»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 moritz on 25 December 2014.
Kristien Is there a nicer way to do the for loop? 00:00
Couldn't find an each function or something like that
masak map?
Kristien map is lazy 00:01
AFAIK
masak troo
but you're feeding things into it, so that shouldn't be a problem, I think.
vendethiel m: : 1..* Z @lines ==> map { "$^i: $^l" } ==> map(&say); 00:02
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/CkJoJ8iYDB␤Invocant colon not allowed here␤at /tmp/CkJoJ8iYDB:1␤------> :⏏ 1..* Z @lines ==> map { "$^i: $^l" } ==␤»
Kristien m: sub each(&f, @xs) { for @xs { &f($_) } }; each (1, 2, 3, 4, 5), &say
camelia rakudo-moar f12020: OUTPUT«Type check failed in binding &f; expected 'Callable' but got 'Parcel'␤ in sub each at /tmp/bboYqen_J_:1␤ in block <unit> at /tmp/bboYqen_J_:1␤␤»
Kristien m: sub each(&f, @xs) { for @xs { &f($_) } }; each &say, (1, 2, 3, 4, 5)
camelia rakudo-moar f12020: OUTPUT«1␤2␤3␤4␤5␤»
vendethiel m: : 1..* Z @lines ==> map { "$^i: $^l" } ==> each(&say);
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/uWM3fyBizI␤Invocant colon not allowed here␤at /tmp/uWM3fyBizI:1␤------> :⏏ 1..* Z @lines ==> map { "$^i: $^l" } ==␤»
Kristien m: map &say, (1, 2, 3, 4, 5)
camelia rakudo-moar f12020: OUTPUT«1␤2␤3␤4␤5␤»
vendethiel m: 1..* Z @lines ==> map { "$^i: $^l" } ==> each(&say);
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/t3qGX7Nwwy␤Variable '@lines' is not declared. Did you mean '&lines'?␤at /tmp/t3qGX7Nwwy:1␤------> 1..* Z @lines⏏ ==> map { "$^i: $^l" } ==> each(&say);␤ expecting any…»
Kristien well it only shows that because I force it to print the result of say
how about
vendethiel m: 1..* Z <a b c> ==> map { "$^i: $^l" } ==> each(&say);
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/QrAAdolFhD␤Undeclared routine:␤ each used at line 1␤␤»
vendethiel oh,yes, you declared it 00:03
Kristien: sink()
Kristien m: sub ignore(Any $x) { }; ignore(map(&say, (1, 2, 3, 4, 5)))
camelia ( no output )
vendethiel which is like ml's sink
m: 1..* Z <a b c> ==> map { "$^i: $^l" } ==> map(&say);
camelia rakudo-moar f12020: OUTPUT«1: a␤2: b␤3: c␤»
vendethiel this works though
Kristien no 00:04
it only works in this case because the output of map is iterated over
I need eager map which returns Nil
00:04 dayangkun left
vendethiel m: 1..* Z <a b c> ==> map { "$^i: $^l" } ==> map(&say); 1 00:05
camelia rakudo-moar f12020: OUTPUT«1: a␤2: b␤3: c␤»
Kristien but sink should work
m: my @lines = "Hello", "World"; sink { 1..* Z @lines ==> map { "$^i: $^l" } ==> map &say }; 00:06
camelia rakudo-moar f12020: OUTPUT«1: Hello␤2: World␤»
Kristien Dankeschön! 00:09
00:10 espadrine_ left
Kristien I like sub each(&f, @xs) { for @xs { &f($_) } } better, though. 00:14
Plays nicely with ==>.
I guess sink is implicit when an expression is used as a statement in non-return position? 00:15
masak aye 00:16
Kristien When a good Perl 6 book comes out I'll be the first one to buy a copy! 00:18
masak "a good Perl 6 book" might be our biggest weakness... :/ 00:19
we tried to do it as a community effort once, and ran into difficulties. 00:20
00:30 zakharyas left
Kristien Methods like join and uc seem to be callable without . 00:35
m: say(join(' ', <1 2 3>))
camelia rakudo-moar f12020: OUTPUT«1 2 3␤»
Kristien however, I can only find documentation of methods 00:36
Is this special behaviour for method calls or am I overlooking something?
m: say(<1 2 3>.join(' ')) 00:39
camelia rakudo-moar f12020: OUTPUT«1 2 3␤»
vendethiel Kristien: it's there if you put the "is export" on the method, IIRC
m: class Foo { method bar is export { 1 } }; bar(Foo.new);
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/RFlM_J46TP␤Undeclared routine:␤ bar used at line 1. Did you mean 'bag'?␤␤»
vendethiel no, so, probably just declared below.
(I know what is export does, but I thought it also did that)
Kristien > Methods also take the "is export" trait: the method will then be exported as a multi sub that takes the object as the first parameter:
vendethiel right :-) 00:40
masak I don't know if that's supposed to work or not, but I can say this: 00:44
the notion that the sub form of built-in methods are an exported form of the method in Perl 6 has been replaced by other things. 00:45
usually, the sub is just defined in the setting, and calls the corresponding method. 00:46
Kristien m: class Foo { method bar is export { 1 } }; import Foo; say bar(Foo.new); 00:47
camelia rakudo-moar f12020: OUTPUT«1␤»
vendethiel Kristien++ # smarter than me
masak hey, it works :)
Kristien :D
masak what I said still holds true, though. 00:48
I don't end up exporting methods a lot.
IIRC, even Perl 5 lore considers it a Bad Thing.
Kristien This is also interesting: 00:51
This works:
m: sub filter-profanity(Str $s) { $s.subst(/Java/, '****', :g) }; "Perl 6 is so much more fun than Java!" ==> filter-profanity() ==> uc() ==> say()
camelia rakudo-moar f12020: OUTPUT«PERL 6 IS SO MUCH MORE FUN THAN ****!␤»
Kristien whereas this fails:
m: sub filter-profanity(Str $s) { $s.subst(/Java/, '****', :g) }; "Perl 6 is so much more fun than Java!" ==> filter-profanity ==> uc ==> say
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/71EN2a_iYq␤Preceding context expects a term, but found infix ==> instead␤at /tmp/71EN2a_iYq:1␤------> fun than Java!" ==> filter-profanity ==>⏏ uc ==> say␤»
00:51 agentzh joined
vendethiel Kristien: you currently need () after a call inside a ==> 00:52
m: 'a' ==> say
camelia rakudo-moar f12020: OUTPUT«===SORRY!=== Error while compiling /tmp/Eod2xKzdcT␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument␤at /tmp/Eod2xKzdcT:1␤------> 'a' ==> say⏏<EOL>…»
vendethiel m: 'a' ==> say() 00:53
camelia rakudo-moar f12020: OUTPUT«a␤»
Kristien "currently" so will it not stay that way?
vendethiel well, it works in std
std: 'a' ==> say
camelia std f9b7f55: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/AWH6U6dwVw line 1:␤------> 'a' ==> say⏏<EOL>␤Check failed␤FAILED 00:00 134m␤»
vendethiel TimToady: can you confirm that's a STD bug?
or maybe masak++ can find the RT ticket associated to this one
Kristien ==> is most shiny. |> was my favourite feature in LiveScript and F#
vendethiel we completly and utterly stole it from f#, yes :P 00:54
Kristien I wonder if Arabs like normal calls better. :)
vendethiel I think elixir did that as well
Kristien In Elixir x |> f(1) is sugar for f(x, 1)
In LiveScript and F# it's sugar for f(1) x 00:55
Well in F# it's slightly different
TimToady well, we're still trying to figure out what feed operators are
Kristien in F# it's sugar for let __1 = x; f(1) x
because of order of evaluation being LTR always
TimToady and we certainly didn't steal it from F#
Kristien uh s/x$/__1/ 00:56
TimToady it's really just the idea of Unix pipes that can transmit objects, is all
I think we thought of it before PowerShell
Kristien Joe the Plumber is proud.
vendethiel TimToady: sorry, my "we" was from LiveScript ;-) 00:57
TimToady as for join, that can't be done by 'is export' because the order of arguments is different
ok
vendethiel Kristien: I'd certainly like f# better with more curry 00:58
Kristien vendethiel: LasagnaScript steals it from Elixir because almost no JS APIs are curried.
vendethiel Kristien: that's why prelude.ls has curry()
Kristien Meh 00:59
I kinda implemented protocols btw :P
masak wait... LasagnaScript?
vendethiel masak: Kristien's altJS :P 01:00
also, ls's |> allows for the "_" placeholder, so that if a API is considered "poxy", you can just work around it. ('foo' |> fn arg1, _)
masak I see.
Kristien vendethiel: I generalised that; % operator takes any term and turns it into a function, turning %n into the nth argument
vendethiel this is like clojure's %, right? 01:01
Kristien jawohl, except it doesn't restrict to calls
E.g. %%1 is the identity function. %(%1 * %2 + %3) and %try { foo(%1) } finally { bar() } are other examples. 01:02
vendethiel oh, pretty good 01:03
Kristien Not sure about nesting of them though. Could disallow or make %n pick arguments from the innermost %-function
masak a little bit like we use whatever and placeholder parameters for in Perl 6.
Kristien If you want a constant function returning a number you have to parenthesise the number though, so %(42) instead of %42 01:05
since the latter will try to yield the 42nd argument :P
01:05 BenGoldberg left 01:06 BenGoldberg joined 01:10 spider-mario left 01:28 telex left 01:30 telex joined 01:42 beastd left
masak 'night, #perl6 01:48
Kristien masak: goodbye 01:50
psch hi #perl6 \o 01:51
vendethiel: you mis-containered %num-to-notes in your PR :)
vendethiel: it's still neater though ;) 01:52
vendethiel psch: mmh?
did I? :p sorry!
psch vendethiel: yes, you scalared it, even though it's a hash
vendethiel: also, there's a tad more in the repo now, including a README, and tritone-in-c.pl has fallen to the wayside 01:53
vendethiel psch: 'twas a typo, sorry :( 01:54
psch vendethiel: it's fine. the repo is a collection of stuff i'm trying and that file in particular was just a check to see if what i'm imagining could be feasible :) 01:55
tony-o_ jnthn: ab5tract: bsd32 fails to build JVM perl6 too with the same error
too weird
01:56 fil_ left
tony-o_ ab5tract: what 686 linux distro did you use? 01:58
02:00 Kristien left 02:08 ggoebel111111111 joined 02:09 retupmoc1 joined, nine_ joined, felher_ joined 02:10 Psyche^ joined, robins joined 02:11 Tux__ joined, xinming_ joined, breinbaa1 joined 02:12 broquain1 joined 02:14 raydiak_ joined 02:16 Pleiades` left, Patterner left, vukcrni left, raydiak left, xinming left, ggoebel111111119 left, breinbaas left, |Tux| left, tadzik left, mattp_ left, avuserow left, robinsmidsrod left, retupmoca left, nine left, broquaint left, lestrrat left, felher left, mattp__ joined, Pleiades` joined 02:18 avuserow joined 02:20 tadzik joined 02:21 vukcrni joined 02:22 cxreg joined, skaufman_ joined 02:23 japhb_ joined 02:24 [particle]1 joined, nebuchad` joined 02:26 simcop2387_ joined 02:27 TimToady_ joined, slavik1 joined, risou is now known as 17SABKHBC, risou joined, VAB17AF joined, lestrrat joined 02:28 gfldex left, gfldex joined 02:29 rmgk_ joined, rmgk left, rmgk_ is now known as rmgk 02:33 daxim_ joined, agentzh left 02:35 cxreg2 left, skaufman left, TimToady left, japhb left, nebuchadnezzar left, petercommand left, [particle] left, 17SABKHBC left, slavik left, simcop2387 left, daxim left, simcop2387_ is now known as simcop2387 02:40 Ben_Goldberg joined, [particle] joined 02:41 kst` joined 02:44 [Coke]_ joined, huf_ joined, bartolin_ joined, Juerd_ joined, cosimo_ joined 02:45 ab5tract_ joined, BinGOs_ joined, osfamero1 joined, krunen_ joined, yeltzooo joined 02:46 Gothmog_1 joined, bcode_ joined 02:49 [particle]1 left, BenGoldberg left, bcode left, dalek left, IllvilJa left, PZt left, [Tux] left, lue left, pnu left, Juerd left, krunen left, jferrero left, BinGOs left, Gothmog_ left, bartolin left, woolfy left, ab5tract left, renormalist left, kst left, osfameron left, huf left, [Coke] left, yeltzooo9 left, dg left, cosimo left, charsbar_____ left, bcode_ is now known as bcode, Juerd_ is now known as Juerd, Gothmog_1 is now known as Gothmog_ 02:50 jferrero joined 02:51 renormalist joined 02:52 dg joined, TimToady_ is now known as TimToady, TimToady left, TimToady joined 02:55 lue joined 02:56 IllvilJa joined 02:57 gfldex left 02:58 [Tux] joined 02:59 charsbar_____ joined 03:00 woolfy joined 03:02 pnu joined, dalek joined, pnu left, pnu joined, ChanServ sets mode: +v dalek 03:15 VAB17AF joined 03:29 dj_goku left, dj_goku_ joined, dj_goku_ left, dj_goku_ joined 03:37 Maddingu1 joined 03:39 kaare__ joined 03:41 ugexe_ joined 03:45 kst` is now known as kst 03:48 stux|RC joined, Timbus_ joined 03:50 silug_ joined, noganex joined 03:52 Timbus left, stux|RC-only left, Alina-malina left, molaf_ left, kaare_ left, silug left, ugexe left, Maddingue left 03:54 noganex_ left, molaf_ joined 03:58 dj_goku_ left 04:18 dj_goku joined, dj_goku left, dj_goku joined 04:35 dj_goku left 04:38 dj_goku joined, dj_goku left, dj_goku joined
skids .tell japhb "The last statement in a module with more than one statement seems to have some extra-sinky context (because it supplies the return value of "use" I think). I just put 1; at the end of the module to prevent that." 04:58
yoleaux skids: I'll pass your message to japhb.
timotimo i don't think we have return values from use
TimToady the last statement of a module file should be sunk just as much as the other statements; if not, it's a bug 05:13
in fact, the 1; should be complaining about a useless use of '1' in a sink context 05:16
skids Well, I can't get it do error anymore so maybe it is fixed? I'll try with the code I needed to put it in way back when. 05:19
TimToady and, in fact, it doesn't complain, so it's a bug I thnk
if the final statement is a for loop, that used to require a more explicit sink 05:20
nowadays we autosink any loop in a statementlist
skids Yeah, no complaints about the final 1; in my Sum module, which uses it in most of the files. 05:22
TimToady but it does seem to sink the final statement anyway, just won't complain about useless use there for some reason 05:23
m: module Foo { 42; 43; 44 }
camelia rakudo-moar f12020: OUTPUT«WARNINGS:␤Useless use of constant integer 42 in sink context (line 1)␤Useless use of constant integer 43 in sink context (line 1)␤Useless use of constant integer 44 in sink context (line 1)␤»
TimToady warns there 05:24
m: module Foo; 42; 43; 44
camelia rakudo-moar f12020: OUTPUT«WARNINGS:␤Useless use of constant integer 42 in sink context (line 1)␤Useless use of constant integer 43 in sink context (line 1)␤Useless use of constant integer 44 in sink context (line 1)␤»
TimToady and there, huh
but use seems to suppress the final one somehow 05:25
skids It used to more than sink it, it somehow extra-double-super-secret-sunk it and punned the role. I have to yank out all the 1s and try all combos of precomp to see if there is still any such problem. 05:26
TimToady weird, if I put a module foo; at the front of my foo.pm, it warns on the last one now
so the whole file is treated as a block that might return a value, but if we switch into module, it enforces sink on the last statement 05:27
skids .tell japhb well, maybe read the backlog before doing what I just said. 05:29
yoleaux skids: I'll pass your message to japhb.
05:30 adu joined
skids m: (role A { }).HOW.WHAT.say; A.HOW.WHAT.say; # scratches head, wishes he had more time to read code. 05:50
camelia rakudo-moar f12020: OUTPUT«(Perl6::Metamodel::ParametricRoleHOW)␤(Perl6::Metamodel::ParametricRoleGroupHOW)␤»
06:02 Alina-malina joined 06:08 atroxaper joined 06:54 espadrine_ joined
japhb_ Thanks for taking a gander at my Role problem, skids. 07:11
07:11 japhb_ is now known as japhb
japhb .botsnack 07:11
yoleaux :D
04:58Z <skids> japhb: "The last statement in a module with more than one statement seems to have some extra-sinky context (because it supplies the return value of "use" I think). I just put 1; at the end of the module to prevent that."
05:29Z <skids> japhb: well, maybe read the backlog before doing what I just said.
moritz \o 07:16
07:30 adu left 07:32 darutoko joined, espadrine joined 07:33 espadrine_ left 07:36 xfix joined 07:52 rindolf joined
FROGGS[mobile] o/ 08:04
08:04 prime left
FROGGS[mobile] mornäng 08:04
08:05 Ben_Goldberg left, Ben_Goldberg joined 08:09 adu joined 08:11 prime joined 08:28 xinming_ left 08:29 xinming joined
moritz .ask jnthn is there a good reason to keep $!started in Proc::Async private? sounds like potentially useful information for the caller 08:37
yoleaux moritz: I'll pass your message to jnthn.
dalek ast: 7eb85d9 | usev6++ | S32-str/substr-rw.t:
Add test for RT #114526 (fudged 'skip' for Parrot)
 08:42
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=114526
dalek ast: f16e55e | usev6++ | S32-str/sprintf.t:
Fudge two tests (todo) since error message would change with NQP PR 219
 08:48
08:52 zby_home joined
dalek c: 1517511 | moritz++ | lib/Type/ (2 files):
Start to documtn Proc::Async
 08:54
09:05 agentzh joined, ptc_p6 joined 09:06 atroxaper left, telex left 09:08 telex joined
dalek p: 878c6ae | usev6++ | src/HLL/sprintf.nqp:
Change message for mismatch between specified arguments in directives and supplied arguments
 09:11
p: 6fa84cc | usev6++ | t/hll/06-sprintf.t:
Adjust error message for directives/arguements mismatch in nqp tests
p: 5a33762 | moritz++ | / (2 files):
Merge pull request #219 from usev6/master

Improve error message for sprintf
09:16 jakesyl left 09:17 jakesyl joined 09:22 nebuchad` is now known as nebuchadnezzar 09:23 vendethiel left 09:24 vendethiel joined 09:27 mr-foobar left 09:30 molaf__ joined 09:33 molaf_ left, Ben_Goldberg left 09:34 Ben_Goldberg joined 09:35 Ben_Goldberg left, Ben_Goldberg joined 09:36 Rounin joined 09:39 kaare__ is now known as kaare_ 09:46 vendethiel left 09:50 BinGOs_ is now known as BinGOs, zby_home left, BinGOs left, BinGOs joined 10:00 espadrine_ joined 10:01 virtualsue joined 10:04 espadrine left 10:09 vendethiel joined 10:11 bjz_ joined, bjz left 10:13 Ben_Goldberg left 10:19 gfldex joined 10:27 sz joined, mvuets joined, sz left 10:29 abraxxa joined 10:30 szabgab joined, breinbaa1 left
szabgab rakudo: say 19+23 10:30
camelia rakudo-{parrot,moar} f12020: OUTPUT«42␤»
masak good antenoon, #perl6 10:37
10:42 breinbaas joined 10:43 zby_home_ joined, zby_home_ left 10:49 agentzh left
moritz good *, masak 10:53
dalek c: 1c9eace | moritz++ | lib/Type/ (2 files):
Document X::Proc::Async::AlreadyStarted
 10:59
szabgab hello ppl 11:01
moritz \o szabgab
szabgab I have a really strange bug on the Perl6Maven site
\o moritz
beside it leaking memory like hell :)
once in a while it seems as if the regex engine stopped working 11:02
ab5tract_ m: my $x = '\E]45;65H'; say $x.substr(/45/,77)
camelia rakudo-moar f12020: OUTPUT«No such method '!cursor_start' for invocant of type 'Int'␤ in sub substr at src/gen/m-CORE.setting:7679␤ in method substr at src/gen/m-CORE.setting:4097␤ in block <unit> at /tmp/161RXUN10t:1␤␤»
moritz szabgab: are you ever using user-supplied strings in regexes?
11:03 Woodi left
szabgab every time someone visits the site, it loads a file, parses it using some regexes 11:03
do you mean the regex part?
I think no
let me check
but I finish the sentence first :)
moritz oh, you weren't
then go ahead :-)
ab5tract_ ah, never mind.. another confusion of subst and substr 11:04
szabgab so the code parses the header that looks like =title Title
but sometimes it seems it stops the parsing
and does not recognize the matching rows
then after the restart (I have to restart the server every 30 requests) it works again 11:05
so far I could not figure out what might trigger the behaviour
pastebot? 11:06
moritz none here
just use gist.github.com or so
szabgab gist.github.com/szabgab/35359cfe441fd0a9f9fb 11:07
those seem to be the only regexes 11:08
but that's just my code, maybe some of the modules I use...
Maybe the template module?
moritz which one do you use?
szabgab Template::Mojo 11:09
Eerything is Rakudo Star 2014.12 except Bailador which I've patched
I thin Bailador builds regexes on-the-fly 11:11
from the routes in the application
moritz but I hope those aren't user-input
szabgab If that's what you were after
not from the end-user
from the source code of Perl6::Maven 11:12
moritz but I think my question as mis-guided. I thought the regexes hangs, but it seems they simply don't match=
s/=/?/
szabgab right
they stop matching
and never resume matching
moritz can you identify the place where the first stop matching? 11:13
11:13 rurban_ joined
szabgab probably not 11:14
I might need to add lots of logging and that might help
moritz well, I haven't heard of this phenomenon before 11:15
szabgab OK, if there is nothing obvious to you, then I'll just add more code to be able to log the requests well
moritz it might or might not help
szabgab and then I might have better input
11:16 espadrine_ left
szabgab is there a logging module ? 11:16
11:17 spider-mario joined 11:20 isBEKaml joined
moritz I'm not aware of any 11:23
dalek kudo/nom: 6429a22 | moritz++ | src/core/Proc/Async.pm:
Proc::Async.write: accept Blob, not Buf
 11:24
szabgab ok, thanks 11:26
11:37 kjs_ joined 11:41 szabgab left
rurban_ got now msvc on win64 working 11:56
dalek kudo/nom: 7e8d8af | moritz++ | src/core/Proc/Async.pm:
Proc::Async: fix exception throwing
 12:00
12:11 rindolf left
dalek line-Python: c26bd6e | (Stefan Seifert)++ | .gitignore:
.gitignore file
 12:13
12:15 abraxxa left 12:18 jluis_ joined 12:19 Kristien joined 12:31 kjs_ left 12:38 rmgk_ joined, rmgk left, rmgk_ is now known as rmgk
Kristien hola #perl6 12:50
masak \o 12:54
vendethiel: which would you say was the first altJS? has someone written a history/chronicle about them? which altJS languages do you think will be in use/remembered 20 years down the line, and why? 12:55
12:56 rurban_ left 12:57 jluis_ left
vendethiel masak: GWT would probably be it 13:00
masak oh, right.
GWT is basically abandoned now, is it? 13:01
13:02 virtualsue left
vendethiel masak: I think gmail still uses it? 13:03
masak oh, mebbe.
vendethiel masak: sorry, apparently it isn't anymore. But it seems chrome web store and google groups use it groups.google.com/forum/#!msg/goog...WzIrZ1vzcJ 13:04
it was from 2011 though, so maybe not up to date
masak ah.
interesting that Gmail moved off it.
I wonder why.
vendethiel masak: I'm not *sure* it did. gwt's website doesn't mention it -- it mentions adwords and wallet 13:05
masak *nod*
vendethiel masak: GWT 2.7 on nov 14. I don't think it's been "abandoned" -- it's just that no one wants to write java
masak fair enough 13:07
isBEKaml masak: Gmail didn't abandon GWT - but AFAICR, they maintain their own version of GWT 13:10
masak: that was in addition to open source releases of GWT (by its own team) 13:11
masak ugh
13:11 jluis_ joined
isBEKaml yeah, as I read it back then - their use of GWT was simply too complex and it cannot be integrated into the main release. 13:11
so they had to maintain their own internal version of GWT 13:12
Kristien masak: Objective-C is also pretty old 13:20
I mean Objective-J
masak heard about it, but not more'n that 13:21
what is it?
Kristien It's Objective-C except with JavaScript instead of C.
masak but... hm. 13:22
JS already has an object system. and a syntax for method calls.
I guess I don't see the point.
Kristien Objective-J has a new separate object system xD
masak o.O 13:23
they must've not liked the first one...
moritz
.oO( how did that happen? :-)
 13:28
13:29 adu left
masak random thought: P6 is a bigger breaking change than py3k because Larry had more to atone for than Guido (language-wise) 13:30
ab5tract_ masak: that's an interesting perspective 13:34
masak maybe that's obvious, I dunno. the Perl 5 line didn't shirk from sinning if it got the job done.
Perl 6 aims to do better. Python somehow always did.
13:37 Kristien left
ab5tract_ for that generation of dynamic languages, it seems to me that python "won" in the sense that it runs the spectrum of teaching to sysadmin to web to science and back again 13:38
and i always take the POV that perl 6 caused some sort of irreperable harm to the reputation of perl 5 with a grain of salt 13:39
because it was pretty clear to me that people were leaving perl 5 to python at a pretty big clip even prior to the announcement 13:40
masak yes, Python sure has mindshare. 13:42
and deservedly, if you ask me. I say that as someone who loves Perl.
(and as one who isn't all that into indentation-as-blocks) 13:43
ab5tract_ it is almost certainly a retcon on my part, but i always took the design of python as directly inspired by perl, but in the sense that guido's design was in realizing the power of exploring a language where DWIM is reinforced by the fact that !(TMTOWTDI) 13:44
moritz as a relative beginner in python, I find that it seldom DWIM 13:45
but it usually gives an error instead of doing the wrong thing 13:46
ab5tract_ moritz: that's because when there is one way to do it, 'do what i mean' takes on a new dimension where you must 'mean what you do'
when TMTOWTDI, meaning what you do can come free-form, even retaining the jazz definition; the multiplicity is a double edged sword, though. that's why part of p5's continued popularity is at least partly a result of the "modern perl" movement, which advocated heavily for trimming down the cuteness and dark magic 13:51
13:56 isBEKaml left
ab5tract_ i'm a huge fan of TIMTOWTDI, but i've always largely attributed python's near-universal appeal (minus the haters-for-life) to guido's rejection of the principle 13:56
14:00 osfamero1 is now known as osfameron
ab5tract_ what makes perl 6 so appealing to me is that it really does feel like it has been carefully designed to host that multiplicity of 'ways to do it', while addressing things like the dangers of building a monkey patching culture by making it basically unnecessary *and* still provide support for it :) 14:02
*providing
dalek ast: 3101a04 | usev6++ | S11-modules/export.t:
Add test for RT #84280
 14:03
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=84280
dalek c: c9d67d3 | moritz++ | lib/Type/ (3 files):
Document a bit more of Proc::Aync, related exceptions
 14:06
14:08 Kristien joined
moritz lizmat: Proc::Async.say doesn't use .gist to stringify its argument. Is that intentential? 14:09
14:12 jluis_ left
moritz m: say $*KERNEL.signal('KILL') 14:16
camelia rakudo-moar 7e8d8a: OUTPUT«9␤»
moritz m: say $*KERNEL.signal('SIGKILL')
camelia rakudo-moar 7e8d8a: OUTPUT«9␤»
moritz m: say $*KERNEL.signal(SIGKILL)
camelia rakudo-moar 7e8d8a: OUTPUT«9␤»
moritz m: say $*KERNEL.signal(9)
camelia rakudo-moar 7e8d8a: OUTPUT«9␤»
14:17 jluis_ joined, Peter_R joined
Kristien m: say $*KERNEL.WHAT 14:18
camelia rakudo-moar 7e8d8a: OUTPUT«(Kernel)␤»
Kristien Interesting.
dalek c: 3b83fb1 | moritz++ | lib/Type/Proc/Async.pod:
mostly finish documenting Proc::Async
 14:22
14:22 rindolf joined
lizmat moritz: Proc::Async.say not gisting is probably an oversight 14:25
14:28 mr-foobar joined
dalek c: 83a1c78 | moritz++ | lib/Type/Proc/Async.pod:
Add a note that only MoarVM supports Proc::Async
 14:28
Kristien m: ObjAt.new 14:29
camelia ( no output )
Kristien Segfaults on my machine :P
moritz :(
14:30 sirdancealot joined
dalek ast: bc2983c | moritz++ | S17-procasync/basic.t:
start to test some exceptions for Proc::Async
 14:31
Kristien seems to be an old bug 14:34
dalek ast: e223c44 | moritz++ | S17-procasync/basic.t:
Proc::Async: test more exceptions
 14:35
Kristien Stuff like Set uses WHICH to determine equality right?
moritz identity 14:36
Kristien right
good
moritz (sorry for being pedantic) 14:37
Kristien And it should be perfectly fine to override WHICH for immutable user-defined types?
lizmat Kristien: yes, that's the idea 14:38
Kristien Nice.
Perl 6 is well-designed. 14:39
moritz can anybody on windows confirm that 'type' on the command line without arguments reads from STDIN?
or do I have to use 'copy con' or such weird magics? 14:40
lumimies moritz: It does not 14:43
type con: seems to 14:44
moritz lumimies: thanks 14:48
Kristien This is interesting: gist.github.com/rightfold/f05c4d4c31e09ca0d9bf 14:53
moritz Kristien: the REPL has known problems with natively typed variables
Kristien OK!
moritz Kristien: basically, it starts a new compilation unit for each line
Kristien: but to make the previous variables available, it copies them into an outer scope 14:54
Kristien: and that doesn't work well yet in some corner cases
Kristien I have to implement a REPL soon. 14:55
moritz +1 14:56
Kristien I think I have an idea of how I'll do it.
14:58 jmark joined 14:59 _4d47 joined
dalek ast: 67bb8b3 | moritz++ | S17-procasync/basic.t:
more tests for Proc::Async

in particular:
  * test .write and .print, open for writing
  * .close-stdin
  * more exceptions		 15:07
Kristien Just a reduction over the lines of stdin :P 15:08
vendethiel moritz: it doesn't either here :P 15:12
15:14 virtualsue joined
[Tux] what is the perl way to do «my Str $foo = function (); is_valid_utf8 ($foo) and "UTF8".say» ? 15:15
where full-ASCII would return False 15:16
15:16 rurban_ joined 15:18 Ben_Goldberg joined
moritz [Tux]: checking of a Str is valid utf-8 is mixing of levels 15:19
[Tux] ?
moritz [Tux]: Str is an encoding-indepenent sequence of codepoints (later hopefully graphemes)
Kristien > Perl 6 is well-designed. 15:20
[Tux] fine :) 15:21
Ben_Goldberg Somewhere hidden behind the Str is a Buf, which will contain some form of utf, but that's a separate level.
moritz [Tux]: what do you want to achieve in the end?
[Tux] So, how can I check id *all* codepoints in a string are inside ASCII?
moritz all($str.ords) < 128 # but probably not most efficient 15:22
[Tux] wants to copy as much as possible of Text::CSV_XS (perl5) to Text::CSV (perl6)
it is lazy, so I do not care about performance there
Kristien encode it to ASCII and check for encoding errors 15:23
that's what I'd do
I don't like hardcoding 128
moritz m: say so 'abc' ~~ /^ <[\x0 .. \x127]>* $ /
camelia rakudo-moar 7e8d8a: OUTPUT«True␤»
moritz m: say so 'abcö' ~~ /^ <[\x0 .. \x127]>* $ /
camelia rakudo-moar 7e8d8a: OUTPUT«True␤»
moritz huh
erm
m: say so 'abcö' ~~ /^ <[\x0 .. \x7F]>* $ / 15:24
camelia rakudo-moar 7e8d8a: OUTPUT«False␤»
[Tux] :)
moritz m: say so defined try 'abcö'.encode('ASCII')
camelia rakudo-moar 7e8d8a: OUTPUT«True␤»
moritz m: say 'ö'.encode('ASCII')
camelia rakudo-moar 7e8d8a: OUTPUT«Blob[uint8]:0x<3f>␤» 15:25
moritz that looks like a bug to me
ab5tract_ [&&] "6kgj".ords.map( * < 128 )
Kristien m: say 'ö'.encode('ASCII').decode('ASCII') 15:26
camelia rakudo-moar 7e8d8a: OUTPUT«?␤»
Kristien moritz: it uses a replacement character instead of throwing an exception
moritz Kristien: and that's a dangerous default, IMHO, because it silently loses information
Kristien Yes.
IMO there should be a way to parameterise on error handling mechanic. 15:27
moritz well, .encode('ASCII', :replacement) could work
ab5tract_ fwiw, i was banging my head on the wall trying to manipulate terminal escape sequences
Kristien I've seen one C++ library do enum ErrorHandling { Throw, Replacement, AssumeOK };
moritz AssumeOK seems pretty dangerous
ab5tract_ yes, adverbs seem like a good approach, and move the current default behavior into one
moritz but Perl 5's Encode has similar options, iirc
Kristien Yeah, it's UB if it's not actually OK 15:28
[Tux] forsees some interesting problems in decoding binary data :)
moritz and we try to avoid undefined behavior where we can
[Tux] images in CSV
Kristien moritz: good
:replacement might also take a parameter: $str.encode('ASCII', :replacement('x')) 15:29
or rather a uint8 not a Str
15:31 ingy1 is now known as ingy 15:32 Kristien left
lizmat seems like usev6 ( bartolin_ ??? ) has forgotten to add a RT84280.pm to roast 15:34
ab5tract_ one thing i've been bumping my head on is how to print an escape sequence once it is in a variable without it getting interpolated 15:35
moritz ab5tract_: because it's not UTF-8?
ab5tract_ i imagine it is related somehow. just printing the variable will "do" the cursror move command i have stored in the variable 15:36
and i can even extract the numbers with regex groups 15:37
moritz ab5tract_: oh, you want it for debugging?
ab5tract_ but so far my efforts to put numbers back in have been fruitless
exactly
moritz say $str.perl; # ?
15:37 zakharyas joined
ab5tract_ damnit moritz, you genius :) 15:38
dalek ast: d591ae5 | moritz++ | S17-procasync/basic.t:
More Proc::Async tests
moritz it would be very welcome if somebody could run S17-procasync/basic.t on windows 15:39
and inform me of any failures
ab5tract_ i don't suppose there is a built in character class for matching escape sequences :) 15:40
moritz iirc there's <print> for printable characters
so something not-<print>able could be an escape sequence
bartolin_ lizmat: oops, you're right. 15:44
ab5tract_ m: "\n" !~~ /<print>/
camelia rakudo-moar 7e8d8a: OUTPUT«Cursor<139843422512584>Cursor<139843422562104>P6opaque: no such attribute '$!pos'␤ in method ACCEPTS at src/gen/m-CORE.setting:14610␤ in block <unit> at /tmp/ASa6x7LUws:1␤␤»
ab5tract_ same error for single-quotes 15:45
dalek kudo/nom: 5dd603b | lizmat++ | src/core/Proc/Async.pm:
Make Proc::Async.say gist, spotted by moritz++
 15:46
FROGGS[mobile] <print> is either NYI or it picks up the print sub
dalek ast: e8465cb | usev6++ | packages/RT84280.pm:
Add missing file for test for RT #84280
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=84280
bartolin_ lizmat++ # spotting the missing file 15:47
FROGGS[mobile] ab5tract_: matching escape sequences is not quite possible when these are interpolated :o)
but you can match a backslash and something afterwards, according to what you wanna see 15:48
ab5tract_ m: "for" ~~ /<print>/
camelia rakudo-moar 7e8d8a: OUTPUT«Cursor<140112800721080>Cursor<140112800770600>P6opaque: no such attribute '$!pos'␤ in method ACCEPTS at src/gen/m-CORE.setting:14610␤ in block <unit> at /tmp/BXPDVP2xWT:1␤␤»
FROGGS[mobile] m: "foo" ~~ /<say>/ 15:49
camelia rakudo-moar 7e8d8a: OUTPUT«Cursor.new()␤Cursor.new()␤P6opaque: no such attribute '$!pos'␤ in method ACCEPTS at src/gen/m-CORE.setting:14610␤ in block <unit> at /tmp/9ocu5YoTtd:1␤␤»
ab5tract_ FROGGS[mobile]: ah.. i am seeing a way through now, actually. but yeah, something there doesn't seem happy 15:50
15:51 Kristien joined 15:53 pnu left 15:54 ggherdov left, Ben_Goldberg left
[Tux] «has %errors = (1 => "foo", 2 => "bar");» can that be typed? «has %errors (int => Str) = (1 => "foo", 2 => "bar");^? 15:54
15:54 pnu joined, Ben_Goldberg joined
masak [Tux]: `has Str %.errors{Int};`, I think. 15:56
FROGGS_ m: my $errors = :{ 1 => "foo", 2 => "bar" }; say $errors.perl
camelia rakudo-moar 7e8d8a: OUTPUT«:{1 => "foo", 2 => "bar"}␤»
FROGGS_ m: my $errors = :{ 1 => "foo", 2 => "bar" }; say $errors.keys[0].WHAT
camelia rakudo-moar 7e8d8a: OUTPUT«(Int)␤»
FROGGS_ m: my %errors{Int} = 1 => "foo", 2 => "bar"; say %errors.keys[0].WHAT
camelia rakudo-moar 7e8d8a: OUTPUT«(Int)␤»
Kristien m: class C { }; my %errors{C} = C.new => 'a', C.new => 'b'; say %errors.keys[0].WHAT 15:57
camelia rakudo-moar 7e8d8a: OUTPUT«(C)␤»
Kristien m: class C { }; my %errors{C} = C.new => 'a', C.new => 'b'; say %errors.WHAT
camelia rakudo-moar 7e8d8a: OUTPUT«(Hash[Any,C])␤»
Kristien Shiny, I thought hashes worked only with string keys.
FROGGS_ m: say "a" => b => c 15:58
camelia rakudo-moar 7e8d8a: OUTPUT«===SORRY!=== Error while compiling /tmp/qmGmFUa_GB␤Undeclared routine:␤ c used at line 1␤␤»
FROGGS_ m: say "a" => b => "c"
camelia rakudo-moar 7e8d8a: OUTPUT«a => b => c␤»
FROGGS_ m: say ("a" => b => "c").keys
camelia rakudo-moar 7e8d8a: OUTPUT«a␤»
FROGGS_ m: say (("a" => b) => "c").keys
camelia rakudo-moar 7e8d8a: OUTPUT«===SORRY!=== Error while compiling /tmp/Sn981WiJk2␤Undeclared routine:␤ b used at line 1␤␤»
FROGGS_ m: say (("a" => "b") => "c").keys
camelia rakudo-moar 7e8d8a: OUTPUT«a => b␤»
FROGGS_ a pair as a key...
everything can be a key nowadays :o) 15:59
so you don't have to serialize something to be used as a key
masak pairs are not good keys, though.
they're mutable. 16:00
m: my $p = a => 5; $p.value = 42; say $p 16:01
camelia rakudo-moar 7e8d8a: OUTPUT«a => 42␤»
Kristien m: say (("a" => "b") => "c")[("a" => "b")]
camelia rakudo-moar 7e8d8a: OUTPUT«No such method 'Int' for invocant of type 'Pair'␤ in sub postcircumfix:<[ ]> at src/gen/m-CORE.setting:2869␤ in block <unit> at /tmp/Sam3vOxEkV:1␤␤»
16:01 ggherdov joined
Kristien oops lol 16:01
m: say (("a" => "b") => "c"){("a" => "b")}
camelia rakudo-moar 7e8d8a: OUTPUT«c␤»
Kristien If pairs are mutable I don't agree with this behavior.
m: say (("a" => "b") => "c").WHAT 16:02
camelia rakudo-moar 7e8d8a: OUTPUT«(Pair)␤»
Kristien Hmm.
16:02 Ben_Goldberg left
Kristien Oh right. I'm a fool. 16:02
[Tux] int is not allowed, Int is
16:02 Ben_Goldberg joined
[Tux] m: Str %error{Int} = :{ 1 => "foo" }; 16:03
camelia rakudo-moar 7e8d8a: OUTPUT«===SORRY!=== Error while compiling /tmp/QgY1kjMr_W␤Undeclared routine:␤ error used at line 1␤␤»
[Tux] m: my Str %error{Int} = :{ 1 => "foo" };
camelia rakudo-moar 7e8d8a: OUTPUT«Saw 1 call to deprecated code during execution.␤================================================================================␤%error = itemized hash called at:␤ /tmp/9_zmuVNxtd, line 1␤Deprecated since v2014.7, will be removed with release v201…»
16:03 ptc_p6 is now known as [ptc]
[Tux] m: my Str %error{int} = :{ 1 => "foo" }; 16:03
camelia rakudo-moar 7e8d8a: OUTPUT«Type check failed in binding key; expected 'int' but got 'Int'␤ in method STORE_AT_KEY at src/gen/m-CORE.setting:10768␤ in block at src/gen/m-CORE.setting:10398␤ in method STORE at src/gen/m-CORE.setting:10391␤ in block <unit> at /tmp/LAnR3Muz…»
16:05 Ben_Goldberg left
Kristien m: my $k = ("a" => "b"); say $k.WHAT; my $h = {("a" => "b") => "c"}; say $h.WHAT; say $h{$k} 16:05
camelia rakudo-moar 7e8d8a: OUTPUT«(Pair)␤(Hash)␤c␤»
16:05 Ben_Goldberg joined
Kristien m: my $x = 'a' => 'b'; my $y = $x; $x.value = 'c'; say $x, $y; 16:06
camelia rakudo-moar 7e8d8a: OUTPUT«a => ca => c␤»
Kristien m: my $x = 'a' => 'b'; my $y = 'a' => 'c'; $x.value = 'c'; say $x === $y;
camelia rakudo-moar 7e8d8a: OUTPUT«False␤»
16:07 [ptc] left
Kristien Does Hash not use WHICH for keys? 16:07
16:08 [ptc] joined
skids Default Hash is Str, Hash[Any] should use WHICH IIRC 16:12
Kristien m: subset Nat of Int where { * >= 0 }; say -1 ~~ Nat
camelia rakudo-moar 7e8d8a: OUTPUT«True␤»
16:12 Ben_Goldberg left 16:13 Ben_Goldberg joined
Kristien weird 16:14
16:14 [ptc] left
[Tux] Saw 1 call to deprecated code during execution. 16:14
================================================================================
%!errors = itemized hash called at:
Deprecated since v2014.7, will be removed with release v2015.7! 16:15
Please use %!errors = %(itemized hash) instead.
masak [Tux]: did you use {} ?
Kristien m: say :{('a'=>'b') => 1, ('a'=>'b') => 2}
camelia rakudo-moar 7e8d8a: OUTPUT«(a => b) => 1, (a => b) => 2␤»
Kristien So far so good.
skids m: subset Nat of Int where * >= 0; say -1 ~~ Nat 16:16
camelia rakudo-moar 7e8d8a: OUTPUT«False␤»
[Tux] I used «has Str %!errors{Int} = :{ 1 => "foo" };
Kristien m: say :{('a'=>'b') => 1, ('a'=>'b') => 2}{('a' => 'b')}
camelia rakudo-moar 7e8d8a: OUTPUT«(Any)␤»
16:16 [ptc] joined
Kristien ah, cool 16:16
skids { * => 0 } is a closure that returns a closure.
Kristien skids: aahh I ee
I see*
16:17 kaare_ left
[Tux] m: my Str %e{Int} = :{ 1 => "foo" }; 16:18
camelia rakudo-moar 7e8d8a: OUTPUT«Saw 1 call to deprecated code during execution.␤================================================================================␤%e = itemized hash called at:␤ /tmp/m5E55mLkbx, line 1␤Deprecated since v2014.7, will be removed with release v2015.7!…»
FROGGS_ m: my Str %e{Int} = 1 => "foo";
camelia ( no output )
FROGGS_ that is redundant when you supply the shape
Kristien Time for dinner. Byebye!
FROGGS_ o/
16:18 Kristien left 16:21 spider-mario left 16:22 [ptc] left
[Tux] FROGGS_++ 16:23
masak yeah, I suspected there was one hash too many in there. 16:25
16:26 jluis_ left
[Tux] so «warn "foo\n"» does not prevent noting the location of the warn as it does in perl5 16:30
16:30 [ptc] joined, [ptc] left
[Tux] does perl6 have a warn that prints to catchable STDERR without noting the location? 16:31
dalek c: bacbf30 | moritz++ | lib/Type/Proc/Async.pod:
Proc::Async: say, write and print return promises
16:31 [ptc] joined
moritz [Tux]: note() prints to STDERR 16:31
but it's not an exception like warn
16:31 [ptc] left
dalek c: b60db22 | moritz++ | lib/Type/Promise.pod:
Try to fix formatting of a code block
 16:32
16:32 [ptc] joined 16:33 [ptc] left 16:34 [ptc] joined
[Tux] bummer 16:34
skids So when browsing doc/ via github, the reaon some files do not display is because github renders with perl5 pod, right?
moritz skids: correct
doc.perl6.org should display them correctly, though :-) 16:35
dalek c: bc676ad | moritz++ | README.md:
README: add a note that github mis-displays the pod files
 16:36
skids Any way to just get github to do plaintext for .pod? 16:39
moritz no idea
rurban_ name it pod6
skids Hmm....
rurban_ they won't interfer with this perl5 vs perl6 fight who got the better/right pod format 16:40
moritz I see no fight
skids I was thinking more a project-local setting.
16:41 Kristien joined
Kristien my %x = :{ ('a' => 'b') => 'c' }; say %x{('a' => 'b')}; 16:42
m: my %x = :{ ('a' => 'b') => 'c' }; say %x{('a' => 'b')};
camelia rakudo-moar 5dd603: OUTPUT«c␤Saw 1 call to deprecated code during execution.␤================================================================================␤%x = itemized hash called at:␤ /tmp/3YhNbc4pTr, line 1␤Deprecated since v2014.7, will be removed with release v201…»
Kristien m: my $x = :{ ('a' => 'b') => 'c' }; say $x{('a' => 'b')}; 16:43
camelia rakudo-moar 5dd603: OUTPUT«(Any)␤»
Kristien Ok, good.
16:44 [ptc] left, [ptc] joined 16:45 [ptc] left
Kristien m: class A { has $.x = 1; submethod BUILD { } }; say A.new.x; 16:48
camelia rakudo-moar 5dd603: OUTPUT«1␤»
[Tux] what is the most readable/most common way to write qq{"abc\nc"} in perl6? I keep getting confused with all Q and <> variations 16:50
moritz qq{"abc\nc"} is good, IMHO 16:51
[Tux] it sees the \n not as newline /
moritz oh
you don't want that?
q{"abc\nc"}
16:51 Peter_R left
[Tux] I want an embedded newline, as in perl5 16:51
moritz m: say q{"abc\nc" 16:52
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!=== Error while compiling /tmp/GYGKJtq0R7␤Couldn't find terminator }␤at /tmp/GYGKJtq0R7:1␤------> say q{"abc\nc"⏏<EOL>␤»
moritz m: say q{"abc\nc"}
camelia rakudo-moar 5dd603: OUTPUT«"abc\nc"␤»
moritz m: say qq{"abc\nc"}
camelia rakudo-moar 5dd603: OUTPUT«"abc␤c"␤»
moritz that one expand the newlne
[Tux] weeeeeeeird
16:52 [ptc] joined
[Tux] looks for another cause then 16:52
dalek c: 7c4d12d | moritz++ | lib/Type/Proc/Async.pod:
More formatting fixes
 16:53
Kristien m: my %h = 1, 2, 3, 4; say %h.WHAT; say map({ .WHAT }, %h.keys); 16:55
camelia rakudo-moar 5dd603: OUTPUT«(Hash)␤(Str) (Str)␤»
Kristien :S
moritz fwiw I've now changed the doc.perl6.org updating cron job to run every 5 minutes, but only if it's not running (via flock (1))
and it only does something if there's a new commit
Kristien m: my %h{Any} = 1, 2, 3, 4; say %h.WHAT; say map({ .WHAT }, %h.keys);
camelia rakudo-moar 5dd603: OUTPUT«(Hash[Any,Any])␤(Int) (Int)␤»
Kristien Ahh, it defaults to Str. 16:56
moritz yes
Kristien Why?
moritz p5 legacy (and often still very useful behavior)
Kristien OK. 16:57
dalek c: 0afbbe8 | moritz++ | lib/Type/Any.pod:
A bit more documentation for class Any
 17:03
[ptc] moritz: hi! Have you got time for a couple of questions? 17:04
moritz [ptc]: yes (though $family might demand some attention)
[ptc] cool! (and if you need to go, no worries) 17:05
um, I was looking at #28 on the docs repo; the one about syntax highlighting the code on the website
and was looking at what people had already mentioned, and found SyntaxHighlighter somewhat wanting in functionality 17:06
namely, it doesn't handle angle brackets nicely...
anyway, I noticed that pygments (pygments.org) handles Perl6.
moritz as does vim
[ptc] would it be ok to investigate that as an option for docs.perl.org for the syntax highlighting/ 17:07
moritz the problem with not doing it in the browser is that Perl 6 pod allows interweaving of code with markup
you can do things like
=being code :allow<B L>
and then B<...> and L<...> tags in the code produce bold/link HTML 17:08
but the rest is taken literally
[ptc] aha
moritz and it's hard to find hilighters that support that
17:08 TimToady_ joined
moritz I guess I should add that to the ticket 17:08
17:09 ab5tract joined
[ptc] dunno if pygments does that; they even advertise that they support perl6 :-) 17:09
17:09 krunen joined
[ptc] even though it's in python... 17:09
flussence the pygments support was adaded to get perl6 to look nice on github (which then promptly dropped pygments entirely, for an in-house library with half the functionality...)
17:09 xfix_ joined
flussence s/adaded/added/ 17:09
17:10 risou_ joined, Kristien_ joined, petercom1and joined
[ptc] I just wondered if it's worthwhile investigating using pygments instead of something like highlight.js or SyntaxHighlight for docs.perl.org 17:10
the problem I see is that one probably has to install pygments on the server, which might not be an option... 17:11
just trying to throw ideas around...
moritz [ptc]: installing stuff on the server is an option
17:11 huf joined
flussence there's a high chance it's already there as part of some other dep... 17:11
moritz [ptc]: I have root there :-)
flussence (asciidoc uses it, and plenty of things use that) 17:12
moritz [ptc]: it just shouldn't be something obscure which I end up having to maintain that
if it's debian-packaged, that makes it easier (but it's not a hard requirement)
Kristien_ m: say [1, 2, 3].perl, set(1, 2, 3).perl
camelia rakudo-moar 5dd603: OUTPUT«[1, 2, 3]set(1,2,3)␤»
[ptc] moritz: which was one thing that sort of made me a bit uneasy when looking at the other solutions
Kristien_ Inconsistent whitespace! :P 17:13
17:13 rmgk_ joined, rmgk left, rmgk_ is now known as rmgk
moritz [ptc]: understandable 17:13
[ptc] moritz: it's in debian
moritz python-pygments, python3-pygments
[ptc] I'll look into it a bit more before you need to install anything
I'll push a feature branch when I'm ready and ping you so for feedback if that's ok 17:14
moritz [ptc]: sure
[ptc] moritz: cool, ta :-) 17:15
moritz [ptc]: fwiw on my blog I use vim + Text::VimColor for hilighting Perl 6 code
17:15 lestrrat left, jakesyl left, krunen_ left, ab5tract_ left, risou left, Kristien left
moritz [ptc]: but I'm not married to that solution by any means 17:15
17:15 jakesyl joined, xfix left, Alina-malina left, huf_ left, kst left, TimToady left
[ptc] hrm, that might be a better option; didn't know about Text::VimColor 17:17
that might be easier, and a Perl-only solution
moritz I've considered porting Text::VimColor to Perl 6 17:18
but haven't done it yet
[ptc] I'll play with a couple of options, and then recommend something; I'd like to keep the admin-overhead as low as possible, and not have to use too many languages in building the html docs 17:19
flussence heh, I just saw the pygments homepage... :) 17:20
(that list)
17:33 IllvilJa left 17:34 jantore left, jantore joined
TimToady_ it's quite possible that szabgab's bug is the one we flushed out of spesh a week or two ago 17:37
that one that involved regex and negations
17:38 IllvilJa joined 17:39 espadrine_ joined 17:42 ugexe_ is now known as ugexe 17:43 sirdancealot left 17:46 sirdancealot joined 17:50 Vlavv_ left 17:55 xfix_ is now known as xfix 17:56 zakharyas left 17:59 telex left 18:00 telex joined 18:02 Vlavv_ joined
dalek c: 13001a6 | paultcochrane++ | lib/Language/operators.pod:
Add missing prepositon
 18:04
c: 0b382b6 | paultcochrane++ | lib/Type/Any.pod:
Merge branch 'master' of github.com:perl6/doc
18:06 espadrine_ left
timotimo could we spec UNDO phasers inside grammars/regexes to fire if we backtrack over a closure? 18:16
18:19 Alina-malina joined
moritz isn'T that already specced 18:26
Kristien_ If you use map in sink context, does it produce a list or is that optimised out?
moritz Kristien_: I think it's optimized out
timotimo dunno 18:31
18:33 FROGGS__ joined 18:34 Sqirrel left 18:35 sirdancealot left 18:36 Sqirrel joined, lestrrat joined 18:37 FROGGS_ left
TimToady_ it's treated as eager 18:43
but it doesn't have to actually return any values
part of the GLR is to make sure more things optimize that 18:44
it does have to run the block though, since it might have side effects
I think P5 goes as far as to put the inside of the map's block into void context 18:45
[ptc] is it ok if I add '# vim: ft=perl6' at the bottom of files I come across, so that at least vim knows the file is perl6? 18:56
Kristien_ Install a proper ft detection plugin.
nwc10 use emacs? :-) 18:57
(couldn't resist. I'm not sure if emacs does it any better.)
[ptc] ok. Just thought I'd ask before I start changing files willy-nilly 18:58
Kristien_: what would you recommend? 18:59
Kristien_ Or make the file extension pm6 instead of pm.
bartolin_ fwiw: most (if not all) test files from roast have '# vim: ft=perl6' as last line 19:01
TimToady_ [ptc]: you should generally also add et to expand tabs to spaces
expandtab and sw=4
[ptc] bartolin_: yeah, that's one reason why I ask; I've seen files with, and several without 19:02
and have had the problem that some files aren't detected correctly as perl6
TimToady_: that's a good idea, I'll do that where there already is a vim coda
19:03 rindolf left
TimToady_ a lot of rakudo files have vim modelines, but not many nqp files yet 19:04
19:04 TimToady_ is now known as TimToady
bartolin_ +1 for adding vim modelines 19:04
itz_ perl6 vim tags generation would be nice too 19:06
19:12 Mouq joined
japhb While glancing through the backlog, I saw a discussion about what encoding high-codepoint strings to ASCII should do. Did anyone suggest it returning Failure? That seems the obvious choice for dial-able complaint level .... 19:12
geekosaur that was the suggested default 19:13
hm, no, that was an exception
Failure might work
Kristien_ should be configurable though 19:14
japhb Kristien_: 'use fatal'
Kristien_ no I mean whether it should fail or use a replacement character 19:15
Mouq Also, would it make sense to remove accents? or is that outside the scope of .encode?
japhb Kristien_: now *that* could be an adverb 19:16
Kristien_ "class FatRat is Cool does Rational[Int, Int]" reads funny
japhb Mouq: I think that's outside the .encode scope -- that's more of a mapping than an encoding
(Insomuch as you expect .encode to be reversible) 19:17
Also, transliteration seems to me a very non-core thing, ripe for modules 19:18
(Because it would likely have to be language-dependent.)
19:19 KCL_ joined
japhb And the core setting provides mechanism and tools, not policy. 19:19
bartolin_ when I looked at the nqp code for sprintf I found that we are very strict (using nqp::die) when the number of used arguments in directives and the number of supplied arguments are different. on the other hand we are much less strict when at least one index is used. (examples follow)
m: say sprintf "%s", "a", "b", "c"; # here we die 19:20
camelia rakudo-moar 5dd603: OUTPUT«Too few directives: found 1, fewer than the 3 arguments after the format string␤␤»
bartolin_ m: say sprintf "%1$s", "a", "b", "c"; # but here not
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!=== Error while compiling /tmp/7vj5HvGMVt␤Variable '$s' is not declared␤at /tmp/7vj5HvGMVt:1␤------> say sprintf "%1$s⏏", "a", "b", "c"; # but here not␤ expecting any of:␤ postfix␤»
bartolin_ m: say sprintf q[%1$s], "a", "b", "c"; # but here not
camelia rakudo-moar 5dd603: OUTPUT«a␤»
bartolin_ m: say sprintf q[%5$s], "a", "b", "c"; # and here neither
camelia rakudo-moar 5dd603: OUTPUT«␤»
bartolin_ I was surprised about the latter two 19:21
japhb bartolin_: I would think that's most likely a NYI, just because harder. 19:24
Mouq moritz: I saw you were trying to parellelize doc/htmlify.p6. I have an old branch sitting around that converts all the processing into safe actions on Supplies. It still failed when I last tried to add start { } blocks, but I'll push the branch to github 19:26
bartolin_ japhb: yes, that's plausible. I wanted to make sure there is no other reason
19:33 Rounin left
Kristien_ I think Rat.nude is quite an appropriate name. 19:33
It removes the wrapper, making the numbers nude :P
dalek c/supplied: 8d8fd30 | Mouq++ | / (3 files):
WIP...
 19:39
c/supplied: 5c25606 | Mouq++ | / (188 files):
Merge branch 'master' into supplied

Conflicts: htmlify.p6
rl6-roast-data: a7f253b | coke++ | / (5 files):
today (automated commit)
 19:40
19:41 agentzh joined
dalek kudo/nom: 243c164 | lizmat++ | src/core/Any.pm:
Remove now superfluous test for < 0
 19:44
japhb clears the last crashing bug out of a chunk of code he's been fighting for weeks, has a successful run, and has one of those "OMG -- I can't believe that actually WORKED!" moments
lizmat japhb++ 19:46
Kristien_ m: for 42 { .say } 19:47
camelia rakudo-moar 5dd603: OUTPUT«42␤»
Kristien_ It'll be fun to implement something like Scala's case class feature.
dalek Heuristic branch merge: pushed 31 commits to specs/newio by lizmat 19:48
19:49 jmark left
vendethiel Kristien_: well, "sealed" would be more interesting to implement :) 19:49
Kristien_: see timotimo's adt.p6
Kristien_ I don't like adt.p6 in that it does parsing stuff.
vendethiel :( 19:51
Kristien_ But having immutable records with automatic .perl and .WHICH is very nice.
moritz Mouq++ 19:53
dalek c/supplied: ada0d13 | Mouq++ | / (2 files):
Fix for latest Rakudo
 19:55
Kristien_ Maybe it can be achieved with just regular class syntax and some extra methods to generate construction, .perl and .WHICH 19:56
m: sub record($class) { $class.^add_method('perl', my method perl() { "TODO" }); }; class A { record(A) }; say A.new.perl 19:59
camelia rakudo-moar 5dd603: OUTPUT«A.new()␤»
Kristien_ Something like that.
Except without the bug, lol.
compose seems to override perl 20:00
moritz m: sub record($class) { $class.^add_method('perl', anon method perl() { 'TODO' }) }; class A { BEGIN record(A) }; say A.new.perl
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!===␤Type check failed in binding $class; expected 'Any' but got 'A'␤»
moritz m: sub record(Mu $class) { $class.^add_method('perl', anon method perl() { 'TODO' }) }; class A { BEGIN record(A) }; say A.new.perl
camelia rakudo-moar 5dd603: OUTPUT«TODO␤»
_4d47 just finished my first module, i'm happy have to share github.com/4d47/Tag 20:01
Kristien_ ah cool
moritz: nice lemme expand on this stuff
20:01 haxmeister joined
moritz _4d47: nice 20:02
Kristien_ btw why is it add_method not add-method? 20:03
moritz _4d47: see doc.perl6.org/language/modules for how to get it into the ecosystem
Kristien_: because jnthn++ came up with all the meta model :-) 20:04
Kristien_: there was a time when we said that low-level stuff should have _ in the name and high-level stuff -
Kristien_ ah OK
masak oh, we've stopped saying that? pity.
Mouq *low_level
moritz Kristien_: that hasn't really panned out, but we don't have too much consistency as the result :(
masak: I'm not sure
vendethiel Kristien_: seems like we could add another declarator instead of "class". like "record" 20:05
and you'd do `record Foo { has $!a; has !b; }`. or maybe `is immutable` on the class
moritz m: multi trait_mod:<is>(Mu $class, :$record!) { say $class.^add_method('marker', anon method() { 'MARKED' } }; class Flurb is record { }; say Flurb.marker 20:07
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!===␤Type 'method' is not declared. Did you mean 'Method'?␤at /tmp/qZfE1aWWFO:1␤------> $class.^add_method('marker', anon method⏏() { 'MARKED' } }; class Flurb is record␤Malformed anon␤at /tmp/qZfE1a…»
moritz m: multi trait_mod:<is>(Mu $class, :$record!) { say $class.^add_method('marker', anon method () { 'MARKED' } }; class Flurb is record { }; say Flurb.marker
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!=== Error while compiling /tmp/_8vwqkKrmN␤Unable to parse expression in argument list; couldn't find final ')' ␤at /tmp/_8vwqkKrmN:1␤------> d('marker', anon method () { 'MARKED' } ⏏}; class Flurb is record…»
moritz m: multi trait_mod:<is>(Mu $class, :$record!) { say $class.^add_method('marker', anon method () { 'MARKED' } ) }; class Flurb is record { }; say Flurb.marker
camelia rakudo-moar 5dd603: OUTPUT«<anon>␤MARKED␤»
moritz (minus the first say)
vendethiel m: class MetamodelX::RecordHOW is Metamodel::ClassHOW { }; my package EXPORTHOW { package DECLARE { constant record = MetamodelX::RecordHOW; } }; record X {}; 20:08
camelia rakudo-moar 5dd603: OUTPUT«===SORRY!=== Error while compiling /tmp/4QOLDnNaGt␤Unexpected block in infix position (two terms in a row, or previous statement missing semicolon?)␤at /tmp/4QOLDnNaGt:1␤------> = MetamodelX::RecordHOW; } }; record X ⏏…»
Kristien_ gist.github.com/rightfold/9d119b8ef0aa5e284b54 20:09
awesome :)
dalek c: de93815 | moritz++ | lib/Type/Mu.pod:
Small improvements to the Mu docs
c: 2148faf | moritz++ | lib/Type/Proc/Async.pod:
Proc::Async.say now calls .gist, just like regular &say. lizmat++
Kristien_ Now only new with positional agruments, and WHICH
and copy to create a new instance with some attributes replaced, e.g. $person.copy(:name<vendethiel>) 20:10
moritz Kristien_: .clone already does that
Kristien_ holy shit! 20:11
moritz Kristien_: doc.perl6.org/routine/clone
Kristien_ oh wait, perl is already there 20:12
only WHICH, then 20:13
moritz isn't it funny when you start to extend stuff, and find it's already there? :-)
Kristien_ yeah 20:14
can't get it to prevent compose from adding WHICH though
moritz call record at BEGIN time
or use a trait, which automatically runs ASAP
irclog.perlgeek.de/perl6/2015-01-25#i_10002270 20:15
Kristien_ oh neat
masak semi-cogent thought about 'is parsed': someone suggested the other day that 'is parsed' be further separated from the macro itself, and the two communicate over an interface or something. 20:16
vendethiel Kristien_: or create a declarator: github.com/jnthn/oo-monitors/blob/...onitors.pm :P
masak but if we look inside the compiler, regex and code are not just combined into one, they are *interleaved*. 20:17
Kristien_ vendethiel: is that the correct link?
vendethiel Kristien_: yes, it's exporting the "monitor" keyword, see at the bottom
Kristien_ oh right
shiny
lizmat m: my $h = :{}; my $p = (a => 42); $h{$p} = 63; say $h{$p}:p; $p.value = 17; say $h{$p}:p; say $h.pairs # is this intended / desired bahaviour ?
camelia rakudo-moar 5dd603: OUTPUT«(a => 42) => 63␤(a => 17) => 63␤(a => 17) => 63␤»
vendethiel Kristien_: github.com/jnthn/oo-monitors/blob/.../t/basic.t then in the test he uses "monitor Classname" 20:18
masak (sometimes I feel that the factoring with actions is a wee bit overgeneralized. I get the "you can have N action classes to 1 grammar!" advantage, but normally I just have 1-to-1.)
vendethiel masak: I think is parsed's syntax should really look like sweetjs :-)
(which is just racket templates)
masak regex and code are interleaved in the sense that we really like to check consistency ASAP, which often means right in the middle of a parse. 20:19
dalek kudo/nom: 403b0f0 | moritz++ | src/core/Proc/Async.pm:
Fix wording of error message

at least I do not think "process" made sense here, when it is about handles/streams		 20:20
lizmat moritz: but but, it's "Proc" async ? 20:21
moritz lizmat: I don't understand your objection 20:22
lizmat IO::Async has a similar error, no? 20:23
IO::Socket::Async I mean
moritz lizmat: haven't document that one yet :-)
so I don't know
lizmat hmmm... doesn't seem to have that check :-( 20:24
it probably should
moritz does the "process" in "Can only tap one of chars or bytes supply for process stdout" make sense?
lizmat and then maybe it's important to see in the message whether it was a process or a socket
moritz stdout is not a process
oh 20:25
but IO::Socke::Async won't have stdout/stderr, will it? 20:26
lizmat it has a chars_supply/bytes_supply
and I also guess we don't want to do both at the same time there either 20:27
moritz yes, but my point is that the error message won't be amibguous, because it doesn't say "stdout" or "stderr" in the error message
lizmat ok :-) 20:28
dalek c: d93a043 | moritz++ | lib/Type/ (2 files):
Document X::Proc::Async::CharsOrBytes
moritz Proc::Async is about 280 lines of code (including the exception classes), and it took me basically all the hacking time of the weekend to document it 20:29
now I think the documentation is complete, in the sense that it covers all features
Kristien_ K.
Mouq masak: macro foo ($a, $b) { #`[code goes here] }; { parsed foo { <.sym> <.ws> (<.num>)(\S+) }; foo 4inches }; parsed foo { (<.num>) <.ws> in unit <.ws> (<.EXPR>) }; 4 in unit "inches"
20:29 robins left
moritz Mouq: you skipped the interesting part: what does the code look like 20:31
20:31 jmark joined, jmark left, robinsmidsrod joined
moritz 'cause it will have to use the captures from parse to be useful 20:31
Kristien_ vendethiel: this works: gist.github.com/rightfold/fefad5f66e7654e1a0e5 20:32
I only have to fix WHICH to return some proper unique value but I can't do that now since ObjAt.new segfaults
moritz I guess we should fix that then :-) 20:33
masak ok, the easiest way to make 'is parsed' an code interleave I guess is just using the code block feature of regexes.
and* 20:34
Mouq moritz: True... :)
moritz m: sub f(str $x) { say $x }; my Str $x = 'foo'; f $x
camelia rakudo-moar 5dd603: OUTPUT«foo␤»
masak yeah, putting the parse after the macro code is surely the wrong way forward. 20:35
20:35 darutoko left
moritz
.oO( why did I just want to write 'tmux attack'? )
 20:35
Mouq m: sub infix:<in unit> (|) { "test" }; say 4 in unit "meters" # nbsp to the rescue
camelia rakudo-moar 243c16: OUTPUT«test␤»
moritz that's supposed to not work 20:36
Mouq Oh, wait.. it works with regular space too
moritz operators my not contain whitespace
Mouq Yeah, that's not so good
dalek kudo/nom: cb5bb33 | moritz++ | src/core/ObjAt.pm:
Make ObjAt.new(str) work

  Kristien_++ for reporting it as not working		 20:38
moritz Kristien_: there you go
Kristien_ shiny, thanks 20:39
will try it tomorrow, no time to reinstall rakudo right now
moritz Kristien_: have the appropriate amount of fun!
20:40 brrt joined
vendethiel Kristien_: nice :-) 20:40
Kristien_ vendethiel: I noticed your comment on GH :P 20:41
20:41 _4d47 left
Kristien_ The overview section is just to make fun of CoffeeScript, though. :P 20:42
20:42 ggherdov left, pnu left
dalek ast: 7d17de6 | moritz++ | S32-str/sprintf.t:
unfudge now passing test, bartolin++
 20:44
20:47 kst joined 20:48 brrt left
bartolin_ moritz: won't that require a nqp bump to work? 20:48
moritz bartolin_: it passed for me, and I'm on the recommended nqp version 20:49
(I think :-)
$ ./install/bin/nqp-m --version 20:50
This is nqp version 2015.01-2-g4589d81 built on MoarVM version 2015.01-5-ga29eaa9
$ cat tools/build/NQP_REVISION
2015.01-2-g4589d81
oh
where did that come from?
bartolin_: spectesting with nqp master now, and if it's fine, I'll bump the NQP_REVISION 20:51
Kristien_ Is it possible to create custom phasers?
bartolin_ moritz++
20:52 Kristien_ is now known as Kristien
moritz Kristien_: I don't think so, because phasers most usually hook in some compiler code or even backend-specific codegen to get fired 20:52
Kristien OK.
moritz Kristien: what phase did you have in mind?
Kristien just wondering
jnthn Well, if you are going to write the code to fire them it's do-able I guess.
yoleaux 08:37Z <moritz> jnthn: is there a good reason to keep $!started in Proc::Async private? sounds like potentially useful information for the caller
Kristien No specific use-case in mind. 20:53
dalek kudo/nom: 00c3fae | moritz++ | tools/build/NQP_REVISION:
Bump NQP revision.

gives us an sprintf error message fix, free beer and world peace
20:53 jluis_ joined
moritz now I did the 'git push' without waiting for the spectest to finish. The muscle memory is strong in this one... 20:54
20:54 haxmeister left
jnthn m: sub foo() { }; &foo.add_phaser('FOO', { say "lol" }); &foo.fire_phasers('FOO'); 20:54
camelia rakudo-moar 243c16: OUTPUT«lol␤»
Kristien make it a pre-push hook
jnthn: lol 20:55
dalek osystem: d73f217 | (Mathieu Gagnon)++ | META.list:
Add Tag module
osystem: c63af83 | moritz++ | META.list:
Merge pull request #39 from 4d47/master

Add Tag module
jnthn You can store any phasers you want on the block. But yeah, a bunch lead to some code-gen in the surroudning scope.
If you can arrnage for that to happen then something should be possible in slang land... 20:56
moritz: No strong reason; do you have a use-case in mind?
moritz jnthn: not really; just that if you juggle with multiple Proc::Async objects, it likely helps to get that piece of information 20:58
or you can write sub f(Proc::Async:D $proc where { .started }) { ... } 20:59
20:59 agentzh left
moritz to declare that you only want started processes there 20:59
just like $.w is public
(which I haven't documented yet!)
jnthn Yeah, no objects to making it publicly readable 21:01
21:04 PZt joined
Kristien m: sub f { END { say "hi" } }; f; f; f; 21:04
camelia rakudo-moar 243c16: OUTPUT«hi␤»
dalek ast: 0630ae1 | usev6++ | S (2 files):
Unfudge passing tests
 21:06
kudo/nom: 225cf21 | moritz++ | src/core/Proc/Async.pm:
Proc::Async: make $!started public
 21:13
c: ccb0946 | moritz++ | lib/Type/Proc/Async.pod:
document .w, .started
ast: 8c487c0 | moritz++ | S17-procasync/basic.t:
Test Proc::Async.w and .started
 21:16
c: d742c13 | moritz++ | lib/Type/Proc/Async.pod:
Fix typo
 21:21
[ptc] moritz: I got the syntax highlighting working with pygments 21:22
the solution is (at present) quite ugly, however can be improved; I just don't yet know how
moritz [ptc]: push it to a branch, please 21:23
[ptc] unfortunately, I had to patch Pod::To::HTML and doc... should I push both branches? Or should I send patches?
is it ok to push to a branch in the main repo, or should I do it on my own github account
(sorry, still learning the ropes of the project) 21:24
moritz [ptc]: pull request for Pod::To::HTML (it's not in the perl6 org)
[ptc]: and a branch for doc
[ptc] moritz: atually Pod::To::HTML is in the perl6 org... I pushed code to it today... 21:25
moritz [ptc]: oh, then a branch is fine
dalek c/add_pygments_highlighting: d7e1907 | paultcochrane++ | html/css/pygments.css:
Add pygments css colour definitions file
c/add_pygments_highlighting: 776d3f4 | paultcochrane++ | template/head.html:
Load pygments style definitions
[ptc] ok
moritz [ptc]: and my memory is out of date :-)
[ptc]: you already know your way around the ecosystem better than me, it seems :-) 21:27
21:27 xfix left
[ptc] moritz: still learning, I'm afraid. But getting there :-) 21:27
moritz: I used `perl6 -Ipath/to/Pod-To-HTML/lib/ htmlify.p6` to test 21:28
21:28 kjs_ joined
[ptc] the output looks good, and I don't know if it slows down the htmlify program that much 21:28
moritz [ptc]: I see what you mean about it being ugly. I'd suggest to patch Pod::To::HTML to accept a callback for code / preformatted text, and then set that callback in htmlify 21:29
[ptc] it basically runs a qqx on pygmentize, the output of which then replaces what used to be simply in a <pre>
moritz [ptc]: because making Pod::To::HTML generally depend on pygmentize is a bad idea
[ptc] agreed
it's really just a proof of concept
21:29 kjs_ left
moritz aye 21:30
21:30 kjs_ joined
b2gills m: sub test ( :named($) where *.defined ){...}; test() 21:32
camelia rakudo-moar 243c16: OUTPUT«Constraint type check failed for parameter '<anon>'␤ in sub test at /tmp/IkGVAotpuW:1␤ in block <unit> at /tmp/IkGVAotpuW:1␤␤»
b2gills m: sub test ( :named($)! where *.defined ){...}; test()
camelia rakudo-moar 243c16: OUTPUT«Required named parameter 'named' not passed␤ in sub test at /tmp/fVCYwjA8VD:1␤ in block <unit> at /tmp/fVCYwjA8VD:1␤␤»
[ptc] I'd like to know how to pass the contents of a pod node out to a qqx without having to store the output in a temporary file first
unfortunately, couldn't work that one out, hopefully someone here as an idea 21:33
b2gills Perhaps you could try Proc::Async instead 21:34
[ptc] anyway, I'll head off
gn8!
21:35 FROGGS[mobile] left
moritz aye, Proc::Async works 21:36
21:38 [ptc] left 21:39 Kristien left 21:40 Kristien joined 21:42 jluis_ left 21:47 agentzh joined 21:51 mvuets left
dalek kudo/nom: e34a9c0 | lizmat++ | src/core/control.pm:
Add CLONE-HASH-DECONTAINERIZED helper sub
 21:51
22:00 Kristien left 22:02 jack_rabbit joined
dalek kudo-star-daily: 50edfb8 | coke++ | log/ (9 files):
today (automated commit)
 22:03
22:06 zby_home joined
Mouq gist.github.com/Mouq/10d5b19f7869add9e9e0 22:07
(bug in Supply.categorize) 22:08
Mouq rakudobugs it 22:14
22:14 kjs_ left
lizmat will look at it tomorrow, unless someone beats me to it 22:14
ab5tract still running into some weirdness with regexes and escapes sequences :( gist.github.com/ab5tract/4961802803d3e5fb489d
moritz Mouq: should it call .done of the sub-supplies before or after the outer? 22:16
ab5tract m: my @a = ("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values; @a.say 22:17
camelia rakudo-moar 243c16: OUTPUT«｢55｣ ｢x｣ ｢77｣ ␤»
22:18 doobi-sham-5008 joined
ab5tract m: my @a = ("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values.map: { ~$_ }; @a.perl.say 22:18
camelia rakudo-moar 243c16: OUTPUT«Array.new("55", "x", "77", "")␤»
22:19 zby_home left
skids r: sub f (|c) { c.perl.say }; f(1, b => 3); f(1, "b" => 3); \(1, "b" => 3).perl.say; # Should Capture literals use Capture syntax rules? 22:19
camelia rakudo-{parrot,moar} 243c16: OUTPUT«Capture.new(list => (1,), hash => {"b" => 3})␤Capture.new(list => (1, "b" => 3,))␤Capture.new(list => (1,), hash => {"b" => 3})␤»
ab5tract what's with the extra element, and i'm guessing there is a better way to get the "just the strings please" of regex captures?
timotimo m: say("55x77" ~~ rx{(\d\d)(x)(\d\d)}) 22:20
camelia rakudo-moar 243c16: OUTPUT«｢55x77｣␤ 0 => ｢55｣␤ 1 => ｢x｣␤ 2 => ｢77｣␤»
Mouq I'd say after, since that seems to be the case for other supply-generating methods, but I don't really know. Although if you're relying on the order that .done is called on Supplies, you're probably doing something wrong :P
timotimo m: say(("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values.perl)
camelia rakudo-moar 243c16: OUTPUT«(Match.new(orig => "55x77", from => 0, to => 2, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "55x77", from => 2, to => 3, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "55x77", from => 3, to => 5, ast => A…»
timotimo m: say(("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values>>.Str.perl) 22:21
camelia rakudo-moar 243c16: OUTPUT«("55", "x", "77", "")␤»
ab5tract ah, yes.. >>.Str, should have thought of that
timotimo is not sure where that final element comes from, something about .values being wonky?
m: say(("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values[*-1].perl)
camelia rakudo-moar 243c16: OUTPUT«[]␤»
timotimo m: say(("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values[*-1]) 22:22
camelia rakudo-moar 243c16: OUTPUT«␤»
timotimo m: say(("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values[*-2])
camelia rakudo-moar 243c16: OUTPUT«｢77｣␤»
timotimo i got nothin'
except bedtime
i definitely gots the bedtime
ab5tract :(
ye
22:22 brrt joined
lizmat good night timotimo! 22:22
ab5tract i'm about there too
it's just weird that i can print the escape sequences after its been processed, and it just prints back. if i write that output back in manually as an argument to print, it moves the cursor 22:26
22:26 doobi-sham-5008 left
ab5tract m: my @a = ("55x77" ~~ rx{(\d\d)(x)(\d\d)}).values>>.Str; @a.join.say; @a.join.comb.elems.say 22:30
camelia rakudo-moar 243c16: OUTPUT«55x77␤5␤»
ab5tract invisible characters :( 22:31
dalek kudo/nom: a1c8b97 | lizmat++ | src/core/ (3 files):
Some more uses of CLONE-HASH-DECONTAINERIZED()
22:32 brrt left
moritz Mouq: I seem to have a fix for you; waiting for spectests... 22:34
22:34 ggherdov joined 22:35 pnu joined
moritz fwiw I've installed mosh on hack.p6c.org 22:36
might interest some of you :-)
22:38 ilbot3 left
dalek kudo/nom: 0c524e3 | moritz++ | src/core/Proc/Async.pm:
Throw an X::OS from Proc::Async if the external command could not be launched
 22:38
kudo/nom: 94501a6 | moritz++ | src/core/SupplyOperations.pm:
Supply.categorize: distribute .done to sub-supplies
22:38 ilbot3 joined
Mouq moritz++ 22:43
Mouq rebuilds
lizmat moritz: wouldn't that not also need to be done for Supply.classify ? 22:46
dalek kudo/nom: 8d7542d | lizmat++ | src/core/control.pm:
Create helper sub CLONE-LIST-DECONTAINERIZED
 22:49
22:52 moomin-aba joined, moomin-aba left 22:53 ken joined, ken is now known as Guest13945
Guest13945 Hey 22:54
22:54 Guest13945 left, brrt joined, fil_ joined
fil_ I've been staring at this code all day: rosettacode.org/wiki/Multifactorial#Perl_6 22:55
specifically sub mfact($n, :$degree = 1) { [*] $n, *-$degree ...^ * <= 0 }
I don't understand the line in the braces
I assume that is a "whatever" star?
masak fil_: [*] means "multiply all these things"
dalek kudo/nom: 371434f | lizmat++ | src/core/Proc/Async.pm:
One more use of CLONE-LIST-DECONTAINERIZED
masak fil_: the `...` creates a sequence of values. 22:56
lizmat m: say [+] 1,2,3,4,5
camelia rakudo-moar 243c16: OUTPUT«15␤»
masak fil_: and the `*-$degree` comes out meaning "subtract $degree from the previous value"
lizmat and on that note, good night, #perl6!
masak fil_: the `^` on the `...^` means "discard the last value"
fil_: the `* <= 0` means "keep doing the sequence until we find a value that less than or equal to 0" 22:57
(and then don't include that value in the sequence, due to `^`)
fil_ so you can use a conditional in place of a limit value?
22:57 Ben_Goldberg left 22:58 Ben_Goldberg joined
fil_ I think part of my confusion is that * is overloaded in the statement 22:58
Ovid_ m: sub f(Int $x) {...}; f(3.0)
camelia rakudo-moar 243c16: OUTPUT«===SORRY!=== Error while compiling /tmp/Y3anEyM3JH␤Calling 'f' will never work with argument types (Rat)␤ Expected: :(Int $x)␤at /tmp/Y3anEyM3JH:1␤------> sub f(Int $x) {...}; ⏏f(3.0)␤»
masak fil_: you can use anything that can be matched against. `* <= 0` just happens to create a little runnable thing that returns a boolean when called.
fil_ within the [*] it means multiply, but then later it is a whatever star
masak m: my $c = (* <= 0); say $c(-4); say $c(0); say $c(7)
camelia rakudo-moar 243c16: OUTPUT«True␤True␤False␤»
masak (parentheses for clarity; not needed) 22:59
fil_: after a while you learn to distinguish by context.
fil_: there's `*` as an operator, which just means multiplication.
Ovid_ Since 3.0 == 3 (not ===), is it fair to say that the hard-coded Rat 3.0 can’t match an Int?
masak fil_: and then there's `*` as a term, which is pronounced "whatever" and has a number of useful semantics.
fil_ ok let me try this. 23:00
{ [*] $n, *-$degree ...^ * <= 0 }
masak Ovid_: I think so, yes.
fil_ [*] multiply what follows.
masak Ovid_: provided I understand your question.
Ovid_ I sort of agree they shouldn’t match, but it looks a touch odd. 23:01
fil_ $n, *-$degree ..^ * <=0 means start with $n, then a term minus degree, then keep subtracting degree until you are <= 0 dropping last term
masak m: my $degree = 5; my $n = 20; say $n, *-$degree ...^ * <= 0
camelia rakudo-moar 243c16: OUTPUT«20 15 10 5␤»
masak fil_: ^^ maybe helps
fil_: yes, you got it just right.
fil_ the part that's a bit confusing is the $n, *-$degree part
those are the first two terms in the sequence?
masak fil_: yes, with the second one being a bit special.
fil_: since it's also a WhateverCode, just like the stopping criterion. 23:02
m: my $degree = 5; say (* - $degree).^name
camelia rakudo-moar 243c16: OUTPUT«WhateverCode.new()␤»
fil_ there's a REPL here! Coolness!
masak we lovez our camelia.
Mouq m: say 20, -> $i { $i - 5 } ... ^ -> $n { $n <= 0 } 23:03
camelia rakudo-moar 243c16: OUTPUT«Cannot call 'Numeric'; none of these signatures match:␤:(Mu:U \v: *%_)␤ in sub prefix:<^> at src/gen/m-CORE.setting:8468␤ in block <unit> at /tmp/vEqOSIpuun:1␤␤»
Mouq Err
m: say 20, -> $i { $i - 5 } ...^ -> $n { $n <= 0 }
camelia rakudo-moar 243c16: OUTPUT«20 15 10 5␤»
fil_ m: 10, 9, *-1 ... 0 23:04
camelia ( no output )
fil_ m: 10, *-1 ... 0
camelia ( no output )
Mouq fil_: "say "
fil_ m: say 10, 9, *-1 ... 0
camelia rakudo-moar 243c16: OUTPUT«10 9 8 7 6 5 4 3 2 1 0␤»
fil_ so the whatever term doesn't have to be the first term
but I assume it has to be the last one?
masak it has to be the term before the `...`
fil_ aha
masak but you can have several `...`, so you can sort of "steer" the sequence if you need to. 23:05
m: say 1, 2 ... 10, 20 ... 100
camelia rakudo-moar 243c16: OUTPUT«1 2 3 4 5 6 7 8 9 10 20 30 40 50 60 70 80 90 100␤»
Mouq m: my @fib := 1, 1, * + * ... *; say @fib[^10]
camelia rakudo-moar 243c16: OUTPUT«1 1 2 3 5 8 13 21 34 55␤»
fil_ how does [*] work? It takes any infix operator?
masak yep. 23:06
m: say [&&] True, False, True
camelia rakudo-moar 243c16: OUTPUT«False␤»
masak m: say [&&] True, True, True
camelia rakudo-moar 243c16: OUTPUT«True␤»
masak fil_: we call them "metaoperators". we have a couple of them, besides [op]
fil_ I'm getting concerned masak that artificial intelligence has been discovered, but it's being kept a secret. You're just too fast to be believable... 23:07
23:07 kaare_ joined
masak fil_: we have >>op<<, "distribute operation elementwise" (also works on prefixes and postfixes, with appropriate modification) 23:07
fil_: I type fast.
fil_: we have Xop, though just X is easier to understand at first.
fil_ so later on in the program it calls the sub as: map &mfact.assuming(:$degree), 1 .. 10; 23:08
masak @a X @b gives you all combinations of @a elements and @b elements.
right.
fil_ what is .assuming?
it's not in learnxiny
masak it takes a sub and fills in one of the parameters with a value.
let's make an example.
m: sub greet($name) { say "howdy, $name!" }; &greet.assuming(:name<fil_>).() 23:09
camelia rakudo-moar 243c16: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in sub greet at /tmp/Mlk1yQ8UA1:1␤ in sub CURRIED at src/gen/m-CORE.setting:3687␤ in block <unit> at /tmp/Mlk1yQ8UA1:1␤␤»
fil_ in this case it's filling in with $_ from map and 1..10?
masak m: sub greet(:$name) { say "howdy, $name!" }; &greet.assuming(:name<fil_>).()
camelia rakudo-moar 243c16: OUTPUT«howdy, fil_!␤»
masak there we go.
I think I just proved I'm not an AI, too ;)
fil_ Just because you can pass the Turing test doesn't mean you're not an AI
masak no, it's filling in with $degree from the for loop. 23:10
fil_: true...
I could have faked that mistake...
gee, you humans have very high requirements.
fil_ there is no for loop
masak on the line before the map
the line before the .assuming, I should say 23:11
23:11 ribasushi left
fil_ what's the difference between &mfact.assuming(:$degree) and &mfact(:degree<$degree>) 23:12
masak it's :degree($degree)
but given that, no difference -- it's a shorthand
fil_ let me try again
masak (it's :name<str> when you mean a literal string and :name($expr) when you mean an expression)
fil_ what's the difference between &mfact.assuming(:$degree) and &mfact(:$degree) 23:13
masak the .assuming produces a new sub, still without calling it.
fil_ oh I see!
masak the *map* calls that new sub, and passes $_
23:13 agentzh left
fil_ that's what the & is doing 23:13
23:13 ribasushi joined
masak which binds to the $n 23:13
fil_ I see 23:14
it's like a "partial"
in clojure
masak yes, the & here means "we're talking about the sub as a thing, not calling it"
m: sub foo {}; say &foo
camelia rakudo-moar 243c16: OUTPUT«sub foo () { #`(Sub|66661536) ... }␤»
masak m: sub foo {}; say &foo.arity
camelia rakudo-moar 243c16: OUTPUT«0␤»
masak m: sub foo($n) {}; say &foo.arity
camelia rakudo-moar 243c16: OUTPUT«1␤»
fil_ I think the writer did a poor job writing that perl6 example 23:15
masak in Perl 6, subs secretly are actually called `&foo` etc. it's just that `foo` is a shorthand for `&foo()`
fil_: it looks like good Perl 6 to me.
fil_: that's not to say there aren't other ways.
fil_ when I read the Java example, it's immediately clear to the reader what's going on 23:16
this one is using some pretty advanced concepts
masak fil_: I probably would've reached for a sub inside the for loop instead. then I wouldn't have to pass $degree with .assuming.
fil_ do we really need a partial!
masak no.
here, let me rewrite it for you without it.
fil_ I think I'm just going to change it for the sake of future generations 23:17
masak m: for 1 .. 5 -> $d { sub mfact($n) { [*] $n, *-$d ...^ * <= 0 }; say "$d: ", map &mfact, 1..10 }
camelia rakudo-moar 243c16: OUTPUT«1: 1 2 6 24 120 720 5040 40320 362880 3628800␤2: 1 2 3 8 15 48 105 384 945 3840␤3: 1 2 3 4 10 18 28 80 162 280␤4: 1 2 3 4 5 12 21 32 45 120␤5: 1 2 3 4 5 6 14 24 36 50␤»
masak there ya go
speedy rewrite! AI wins again \o/
ooh, and I can show off postfix >> 23:19
m: for 1 .. 5 -> $d { sub postfix:<mfact>($n) { [*] $n, *-$d ...^ * <= 0 }; say "$d: ", (1..10)>>mfact } 23:20
camelia rakudo-moar 243c16: OUTPUT«1: 1 2 6 24 120 720 5040 40320 362880 3628800␤2: 1 2 3 8 15 48 105 384 945 3840␤3: 1 2 3 4 10 18 28 80 162 280␤4: 1 2 3 4 5 12 21 32 45 120␤5: 1 2 3 4 5 6 14 24 36 50␤»
masak not much longer, and kinda cute :)
23:20 _4d47 joined
masak note how the `>>` means "distribute this operator (that we just defined) over all the elements" 23:20
Mouq m: for 1 .. 5 -> $d { sub mfact($n) { [*] $n, *-$d ...^ * <= 0 }; say "$d: ", (1..10)>>.&mfact } # works too
camelia rakudo-moar 243c16: OUTPUT«1: 1 2 6 24 120 720 5040 40320 362880 3628800␤2: 1 2 3 8 15 48 105 384 945 3840␤3: 1 2 3 4 10 18 28 80 162 280␤4: 1 2 3 4 5 12 21 32 45 120␤5: 1 2 3 4 5 6 14 24 36 50␤»
masak ooh, yes.
fil_ I think your first approach was clearer 23:21
masak Mouq++ is here using the fact that any sub can be called as a method: `$obj.&mysub`
fil_: yes, possibly.
fil_: though it depends what you're used to.
I agree not much is gained by making it an operator here.
sometimes it makes a world of difference. when you can use [op], for example. 23:22
fil_ I've updated rosettacode.org/wiki/Multifactorial#Perl_6 with your simplified solution. Thanks for walking me through this! 23:26
masak my pleasure. 23:27
fil_ I understand that { [*] $n, *-$degree ...^ * <= 0 } is the correct way to write it, but it troubles me that perl6 allows for three completely different meanings of * in the same statement.
masak two are the same, though.
fil_ well kinda
but they're not the same instance
If I saw two $n in the same statement, chances are they referred to the same thing. But here the first one refers to "previous term" and the second one refers to "current term" 23:28
masak would it reassure you if I said that Perl 6 itself is never confused as to what kind of `*` you use, and that with (minimal) practice it's a breeze to tell the two different meanings apart?
fil_ fair enough
masak oh, but the `*` means "whatever".
fil_ I just started studying it today
masak that you *have* to know, otherwise it will not make sense :) 23:29
fil_ like what's the difference between $_ and *
masak the thing with "whatever" is that the surrounding context/callee fills in the appropriate meaning.
easiest to show the difference. hold on.
23:29 _4d47 left
masak m: my &t = { $_ == 7 }; say &t(5); say &t(7) 23:29
camelia rakudo-moar 243c16: OUTPUT«False␤True␤»
masak m: my &t = * == 7; say &t(5); say &t(7) 23:30
camelia rakudo-moar 243c16: OUTPUT«False␤True␤»
masak so `*` is kind of `$_` plus invisible `{ }`
fil_ ah
ok that is a HUGE help
it means that spaces are significant in perl 23:31
or at least moreso than in other languages like Java or C.
masak correct.
it's used to differentiate whether something is a postfix or an infix, for example. 23:32
m: my @a; @a[0]
camelia ( no output )
masak m: my @a; @a [0]
camelia rakudo-moar 243c16: OUTPUT«===SORRY!=== Error while compiling /tmp/DItW6NY4u5␤Two terms in a row␤at /tmp/DItW6NY4u5:1␤------> my @a; @a [⏏0]␤ expecting any of:␤ infix stopper␤ infix or meta-infix␤ bracketed…»
masak ok, that last one wasn't an infix, but still.
fil_ ya
interesting
masak m: my $x; $x ++
camelia rakudo-moar 243c16: OUTPUT«===SORRY!=== Error while compiling /tmp/W7u86gAYMY␤Preceding context expects a term, but found infix + instead␤at /tmp/W7u86gAYMY:1␤------> my $x; $x ++⏏<EOL>␤»
fil_ I actually kind of like spaces being significant
results in code with more consistent styling 23:33
23:33 brrt left
fil_ hey I have another question for you. 23:34
Mouq fil_: Another rational is that, with the complex grammar of Perl 6, it's super useful to be able to easily tell the user that there's a typo. Having significant space means that typos are much less likely to be interpreted as valid code
masak fil_: the thing that some people don't like is that they can't put spaces between `mysub` and `(...)`
fil_ I submitted an issue on learnxiny
github.com/adambard/learnxinyminut...t-71391089
masak heh 23:35
m: sub named-def(:$def) { say "\$def is $def" }; named-def(:10def)
camelia rakudo-moar 243c16: OUTPUT«$def is 10␤»
masak there's a backstory to why this weird feature exists. 23:36
fil_ so at least you agree with me that it doesn't look right... :-)
23:36 rurban_ left
masak yes, but I'm also showing that it works. 23:36
the reason it's there is to support this feature: 23:37
m: $_ = "ABCDE"; m:4th/./; say ~$/
camelia rakudo-moar 243c16: OUTPUT«D␤»
fil_ Does the second m: refer to camelia or match? 23:39
masak m: $_ = "ABCDE"; m/./; say ~$/ 23:40
camelia rakudo-moar 243c16: OUTPUT«A␤»
masak it's `m//` for "match using this regular expression"
fil_ I see
that's pretty freekin obscure
so someone has created :nd, :st, :rd etc...
masak note how I'm not arguing back.
yep.
fil_ sigh
masak when it's useful, it's very useful. 23:41
fil_ That's Larry and his desire for the language to match English
masak but it's rare.
fil_ is there a :me for French?
masak I could have lived with just :nth($n), actually
fil_: only in the expansion pack.
fil_ lol
It feels like cuteness that will help ppl win obfuscation awards 23:42
masak and the 'm' and the 'e' have to be written with Unicode superscript.
fil_ but it doesn't add anything to the clarity of the language
masak oh, I don't know about that.
m:2nd/ <$regex> /
feels pretty clear to me. 23:43
fil_ I assume then that named arguments can't start with numbers for that reason
does it only work for numbers
or can you use <strings>, $vars or other things? 23:44
It feels like an entire language construct just added for the benefit of the match operation
masak basically. 23:46
named arguments can't start with numbers because identifiers can't.
there's prior art there from basically all other languages.
if identifiers could start with numbers, you'd have to backtrack once you distinguished an identifier from a number literal. 23:47
and we prefer to keep our languages backtrack-free.
fil_ so masak, besides being an amazing perl6 tutor, what's your role in the perl6 community?
masak I like to break things.
fil_ :-) 23:48
masak no; correction. I use stuff, and it breaks.
then I submit literally hundreds of bug reports.
fil_ nice
are you a core perl6 contributor?
masak on a good day, yes.
fil_ cool!
how many of you are there?
masak looks around
looks like just me :(
fil_ lol
I mean core perl6 code contributors 23:49
masak it's a fuzzy number, for sure.
fil_ is Larry writing compiler/interpretor code?
masak there are scores of names in the Rakudo commit log, for example.
fil_: yes, nowadays he is.
fil_ that's cool
have you ever met him in person?
masak fil_: he had a self-imposed embargo for several years.
something about Perl 5 ending up a bit messy in the implementation. 23:50
fil_ I've read through perl5 source code.
It was seriously terrible
I had a bug I needed to deal with
Some of the worst code I've ever seen
masak fil_: I have met him in person. I've done the whole star-struck thing, and "omgomg I'm sitting next to Larry at a hackathon", and gotten yelled at by him, and had amazing technical discussions (mostly on IRC).
fil_: what I've seen has made my hair curl a bit, too :) 23:51
fil_ Larry's a visionary, and I'm loving what I see on perl6. Just goes to show you can write an inspirational spec without being able to write a great compiler for it
masak see, for example, strangelyconsistent.org/blog/dash-n-and-dash-p
fil_: if you want to see something great that Larry wrote, though, check out STD.pm6 23:52
fil_: that's inspirational. and it's running code, too.
back when the rest of us couldn't imagine how to make a Perl 6 parser work, Larry wrote one in Perl 6, then wrote "just enough" shimming to translate that Perl 6 back to Perl 5, so it could run. 23:53
fil_ falls prostate 23:55
masak considering how messy Perl 5's whole lexer/parser is -- to the point where people talk about "X ways a parser can lie to itself" when they describe it -- STD.pm6 just stands out as an awesome exemplar of clarity and intent. 23:56
fil_ gotta go make dinner
thanks for the inspiration
masak nice talking to you
fil_ l8ras
masak o/
fil_ before I go
nah
another time
masak :)
fil_ l8r
23:56 fil_ 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!