»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
derp_commander so for example, "func[x_] := 2*x-1" is actually the creation of a rule. x_ is the pattern for an arbitrary expression, which is to be scoped as "x", ":=" is the SetDelayed operator, which creates a rule whose evaluation is deferred, and the rest of the expression is the thing whatever matches that pattern gets replaces with, after scoping 00:01
s/replaces/replaced/g
TimToady that's pretty much just a lambda 00:02
derp_commander in its simplest use case
TimToady one could certainly picture extending it into macroland 00:03
derp_commander reference.wolfram.com/language/ref/...layed.html explains it in more detail 00:04
00:04 Actualeyes joined, spider-mario left
derp_commander note that because, as a mathematical language, undefined variables are actually still valid objects (which can be used in expressions and so forth), you can do the substitution after the fact by using the replace operator /. on a rule set (an associative array of patterns) 00:06
psch m: my $func = 2 * * - 1; say $func(3) 00:07
camelia rakudo-moar 361448: OUTPUT«5␤»
psch from a perl-ish perspective (well, and C-ish, and Java-ish too i suppose) i'd dislike overloading [] like that 00:08
derp_commander psch: Mathematica uses [] as function parentheses
psch from the same perl-ish perspective, i'd readily accept e.g. ´` or `´
TimToady that's just cultural differences 00:09
psch yeah, i'm aware :)
TimToady I worry more about mental traps :)
psch i do agree, fundamentally it just seems like a lambda
derp_commander m: my &func($x) = 2+$x; say &func($x);
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/3qjyjlBqtq␤The () shape syntax in routine declarations is reserved (maybe use :() to declare a longname?)␤at /tmp/3qjyjlBqtq:1␤------> 3my &func($x7⏏5) = 2+$x; say &func($x);␤ expecting any of:…»
00:10 grondilu joined
psch m: my &func = -> $x { 2 + $x }; say func(2) 00:10
camelia rakudo-moar 361448: OUTPUT«4␤»
TimToady note this is not what mathematica is doing 00:11
derp_commander right
TimToady that's more of a macro rewrite system
for algebra and such
assuming that by default is less confusing to mathematicians, and more confusing to mere mortals :) 00:12
so P6 separates those concerns
derp_commander TimToady: assuming what? That undefined variables are symbols which can have their values substituted freely? 00:13
00:13 kurahaupo left
dwb Hrumph, I'm dumb: gist.github.com/anonymous/9dcf4e0d...a03669bd9f 00:14
TimToady rewrite rules tend toward surprise for people who think in terms of data mutation, since the referentiality is deferred
so P6 defaults to a more evaluate-it-by-default model
deferred evaluation then takes more work 00:15
one of those tradeoffs I keep mentioning :)
00:15 yqt left
derp_commander TimToady: right, and the other problem is Mathematica doesn't really leave much room for OO 00:15
TimToady sure, it's gonna be pretty thoroughly FP 00:16
we rather like the balance we've struck there
psch m: gist.github.com/anonymous/9dcf4e0d...a03669bd9f 00:17
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/8K9AIlZ5CQ␤Two terms in a row across lines (missing semicolon or comma?)␤at /tmp/8K9AIlZ5CQ:2␤------> 3Could not find symbol '&foo'7⏏5<EOL>␤ expecting any of:␤ infix␤ infix st…»
psch actually looks at the gist now 00:18
derp_commander TimToady: the one exception in Mathematica, is you can use the TagSet and TagSetDelayed opers to attach a rule to a symbol
as opposed to the global lexical scope
psch dwb: subs are 'my' (i.e. lexically) scoped by default, which means they aren't visible outside their own package
m: module A { our sub foo { say "ok" } }; A::foo; module B { sub foo { say "never gonna work" } }; B::foo 00:19
camelia rakudo-moar 361448: OUTPUT«ok␤Could not find symbol '&foo'␤ in block <unit> at /tmp/Zl9qyjpROz line 1␤␤Actually thrown at:␤ in block <unit> at /tmp/Zl9qyjpROz line 1␤␤»
TimToady Larry's Law: 90% of any given programming paradigm does not forbid interop with other programming paradigms. It's only that last 10% that turns it into a religion.
dwb psch: derp, thanks! 00:20
derp_commander TimToady: well sure, I can have a lot of fun with the /: operator, but that still is only a binding to a symbol
dwb: what did I do?
AlexDaniel m: sub infix:<-> { $^a + $^b }; say 2 − 2 00:21
camelia rakudo-moar 361448: OUTPUT«0␤»
AlexDaniel ZoffixWin: ↑ :P
dwb derp_commander: You responded to my herp derp edness.
derp_commander dwb: wasn't me
dwb Apologies for lighting up your IRC client.
ZoffixWin You're cheating
AlexDaniel ZoffixWin: if “using Perl 6” is cheating, then yes 00:22
derp_commander ZoffixWin: define cheating, please
TimToady derp_commander: in P6 it's actually impossible to attach a rule outside of its intended lexical scope, so you can't rewrite other people's code for them on the fly
ZoffixWin derp_commander, I think you're involving yourself in too many conversations :) neither dwb nor I were replying to you :)
m: sub infix:<⁤> { $^a + $^b }; say 2⁤2
camelia rakudo-moar 361448: OUTPUT«4␤»
ZoffixWin AlexDaniel, ^ :)
psch m: { CALLER:('&infix:<+>') := sub { $^a - $^b } }; say 1 + 2 00:23
camelia rakudo-moar 361448: OUTPUT«===SORRY!===␤This type (QAST::WVal) does not support positional operations␤»
ZoffixWin .u -−
yoleaux U+002D HYPHEN-MINUS [Pd] (-)
U+2212 MINUS SIGN [Sm] (−)
psch m: { CALLER::('&infix:<+>') := sub { $^a - $^b } }; say 1 + 2
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Gwb4CKq4yf␤Cannot use bind operator with this left-hand side␤at /tmp/Gwb4CKq4yf:1␤------> 3ER::('&infix:<+>') := sub { $^a - $^b } 7⏏5}; say 1 + 2␤»
derp_commander ZoffixWin: no, I think he confused who was the one helping him
ZoffixWin derp_commander, no, "derp" is a word and he used it when his mistake was pointed out -_-
derp_commander OHHHHHHH
ZoffixWin :)
AlexDaniel ZoffixWin: I don't know what to say now 00:24
derp_commander derp ;-P
ZoffixWin AlexDaniel, about what? :)
TimToady thinks of "derp" as an abortive facepalm
derp_commander TimToady: it's more like a self-facepalm
AlexDaniel ZoffixWin: about ⁤
ZoffixWin m: sub infix:<⁤> { $^a + $^b }; sub infix:<⁤⁤> { $^a - $^b }; say 2⁤2⁤42 00:25
camelia rakudo-moar 361448: OUTPUT«46␤»
00:25 slobo left
ZoffixWin ^_^ 00:25
derp_commander a facepalm is "you're being stupid", derp is "I'm being stupid"
TimToady ZoffixWin: now be nice
invisible characters are just evil
psch always these spooky invisible operators :/
derp_commander ZoffixWin: I don't even know much Perl 6 and I know that's just plain nuts 00:26
dwb Then my derp was a multi, I suppose.
psch .u ⁤⁤
yoleaux U+2064 INVISIBLE PLUS [Cf] (<control>)
ZoffixWin I wonder why that gave 46 and not -38
psch .u ⁤
yoleaux U+2064 INVISIBLE PLUS [Cf] (<control>)
00:26 kurahaupo joined
psch .u ⁤ 00:26
yoleaux U+2064 INVISIBLE PLUS [Cf] (<control>)
psch ...i copied both /o\
derp_commander .u INVISIBLE CHEESEBURGER
psch seriously ZoffixWin
yoleaux No characters found
ZoffixWin Oh... fat fingers again >_< 00:27
m: sub infix:<⁤> { $^a + $^b }; sub infix:<⁤⁤> { $^a - $^b }; say 2⁤2⁣42
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MLIHx_weDu␤Bogus postfix␤at /tmp/MLIHx_weDu:1␤------> 3}; sub infix:<⁤⁤> { $^a - $^b }; say 2⁤27⏏5⁣42␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ …»
ZoffixWin Eh, whatever. :)
psch, the other one is invisible separator
AlexDaniel ZoffixWin: guess who got caught in his own trap!
ZoffixWin :)
TimToady He that diggeth a pit shall surely fall into it. 00:28
AlexDaniel well, not necessarily
TimToady We're talking roadrunner ethics here.
derp_commander "It is a fact of history that classes give birth to their own grave-diggers." --Marx
TimToady which one was that, Groucho? :P 00:29
derp_commander lolno
00:29 ShimmerFairy left
psch m: class C { method new { class grave-digger { }.new } } # oh god 00:30
camelia ( no output )
psch m: class C { method new { class grave-digger { }.new } }; my C $c .= new; 00:31
camelia rakudo-moar 361448: OUTPUT«Type check failed in assignment to $c; expected C but got C::grave-digger (C::grave-digger.new)␤ in block <unit> at /tmp/CVeawMx2Oa line 1␤␤»
psch that is perfectly reasonable though, i'd say :)
AlexDaniel m: multi class { } # hey look! a multi class! :PP 00:32
camelia ( no output )
geekosaur classic :p
TimToady anyway, as you can see, we have no trouble adding new Unicode operators :) 00:33
derp_commander m: class bourgeoisie { method new { class proletariat { method new {say 'Workers of the world, unite!';} }.new } }.new;
camelia rakudo-moar 361448: OUTPUT«Workers of the world, unite!␤»
00:33 ShimmerFairy joined
TimToady FP is the opiate of the nomenclatura... 00:34
derp_commander lol()
AlexDaniel what I really like about Perl 6 is that it allows me to write nonsense. Sometimes it makes my day brighter 00:35
though sometimes people say that such code should produce errors like “Really stupid variable name! Don't do things like that”
derp_commander AlexDaniel: anyone can write nonsense, but it takes an expert to write nonsense that will compile 00:36
AlexDaniel fools! :)
aw, that message got delayed a bit
derp_commander: have you ever tried doing that in perl 6? 00:37
00:37 dwb left
derp_commander also, warnings about variable names sound more like the proper provenance of an IDE 00:37
psch m: class bourgeoisie { method defenestrate { say "<wilhelm scream>" }; method new { class proletariat { method new {say 'Workers of the world, unite!'; CALLER::<self>.defenestrate() } }.new } }.new;
camelia rakudo-moar 361448: OUTPUT«Workers of the world, unite!␤<wilhelm scream>␤»
AlexDaniel derp_commander: the error messages are so great that they will guide you step by step to something compilable
00:37 ssotka left
derp_commander AlexDaniel: that class warefare joke was the first non-trivial thing I've ever written in Perl 6 :-) 00:38
AlexDaniel m: (?"(?{noEUSHOUAuC:4}3C13∞]
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/eDVTZMHBRv␤Confused␤at /tmp/eDVTZMHBRv:1␤------> 3(?"(?{noEUSHOUAuC:7⏏054}3C13∞]␤ expecting any of:␤ colon pair␤»
AlexDaniel m: (?"(?{noEUSHOUAuC:k(4)}3C13∞]
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/V2GlZwayxP␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/V2GlZwayxP:1␤------> 3(?"(?{noEUSHOUAuC:k(4)}3C13∞]7⏏5<EOL>␤ expecting any of:␤ horizontal wh…»
AlexDaniel m: (?"(?{noEUSHOUAuC:k(4)}3C13∞]" 00:39
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/EJUKFbinrV␤Unable to parse expression in parenthesized expression; couldn't find final ')' ␤at /tmp/EJUKFbinrV:1␤------> 3(?"(?{noEUSHOUAuC:k(4)}3C13∞]"7⏏5<EOL>␤ expecting any of:␤ s…»
AlexDaniel m: (?"(?{noEUSHOUAuC:k(4)}3C13∞]")
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pK6U9axXU1␤Undeclared routine:␤ noEUSHOUAuC:k<4> used at line 1␤␤»
00:39 pierre_ left
TimToady m: say foo 00:40
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/M1gcEQgU4l␤Undeclared routine:␤ foo used at line 1␤␤»
TimToady m: say bar
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/BZ2iCpDuzu␤Undeclared routine:␤ bar used at line 1. Did you mean 'VAR', 'bag'?␤␤»
TimToady m: say bag
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5===␤Argument to "bag" seems to be malformed␤at /tmp/9b9ZRA0OVC:1␤------> 3say bag7⏏5<EOL>␤Other potential difficulties:␤ Function "bag" may not be called without arguments (please use () or whitespace to denote arguments, …»
TimToady m: say bag()
camelia rakudo-moar 361448: OUTPUT«bag()␤»
TimToady there you go :)
sortiz m: my $foo:::_ = 42; say ::<$foo:::_>; MY::.keys.say; # btw AlexDaniel.
camelia rakudo-moar 361448: OUTPUT«42␤($=pod !UNIT_MARKER EXPORT $_ $! ::?PACKAGE GLOBALish $¢ $=finish $/ $foo:::_ $?PACKAGE)␤»
00:40 wamba left
derp_commander so anyway, I like Mathematica's idea of defining operators by way of rulesets, as opposed to Python which uses doubly-underscored special method names, but I like how Python leaves the behavior up to the objects themselves 00:41
AlexDaniel m: sub noEUSHOUAuC {}; (?"(?{noEUSHOUAuC :k(4)}3C13∞]")
camelia rakudo-moar 361448: OUTPUT«WARNINGS for /tmp/B_XWhVmmHi:␤Useless use of "?" in expression "?\"(?{noEUSHOUAuC :k(4)}3C13∞]\"" in sink context (line 1)␤Unexpected named parameter 'k' passed␤ in sub noEUSHOUAuC at /tmp/B_XWhVmmHi line 1␤ in block <unit> at /tmp/B_XWhVmmHi l…»
AlexDaniel look! A run-time error
so it compiles
TimToady derp_commander: we're fine with delegating methods to objects, but we delegate language tweaking only to lexical scopes 00:42
confusing those is a recipe for disaster 00:43
derp_commander TimToady: I'm aware of the decisions the Perl 6 community has made, I'm just musing on other possibilities
AlexDaniel unexpected parameter, let's fix that. sink context, let's fix that too. Now it wars about nil in string context, but that's fixable
m: sub noEUSHOUAuC(:$k) {‘’}; say (?"(?{noEUSHOUAuC :k(4)}3C13∞]")
camelia rakudo-moar 361448: OUTPUT«True␤»
AlexDaniel \o/
warns*
there was one error message that I didn't like, but then again, it's about our lovely colons 00:44
m: sub noEUSHOUAuC() {}; noEUSHOUAuC:k(4)
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/0g456e5xlc␤Undeclared routine:␤ noEUSHOUAuC:k<4> used at line 1. Did you mean 'noEUSHOUAuC'?␤␤»
derp_commander as for implementation, one way is to scan the expression for objects, load their rules, and then apply them, one-by-one, in some deterministic order 00:45
00:46 dvinciguerra left
TimToady but if you've already scanned the expression, you can't change the parse without reparsing, which is evil 00:47
it tends to fall into "One True Language" kinds of systems
like Tcl
or Lisp, for that matter
derp_commander TimToady: I think you're confusing the lexer with the parser 00:48
geekosaur no difference here, or in other modern parsers 00:49
TimToady P6 is a postmodern parser--there is no lexer.
geekosaur splitting them is a holdover from PDP11 64k address spaces
TimToady or looking at it the other way, a lexer is merely epiphenominal to the parser
derp_commander geekosaur: except if we do split them up, we can load the ruleset at the lexing stage, then apply them at the parsing stage
TimToady then your lexer must understand the One True Lexing Language 00:50
derp_commander TimToady: um, you mean it has an idea what *delimiters* are? 00:51
TimToady amongn other things
we are very serious when we say that those things are all derived concepts in P6 grammars
when you quote anything, you are going into a specific sublanguage that knows about those delimiters 00:52
derp_commander TimToady: yes, I know, everything's a regex to you folks
TimToady and everything's a grammar, which is a class, and everything is a rule, which is a method, which is multidispatched by the lexer 00:53
regex is a very small part of it
psch TimToady++
TimToady (by the classical definition)
derp_commander TimToady: Mathematica allows you to create new tokens, hence altering the lexer, but that comes via a separate package, and the way it works is kinda wonky. 00:54
TimToady and by the classical definition, the regex bits are the parts we can deduce a lexer from
and we deduce a new lexer every time you tweak the language in a given lexical scope
m: say Q/$x/ 00:55
camelia rakudo-moar 361448: OUTPUT«$x␤»
TimToady m: say Q:s/$x/
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/wRjjbSER9N␤Variable '$x' is not declared␤at /tmp/wRjjbSER9N:1␤------> 3say Q:s/7⏏5$x/␤»
TimToady here we start off with the root Q language from the current language braid, and mix in the grammar rule that recognizes $, and deduce a lexer that recognizes $ as a pure regex prefix to some other rule, and then parse using that new language in the lexical scope of the // 00:56
we really dislike arguments from "tradition" here :)
at the end of that, we pop back to the language outside of the Q:s// 00:57
(we do some caching so we don't redo this work, of course) 00:58
TimToady should probably pop his scope and go get dinner...
derp_commander TimToady: if I didn't know better, I'd say either you, or the whole Perl 6 community, was high on something 00:59
psch butterflhigh, maybe... :)
TimToady we've just been thinking about how to do things righter for 15 years 01:00
ZoffixWin ehehe :)
derp_commander "It's not a language, it's a braid of languages! With sublanguages! You can make your OWN language! Wowww, pretty butterflies!" 01:01
ZoffixWin :D
skids But, it works.
TimToady admittedly, it's not fast...yet...
derp_commander I mean, you're literally at the point you could turn /dev/urandom's output into a valid Perl sublang 01:02
TimToady but the first time you use " in a program, it actually does derive Q:qq for you
that's easy, using token { .*? } :) 01:03
AlexDaniel derp_commander: is there anything bad about it?
01:03 Sqirrel left
skids derp_commenader: that's pretty close to the Q:to<some_very_unlikely_char_sequence> sublang :-) 01:03
TimToady we do expect people to think about it a little before choosing to shoot themselves in the foot
derp_commander AlexDaniel: I could easily spend the next hour enumerating various security holes that can introduce 01:04
TimToady no security hole till you assign some semantics to .*?
well, maybe DoS...
anyway, dinner &
derp_commander sounds like a great language... a great domain-specific language 01:05
psch i don't see how "turn /dev/urandom output into <something that compiles>" depends on sublanges
any language that does text substitution can do that
sure, there's an extra step 01:06
derp_commander psch: it was something of a joke
psch derp_commander: oh.
ha ha
psch goes to bed :)
g'nite #perl6 o/
derp_commander re.sub(r'.*',r'print "OHAI"')
sortiz psch o/
skids anyway, that's one good reason why such tweaking is contained by scoping. 01:07
derp_commander re.sub(string,r'.*',r'print "OHAI"')
grondilu m: say r{}.WHAT
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/trshFSEbEM␤Undeclared routine:␤ r used at line 1␤␤»
grondilu is not sure what r'' is 01:08
m: say r''.WHAT
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/5jVhvYG7dP␤Two terms in a row␤at /tmp/5jVhvYG7dP:1␤------> 3say r7⏏5''.WHAT␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statem…»
derp_commander grondilu: python's syntax for raw strings
it automatically escapes all backslashes
grondilu oh yeah that was python code 01:09
grondilu wonders how he could have thought otherwise
derp_commander dunno, lack of sleep?
AlexDaniel lack of perl 6
grondilu funny thing is I was about to say that r'.*',r'print "OHAI"' looked very confusing 01:10
I was about to suggest using {} instead of '', like r{.*}
derp_commander grondilu: the syntax of prefixing the string with a character is probably Perl-derived. The same thing is done with u/b 01:11
grondilu it's funny because people often say that python code is cleaner than perl code, in that case it was quite the opposite
derp_commander though some of that is because I omitted the import for re, and never defined the string variable 01:12
01:13 pierre_ joined
skids Actually I think that harkens back to sh's test 01:13
derp_commander skids: wherever it's from
grondilu quite probably, there is not much genuinely Perlish, since Perl was a mix of sh/C/awk or something. 01:14
sortiz In C++ the strings can be prefixed too.
jdv79 ZoffixWin: what's your next post about? 01:15
ZoffixWin jdv79, I don't have one 01:16
geekosaur sh, c, csh, awk, a bit of sed and a mainframe basic (iirc for the `foo while bar` syntax which I think I first saw in a basic for tops/20)
derp_commander some English, maybe a little Lisp, and not even Larry Wall knows what else 01:17
oh, another thing: while I know it's also used for comments, I kinda think # is a more appropriate sigil for hashes, since it's literally called (among other things) a hash mark 01:20
jdv79 ZoffixWin: btw, nice tests in your irc module
ZoffixWin Thanks. I always pride myself on doing thorough testing :P 01:21
skids "hash" is really using a specific implementation to name a general category.
01:22 cpage_ left
derp_commander skids: yeah, but it's still the most common implementation thereof 01:22
jdv79 is there a glyph for map or dictionary?
ZoffixWin .u map
yoleaux U+05BC HEBREW POINT DAGESH OR MAPIQ [Mn] (◌ּ)
U+26EF MAP SYMBOL FOR LIGHTHOUSE [So] (⛯)
U+A0B8 YI SYLLABLE MAP [Lo] (ꂸ)
ZoffixWin I always thought % looked like an 'H' 01:23
01:23 cpage_ joined
derp_commander .u dictionary 01:24
yoleaux No characters found
sortiz And in fact % means Associative, with a default 'is Hash'
derp_commander .u book
yoleaux U+1F4D1 BOOKMARK TABS [So] (📑)
U+1F4D5 CLOSED BOOK [So] (📕)
U+1F4D6 OPEN BOOK [So] (📖)
01:25 BenGoldberg left 01:26 BenGoldberg joined
ZoffixWin m: sub term:<📑> is rw {state %h}; 📑<foo> = 42; say 📑 01:26
camelia rakudo-moar 361448: OUTPUT«{foo => 42}␤»
01:26 ssotka joined
ZoffixWin ^_^ 01:26
01:27 BenGoldberg left 01:28 huggable left, BenGoldberg joined
ZoffixWin m: sub term:<📑> is rw {state %h}; sub postcircumfix:<⇦📄 ⁤> is rw { $^a{$^b} }; 📑⇦📄 'foo'⁤ = 42; say 📑 01:33
camelia rakudo-moar 361448: OUTPUT«{foo => 42}␤»
01:33 pierre_ left
ZoffixWin eehehe 01:33
m: sub term:<📑> is rw {state %h}; sub postcircumfix:<⇦📄 ⁤> is rw { $^a{$^b} }; 📑⇦📄 'foo' = 42; say 📑 01:34
camelia rakudo-moar 361448: OUTPUT«5===SORRY!5=== Error while compiling /tmp/HxxTRv8SML␤Unable to parse expression in postcircumfix:sym<⇦📄 ⁤>; couldn't find final $stopper ␤at /tmp/HxxTRv8SML:1␤------> 3⇦📄 ⁤> is rw { $^a{$^b} }; 📑⇦📄 'foo' = 427⏏5; say …»
01:38 pierre_ joined 01:39 pierre_ left 01:46 itcharlie left, ilbot3 left
ZoffixWin m: sub term:<🏠> is rw {state @}; sub postcircumfix:<🚪 ⁤> is rw {$^a[$^b]}; 🏠.push: 42; 🏠🚪1⁤ = 72; say 🏠 01:46
camelia rakudo-moar 361448: OUTPUT«[42 72]␤»
01:47 ilbot3 joined 01:53 BenGoldberg left 01:54 BenGoldberg joined
ZoffixWin m: sub term:<⁢> is rw {state @}; sub postcircumfix:<⁣ ⁤> is rw {$^a[$^b]}; ⁢.push: 42; ⁢⁣1⁤ = 72; say ⁢ 01:56
camelia rakudo-moar 361448: OUTPUT«[42 72]␤»
ZoffixWin ahaha. My job here is done :P
grondilu empty terms should be forbidden imho. I vaguely remember TimToady allowed them some time ago, can't remember why. 01:58
sortiz grondilu, isn't empty, ZoffixWin used an invisible char. 01:59
ZoffixWin grondilu, it's not really empty :P
grondilu of course it's not, but it's just as if
ZoffixWin The 1 = 72 bit is what I'm especially proud of :D
.oO( modify Brainfuck operators to use invisible Unicode chars instead.... ??? PROFIT! )
 02:05
I just need a catchy name
sortiz lol 02:06
geekosaur sounds like that belongs in Whitespace rather than Brainfuck >.>
02:07 noganex_ joined
ZoffixWin But those chars aren't whitespace! 02:07
m: '⁣'.uniprop.say 02:08
camelia rakudo-moar 361448: OUTPUT«Cf␤»
ZoffixWin Oh, there's a language named Whitespace :(
02:10 noganex left
derp_commander indeed there is! 02:12
ZoffixWin came up with the name for his new lang: "Anguish" 02:15
derp_commander haha funny
ZoffixWin perl6 -MAcme::Anguish -e '' <-- all programs will look like that 02:16
02:16 sufrostico left
derp_commander for mine, I was thinking of "Siga" 02:17
02:17 cpage_ left 02:24 cpage_ joined 02:32 BenGoldberg left 02:34 BenGoldberg joined
ZoffixWin jdv79, why were you asking about my next article? 02:36
(it's just now I know what it'll be, and I'll probably churn it out tomorrow)
jdv79 just curious. looking for inspirations. 02:37
02:41 cpage_ left 02:46 kid51 left 02:51 |2701 left 02:57 mr-foobar left 02:58 mr-foobar joined 03:00 pierrot left 03:05 derp_commander left 03:13 mr-foobar left 03:14 mr-foobar joined 03:16 ShimmerFairy left 03:17 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 03:21 AlexDaniel left, ShimmerFairy joined 03:27 Xliff joined 03:29 cpage_ joined
TEttinger ZoffixWin: make all the numbers use factoradic representation, no other way to enter an integer. 03:34
alternately, the numbers use balanced pentary (or whatever base 5 is), the keywords are all emoji, and the math symbols change in meaning depending on the checksum of your source code to that point 03:37
03:38 Sqirrel joined
jdv79 is it possible to use a modifier after a sub call using the colon form? 03:41
m: say: "asdf" if 1 # anyway to make this work? 03:42
camelia rakudo-moar 361448: OUTPUT«WARNINGS for /tmp/_jd51WTq3K:␤Useless use of constant string "asdf" in sink context (line 1)␤»
03:46 kaare_ joined
sortiz m: (say: "asdf") if 1; # Make priority explicit 03:50
camelia rakudo-moar 361448: OUTPUT«WARNINGS for /tmp/fts1ggqF0l:␤Useless use of constant string "asdf" in sink context (line 1)␤»
03:52 khw left
sortiz :P 03:53
04:02 molaf joined 04:06 cpage_ left 04:10 pierrot joined
BenGoldberg TEttinger, perl6 is designed with -Ofun, not with -Oabsurd. 04:11
TEttinger not for perl6, for ZoffixWin's language that needs more horror to be called Anguish 04:12
BenGoldberg Hmm...
04:13 cpage_ joined 04:14 Actualeyes left
BenGoldberg Well, it needs to be eviller than Malbolge, for starters :) 04:14
04:20 mr-foobar left 04:21 mr-foobar joined 04:28 BenGoldberg left 04:30 BenGoldberg joined
TEttinger ah, better. the keywords are all a combination of an APL character (with different meaning than they normally have), one or more Malayalam diacritics applied to an emoji, then a Cyrillic letter. designed to make text entry as hard as possible 04:46
04:47 ggoebel115 joined
TEttinger the only acceptable whitespace is the runic one 04:47
04:48 ggoebel114 left, mr-foobar left
TEttinger hm,   does this work? 04:49
04:50 mr-foobar joined, Cabanossi left 04:52 Cabanossi joined 04:55 jack_rabbit joined 05:02 ZoffixWin left 05:11 BenGoldberg left 05:13 BenGoldberg joined 05:18 mr-foobar left 05:19 mr-foobar joined 05:22 pierre_ joined 05:26 CIAvash joined, mr-foobar left 05:28 huggable joined, mr-foobar joined 05:29 dontTrackme joined 05:34 kurahaupo left 05:36 mr-foobar left 05:37 mr-foobar joined 05:40 pierre_ left 05:44 loren joined 05:45 kurahaupo joined 05:58 mr-foobar left
loren Hi, Perl6 everyone 05:59
05:59 mr-foobar joined
loren Does anyone can help me ? 06:01
sortiz \o loren. 06:03
loren sortiz, \o
I found a problem about class or role reference each other, there is a test at gist.github.com/araraloren/54eab01...dedf30cf73
Can u have a look this problem ? 06:04
sortiz At first seems weird that inside 'something.pm6' you 'use something'. 06:05
06:06 cpage_ left, Actualeyes joined
sortiz Seems the names are inverted. 06:07
loren sorry, wait me
06:07 cpage_ joined
loren the file name is wrong .. re look gist.github.com/araraloren/54eab01...dedf30cf73 06:08
Should be like this . 06:09
sortiz, en 06:10
06:11 buharin joined
buharin hiho 06:11
Xliff \o #perl6
loren \o .
buharin ?? 06:12
sortiz loren, the other problem is that if 'something.pm6' uses reference, 'reference.pm6' *can not* 'use something'. 06:14
loren so i can not use like that ? 06:15
sortiz No, no circular uses allowed.
loren but if i want use *Something* under reference.pm6, what should i do .?
timotimo there's always the "put the things into the same module" and "pull the common dependencies out" 06:17
06:17 broquaint joined 06:18 cpage_ left
timotimo as well as "allow a variable to keep that dependency and resolve it later on when things have settled down" 06:18
raydiak didn't there used to be some helpful error message about circular module loading instead of running until it blows up?
timotimo via the GLOBAL:: namespace
loren yeah, i think compiler should warning me, or report a error 06:19
06:19 cpage_ joined
timotimo yeah, i suppose so 06:19
it's probably just a small matter of programming
that nobody got a round tuit for yet
sortiz Yep, that a can be expected but NYI. 06:20
06:23 pierre_ joined
raydiak github.com/rakudo/rakudo/blob/nom/.../ChangeLog says circular module loading detection was initially added in 2012.08, broke at some point, was fixed in 2015.12, and apparently is now broken again... 06:28
06:29 pierre_ left 06:35 mr-foobar left 06:36 mr-foobar joined
sortiz There is a ticket open: RT#126688 06:38
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=126688
06:39 kurahaupo left 06:45 mr-foobar left 06:46 mr-foobar joined 06:49 cpage_ left 06:50 cpage_ joined 06:51 pierre_ joined, cpage_ left 06:52 pierre_ left 06:55 pierre_ joined 06:56 wamba joined 07:01 mr-foobar left 07:02 mr-foobar joined 07:06 mr-foobar left 07:12 mr-foobar joined 07:13 loren left 07:15 buharin left
timotimo do you folks know how "unscrupulous" (english) is not the opposite of "skrupellos" (german)? 07:16
if we want perl6 to feel more like a natural language, we clearly need something that's exactly spelled like something in another programming language, but actually means the opposite 07:17
huf perl5's continue was p good at this
07:17 dontTrackme left
timotimo like, how about this, we have "unnumeric" mean the same thing as "is this like a number?" 07:17
07:18 dontTrackme joined, buharin joined, firstdayonthejob joined 07:19 skids left 07:22 buharin left 07:27 mr-foobar left 07:28 mr-foobar joined 07:29 buharin joined 07:34 domidumont joined, domidumont left, domidumont joined
sortiz In german the postfix particle 'los' negate, I guess. 07:35
07:35 rindolf joined
timotimo yup, whereas in english postfixparticle "lous" is just "this word be an adjective" 07:36
07:38 espadrine joined 07:39 domidumont left, ssotka left 07:40 domidumont joined
timotimo also, there's a postfix particle "lös" that's not used very often - i think it's actually mostly colloquial perhaps? - that is the opposite of 'los' %) 07:41
skrupulös. fantastic
well, it'd be "skrupellös", but that doesn't sound right at all
07:43 dontTrackme left, firstdayonthejob left 07:46 cpage_ joined 07:50 BenGoldberg left 07:51 BenGoldberg joined 07:56 rodarmor left, BenGoldberg left, rodarmor joined 07:57 BenGoldberg joined, mr-foobar left 07:59 mr-foobar joined 08:10 buharin left 08:11 avalenn left, darutoko joined
masak timotimo: in Swedish, -lös would mean "without ..." 08:12
good antenoon, #perl6
08:20 pierre_ left 08:24 mr-foobar left 08:26 mr-foobar joined
masak how much of an effort do people think it'd be to write a C compiler in Perl 6? 08:27
it feels bigger than the average mini-challenge... but how much bigger?
timotimo ugh. if you can get away with only supporting a C from, i don't know, 1980, perhaps? 08:28
masak I don't think I'm thinking "feature complete" here 08:29
I'm thinking "can compile 'hello world' into a runnable a.out"
timotimo *shrugs*. if you only need hello world, you can cut *all* the corners :D 08:30
08:30 pierre_ joined
timotimo i haven't thought about building a c compiler. but maybe the tcc is a good place to look for inspiration? 08:30
masak url? 08:31
08:31 domidumont left, BenGoldberg left
timotimo not sure if it's internally nice and clean or if it's golf-class code 08:31
08:31 BenGoldberg joined
timotimo bellard.org/tcc/ - this one 08:31
08:31 domidumont joined
timotimo i heard there's a linux livecd that comes not with a compiled kernel but with tcc and the kernel source code 08:31
and it just compiles a kernel for you in a few seconds upon boot
masak cute 08:32
timotimo turns out your code really runs rather ridiculously fast if it fits completely into the cpu's cache or something crazy like that 08:33
29.6 MBytes/second on a 2.4ghz pentium 4 ... that might actually be as fast as the RAM would go in a system that old :P :P 08:36
08:37 cpage_ left, rindolf left 08:40 TEttinger left 08:45 rindolf joined 08:56 pmurias joined
pmurias masak: would a miniature C compiler be mostly a grammar generating assembly? 08:58
* wouldn't
09:00 domidumont left
pmurias masak: seems like for a subset you could squeeze it into something like a pagefull of code 09:00
09:00 domidumont joined
masak pmurias: right, that's what I'm thinking 09:06
could be a cute minimalistic exercise 09:07
I've seen people write minimal compilers online. looks like fun
09:08 spider-mario joined 09:09 Emeric joined 09:12 sortiz left 09:14 BenGoldberg left, pmurias left
timotimo "see people write minimal compilers online" sounds like some livestream kind of deal :) 09:15
09:17 RabidGravy joined
RabidGravy bam! 09:24
09:28 huggable left, huggable joined
Woodi damn, even me is thinking about doing some nice lang :) WhereYouAreMakeWhatYouWant probably ;) 09:30
but last night backlog and yesterday interview with Turbo Pascal author is realy cold water bucket :) 09:31
masak Woodi: why? 09:34
09:34 Ven joined 09:35 Ven left
Woodi masak: if someone want to toy a bit or learn than it's ok. but if someone want to make something usable for others then it's much harder... 09:36
09:37 Ven joined
masak this was true even before the Hjelsberg interview :) 09:38
Woodi and so many things on the way :) dragon book is outdated, lexing is not necesary end evil even, generating asm requires knowing asm :) etc...
masak: yea, probably just my ears got that :) 09:39
anyway, what is so wrong with many pass parsing ? inevitable backdors or something ? 09:40
masak: but making perfect/secure C compiler is nice to have, even if a bit simpler :) 09:41
09:42 pierre_ left, cpage_ joined 09:43 pierre_ joined
tadzik ...lexing is evil? 09:44
moritz premature lexing is the root of all evil! 09:45
Woodi yes, TimToady++ was talkig abou that today... but exactly what that means and how it (don't) work ? and I djb wroted same thing somewhere... 09:46
09:46 grondilu left
Woodi err, similiar thing, djb said "don't parse!"... 09:47
geekosaur do you know about the typedef problem in C? 09:48
to parse something using a typedef, the parser for typedefs has to feed back into the lexer 09:49
otherwise you'll lex a type defined via a typedef as a variable/function/struct field name
if lexing is part of parsing, this can be handled cleanly. if lexing is a separate pass, it gets ugly 09:50
and that's one of the simpler cases
RabidGravy lexing luthor 09:51
tadzik hah, interesting
Woodi lexer is just a tokenizer ? 09:52
geekosaur yes 09:53
Woodi so recipe is to make lexer simpler ?
RabidGravy there was a time that 'lex' and 'yacc' were available on nearly all Unix-like systems, can't remember the last time I used either explicitly
tadzik I do, for my compilers class :P 09:54
arnsholt I used lex and yacc (flex and bison, really) just last week!
Woodi RabidGravy: when you compile something bigger ? :)
at least ./configure search for it :)
geekosaur well, the only reason they got split in the first place is that PDP11s didn't have a large enough address space for a proper parser
arnsholt I had some Python code that had bitrotted (again >.<) and decided to try my hand at implementing it in C
Used lex and yacc to do the input file compilation 09:55
geekosaur so they split the lexer and parser so each one would fit in a 64k address soace
RabidGravy yeah, but I haven't actually *written* anything for lex or yacc since the middle of the nineties I don't thing
geekosaur (and a large part of why lex and yacc existed was to optimize the state tables down to fit that same address space limit) 09:56
masak Woodi: what's wrong with multi-pass parsing? are you familiar with some of the "fun" consequences of multi-pass parsing in Perl 5?
Woodi: they're worth studying, just to realize how wrong things can go.
geekosaur this would be why I went with a simpler case :) 09:57
Woodi masak: not realy, just about that problems exist. but probably multi-pass was more problem for regexes then language ? 09:58
10:00 CIAvash left
masak it's most clear in the regex slang, and Perl 6 properly learned from that 10:02
but I'm pretty sure it shows up elsewhere, too
the canonical example for regexes is... you have //x so you can put in whitespace and even comments 10:03
and then when you put a `/` in a comment... boom 10:04
Woodi x was just promoted by some top peoples, nobody realy used it ;) 10:05
geekosaur uses it a fair bit
nine uses it everywhere
Woodi damn... :)
nine Does that mean I'm one of the top people? :) 10:06
geekosaur regexes are hard enough to understand as it is, if you want other people to be able to figure out what you're doing it helps a lot
ShimmerFairy That description of /x sounds like the default (only?) choice for Perl 6 regexen :P
geekosaur that was a design goal, yes
Woodi so what's so evil in lexing/parsing domain ? 10:09
10:10 kid51 joined
geekosaur basically, whenever you find yourself having to do more and more feedback between the lex and parse phases (which complicates both and makes both harder to understand), it's a strong clue that they shouldn't have been separated in the first place 10:10
the typedef issue I mentioned is the only place this comes up in C. it comes up a lot more in C++, and in perl 5 it turns both phases into crawling horrors 10:11
in perl 5, you have to exhaustively test any change to either one because it's very likely that you will break the whole thing 10:12
Woodi geekosaur: so things into some format like ast ASAP ? 10:13
10:13 CIAvash joined
geekosaur oh, masak, an example of a non-regex thing that breaks stuff is prototypes, which mutate the language at both lexer and parser levels 10:13
in p5 10:14
Woodi probably perl 5 backward compatibility policy is more problematic then parsing... but it's good thing.
10:14 xiaomiao left
geekosaur no, backcompat is why this can't be fixed in perl 5, it's not why the problem is there in the first place 10:14
woodi, more the opposite --- tokenizers are generating a simple AST at a lower level than the parser's output AST 10:15
but you are in trouble as soon as the parser needs to modify the AST that the tokenizer will produce for subsequent input, which happens a lot in real languages 10:17
Woodi geekosaur: this is probably part I don't understand... 10:18
do tokenizer/lexer use modified ast for tokenizing/parsing ? 10:19
...like in typedef example
geekosaur I'm not quite sure what you're asking 10:20
10:20 xiaomiao joined
Woodi parser modifies tokenizer output, why it is problem for tokenizer ? 10:20
geekosaur in the typedef case, something that would normally be lexed as <ident> has to be lexed as <type>, or you will do the wrong thing 10:21
you are not modifying the tokenizer output. you are modifying the language the tokenizer recognizes, out from under the tokenizer
Woodi so, lexer doing parsing is not needed, just tokenizer ? 10:23
probably EDEFINITIONSAREBLURRY here... 10:25
geekosaur in effect, parsing a typedef `typedef unsigned long size_t;` produces a new variant of C that has a new type in it (size_t). if you don't mutate the lexer then a subsequent `extern size_t foo()` produces a lex AST [T_EXTERN, T_IDENT<size_t>, T_IDENT<foo>, '(', ')'] which doesn't match the parse rule [T_EXTERN, OPTIONAL T_TYPESPEC, T_IDENT '(']
er, missing comma in that last part: `, T_IDENT, '(']` 10:26
(lexer ASTs aren't very treelike normally. if needing to mutate the lexer from the parser is a strong clue you've made a mistake, outputting a tree from the lexer is a flaming red flag) 10:27
(because it means you've put part of the parser in the lexer to try to compensate for having split the lexer from the parser) 10:28
Woodi geekosaur: I think I understand something now :) 10:31
btw. modern languages don't use something like extern ? 10:32
geekosaur they usually have something like it. it's a forward declaration
Woodi ook :) are there things in academia that can be used instead of ast ? 10:33
geekosaur C was originally intended to be very simple (ANSI C has changed this significantly) so they used a special keyword
Woodi geekosaur: yea, I was thinking few times: WHERE IS TypeC ?? 10:35
geekosaur and this was pretty much just to simplify the parsing rules to fit in a tiny address space
academia still uses ASTs. I'm not aware of a replacement; about the only viable one is to do it the way neurons do, which is highly unstructured and difficult for people to understand. (admittedly perl 5 has more or less ended up this way p ) 10:36
:p
10:37 mr-foobar left, pierre_ left
RabidGravy dear CSS if you don't stop screwing me around I'm going to go back to 1997 and use tables for layout 10:39
geekosaur noooooooooooo 10:40
10:44 lue joined, ShimmerFairy left, lue is now known as ShimmerFairy 10:47 Ven left 10:49 wamba left 10:51 wamba joined
Woodi geekosaur: what it means perl 5 is working like neurons ? google don't answered me for that... 10:55
10:58 cpage_ left 10:59 lizmat left
geekosaur the brain's neural networks have both forward and backward feedback, over multiple channels 11:02
11:04 Ven joined
RabidGravy goes with perl6.noisegang.com/ and damn the torpedoes 11:04
11:05 domidumont left
geekosaur so, for example, researchers recently discovered that if the higher level brain "circuits" don't expect to see something, the lower level circuits will "erase" it completely. you can literally not see something right in front of you, as a result. 11:05
Woodi RabidGravy++ :) 11:06
geekosaur: so, chemicals/neurotransmitors need to be nullified somehow... 11:09
geekosaur www.sciencedaily.com/releases/2016...124822.htm
DrForr Any docker users handy?
RabidGravy .tell lizmat if you get a minute is there any chance of giving the "Noise Gang" a mention in the p6w next week? 11:10
yoleaux RabidGravy: I'll pass your message to lizmat.
geekosaur neurotransmitters aren't the only channel. electric fields are used for feedback, it turns out 11:13
timotimo evolution tends to be sensitive to basically every single little detail 11:14
like when that genetic algorithm designed circuit boards where there was elements not connected to anything, but removing them caused things to no longer work 11:15
Woodi just wants faster cars... 11:16
or maybe less damn road signs.... 11:17
stmuk_ escaped tigers in Holland! perl6 team there beware! 11:35
Woodi funny thing, I was taught that hearing is more important then hearing :) "שמע ישראל"
*then vision
11:42 mr-foobar joined 11:47 mr-foobar left 11:48 mr-foobar joined 11:53 |2701 joined 11:59 _4d47 joined 12:03 kid51 left 12:05 kid51 joined 12:09 Actualeyes left 12:10 Actualeyes joined 12:11 kurahaupo joined 12:14 domidumont joined, domidumont left, brabo left 12:15 brabo joined, domidumont joined 12:21 kurahaupo left, kurahaupo joined 12:23 chenryn joined 12:26 Emeric left, Emeric joined 12:31 dogbert17 left 12:45 araujo_ joined 12:46 araujo_ left 12:47 araujo_ joined 12:48 araujo left, araujo_ left 12:49 araujo_ joined 12:50 _mg_ joined 12:53 Sqirrel left 12:54 pmurias joined
pmurias ShimmerFairy: you can turn on significant white space in Perl 6 12:55
12:56 wamba left 12:58 darutoko- joined, _4d47 left
pmurias Woodi: the problem with multi pass parsing is that you wast time parsing things multiple times 13:00
13:01 darutoko left 13:02 Ven left
pmurias Woodi: it's a problem for both compilers and human readers 13:04
13:14 dogbert17 joined 13:17 tbrowder left
|2701 is distributed computing something that perl6 is able/going to be able to do? 13:21
pmurias |2701: there shouldn't be anything stopping perl6 from doing it 13:23
|2701: isn't sockets enough to do that?
* aren't
|2701 I was envisioning something like Ruby's drb, where objects can be sent and instantiated on a client device 13:25
13:26 itcharlie joined, jdv79 left 13:28 huggable left, huggable joined 13:32 jdv79 joined 13:37 tbrowder joined 13:43 tbrowder left 13:46 kaare_ left
pmurias |2701: to familiar with drb besides looking at the tutorial but it looks like something that should be possible to make 13:57
14:02 Ven joined 14:10 Sgeo joined 14:12 brabo left, Sgeo_ left, dmc_ left 14:14 Xliff left 14:16 dmc00 joined 14:21 notbenh__ joined 14:23 notbenh_ left, notbenh__ is now known as notbenh_, SmokeMachine____ left 14:24 kipd joined, SmokeMachine____ joined, Lucas_One joined 14:26 dmc00 left 14:27 tbrowder joined
tbrowder hi perl 6 people 14:27
14:28 dmc00 joined 14:30 obarb joined
tbrowder a bomb: I have been spending time with Perl 6 pod and wonder why we don't toss it for embedded markdown or some other well-supported markup language? 14:31
14:31 obarb is now known as brabo 14:32 Ven left
tadzik markdown would have to be extended in quite a few ways to reach Pod's featureset 14:34
14:34 Ven joined
tbrowder Maybe a kinder thing to do is not to use Perl 6 pod for stand-alone docs as is being done now on doc.perl6.org. 14:35
tadzik why not?
is it deficient in any way?
tbrowder For example, tables are ver rudimentary at the moment (formatting codes not recognized). 14:37
masak tbrowder: I don't think Perl 6's Pod format is mandatory in any way. I can definitely see the attraction in writing Markdown instead. and I don't see the problem with that.
14:37 Ven left
tadzik well, a switch to markdown would give us the complete lack of any tables, wouldn't it :P 14:37
masak dunno. does CommonMark have tables? don't remember.
Github's flavor does, for example 14:38
14:38 rvchangue joined
tadzik oh, I didn't know 14:38
tbrowder There are other things like sphinx or asciidoc I think can do that.
masak I suspect that if the Perl 6 effort started in 2015 instead of 2000, Markdown would have been at least considered as the docs format.
I'm not all that sure it would've been adopted, but it would've been considered.
anyway, there's nothing that prevents a driven individual from creating excellent tooling for Perl 6 + Markdown rather than Perl 6 + Pod 14:39
tbrowder Can they not co-exist? Use pod for embedded stuff and something else for stand-alone. 14:40
tadzik I'm pretty sure you can have a slang of Perl 6 that uses MD instead of Pod, would be a very interesting experiment indeed :)
14:40 cbk joined 14:41 pmurias left
masak in fact, that's what slangs/braids are there for 14:41
dogbert17 afternoon #perl6
masak if we don't support that use case, it simply means that we are Not Done Yet
tbrowder I never want to reinvent wheels but I think using pod for stand-alone is kind of like that.
14:42 pmurias joined
arnsholt has been too long in academia. The only acceptable text formatting language is LaTeX (with a liberal sprinkling of extension packages on top) 14:42
tadzik just got a year older thinking about his thesis that was ~4 phases of TeX processing 14:43
dogbert17 does anyone want to take a look at an attempt to document X::Promise::Vowed and shout if there are any errors? gist.github.com/dogbert17/cce80367...5bc2d71d91
14:43 kid51 left
tbrowder Hm, can't think of name at the moment but there is a program that handles all kinds of filtering between markip languages. 14:44
arnsholt tadzik: I have a 77-line Makefile driving the compilation process, and 6490 lines of TeX
geekosaur pandoc?
14:44 khw joined
tadzik oh gods 14:44
arnsholt The main thesis.tex document is 301 lines long
There are 29 \usepackage statements, and a pile of different customizations 14:45
14:45 Xliff joined
tadzik I had lilypond first, then latex twice (because index ¯\_(ツ)_/¯) and bibtex somewhere in the middle 14:46
I don't even want to know
arnsholt Yeah, the whole multiple compilation thing is kind of annoying
tadzik I think I lost the original .tex files when my hard drive encryption died
arnsholt I use biber and biblatex which enables even more hilarious customization
pmurias tbrowder: is the problem with Pod 6 the tooling or the language itself? 14:47
arnsholt Like on-the-fly changing of article titles between the .bib file and the bibliography
tadzik tbrowder: is the problem with Pod 6 the tooling or the language itself?
(sorry)
arnsholt (My boss, for some bizarre reason, changes colons between title and subtitle to periods. So I can apply s/\.(?=\s)/:/ via biber =D 14:48
)
14:48 tbrowder left
arnsholt In case it's not abundantly clear, I've been using LaTeX for way (*way*) too long 14:49
pmurias we need to switch from using github to display our module docs to a proper metacpan equivalent at some point
14:52 cdg joined 14:53 cbk left, skingsbu joined 14:54 brabo left 14:57 cbk joined, chenryn left 14:58 chenryn joined 15:00 cbk left 15:01 brabo joined 15:03 CIAvash left
dalek c: bafa6e5 | (Jan-Olof Hendig)++ | doc/Type/X/Promise/Vowed.pod:
Added docs for X::Promise::Vowed
 15:05
15:10 brabo left, travis-ci joined
travis-ci Doc build errored. Jan-Olof Hendig 'Added docs for X::Promise::Vowed' 15:10
travis-ci.org/perl6/doc/builds/130230848 github.com/perl6/doc/compare/edf96...fa6e5f87e3
15:10 travis-ci left
dogbert17 WAT 15:12
Hmm, 'The command "rakudobrew build-panda" failed and exited with 2 during .' according to travis-ci 15:13
can't say I know how to fix that 15:15
15:17 zakharyas joined
Woodi do pdf have tables ? if not then using it as archiving format is DUMB... will produce similiar problems like NASA (probably) have with recovering data from archaic tapes without a archaic reader 15:19
15:21 brabo joined
geekosaur pdf is a stack-based programming language 15:22
it doesn't have much of anything "built in" 15:23
Woodi geekosaur: so just asci(2|to)html ommits tables and just uses raw text location ? 15:24
geekosaur hopefully it does that much of rendering. it could just extract the strings in whatever order they are in the PDF program, and hope it's something sensible 15:25
Woodi text is :) just sequencing not :) 15:26
15:26 cbk joined
geekosaur yes, sequencing gets lost usually 15:27
15:27 hanekomu joined
geekosaur to keep it, you'd have to be able to interpret enough PDF to call any functions that are defined in the file (this is very common) and find the moveto-s, and then render text into a buffer according to those positions. and then rationalize it for lines instead of pixels 15:29
Woodi but maybe we wil have smarter OCRs some time bedore XXII century :) 15:30
but placing data into format without some container notation is just stupid...
15:33 wamba joined
dogbert17 is moritz around? 15:33
.seen moritz 15:34
yoleaux I saw moritz 09:45Z in #perl6: <moritz> premature lexing is the root of all evil!
15:35 _4d47 joined
arnsholt Woodi: To follow up on geekosaur's point, PDF is fundamentally a graphics format. There's a fair amount of support for text, since it was designed for mostly textual documents, but it's still graphics 15:38
RabidGravy boom
15:42 cbk left
moritz dogbert17: on and off 15:42
15:43 brabo left
dogbert17 moritz: I have a question or two about the type-graph.txt file 15:43
I have added a few missing exceptions to the docs lately, am I correct in assuming that I should add them to the type-graph.txt as well? 15:44
moritz dogbert17: yes 15:45
(unless it's already there, of course, which is the case for some of the undocumented types)
dogbert17 moritz: great, I also noticed that X::Obsolete occurs twice. Is that one time too many? 15:46
moritz dogbert17: yes 15:47
dogbert17 moritz: I'll remove the duplicate then
_4d47 does anyone have time to review github.com/perl6/mu/pull/14 ? i'm not confident to commit
dogbert17 moritz: X::Augment::NoSuchType also occurs twice under two different headings, but I can only find one such type in the source 15:48
15:50 brabo joined, chenryn left
dogbert17 moritz: this one I cant find 'class X::Augment::NoSuchType does X::Syntax' 15:50
dalek c: a385760 | (Jan-Olof Hendig)++ | type-graph.txt:
Added X::Promise::Vowed, X::IO::Move and X::IO::Symlink to type-graph.txt and removed a duplicate of X::Obsolete from same
 15:59
16:13 zakharyas left 16:19 travis-ci joined, firstdayonthejob joined
travis-ci Doc build passed. Jan-Olof Hendig 'Added X::Promise::Vowed, X::IO::Move and X::IO::Symlink to type-graph.txt and removed a duplicate of X::Obsolete from same' 16:19
travis-ci.org/perl6/doc/builds/130241760 github.com/perl6/doc/compare/bafa6...8576051459
16:19 travis-ci left 16:27 jack_rabbit left 16:32 Sunbeam_ joined 16:37 aries_liuxueyang left 16:43 aries_liuxueyang joined 16:46 buharin joined 16:47 pmurias left, dmc00 left 16:48 dmc00 joined, dmc00 left, dmc00 joined 16:50 chenryn joined 16:53 dmc00 left, dmc00 joined 16:59 pmurias joined
pmurias Woodi: I don't think pdf readers getting lost is a serious problem 17:00
geekosaur yeh, I was thinking that -- if Adobe were the only source there'd be a problem, but there's lots of open source pdf readers/renderers 17:01
17:03 jack_rabbit joined
pmurias there seems to be a standard version for archiving documents 17:06
* standardized version of the format
moritz the biggest problem with PDF is probably that the newer standards are so friggin huge that nearly nobody implement them in full 17:11
with specs for 3D models and embedded videos and stuff 17:12
(and the fact that it's hard to extract text out of a PDF, but that's a totally different tangent, and probably an intentional tradeoff)
17:15 jack_rabbit left 17:19 pmurias left, dmc00 left 17:20 dmc00 joined 17:22 CIAvash joined, dmc00 left 17:23 dmc00 joined, dmc00 left 17:25 dmc00 joined 17:28 huggable left, huggable joined 17:33 CIAvash left 17:37 kurahaupo left
buharin hiho :) 17:41
17:46 sortiz joined 17:47 CIAvash joined 17:48 Sunbeam_ left
sortiz \o #perl6 17:48
17:50 kurahaupo joined 18:06 chenryn left 18:07 dmc00 left
dalek c: 80d6804 | (Jan-Olof Hendig)++ | doc/Language/concurrency.pod:
Fixed a broken link
 18:11
18:18 mr-foobar left, travis-ci joined
travis-ci Doc build errored. Jan-Olof Hendig 'Fixed a broken link' 18:18
travis-ci.org/perl6/doc/builds/130263316 github.com/perl6/doc/compare/a3857...d68043e9c7
18:18 travis-ci left 18:19 mr-foobar joined 18:21 CIAvash left 18:31 BenGoldberg joined 18:33 ssotka joined 18:34 dmc00 joined 18:35 kaare_ joined 18:37 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 18:40 itcharlie left 18:41 araujo_ left 18:42 araujo_ joined 19:03 cpage_ joined 19:07 BenGoldberg left 19:09 BenGoldberg joined 19:13 mr-foobar left 19:14 Ven joined, mr-foobar joined 19:22 buharin left 19:24 kaare_ left, kaare_ joined
masak has been drawing SVG paths using Perl 6 all day 19:24
while also fantasizing about a tool that will DWIM a lot of what I'm doing for me
TimToady WAT?!? 19:26
moritz TimToady++ 19:27
19:28 buharin joined
buharin huh 19:29
19:29 Ven left
buharin someone use emacs here? 19:29
19:33 travis-ci joined
travis-ci Doc build errored. Nick Logan 'Fix and improve CI testing' 19:33
travis-ci.org/ugexe/doc/builds/130274050 github.com/ugexe/doc/compare/less-...firetravis
19:33 travis-ci left, kaare_ left
masak yeah. with the amount of DWIM I'm picturing, there's bound to be some WAT too. 19:37
but (naturally) I imagine it'd be worth it ;)
timotimo i draw svg paths with graphviz often 19:38
masak one always does, I guess...
timotimo it's not always perfect
masak timotimo: I've been known to use Inkscape a lot
timotimo: today I've been longing for something that's a bit more of a hybrid between Inkscape and programming
timotimo tikz? :)
arnsholt Yeah, TikZ is nice =) 19:39
geekosaur buharin, I use emacs
timotimo i wrote a python program at some point that output'd tikz 19:40
masak tikz looks nice, indeed 19:41
timotimo comes with the usual properties tex/latex has 19:42
wamba buharin, I use emacs too 19:44
buharin I tried emacs right now 19:45
with prelude
timotimo i have heard fantastic things about spacemacs 19:46
buharin it is fine but pretty strange ;P
mst I installed that then never got around to trying it properly
arnsholt Speaking of TeX/LaTeX, this cropped up on my Facebook just now: pbs.twimg.com/media/CiMNh9rWEAAdM6Q.png:large =D =D =D
buharin do you use project tile?
timotimo oh yes
geekosaur just uses plain emacs with standard major modes tbh 19:48
timotimo twitter.com/loltimo/status/474577064503554048
moritz arnsholt: :-) 19:49
buharin you know
after I used atom for programming last time
I don't see a advantage
about using emacs for devloping
specially I am using my customized vim over ssh 19:50
mst well, fine, keep using vim then
editor wars are boring
buharin yah 19:51
geekosaur use what works for you
buharin but the essence
moritz arnsholt: misses "badness: 1000"
buharin is if ui editor are better than console
geekosaur played with atom, would like to see it address memory leaks and poor coexistence with chrome's v8 before trying it again
mst buharin: uh. then why bring emacs into it? 19:52
geekosaur (but hey, at least it's worth trying again someday. lots of editors don't make that cut for me)
mst emacs is both
arnsholt moritz: True, true =D
buharin ;-)
geekosaur on most systems it is. os x ships a terminal-only emacs 19:54
mst right, because OS X GUI apps aren't delivered that way
there's also an emacs-nox package on debian
but that's not really the point
19:55 cpage_ left
buharin actually I am using virtual debian on os x 19:55
:P
19:55 yqt joined
buharin but there is cocoa emacs 19:56
masak arnsholt: autopun spotting on that pbs.twimg.com/media/CiMNh9rWEAAdM6Q.png:large link ;)
20:02 chenryn joined
arnsholt masak: Definitely! 20:04
20:04 kid51 joined 20:06 chenryn left
arnsholt masak: Doesn't look like its set in Times New Roman though =) 20:07
moritz tex defaults to Computer Modern, no? 20:08
mspo visual studio code?
20:08 Actualeyes left
arnsholt moritz: Yeah. But I don't think it's CM either 20:09
moritz arnsholt: right
20:19 ssotka left 20:20 travis-ci joined
travis-ci Doc build passed. Nick Logan 'Travis: Install Pod::To::HTML' 20:20
travis-ci.org/ugexe/doc/builds/130279251 github.com/ugexe/doc/compare/69cf7...d16e5acb40
20:20 travis-ci left 20:25 domidumont left 20:28 darutoko- left 20:31 buharin left, _dolmen_ joined 20:39 _dolmen_ left 20:41 _mg_ left, cpage_ joined 20:42 cpage_ left 20:44 hotel joined 20:46 buharin joined 20:48 cpage joined
dalek c: d8d67cf | (Jan-Olof Hendig)++ | doc/Language/functions.pod:
Another five broken links fixed
 20:49
20:49 Zoffix joined 20:58 wamba left 21:03 TEttinger joined 21:04 travis-ci joined
travis-ci Doc build passed. Jan-Olof Hendig 'Another five broken links fixed' 21:04
travis-ci.org/perl6/doc/builds/130285143 github.com/perl6/doc/compare/80d68...d67cfc5f8e
21:04 travis-ci left, mr-foobar left 21:05 mr-foobar joined 21:08 chee is now known as italy 21:09 italy is now known as chee, aries_liuxueyang left 21:10 aries_liuxueyang joined 21:12 Zoffix left 21:19 crucialr_ joined 21:20 crucialr_ left, skink joined 21:22 Zoffix joined 21:25 Zoffix left 21:27 ssotka joined 21:28 huggable left, huggable joined 21:31 BenGoldberg left 21:32 BenGoldberg joined 21:33 BenGoldberg left 21:34 BenGoldberg joined 21:35 BenGoldberg left 21:36 BenGoldberg joined 21:40 _dolmen_ joined 21:41 BenGoldberg left
ugexe stmuk_: github.com/perl6/doc/pull/514 this should allow p6doc to discover any installed module, and not depend on reading the panda cache file 21:41
21:42 BenGoldberg joined
masak MadcapJake: I finally backlogged over the "transpiler" thing :) 21:42
MadcapJake: here's my take, fwiw -- a compiler is something that turns a language-A program into "the same" language-B program, fsvo "the same" 21:43
most of the time language B is machine code, but not always
thus there is a perfectly good word for transpiler already: "compiler"
even the "source-to-source" feels a little bit unnecessary in my mind: what's human-readable source code and what's machine-executable target code is partly a matter of context 21:44
mst transpiler mostly means to mean "compiler, except targeting something originally designed for being written by humans rather than targeted by compilers" 21:45
timotimo gcc has a flag to keep .s files around :P
masak mst: I buy that.
timotimo assembly code is definitely designed for being written by humans
21:45 Zoffix joined
timotimo that's why it's for punchcards 21:45
masak timotimo: most compilers don't compile to assembly, though. they compile to machine code. 21:46
mst masak: so, like, Transpiler isa Compiler but the word still provides additional context that's sometimes useful
skink timotimo, I'm still trying to figure out fast thread-safe reads
21:47 travis-ci joined
travis-ci Doc build passed. Nick Logan 'Discover all installed modules + Less deps 21:47
travis-ci.org/ugexe/doc/builds/130290094 github.com/ugexe/doc/commit/8bdd3ff54cde
21:47 travis-ci left
skink Using start {} is also extremely slow and I don't think it was solving the problem anyway 21:47
21:48 BenGoldberg left
masak mst: I... think I grudgingly agree. 21:48
mst: only my dislike of the word and the contexts in which it's used now remain ;)
21:49 BenGoldberg joined
mst masak: I have no good answer for that since I don't honestly like the word much either, I just found that it can actually have semantic value 21:51
well, and also, sometimes it provides the semantic value of "the way this was used in context means I can close the tab now and not lose anything"
timotimo skink: you're not getting very far as long as we're trying to allocate multiple stack frames per second 21:52
dalek c: 8bdd3ff | ugexe++ | / (3 files):
Discover all installed modules + Less deps

Discover all installed modules, not just those listed in pandas cache. Also remove dependencies on panda altogether.
c: 8ec85d7 | (Steve Mynott)++ | / (3 files):
Merge pull request #514 from ugexe/discover-installed

Discover all installed modules + Less deps
stmuk_ ugexe: thanks
masak mst: twitter.com/pmrondon/status/648251230804946944 21:54
21:54 BenGoldberg left
timotimo skink: you can try doing stuff with nqp:: ops and try to use as few curlies as you can 21:54
and you just might see your performance start scaling properly with the number of cors 21:55
21:55 BenGoldberg joined
mst masak: yep. I just only *sometimes* regard that as true 21:56
sometimes I grit my teeth and grudgingly admit their usage was useful
21:58 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg, _dolmen_ left, BenGoldberg left
stmuk_ australia? 21:58
21:59 BenGoldberg joined
masak stmuk_: shiny new Australia. 22:01
22:03 chenryn joined, kjk joined 22:04 Emeric1 joined, Emeric left, Emeric1 is now known as Emeric 22:05 Emeric left
kjk is there a way to reload a module from REPL? say I 'use MyModule' from REPL and then make some changes to the source code of MyModule, how can I easily "reload" it in the REPL so I can test it without restart the REPL and load my module again? 22:06
22:07 chenryn left
dalek c: bc8f5b2 | (Jan-Olof Hendig)++ | doc/Language/operators.pod:
A few more broken links
 22:07
22:07 BenGoldberg left 22:09 BenGoldberg joined
masak kjk: I'm not aware that's a use case that's covered by the REPL, or module loading. 22:10
though it would be kinda nice to have a module hotpatching story.
22:12 buharin left, BenGoldberg left 22:14 BenGoldberg joined 22:15 buharin joined
kjk masak: I see..., yeah, it would be quite nice for trying things out in REPL. Thanks 22:15
buharin hi 22:16
22:17 BenGoldberg left 22:19 BenGoldberg joined 22:26 Zoffix left 22:29 BenGoldberg left
skink timotimo, Holy hell 22:30
25,000 open/read/closes in NQP takes 0.4s
22:31 BenGoldberg joined
skink In P6 it's 46.3s 22:31
22:34 BenGoldberg left 22:35 buharin left 22:43 Ven joined 22:57 kjk left 22:59 chenryn joined 23:09 skingsbu left 23:10 tbrowder joined 23:12 Zoffix joined 23:14 tbrowder left 23:19 rindolf left 23:22 skingsbu joined, spider-mario left 23:23 firstdayonthejob left 23:24 kanishka joined 23:26 Zoffix left 23:31 Zoffix joined
Zoffix m: 'test'».uc.say 23:47
camelia rakudo-moar ad8265: OUTPUT«(TEST)␤»
Zoffix Oh, god. This is awesome!! I just got myself a Perl 6 keyboard!
masak .oO( useless use of » ) 23:48
m: say 'test'.uc
camelia rakudo-moar ad8265: OUTPUT«TEST␤»
Zoffix 18 extra keys to use for Unicode ops :) twitter.com/zoffix/status/731631393164414976
masak nice, though :)
Zoffix masak, yeah, I was just testing my shiny (in pink!) keyboard :D 23:49
gfldex you possibly couldn't miss a key on that keyboard 23:50
any typo will make you look like an imposter 23:51
Zoffix heh. I'm kinda still getting used to it. After my MS Curve 2001 with giant keys, this one feels tiny. And I think they actually made the keys SMALLER than regular keys to fit the 18 extra ones :/
I'm enjoying the RGB lights more than I thought I would though lol :D 23:52
masak 'night, #perl6
Zoffix night \o 23:54

Powered by Cro and the Raku® Programming Language.

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

Thank you!