimo thinking it'd be a constructor, right? 06:06
Aearnus yeah
Kaiepi reading this part of BOOTSTRAP.nqp answers so many questions i had about how the MOP works 06:09
06:09 jaldhar_ left 06:10 jaldhar_ joined
Kaiepi like how classes get treated like Any 06:10
i didn't know roles were Cool 06:12
Aearnus they're pretty cool indeed
(i didn't know they were Cool either)
timotimo hm, can you even get at the Cool methods in a role? it'd immediately pun, i'd imagine 06:13
Kaiepi m: say role {}.^mro 06:14
06:14 cpan-p6 left
camelia ((<anon|1>)) 06:14
06:14 cpan-p6 joined, cpan-p6 left, cpan-p6 joined
Aearnus unrelated, but here's a cool trick i just found out to do flatmap/concatmap/monadic bind on lists using slip 06:14
p6: (1, 2, 3, 4).map: { |($^a xx $^a) }
camelia ( no output )
Kaiepi m: use nqp; say nqp::hllizefor(role {}.^mro, 'perl6')
camelia ((<anon|1>))
Kaiepi oh right i'm forgetting it's anonymous 06:15
m: say role {}.^methods
camelia ()
06:15 sjm_uk joined
Kaiepi m: say role {}.HOW.^name 06:16
camelia Perl6::Metamodel::ParametricRoleHOW
Kaiepi m: say role {}.^find_method('WHICH').name 06:17
camelia WHICH
Kaiepi so Mu methods can be looked up at least 06:18
m: say role {}.^find_method('is-prime').name
camelia No such method 'name' for invocant of type 'Mu'. Did you mean any of these?
note
take

in block <unit> at <tmp> line 1
timotimo it's quite possible that you're actually getting punned object's methods with .^find_method
probably want .^lookup instead
Kaiepi m: say role {}.^lookup('is-prime').name 06:19
camelia No such method 'name' for invocant of type 'Mu'. Did you mean any of these?
note
take

in block <unit> at <tmp> line 1
Kaiepi yeah, find_method is a Metamodel::RolePunning method 06:20
m: use nqp; say nqp::istype(role {}, Cool)
camelia 1
Kaiepi weird
timotimo roles are slippery 06:21
06:22 jaldhar_ left
Kaiepi good ig, i can't cargo cult their code for monads then 06:22
06:23 uid9164 joined
Kaiepi at least not completely, they're a decent base for working out how to implement parametric types 06:25
Aearnus is there an easy built-in to repeat until something stops changing? 06:31
timotimo m: say 1, 2, { ($_ + 1, $_ + 2, $_).pick } ... * == *
camelia (1 2 2)
Aearnus trying to implement exploding dice 06:32
timotimo m: say 1, 2, { ($_ + 1, $_ + 2, $_).pick } ... * == *
camelia (1 2 4 5 6 8 10 12 12)
Aearnus m: sub infix:<d>(Int $n, Int $max) is assoc<left> { (^$max).pick xx $n }; sub infix:<b>(@ds, Int $max) is assoc<left> { @ds.map: -> $d { $d ~~ ($max - 1) ?? |($d, |(1 d $max)) !! $d } }; say 3 d 3 b 3
timotimo here's one
camelia (0 2 0 1)
Aearnus huh. that's a clever trick, but i'm not sure if it'll work in this situation
(i'm trying to iterate the `@ds.map` statement) 06:33
unless.. 06:34
m: sub infix:<d>(Int $n, Int $max) is assoc<left> { (^$max).pick xx $n }; sub infix:<b>(@ds, Int $max) is assoc<left> { ({ @ds.map: -> $d { $d ~~ ($max - 1) ?? |($d, |(1 d $max)) !! $d } } ... * == *).tail }; say 3 d 2 b 2 06:35
timotimo .o( how do i input the looking eyes emoji )
camelia (1 1 0 0) 06:36
Aearnus hey it works :D 06:40
m: sub infix:<d>(Int $n, Int $max) is assoc<left> { (^$max).pick xx $n }; sub infix:<b>(@ds, Int $max) is assoc<left> { @ds, { .map: -> $d { $d ~~ ($max - 1) ?? |($d - 1, |(1 d $max)) !! $d } } ... * == * }; say 3 d 4 b 4
camelia ((1 1 3) (1 1 2 2) (1 1 2 2)) 06:41
Aearnus can't think of a way to make that `3 d 4 b 4` syntax more elegant. in an actual tabletop rpg it'd be written `3d4!` (but `!` was taken) 06:42
timotimo so the ! basically means you take the 4 twice, that's why you have 4 b 4? 06:44
m: sub postcircumfix:<d !>($a, $b) { say $a ~ "d" ~ $b ~ "!" }; 3d4!
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3($a, $b) { say $a ~ "d" ~ $b ~ "!" }; 3d7⏏054!		 06:45
Aearnus the ! means "reroll and sum your dice if you rolled the max value"
timotimo ok, it doesn't like that
m: sub postcircumfix:<d b>($a, $b) { say $a ~ "d" ~ $b ~ "!" }; 3d4b
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3($a, $b) { say $a ~ "d" ~ $b ~ "!" }; 3d7⏏054b
06:46 lizmat joined
Aearnus hold on, idea. 06:46
timotimo it has to be a different unicode class of symbols than letters i believe
Aearnus m: sub infix:<d>(Int $n, Int $max) is assoc<left> { (^$max).pick xx $n }; sub postcircumfix:<d !>($a, $b) { say $a ~ "d" ~ $b ~ "!" }; say "dice 1: " ~ (3 d 5); (3 d 5 !) 06:47
*crosses fingers*
timotimo i do believe postcircumfix has to cuddle at least the thing in front 06:48
camelia (timeout)5===SORRY!5=== Error while compiling <tmp>
Negation metaoperator not followed by valid infix
at <tmp>:1
------> 3" }; say "dice 1: " ~ (3 d 5); (3 d 5 !7⏏5)
expecting any of:
infix
infix stopper
06:48 [Sno] left
timotimo i wonder what's up with the box camelia's running on, it seems oddly slow 06:48
20 minutes until jonathan's keynote in riga 06:51
i also just watched masak's talk 07:00
07:02 lizmat left
timotimo www.youtube.com/watch?v=tenl6JEum5k - it starts RIGHT NOW 07:05
\o/
07:22 Sgeo__ joined
Aearnus aearnus.github.io/2019/08/07/d-d-r...-in-perl-6 :D 07:23
07:25 Sgeo_ left
SmokeMachine hi there! I have a script that rewrite some files... but it's changing the line ending of this files... is it possible to make it do not do that? 07:27
timotimo yes, you'll want to turn chomping off on the input side and use print instead of put/say on the output side
SmokeMachine timotimo: I'm using .IO.slurp and .IO.spurt... how can I turn off the chomping? is there any .slurp option for that? 07:28
timotimo oh, slurp and spurt don't do any chomping
in that case you'll probably be using lines? 07:29
the lines method on Str doesn't seem to have a "don't chomp" option
so you'll have to .split, i'm afraid 07:30
SmokeMachine no... just .subst the whole file
timotimo oh
so perhaps it's already changing line endings when it's utf8-decoding
SmokeMachine timotimo: yes, maybe... 07:45
timotimo you can configure decoders to not do newline translation
SmokeMachine timotimo: any idea of how to fix that?
timotimo it's been some time since that change landed and i really knew how to work it
this is related to \n vs \r\n vs \r i expect?
SmokeMachine yes... 07:46
timotimo right
well, there is a way
i just forgot what it is
SmokeMachine :) 07:47
07:49 rindolf joined
SmokeMachine timotimo: I changed to .slurp: :bin and decode it... it seems to work... 07:54
timotimo docs.perl6.org/language/newline - this ought to do it
SmokeMachine timotimo: in my case, I don't know what's the line ending before opening the file... 07:56
timotimo ah
SmokeMachine the :bin seems to be working...
timotimo fair enough
07:56 antoniogamiz joined
antoniogamiz o/ 07:56
SmokeMachine timotimo: thank you very much! 07:57
timotimo you're quite welcome
07:59 Sauvin left 08:00 sauvin joined 08:06 chloekek joined, Aearnus left
chloekek wiki.haskell.org/IRC_channel still mentions #perl6, but does this still make sense? I rarely see people talk about functional programming or Pugs here. 08:10
Elronnd I thought perl6 was supposed to be a bad lisp, a bad haskell, a bad ml, a bad python, bad javascript, bad whatever-else? 08:12
antoniogamiz Kaiepi: thanks for your last PR to the documentation, I did not know I could do that with enums 08:16
discord6 <Aearnus> Elronnd: not bad, but the worst /s 08:17
<Aearnus> sorry about that reputation ahaha 08:18
<Aearnus> definitely a bit of type systems/functional programming buzz here though. i'd say it makes sense. there's a huge overlap in the perl6 and haskell communities 08:19
cpan-p6 New module released to CPAN! Supply::Timeout (0.0.2) by 03CONO 08:20
Kaiepi np antoniogamiz 08:21
m: use MONKEY-TYPING; class Foo {}; role Bar { has $.bar = 1 }; augment class Foo does Bar {}; say Foo.new.bar 08:35
camelia P6opaque: no such attribute '$!bar' on type Foo in a Foo when trying to get a value
in method bar at <tmp> line 1
in block <unit> at <tmp> line 1
antoniogamiz can I override the gist method of a enum? 08:36
08:37 [Sno] joined
Kaiepi if you use a regular method, i think you can 08:37
antoniogamiz mmm ty
Kaiepi m: class Foo does Enumeration { method gist(Foo:D: --> Str:D) { 'ayy lmao' } }; say Foo.new.gist 08:38
camelia 5===SORRY!5=== Error while compiling <tmp>
Cannot have a multi candidate for 'gist' when an only method is also in the package 'Foo'
at <tmp>:1
Kaiepi or not
antoniogamiz ups haha
08:38 zakharyas joined
Kaiepi m: class Foo does Enumeration { multi method gist(::?CLASS:D:) { 'ayy lmao' } }; say Foo.new.gist 08:39
camelia ayy lmao
antoniogamiz gooood
:D
08:39 satori__ left 08:43 dolmen joined 08:47 epony left
uid9164 I know little about python, ruby and lua. And I often define a variable and declare its type, though sometimes perl6 reports some type error and I have to delete the declartion to let the script run. 08:50
The last time I gave up writing a perl5 script is because I cannot search a method to change a string variable into a int. The subroutine is a bit long and I has no way but to change the variable type... 08:52
08:53 mahafyi left, Ven`` joined
uid9164 I said something in a wrong way just now. I was going to print a string as an int or inversely, but I couldn't find out a method on all common search engines. 08:55
08:56 mahafyi joined 08:57 Ven`` left 09:02 Aearnus joined
Aearnus uid9164: in perl6 you can use the `.Int` or `.Num` methods 09:02
m: say <10>.Int; say <10>.Int.WHAT 09:03
camelia 10
(Int)
09:06 [Sno] left 09:08 Ven`` joined
chloekek Or the + prefix operator. 09:09
uid9164 The problem I met was not just that simple..
<perl5 code> use strict; use warnings; use Math::Round; my $a = '2.603';print round($a,1); 09:12
09:12 cpan-p6 left
uid9164 <output> 31 (???) I had no idea then. So I chose another language. 09:12
09:12 cpan-p6 joined, cpan-p6 left, cpan-p6 joined 09:13 Black_Ribbon left
Aearnus you might have better luck getting help in #perl if perl 5 specifically is what you need 09:14
09:17 epony joined
Ven`` tadzik: we got `is succinct` to work for the non-side-effecting use cases 🎉 09:18
09:25 pilne joined 09:28 Ven`` left 09:29 lizmat joined
tyil can I use `perl6 --target` with a specific target to just generate the .precomp stuff without actually running the application itself? 09:29
09:29 satori__ joined
tyil I want to load the precomp stuff in a Docker image for a program, so I can hopefully improve startup times of the application in docker 09:30
09:31 Ven`` joined
chloekek Ven``: wat is dat 09:31
Ven`` chloekek: close to scheme/racket syntax-rules, auto-quoted, etc 09:33
`@succinct macro swap(a, b) { my t = a; a = b; b = t; }` i.e.
SmokeMachine Ven``: I'm curious... could you give a perl6 example of use? 09:36
Ven`` SmokeMachine: well, see swap^ right here. then you can do `my a = 1; my b = 2; swap(a, b);`
09:37 pat_js joined
SmokeMachine Ven``: ok, but what about th `is succinct`? 09:37
Ven`` SmokeMachine: if you didn't h ave `is succinct`, here's how you'd write it: `macro swap(a, b) { quasi { my t = {{{a}}}; {{{a}}} = {{{b}}}; {{{b}}} = {{{a}}}; } }` 09:38
SmokeMachine Ven``: oh! 007?! now I think I got it! 09:39
Ven`` yes, 007, hence why the sigils are missing. sorry, the original message was for tadzik with whom we'd talked about it already :P
SmokeMachine Ven``: sorry, I was looking for the `is succinct`... I didn't realise that it was the `@succinct`... 09:40
Ven`` yeah, the syntax is gonna change in 007 (there's an open PR...) 09:41
SmokeMachine Ven``: thanks! very good! :)
Ven`` 😊
SmokeMachine Ven``: how does it work? do it uses the AST generated by the macro body as the macro return? 09:43
Ven`` SmokeMachine
oops, new keyboard, not used to it yet, sorry.
SmokeMachine :) 09:44
09:44 Aearnus left
Ven`` it wraps the succinct macro's body inside of an unquote, and then walks the code to find and replace references to the macro's args 09:44
09:44 lizmat left
Ven`` and unquote them (for now it's pretty dumb about it) 09:44
chloekek Ven``: ah I see. 09:46
09:46 lizmat joined
chloekek I haven't used macros yet. 09:46
But I have a use for them (Template::Classic).
SmokeMachine Ven``: hum... that's what masak was referring when he was talking about different kinds of variables that should or should not be quoted some time ago?
Ven`` SmokeMachine well, I suppose? in succinct macros, only the arguments need to be unquoted 09:47
in normal macros you might have generated ast nodes from outside the quasi which you woukd then need to unquote in the body of the quasi
Geth ¦ problem-solving: AlexDaniel assigned to rba Issue pl6anet.org will be discontinued breaking "Recent Blog Posts" on perl6.org github.com/perl6/problem-solving/issues/86 09:49
¦ problem-solving: AlexDaniel assigned to maettu Issue pl6anet.org will be discontinued breaking "Recent Blog Posts" on perl6.org github.com/perl6/problem-solving/issues/86
SmokeMachine Ven``: won't succinct macros be more common then non succinct macros? 09:50
Ven`` SmokeMachine: maybe? it's hard to tell right now. if you want to manipulate the AST you've been passed, then you're gonna need non-succinct 09:52
so simple macros are probably gonna be succinct yes, the rest probably nto
scheme/racket's syntax-rules are very powerful however, but we don't (yet?) have multi-macros, so..:)
(when you define a syntax-rule in scheme/racket you can pattern match on several different patterns and generate different results based on that) 09:57
09:57 lizmat left 09:58 AlexDaniel joined, pilne left 10:00 pilne joined 10:04 [Sno] joined 10:11 mowcat joined 10:12 [Sno] left 10:15 leont joined 10:19 Ven`` left 10:23 Ven`` joined 10:25 pilne left 10:26 uid9164 left 10:36 ravenousmoose left 10:38 tilpner joined, tilpner_ left 10:40 uid9164 joined 10:41 dolmen left 10:49 ravenousmoose joined 10:53 lizmat joined
uid9164 Hi, I think of a problem about math and programming. If anyone is interested in it ? My native language isn't English so I may express not quite exactly. 10:56
a is a decimal with limited digits. M is a set of (x,y) when |x/y-a| < E, and the E is a fixed decimal. '||' stands for absolute value. Please find out the specific (x,y) that let (x+y) get to the minimum value.
I've think out a method but I'm not confident to solve the problem. And I don't know how to test it automatically. 10:58
antoniogamiz is M finite? 11:00
and what do you mean by "decimal"? a digit in the decimal part of a numbert? or a rational number? 11:01
uid9164 Wait a minute. 11:02
a is a number with limited digits and a < 0. M is a infinite set of (x,y) when |x/y-a| < E, and the E is a fixed small value. '||' stands for absolute value. Please find out the specific (x,y) that let (x+y) get to the minimum value. 11:03
For example, when I want to find the (x,y) when a is 0.11688. 11:04
antoniogamiz mm are you sure this problem is computable? 11:05
uid9164 When using the method provided from rosettacode.org/wiki/Convert_decim...ational#C, it returns 475/4064
But I want the result as 9/77. In fact, 9/77=0.11688312，475/4064=0.11687992 11:06
I am aware of the difference between a native English speaker and a foreigner... 11:08
antoniogamiz if you want the 0.2939 form, try using .Num 11:10
m: say (9/77).Num
camelia 0.11688311688311688
uid9164 pastebin.com/MS87pr83 11:13
11:19 MARTIMM joined 11:23 zakharyas left 11:25 uid9164 left 11:27 pat_js left, [Sno] joined 11:32 lizmat left 11:37 lizmat joined 11:48 kentnl left 11:53 antoniogamiz left 11:55 lizmat left 11:56 dolmen joined 12:00 reportable6 left, reportable6 joined, ChanServ sets mode: +v reportable6 12:01 Ven`` left 12:06 Ven`` joined 12:07 thundergnat joined
thundergnat m: my $a = .11688; .say for $a.nude, $a.Num.Rat.nude; 12:07
camelia (1461 12500)
(484 4141)
12:13 lucasb joined 12:14 thundergnat left
moritz m: my $a = .11688; .say for $a.nude, $a.Num.Rat(1e-9).nude 12:20
camelia (1461 12500)
(1461 12500)
12:22 mensvaga joined 12:23 pat_js joined 12:29 molaf left 12:37 sena_kun joined, sena_kun left, [Sno] left 12:38 Ven`` left, jkramer joined
jkramer Hello 12:38
12:39 Ven`` joined 12:40 Ven`` left
moritz hi jkramer 12:41
jkramer Is there in a regex to try all possible alternatives and not just go with the longest? 12:44
m: my @s = <a aa abb>; say 'aabb' ~~ m/^ (@s)+ $/
camelia False
jkramer Like in this example, it would match if it tried a + abb, but it goes with aa first and then abb won't match.
m: my @s = <a aa abb>; say 'aabb' ~~ m:exhaustive/^ (@s)+ $/ 12:45
camelia ()
timotimo m: my @s = <a aa abb>; say 'aabb' ~~ m:ov:ex/^ (@s)+ $/
camelia () 12:46
jkramer m: say 'aabb' ~~ m:exhaustive/^ (a || aa || abb)+ $/
camelia (｢aabb｣
0 => ｢a｣
0 => ｢abb｣)
jkramer With || it works but @ interpolation seems to use | which does LTM
timotimo m: my @s = <a aa abb>; say 'aabb' ~~ m:ov:ex/^ (||@s)+ $/ 12:47
camelia ()
timotimo that is supposed to use || with interpolation i think
jkramer Is there a pretty way to use || with an array, ie. without typing all elements of the array into the regex manually? :)
12:49 El_Che joined
jkramer m: say 'aabb' ~~ m:exhaustive/^ (a || aa || abb || ab || b || bb)+ $/ 12:50
camelia (｢aabb｣
0 => ｢a｣
0 => ｢a｣
0 => ｢b｣
0 => ｢b｣ ｢aabb｣
0 => ｢a｣
0 => ｢a｣
0 => ｢bb｣ ｢aabb｣
0 => ｢a｣
0 => ｢abb｣ ｢aabb｣
0 => ｢a｣
0 => ｢ab｣
0 => ｢b｣ ｢aabb｣…
12:51 mowcat left
jkramer timotimo: Can't find anything about ||@ in the docs 12:51
timotimo m: my @s = <a aa abb>; say 'aabb' ~~ m:ov:ex/^ (|| @s)+ $/
camelia ()
timotimo hmm
El_Che lo 12:52
12:55 Doc_Holliwood joined 13:04 lizmat joined, Ven`` joined, Ven`` left 13:08 aaronwells joined 13:09 chloekek left 13:12 Ven`` joined, curan left 13:13 Ven`` left 13:15 AlexDaniel left 13:16 satori__ left 13:21 zakharyas joined 13:24 dolmen left 13:27 dolmen joined 13:28 AlexDaniel joined, dolmen left 13:31 Ven`` joined 13:33 aaronwells left
tyil can anyone tell me why `prove6 -l` complains about a missing plan in my subtest (gitlab.com/tyil/perl6-app-assixt/-...69462200), even though it looks to me like there's clearly a plan in there? (gitlab.com/tyil/perl6-app-assixt/b...fig.t#L24) 13:35
it's surely to be something silly that I'm staring too hard at to notice :I
timotimo is something outputting on stdout perhaps?
13:35 aaronwells joined, ChoHag joined
tyil does that interfere with prove6? because there is some output, yes 13:36
timotimo BEGIN plan :skip-all<set AUTHOR_TESTING=1 to run bin tests> unless %*ENV<AUTHOR_TESTING>;
tyil I could work around it by adding a silent option, though
timotimo BEGIN may be the wrong choice here?
tyil I'm unsure, I copied that ages ago from someone else :p
timotimo well, it's not really like test files would be precompiled
tyil but if stdout stuff interferes with prove6, I can work around that 13:37
with some effort, but at least I have some possible pointer now
timotimo and i don't think skip-all will cause the rest of the file from being run
prevent*
tyil it causes the rest of the tests to be ignored, which is Good Enough for me here 13:38
13:38 dolmen joined
timotimo OK 13:38
13:38 Ven`` left
timotimo i'm not sure what's going wrong with that test :( 13:38
tyil I'm going to add a --silent option and see if that helps 13:39
it seems to indeed resolve the issue for the small test I just re-ran with all output commented 13:40
so I have something to improve, thanks :D
timotimo yeah, it could be confusing the tap parser when there's output on stdout
13:42 pat_js left 13:45 Ven`` joined, mowcat joined
cpan-p6 New module released to CPAN! Gnome::GObject (0.13.16) by 03MARTIMM 13:45
13:46 lizmat left 13:49 natrys joined 13:55 Ven`` left 13:57 Ven`` joined 13:59 Ven`` left, lizmat joined 14:01 [Sno] joined 14:02 lizmat left 14:04 wamiks joined 14:07 ravenousmoose left 14:08 mowcat left 14:12 Ven`` joined
wamiks Hi, I am confused. I created a unit module within which are both subs and classes. I can't seem to call new on one of the classes from within one of the subs. This is true even if I use its full name:ie Mdl1::Mdl2::myClass.new() I get "can't find symbol &myClass". OTOH, I have calling one of the subs works. Is this expected? How do I debug why my sub can't find my class? 14:14
14:15 dolmen left
timotimo "cannot find symbol &myClass" means it's trying to find a sub by that name 14:16
there's clearly something wrong there
14:16 [Sno] left
timotimo without seeing the code, i don't think i'll be able to help 14:16
wamiks Thx. If I paste the same call in the REPL it works. 14:17
timotimo could be it's complaining about &myClass only because it tried to find myClass without & earlier and failed that too? 14:18
wamiks Yeah, I would have to pull that code out to show you, it's a bit involved. Is there something debuging you can suggest to get some visibility what that function sees? 14:19
timotimo you could try ".say for MY::.keys" 14:20
14:21 dolmen joined
wamiks :) I did try something similar. 14:21
It seems to just be the variables defned in that scope. 14:22
14:22 Ven`` left
timotimo could also look at OUR:: 14:23
14:24 dolmen left
wamiks The class is in OUR::.keys 14:26
timotimo if worst comes to worst, you can get at it that way
wamiks ahh, ok. just 'myClass.new()' give 'undeclared name' 14:27
haha, OUR did it. Cool! Thx! 14:28
14:30 AlexDaniel left
wamiks Any thoughts why it is this way? 14:30
14:30 zakharyas left
timotimo i mean, if the class is inside a different pair of curly braces, it wouldn't be lexically available where you're trying to use it 14:31
but your question doesn't suggest that is the case
14:34 Ven`` joined 14:35 ravenousmoose joined
jdv79 whatever or start aren't bad names either:) 14:35
*star
camelia being 4 syllables is meh but whatever 14:36
14:36 aaronwells left, Ven`` left
timotimo let's call it "What" 14:36
"what programming language are you going to use for this project?" - "yes."
jdv79 haha 14:37
timotimo "what ain't no language i ever heard of"
Doc_Holliwood module XX { sub xx { "X" } }
When I try to "use" this rakudo tells me the symbox xx not being found
*symbol. 14:38
timotimo you have to "is export" the sub for it to become visible when "use"d
Doc_Holliwood tried that
14:38 dolmen joined
jdv79 star is 4 chars and one syllable though and isn't a stopword;) 14:39
14:40 antoniogamiz joined
wamiks @timotimo, braces-wise it's a sibling to the function. it 'does' a role which is also at the same scope, not sure if that matters. 14:40
timotimo there's already a programming language called "star"
14:40 tokomer joined
Doc_Holliwood you mean as in < module XX { sub xx is export { "X" } } >? 14:41
jdv79 oh:(
Doc_Holliwood same error
jdv79 i googled star comp lang and skimmed and didn't notice
timotimo oh, you'll perhaps want to have it an our sub and then XX::xx
Star seems to be made by the same person who made Go! 14:42
14:42 dolmen left
Doc_Holliwood Ah, progress. Now I get "Cannot import symbol $xx from XX, because it already exists in this lexical scope" 14:43
timotimo you mean &xx perhaps?
14:43 [Sno] joined
Doc_Holliwood No, it's referring to the $xx in < sub EXPORT (*@args) { %( '$xx' => XX::xx() ) } > 14:44
timotimo oh, you have an EXPORT sub
Doc_Holliwood module XX { our sub xx { "X" } } 14:45
still complains about $xfoox already existing in the scope
which is weird
14:45 ravenousmoose left
timotimo without the actual full code i won't be able to help :( 14:45
Doc_Holliwood That is the full code 14:46
That is XX.pm I run it using perl6 -MXX -e "use XX"
and get the aforementioned error
timotimo why do you import XX twice?
Doc_Holliwood right. 14:47
right, right. *ahemm*
14:53 wamiks left 14:54 MARTIMM left 14:55 dolmen joined 14:57 Sgeo_ joined 14:59 aaronwells joined 15:01 Sgeo__ left 15:06 sjm_uk left 15:07 Sgeo__ joined 15:10 Sgeo_ left 15:14 [Sno] left 15:19 Sgeo joined
Kaiepi m: my constant Foo = role Bar:ver<0.0.1>:auth<github:Kaiepi>:api<6.d> {}.^specialize(Mu, Bar); say Foo.^ver 15:21
camelia BEGIN TIME EXCEPTION IS AT <tmp> at 1
EXCEPTION: Could not instantiate role 'Bar':

5===SORRY!5=== Error while compiling <tmp>
An exception occurred while evaluating a constant
at <tmp>:1
Exception details:
5===SORRY!5=== Error…
Kaiepi m: my constant Foo = role Bar:ver<0.0.1>:auth<github:Kaiepi>:api<6.d>[::T] {}.^specialize(Mu, Bar); say Foo.^ver
camelia (Mu)
15:21 Sgeo__ left
Kaiepi m: my constant Foo = role Bar:ver<0.0.1>:auth<github:Kaiepi>:api<6.d>[::T] {}.^specialize(Mu, Bar); say Foo.^auth 15:21
camelia
Kaiepi m: my constant Foo = role Bar:ver<0.0.1>:auth<github:Kaiepi>:api<6.d>[::T] {}.^specialize(Mu, Bar); say Foo.^api
camelia 15:22
Kaiepi should these be set on Perl6::Metamodel::ConcreteRoleHOW like its name is? the class already does the role for them
*roles 15:23
nah wait, role is right
15:23 Sgeo_ joined 15:25 zakharyas joined 15:26 Sgeo left 15:27 sjm_uk joined 15:32 zakharyas left, aaronwells left
Kaiepi shit, wrong channel 15:33
15:33 zakharyas joined
gfldex m: package P { our regex r { .+ } }; say 'match' ~~ /<P::r>/; 15:33
camelia No such method '!cursor_init' for invocant of type 'P'
in block <unit> at <tmp> line 1
gfldex how do I access r ?
timotimo gfldex: i'd just bind it to an alias 15:35
gfldex timotimo: should what I do work? 15:36
timotimo i think that's syntax for when you're taking a grammar's regexes
so it'll probably try to .^find_method or something
gfldex if I put it into a grammar I get a different error. 15:37
15:38 satori__ joined
Doc_Holliwood m: VM.osname.say 15:38
camelia linux
Doc_Holliwood I'm thinking about adding Win32 support to my github.com/holli-holzer/perl6-Test-Color 15:42
15:42 wildtrees joined
gfldex m: package P { our regex r { .+ } }; say 'match' ~~ P::r; 15:43
Doc_Holliwood That would require bundling some C-code though
camelia Too few positionals passed; expected 1 argument but got 0
in regex r at <tmp> line 1
in block <unit> at <tmp> line 1
gfldex m: package P { constant r = regex { .+ } }; say 'match' ~~ P::r;
camelia ｢match｣
gfldex o.O
15:43 MARTIMM joined
Doc_Holliwood So, how would I write an "XS.Module" if yiz will? Can zef make a C-project? 15:43
15:43 wildtrees left
timotimo we have LibraryMake 15:44
15:44 wildtrees joined
timotimo what do you need that NativeCall doesn't provide, out of curiosity? 15:44
Doc_Holliwood The windows console does not support ANSI escape sequences 15:46
But there is a project named ansicon 15:47
15:48 wildtrees left, wildtrees joined
Doc_Holliwood So i was thinking I could bundle that and have zef build it 15:48
Kaiepi Doc_Holliwood, you can use LibraryMake 15:49
Doc_Holliwood Yes, tx. Looking into it
16:05 Sgeo__ joined 16:06 dolmen left 16:09 Sgeo_ left 16:15 antoniogamiz left 16:22 antoniogamiz joined 16:34 Sgeo_ joined 16:37 Sgeo__ left 16:38 scotticles joined 16:41 mowcat joined 16:45 EuAndreh[m] left, sergiotarxz[m] left, rba[m] left, matiaslina left, Matthew[m] left, unclechu left, lance_w[m] left 16:46 Demos[m] left, Seance[m] left, Guest48765 left, folex left, mack[m] left, epony left, TravisRt2botio[m left, AlexDaniel` left 16:48 zingzhang left 16:49 rindolf left
Doc_Holliwood I am trying to find the source of the prove6 script. but when I open the one in bin i see "CompUnit::RepositoryRegistry.run-script("prove6", :$name, :$auth, :$ver)" 16:50
where to go from there?
16:50 leont left 16:52 AlexDaniel` joined
El_Che Hi AlexDaniel` 16:55
16:58 zingzhang joined 17:05 EuAndreh[m] joined, TravisRt2botio[m joined, lance_w[m] joined, Matthew[m] joined, rba[m] joined, Guest35100 joined, sergiotarxz[m] joined, mack[m] joined, Demos[m] joined, Seance[m] joined, folex joined, unclechu joined, matiaslina joined 17:12 Sgeo_ left, Sgeo_ joined 17:16 Xliff joined 17:25 xinming_ left
Xliff \o 17:29
17:31 Black_Ribbon joined 17:35 mahafyi left 17:36 Aearnus joined
vrurg Xliff: o/ 17:42
Xliff vrurg: How goes 3113? 17:44
vrurg Which 3113? Ah, that old one... ;) 17:45
Playing with your problem. For now the problem is I don't understand yet what it should be. 17:46
Xliff Oh. Yay?
I've boiled it down to this: If there is an error why isn't it being reported? 17:47
Said error must occur if another compunit is being dynamically loaded because it works when loaded on its own.
vrurg That's what I wanna see. First of all, your `try require` masks the problem.
Xliff Yeah. Everybody says that. However, that's the only way I can get it to assign. 17:48
vrurg Xliff: you don't need to assign it this way. ::("Module") is what you need to see that the module is not loaded.
Xliff Still results in the same thing. 17:49
Tell you what. I will rework the bug case your way and see if it helps.
vrurg But the CATCH would step in then. Because currently it caches exploded Failure from symbol resolution. 17:50
Xliff Yes, but wouldn't that at least explain the problem? 17:52
vrurg I'd like to test on 2019.03 and on the HEAD, but the former is still installing dependencies. So, I can't test my hypothesis yet.
Xliff OK. I will hold off until you need me, then.
vrurg Xliff: what happens now (as I see it): import fails, an exception gets thrown. ::("Module") returns a Failure with "NoSuchSymbol" by default – and this is what you get when require fails. try throws away the actual exception, it's not stored in ::("Module") Failure as one might expect. 17:54
2019.03 behaves the same. Ok. 17:55
Xliff vrurg: The one problem I see with that explanation is this: If that were so... it would NOT work in isolation. 17:56
The module loads by itself, so it can't be a "NoSuchSymbol" issue.
18:00 reportable6 left 18:01 reportable6 joined, ChanServ sets mode: +v reportable6 18:03 zakharyas left
vrurg Xliff: something is definitely going on there. I'm just not sure yet where exactly it happens. If you make a deliberate error in CopySnapshot it is reported back normally. 18:06
18:07 irced joined
Xliff Is CopySnapshot the only module loaded? 18:08
I'm still wondering if the root cause is GetTemplateLaunchData + anything else.
vrurg Xliff: looks like the cause is in chain of uses. I wonder at which depth level it happens. 18:10
Xliff Oh, wow!
A missed circular use case? 18:11
vrurg No, nothing is being lost. I was changing a wrong module. 18:16
So, I went as deep as to LaunchTemplateTagSpecification inserting "die" – and it always bounces back if I use CATCH { }; require ::($module).
18:18 rindolf joined
vrurg Xliff: so, get rid of try, do just require, use CATCH in the same scope and at the end bind the ::($module) to a hash key. That works. 18:18
Xliff \o/ 18:19
vrurg: Give me 30 minutes max to get my hands out of the guts of another project and I will verify.
vrurg Xliff: I would comment on the issue, you'll find the corrected script code in it. 18:20
BTW, I'm amazed by the number of modules in the project. Did you use an automation to produce them? Because if it's all manual work then you're one of the craziest and most patient people I ever seen. :) 18:22
Xliff :D 18:23
Automation to get them created, then hand corrected.
So a bit of both?
Dude.. that's my groove. Just look at all of the projects on github. :>
18:24 Sgeo__ joined
irced what project are you guys discussiong, Xliff & vrurg ? 18:26
vrurg irced: Amazon::AWS::EC2
Xliff github.com/Xliff/p6-Amazon-AWS-EC2
irced vrug, Xliff: thankws
Xliff Also... github.com/Xliff/p6-GtkPlus 18:27
Also... github.com/Xliff/p6-GooCanvas
... github.com/Xliff/p6-Clutter
:>
... github.com/Xliff/p6-WebkitGTK
... github.com/Xliff/p6-SourceViewGTK
18:28 Sgeo_ left
vrurg Aha, sure! (can't find approriate emoji for being shocked :) ) 18:28
irced Xliff: nice
Xliff And the blocker... the grandpappy... 18:29
... github.com/Xliff/p6-Pango
vrurg Xliff: BTW, how hard is it to do the same for Amazon MWS?
Xliff MWS?
linky?
irced xliff: thanks for mentioning unpinned 18:30
Xliff irced?
vrurg Xliff: Marketplace API.
irced Xliff: i didn't see p6-Pango on the pinned at github.com/Xliff
Xliff Oh. Yeah. I wish they had more pinned slots. 18:31
All of this would be pinned.
vrurg: Actually... not that difficult for MWS
vrurg Ok, I know who to ask then if I need it. :)
Xliff LOL! Sure thing. 18:32
irced vrurg: which project impressed you with the number of modules? i am interested in seeing how big projects are glued together :-)
😃
Xliff When I say "not that difficult" I mean "slightly difficult unless I can pull out the data from docs.developer.amazonservices.com/e...estDetails for automation"
El_Che AlexDaniel`: going for consensus (everyone) instead of a majority or even consent (most people agree and those who don't don't veto it) for PR is a bold step :)
Xliff irced: The largest, in terms of number of modules is Amazon-AWS-EC2. 18:33
p6-GTKPlus is second.
vrurg irced: Amazon. Scrolling through the list of Types modules makes bored at thrid swipe up... :D
irced laughs.
Xliff github.com/Xliff/p6-Amazon-AWS-EC2...ration.pm6 18:34
Main script: github.com/Xliff/p6-Amazon-AWS-EC2...eClass.pl6
irced vrurg: so i'm looking at the META6.json file, sorry if i am interrupting your guyz investigation btw 18:35
Xliff vrurg: Actually... weirdness still persists.
irced vrurg: which file were you looking at (scrolling)
Xliff After making your changes... if I use only GetLaunchTemplateData, it works just fine.
If I add CopySnapshot, I get: No such symbol 'Amazon::AWS::EC2::Action::GetLaunchTemplateData' 18:36
vrurg Xliff: BTW, I would suggest using `unit class` in favor of `class ... is export`. It's cleaner in many way. `is export` may have obscure side effects when used outside of a package.
irced: we're over with it, I think. Luckily, it's not something in rakudo, after all.
Xliff unit class uses OUR semantics?
vrurg Xliff: yes. 18:37
Xliff k
vrurg And it makes code cleaner and more, to say, logical. 18:38
18:38 tokomer left
Xliff OK. 18:39
irced vrurg: so thanks for taking time to share this with me. what file had the long list of modules ?
Xliff irced: github.com/Xliff/p6-Amazon-AWS-EC2.../EC2/Types
vrurg: After moving GetLaunchTemplateData to "unit class..." I get "No such symbol 'Amazon::AWS::EC2::Action::GetLaunchTemplateData'" 18:40
That happens ONLY if I load CopySnapshot prior.
18:40 Sgeo_ joined
Xliff If I load CopySnapshot AFTER GetLaunchTemplateData... it works! 18:41
wtf?
vrurg Xliff: I get :Malformed has at /Users/vrurg/src/Perl6/p6-Amazon-AWS-EC2/lib/Amazon/AWS/EC2/Response/GetLaunchTemplateDataResponse.pm6 (Amazon::AWS::EC2::Response::GetLaunchTemplateDataResponse):14
irced vrurg: i see. yes, it is impressive. thanks again. and thanks Xliff.
vrurg irced: I was actually scrolling in Atom's file panel. ;) 18:42
Xliff vrurg: Perl6 --version?
irced vrurg: i don't know atom from Adam but it is just a midnight commander type deal right?
vrurg Xliff: This is Rakudo version 2019.07.1-130-ga7302f0a4 built on MoarVM version 2019.07.1-72-g352ae27e4 18:43
irced vrurg: or like vim's Explore
18:43 Sgeo__ left
Xliff vrurg: OK, thanks. I am rebuilding moar-blead 18:43
vrurg irced: images.app.goo.gl/ST5FSa6rjKsxK6tw8 – that rectangle on the left. ;) 18:44
Xliff: I don't think it's related.
18:44 empee0 joined
vrurg Xliff: can you commit your changes to the test branch? 18:44
Xliff Sure. One sec.
irced vrurg: oh i see, so the list length really made an impression there. got it. 😄 18:45
vrurg Xliff: I have changed all declaration in the files I touched to units. 18:46
Xliff Oh, LOL! 18:47
OK. Pushed.
vrurg And I see what the problem is. With unit you must use the full class name. `is export` exports it's short form.
irced Xliff: so you're a big fan of gtk. seems to be true to its cross-platform mission only lately imho. 18:48
Xliff irced: Heh!
vrurg Either you must regenerate things to utilize full names or stick to 'class is export'
Xliff vrurg: Ah. OK.
irced Xliff: which is of course good. and i guess you considered Qt ?
Xliff Kinda depending on the short form.
vrurg Full names are preferable unless you're 100% sure there won't be name clashes in short names. 18:49
Xliff "Amazon::AWS::EC2::Action ...." gets kinda long in the tooth after a while. Yes, I am tracking the short names to minimize (if not eliminate) clashes.
irced Xliff: err, i think that is c++ as opposed to C so guessing not
Xliff irced: Yes, I am considering Qt. That's a while off, though. 18:50
irced Xliff: ambitious. why stop there? GLUT! 18:51
Xliff vrurg: Of course, there is still the issue of order.
Why does GetLaunchTemplateData work fine if it is loaded before CopySnapshot?!?
vrurg Xliff: that's a mistery. And it doesn't matter wether it's unit or is export declared. Weird. 18:52
Xliff Yep! Hence the bug description.... 18:53
irced Xliff: looks like i lost you on GLUT 18:54
Xliff irced: Hunh? Did you say something? 18:55
github.com/Xliff/p6-COGL 18:56
My current WIP
irced examines.
Xliff This might be a better starting point: www.cogl3d.org/ 18:57
irced Xliff: ah, yes. well, seems to me GLUT is all about the graphics and interfaces (inputs) (whereas gtk and qt are that and so much more) but cogl is specific about the graphics (sans the interfacing). 18:59
Xliff: but the emphasis is on 3D of course in GLUT and cogl. 19:00
Xliff Yep.
So COGL should tied you over until someone does GLUT.
8-)
irced Xliff: I see, there is no specific thing GLUT would phenotypically provide when you have those two working. 19:02
19:02 epony joined
timotimo what'd you need glut for when we have sdl2? 19:04
Freeglut 3.0.0 [Released: 7 March 2015]
heh.
irced thanks for chiming in timotimo, and for sharing that library! 19:05
Xliff timotimo: Don't you start.
Between you and vrurg, my dance card is punched!
irced now you can give Xliff two more projects to divide his time with. and that's not a problem if he can divide time itself. 19:06
timotimo sdl is already there, though
Xliff irced: Nope. Don't have a knife sharp enough. Wanna make one for me? ;)
irced Xliff: i dunno, if i ever get to mt olympus perhaps someone can fashion one for me and one for you 19:07
Xliff: so i'll do my best
timotimo am i not removing projects from xliff's theoretical list of stuff to make? 19:09
Xliff I saw a something get tossed onto the pile, only to be yanked off. Methinks it was made of straw. 19:10
timotimo yeah, feel free to ignore glut
19:11 Aearnus left
irced timotimo: depends on the goal. gtk+ has a + in it for a reason, it provides to much. but if one were to just focus on the symmetric set difference of {gtk+} (-) {sdl} then you are actually removing elements from the gtk+ but the difference is probably added back 19:12
timotimo i'm not sure what you're trying to tell me here :) 19:13
irced timotimo: it's ELEMentary. anyway...
modules.perl6.org/dist/SDL2::Raw:cpan:TIMOTIMO 19:14
looks like something i may want to get into! thanks again!
19:20 mowcat left
Xliff CoglQuaternion?!? WTF? 19:20
timotimo what's wrong with that?
quaternions are used a whole lot in 3d math for rotation-y things 19:21
Xliff "Use the word 'Quaternion' in an object." -- Me crosses off bucket list.
irced oops, i used the wrong set notation to describe this.. .. 19:25
irced thinks
basically, timotimo, xliff has gtk+ for UI+ and COGL for 3d but if you add sdl you get the same as COGL plus more. so you can remove COGL but work is replaced and even more might be added. 19:28
but you remove some of gtk+ ui functions
with sdl ui 19:29
so theoretically i think the work would be about the same
ignoring the "even more might be added" for now. 19:30
timotimo cogl goes on top of sdl nicely 19:31
if you bind cogl, you don't have to write bindings for opengl
irced actually, i might revise what i said, adding sdl can replace cogl. and gtk+ UI can be accomplished by cogl. so, i am leaning towards your first assertion.
vrurg Xliff: this is really funny. If CopySnapshot is require'd first then seond require leaves us with... CopySnapshot in the symbol table! And no ides where GetLaunchTemplateData is imported into then... 19:32
timotimo i've got a particle system in SDL2 and i have to either loop over every point and set color and render a dot or i can only have a single color for all dots
in cogl i can just set up a few buffers, say "one of them has colors in them" and fire off the whole bunch at once 19:33
irced anyway, since you're working on interfacing with sdl in perl6 then Xliff definitely doesn't need to add that to his list. 19:34
19:35 ufobat joined
ufobat within a regex i can have { } and inbetween there could be code that gets invoked, but the return value does not mean anything, there are not really sideeffects? 19:36
irced timotimo: so great, Xliff's cogl and timotimo's sdl can work together. thanks for sharing! 19:37
19:37 netrino joined
ufobat and i could also have {} :codegoeshere; the ';' ends the code. 19:38
i wasnt able to find an explanation for this in the docs :-(
Xliff vrurg: Should change back to the class block form and see what happens? 19:39
vrurg Xliff: it doesn't matter. I did already.
Xliff OK 19:40
vrurg Xliff: though I'm afraid `is export` without a wrapping package might contribute to the problem. It's only a guess, though. 19:41
I generally dislike package-less units. They make me suspicious. ;) 19:42
Xliff Yeah. I will have to change things around.
ufobat ah jj had already opend issues about that on the docs repo 19:43
irced ufobat: have you seen any side effects from the return value? 19:44
ufobat: anyway, block code doesn't recognize return (pure block)
ufobat: so the parser can consider the context and determine whether to do the block or not "do" the block into its final evaluated expression 19:45
ufobat: so in the context of a regex, i don't think it looks for what the {} evaluates to. 19:46
ufobat: it just calls a block
ufobat i haven't seen sideeffects, no
19:47 sjm_uk left
ufobat in such a block I could also call $/.made() to avoid having an action class? 19:47
irced ufobat: i'll see if i can address that (new that i am) but let me illustrate firstly a block code routine vs sub 19:50
m: my &blockcode := { 2 + 2}; say "{blockcode} is equal to -42"; 19:51
camelia 4 is equal to -42
irced m: my &blockcode := { return 2 + 2}; say "{blockcode} is equal to -42";
camelia Attempt to return outside of any Routine
in block <unit> at <tmp> line 1
irced my &subroutine := sub routinecode { return 2+2+ }; say "{routinecode} is equal to -42"; 19:52
m: my &subroutine := sub routinecode { return 2+2+ }; say "{routinecode} is equal to -42";
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing required term after infix
at <tmp>:1
------> 3outine := sub routinecode { return 2+2+ 7⏏5}; say "{routinecode} is equal to -42";
expecting any of:
prefix
term
irced m: my &subroutine := sub routinecode { return 2+2 }; say "{routinecode} is equal to -42"; 19:53
camelia 4 is equal to -42
irced there might be some superfluous code there but you see a block is different from a return
from a routine
m: my &subroutine := sub routinecode { return 2+2 }; say "{subroutine} is equal to -42"; 19:54
camelia 4 is equal to -42
irced as for $/.made() to avoid having an action class that is currently beyond me, anyone else? 19:56
ufobat: i'll see if i can address that (new that i am) but let me illustrate firstly a block code >---routine----< err not routine vs sub 19:57
ufobat thanks :)
irced hth 19:58
ufobat m: say "AABCDE" ~~ / (\w) {} $0 / 20:00
camelia ｢AA｣
0 => ｢A｣
ufobat say "AABCDE" ~~ / (\w) {} <?before A> /
m: say "AABCDE" ~~ / (\w) {} <?before A> /
camelia ｢A｣
0 => ｢A｣
ufobat m: "AABCDE" ~~ / (\w) {} <?before $0> /
camelia ( no output )
ufobat is this a bug in the regex engine?
irced and just to say a little more about block code vs routine code, when parsing say "{blockcode}", say will look at the last statement in blockcode, not a return value (in blockcode). it doesn't have to, but it is does. regex on the other hand doesn't have to and doesn't. 20:01
20:07 antoniogamiz left
irced disclaimer: i am still learing the regex ropes, so the blockcode vs subroutine code may only serve to clarify some things not explain everything. as for the bug question 20:09
m: "AABCDE" ~~ m:global/ (\w:) {} <?before $($0)> /
camelia ( no output )
irced m: "AABCDE" ~~ m/ (\w:) {} <?before $($0)> / 20:10
camelia ( no output )
irced m: "AABCDE" ~~ m/ (\w) {} <?before $($0)> /
camelia ( no output )
irced m: say "AABCDE" ~~ m/ (\w) {} <?before $($0)> /
camelia ｢A｣
0 => ｢A｣
irced m: say "AABCDE" ~~ m/ (\w) {} <?before $0> /
camelia False
irced m: say "AABCDE" ~~ / (\w) {} <?before $0> / 20:11
camelia Nil
20:11 jtpalmer joined
irced m: say "AABCDE" ~~ m/ (\w) {} <?before $($0)> / 20:13
camelia ｢A｣
0 => ｢A｣
irced ufobat: see that?
20:13 [Sno] joined
ufobat m: say "AABCDE" ~~ m/ (\w) {} <?before <$0>> / 20:15
camelia ｢A｣
0 => ｢A｣
ufobat ahh nice, irced!
irced There are four ways you can interpolate a string into regex as a pattern. That is using $pattern, $($pattern), <$pattern> or <{$pattern.method}>.
^ docs.perl6.org/language/regexes
ufobat yeah found that on the docs too :D
thank you
irced ufobat: hth 20:16
of course, one of the four ways states $pattern, but when referring to a match, it looks like one of the alternatives apply. 20:23
you could even do this 20:25
m: say "AABCDE" ~~ m/ (\w) {} <?before {$0}> /
camelia ｢A｣
0 => ｢A｣
irced but that seems like undefined regex behavior to me 20:26
ufobat that confuses me, because why does it now check the content of the last expression of that block
irced outside of <?before ...> using the backreference like that won't work 20:29
20:29 MARTIMM left
ufobat but that is an inconsistency? 20:30
not saying that i would expect it to work outside of <?before> 20:31
irced i think the assertion does not allow for ALL patterns
so it is in a "class" of its own
checking the docs for something more substantial than that :-p 20:32
20:34 pilne joined 20:38 [Sno] left
irced well, some experimentation shows that backreferences such as {$0} are as expected ignored in a m/ ... / 20:38
20:39 [Sno] joined
irced not the backreferences themselves mind you, just the block 20:42
20:43 noisegul_ left
irced is experimenting. 20:44
20:51 ufobat_ joined 20:55 ufobat left 21:06 scotticles left, spider_apocalyps joined 21:43 zingzhang left
irced ufobat_: it is inconsistent. 21:44
ufobat_: but it is consistently inconsistent. 😃 21:45
ufobat_ :D
21:47 natrys left
irced btw, a do subroutine accepts a block as an argument so that say "{do {blockcode}} is equal to 1" happens to work. but again, do itself is a routine that returns the last evaluated expression of the blockcode... just to show there is at least consistency in that part of the language 😃 21:49
though do is not a traditional subroutine, at any rate, it behaves like a "routine" vs "block" 21:50
m: my &blockcode := { 2 + 2}; say "{do {blockcode} is equal to -42"; 21:51
camelia 5===SORRY!5=== Error while compiling <tmp>
Strange text after block (missing semicolon or comma?)
at <tmp>:1
------> 3ckcode := { 2 + 2}; say "{do {blockcode}7⏏5 is equal to -42";
expecting any of:
infix
infix…
irced m: my &blockcode := { 2 + 2}; say "{do {blockcode}} is equal to -42";
camelia 4 is equal to -42
irced needs a vision exam
21:52 Summertime left
irced m: my &blockcode := -> $a, $b { $a + $b}; say "{do {blockcode(2,2)}} is equal to -42"; 21:52
camelia 4 is equal to -42
irced m: my &blockcode := -> $a, $b { $a + $b}; say "{blockcode(2,2)} is equal to -42";
camelia 4 is equal to -42
21:53 Summertime joined
irced m: my &blockcode := -> $a, $b { return $a + $b }; say "{blockcode(2,2)} is equal to -42"; 21:54
camelia Attempt to return outside of any Routine
in block <unit> at <tmp> line 1
irced m: my &blockcode := -> $a, $b { return $a + $b }; say "{do {blockcode(2,2)}} is equal to -42"; 21:55
camelia Attempt to return outside of any Routine
in block <unit> at <tmp> line 1
irced oops
scratch that statement about do.
21:56 eseyman left
irced or revise it anyway. 21:56
m: my &blockcode := -> $a, $b { return $a + $b }; say "{do {2 +2}} is equal to -42"; 21:58
camelia 4 is equal to -42
irced so do does not call return after all
but behaves like a subroutine in that the block can return 21:59
22:00 eseyman joined
irced but that's not blockcode, which itself rejects the return keyword 22:00
22:00 Summertime left
irced leaving it again up to the parser to read the last evaluated expression in a "blockcode" 22:01
so blockcode passed into do is still blockcode, which will never have a return statement. of course, a do statement can return the last evaluated expression of the blockcode. 22:03
so there is still consistency
irced whew
22:05 zingzhang joined, Sgeo_ left 22:06 Sgeo_ joined
irced and talking about do as a subroutine was probably wrong. it instead takes a block of code (not a parameter that is blockcode :p) that can return 22:07
but in the end it's what the outer blockcode/routine is that matters 22:10
22:10 spider_apocalyps left
irced so still a regex although it can have routines and do's nested and handle returns, the code block itself does not "return" and errors if a return statement is seen in its outermost part 22:10
so regex doesn't look for a return value and doesn't care about the last evaluated expression. though as we observed, a { } run in an interpolated assertion will read the last evaluated expression. nevertheless, it won't look for a return value since { return ... } is invalid 22:15
22:16 rindolf left 22:17 cpan-p6 left 22:18 cpan-p6 joined, cpan-p6 left, cpan-p6 joined
irced so in an interpolated assertion, we are no longer in the main pattern, but we are in a pattern with a slightly different role it seems, one where { } is run and its last evaluated expression is substituted into the interpolation 22:18
so it's best to think that anything "quoted" can be adverbially modified and an interpolated assertion is a special kind of "quote" is my stance 22:20
john_parr Has rakudo given up on targetting the JVM? It’s hard to find any recent developmentsin that area. 22:22
irced i am using interpolated assertion to refer to lookbehind etc 22:23
i suppose only Rakudo folks can answer that? 22:24
john_parr I imagine some of them are here and aren’t pedantic autistic cunts?
irced who are you calling a cunt? 22:25
anyway, why is openjdk not in active development?
john_parr nobody in particular
yeah, sure, openjdk is.
irced i asked a similar question here and learned that the jvm uses openjdk 22:26
Xliff Java™ is going closed source, so OpenJDK is the only way left.
So expect interest in the JVM to wane a bit. 22:27
irced therefore, Rakudo i assume is using OpenJDK
john_parr I’m sure the same opcodes would be emitted either way, why would it matter about the licensing around any particular java implementation?
Xliff Coz ppl don't like being sued?
john_parr LOL. I’ll come back on Monday. 22:28
john_parr slams door shut behind him.
irced hmmph
Xliff I would suggest everyone stop feeding john_parr if he continues to act like a troll. :( 22:29
tobs john_parr: I know that the continuous integration keeps testing the JVM and people try to fix build errors there, but I haven't heard of dedicated efforts to improve it. I'm not paying attention to it, though, so sorry if I overlook someone's work on it. 22:30
john_parr tobs: ah! thanks for the interesting answer. good to know!
irced tobs: thanks tobs for putting it into perspective 22:31
22:34 lizmat joined 22:40 empee0 left
irced revision 100000 do blocks also do not allow "return" statements. 22:41
22:42 Sgeo__ joined
irced one do's or not do's blockcode. one calls routines may "return" and if no return is specified, returns True, regardless of the last evaluated expression. 22:44
22:46 Sgeo_ left
irced err, returns last evaluated expression or expression explicitly returned 22:51
22:59 lucasb left 23:10 AlexDaniel joined 23:31 _jrjsmrtn joined 23:32 __jrjsmrtn__ left 23:53 leont joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!