»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
[Coke] Am I the only one who heard that read in the style of monty python? 00:04
00:11 Chillance left
lichtkind good night 00:12
kthakore good night
masak good ♞
lichtkind hope you can use this tommorrow gist.github.com/4341223 00:13
00:13 lichtkind left 00:23 LV left, LordVorp joined
kthakore Hello I am trying to understand a bit of teh reasoning being Perl6 syntax. 00:25
when I use for @foo -> $bar { }
japhb_ Erg! I am really tired of Bug #115390! 00:26
kthakore is that running as a for function?
japhb_ kthakore, -> $arg1, $arg2 { } is a "pointy block"
It takes arguments like a Routine, but is not quite a full Routine (it is transparent to return, for example) 00:27
So essentially you have "for @array &function", or rather [email@hidden.address] &function", because for is really just syntactic sugar for map
00:29 cognominal left, cognominal joined
TimToady that's the one syntax in Perl 6 that is allowed to have two terms in a row, by treating -> as an expression terminator 00:29
we do not (yet) allow users to tap into that syntax by mere function declaration 00:31
dalek rl6-bench: 3d37b43 | (Geoffrey Broadwell)++ | perl5/rc-man-or-boy-test:
Make sure perl5/rc-man-or-boy-test is run with 5.010 features enabled
 00:32
rl6-bench: 1a6126e | (Geoffrey Broadwell)++ | / (2 files):
Build Perlito as a component, and switch to 'nodejs' instead of 'node' to call node_js VM
kthakore cool 00:33
oh cool 00:34
kthakore is just reading book a lot 00:35
thanks
00:40 cognominal left 00:50 cognominal joined 00:55 cognominal left 00:56 japhb_ left 00:57 bowtie left 01:01 larks left 01:03 larks joined 01:04 slayer101_ joined 01:06 cognominal joined 01:12 hypolin joined 01:14 cooper_ left, cognominal_ joined, cognominal left 01:15 anuby joined 01:23 cognominal_ left 01:30 cognominal joined 01:32 gdey left 01:35 cooper joined, cooper left, cooper joined 01:56 skids joined 02:33 arlinius left 02:58 FROGGS_ joined 03:02 FROGGS left 03:07 noggle joined 03:11 noggle_ left 03:14 cognominal left, hypolin left 03:16 hypolin joined 03:40 arlinius joined 03:59 djanatyn left 04:01 djanatyn joined 04:12 slayer101_ left 04:14 slayer101_ joined 04:28 ilogger2 joined, ChanServ sets mode: +v ilogger2, groky joined 04:30 popl joined 04:31 mathw joined 04:32 noggle joined
diakopter waves at jnthn waking up to go to the airport 04:43
04:51 MikeFair_ joined 05:11 benabik joined 05:16 adu joined 05:22 SamuraiJack joined 05:29 benabik left 05:30 telex joined 05:37 benabik joined 05:40 rummik joined 06:26 donghongbo joined 06:45 donghongbo left 06:46 adu left 06:47 kaleem joined 07:05 sisar joined
moritz o\ 07:08
sorear \o 07:18
popl ^o 07:22
popl sucks
07:50 FROGGS_ joined 08:08 FROGGS_ is now known as FROGGS
FROGGS does the Perl6::Metamodel::MultiDispatcher.vivify_for method is meant to bring the dispatcher to life? I don't get the sense of "vivify"... 08:14
morning all you happy (still living) perl 6 hacker 08:15
08:16 brrt joined
timotimo oh, yes, what kind of world-ending was that? 08:17
brrt its not done yet
timotimo i guess the writers decided there was too much still hanging in the air
and didn't want it to be a total cliffhanger
FROGGS maybe the release is just late 08:21
timotimo well, they've had *lots* of time to prepare 08:22
a few thousand years, isn't that right?
brrt its the last time i rely on calendar writers for anything 08:24
timotimo no calendar post yet for today? 08:29
FROGGS I heard colomon had one... 08:30
it's somewhere in the backlog
ahh, about 10 hours ago he said he has a draft 08:31
08:33 kresike joined
kresike hello all you happy perl6 people 08:33
brrt hi kresike
kresike brrt, o/ 08:34
timotimo well, there's still lots of time
08:34 Kharec joined 08:44 ashleydev left, rjbs left, nuba joined 08:46 Timbus joined, sahadev joined, sivoais joined, betterworld joined, Tene joined, ashleydev joined, rjbs joined, sunnavy joined
FROGGS jnthn: you remember the trait_mod:<is> import problem if we import two of them from different modules? 08:58
jnthn: I might know the solution and I'd like to have a comment: 08:59
instead of importing the trait_mod to our lexpad directly, we should instead add a proto and the two trait_mods as dispatchers, right?
moritz erm, isn't the proto a dispatcher? 09:03
I'd rephrase
add a proto, and add the candidates from both to-be-imported trait_mod:<is> to the new proto
and make sure not to add the same thing twice over two routes 09:04
(not that I'm jnthn or so :-)
hoelzro morning #perl6 09:07
moritz \o hoelzro 09:10
FROGGS thanks moritz 09:13
so I'm not on the wood way *gg* 09:14
09:15 bapa joined
moritz FROGGS: I don't think that idiom works in English :-) 09:17
phenny: "Holzweg"?
phenny moritz: "Holzweg" (de to en, translate.google.com)
moritz "auf dem Holzweg sein" - "to be on the wrong track", blind alley, whatever 09:18
FROGGS was a joke actually
moritz I know, just rambling for the benefit of our non-German speakers 09:21
FROGGS r: use Test; say Test.WHO 09:26
p6eval rakudo 55dbc0: OUTPUT«("EXPORT" => EXPORT).hash␤»
FROGGS r: use Test; say Test::EXPORT.WHO
p6eval rakudo 55dbc0: OUTPUT«("ALL" => ALL, "DEFAULT" => DEFAULT).hash␤»
FROGGS r: use Test; say Test::EXPORT::ALL.WHO
p6eval rakudo 55dbc0: OUTPUT«("\&plan" => sub plan(Mu ) { ... }, "\&pass" => sub pass(Mu ) { ... }, "\&ok" => sub ok(Mu ) { ... }, "\&nok" => sub nok(Mu ) { ... }, "\&is" => sub is(Mu ) { ... }, "\&isnt" => sub isnt(Mu ) { ... }, "\&is_approx" => sub is_approx(Mu ) { ... }, "\&todo" => sub tod…
FROGGS hmmm
moritz: is there a module/grammar installed that has multis? 09:27
ohh, Test.plan() is a multi 09:28
09:30 bapa left 09:32 bapa joined 09:41 Psyche^ joined 09:43 sisar left
hoelzro how do src/stage* relate to each other in NQP? I'm guessing that stage0 is bundled with NQP, and is used to compile NQP itself into stage1? 09:43
09:43 bapa left 09:44 Psyche^ is now known as Patterner 09:45 bapa joined
moritz correct 09:45
and then stage1 compiles the sources into stage2
which is the compiler we install
hoelzro moritz: the same sources that stage0 compiles into stage1? 09:46
moritz hoelzro: yes, there's only one set of sources for nqp
hoelzro I figured; just checking =)
moritz and 'make bootstrap-files' copies the stage2 files into stage0
09:47 dakkar joined
hoelzro ths is probably a stupid question, but why do you have to compile stage1 into stage2, instead of just using stage1? 09:47
09:49 sdo joined
sdo happy birthday perl 09:49
moritz hoelzro: it's not stupid at all, and I'm not sure about the answer 09:50
hoelzro: I think it's a kind of extra test
hoelzro I see
moritz compiling NQP itself is a better test than running the nqp tests (which aren't nearly as comprehensive as Perl 6 test suite) 09:51
there could also be some bootstrap problems when not having enough stages; I'm not sure about that one though 09:56
09:57 SmokeMachine joined, sdo left
hoelzro well, I think GCC does that; it uses a bootstrap compiler to build itself, and then it uses the fresh compiler to build itself once again 09:59
probably to make sure new optimizations are present, or something
moritz GCC also has a build command which makes a binary comparison of the results of the different bootstrapping changes 10:00
NQP can't do that, because it generates identifiers that have timestamps attached 10:01
hoelzro ah ha
moritz so the results will never be binary identical
bbkr__ \o/ Warszawa release 10:06
hoelzro so what exactly does "pir::compreg__Ps('QAST').post($source)" do? something like create a new QAST object and assign $source to the post attribute? 10:12
timotimo i'm trying to use jnthns optimiser transformation code, but it doesn't seem to register nt^H^Hthe nqp compilr 10:13
i don't know how to work this
also my ssh is laggy like hell :|
i just perl Configure.pl --gen-nqp=the-correct-commit and it seems to have worked. then i makee'd rakudo and make installed it
what else would i have to do? 10:14
moritz hoelzro: parrot has a global registry for compilers
hoelzro so 'QAST' is a compiler object name?
moritz hoelzro: and pir::compreg__Ps($some_string) gets the compiler registered for $some_string
right
and then .post($source) just calls the .post method on it
hoelzro oh, ok!
moritz timotimo: that should be enough 10:15
hoelzro moritz: thanks for walking me through this =)
moritz you're welcome
timotimo oh, excuse me, i meant to say: pir::load_language__Os didn't work 10:16
it ^H ^H^H^H^H^H^H^H^H^H^H^H^H^H^H^Hcompreg does work
10:16 fgomez joined
timotimo the argument to load_language is NQP, is that correct? or maybe it sokuld be nqp? 10:16
moritz probably same case as compreg
but iirc the install paths need some adjustment for load_language to work 10:17
timotimo how it works with nqp instead of NQP,, cool! 10:18
now let's transform some QASTs :)
10:19 fhelmberger joined
timotimo i would love multiline editing in perl6 :~ 10:20
i mean in the default rakudo interpreter
10:21 fhelmberger left 10:32 bowtie joined
moritz timotimo: that's a hard one to crack 10:33
10:35 popl left 10:37 rindolf joined
bbkr__ LWP::Simple fails almost every test with "could not parse headers" on latest Rakudo release 10:37
moritz last I looked, LWP::Simple fetched too little data 10:38
so it tries to parse the headers, but hasn't received them all 10:39
so it complains that the headers are incomplete
timotimo i've copypasted the "visit node" method from Optimizer.pm to get a feel for what the astlooks like for simple cases (code: sprunge.us/CfOC ) but i get a surprising error message: Error while compiling block : Can only use get_how on a SixModelObject
the lowest line on the stack trace is "in any as_post at src/stage2/QAST.nqp:2189" 10:40
arnsholt timotimo: Sounds like you somehow get a native object passed into your code 10:41
moritz which is weird, considering that the qast tree isn't changed at all
timotimo yeah, and it's nqp code, too :| 10:45
10:46 GlitchMr joined
timotimo if i put all the things into the sub itself, it works 10:49
oh, perhaps i know what i did wrong
the way the transform sub is called isn't by name
so probably relies on the order of subs in the NQPCODE below
that seems to also mean i can't call transform from within transform :( 10:56
makes it kind of hard to experiment properly without recursion - don't really want to implement my own stack. then again, i *could* 10:57
11:01 noggle_ joined, noggle left
timotimo this is all less obvious than i thought it would be, but if seen as a puzzle, i guess it's kind of fun :) 11:06
r: sub any($x, $y) { say "test"; } say any(1, 2) + 5; 11:10
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/GKnFRiiSde:1␤------> sub any($x, $y) { say "test"; } ⏏say any(1, 2) + 5;␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ …
FROGGS it's all fun in case you make at least a tiny progress within hours... I lost the fun a bit hunting a bug for ~25 hours.... now I try again and I see clearer now
timotimo r: sub any($x, $y) { say "test"; }; say any(1, 2) + 5;
p6eval rakudo 55dbc0: OUTPUT«test␤6␤»
FROGGS r: say "hello".Int
p6eval rakudo 55dbc0: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '⏏hello' (indicated by ⏏)␤ in method Int at src/gen/CORE.setting:10298␤ in method Int at src/gen/CORE.setting:3803␤ in block at /tmp/Hxs3YYiRQ1:1␤␤»
FROGGS r: say "hello".Num
p6eval rakudo 55dbc0: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '⏏hello' (indicated by ⏏)␤ in method Num at src/gen/CORE.setting:10299␤ in method Num at src/gen/CORE.setting:3804␤ in block at /tmp/a9B7ftCPhZ:1␤␤»
FROGGS r: say "hello".Bool
p6eval rakudo 55dbc0: OUTPUT«True␤»
FROGGS r: say "hello".Bool + 1 11:11
p6eval rakudo 55dbc0: OUTPUT«2␤»
timotimo so, in the optimizer, how do i get to figure out if the call to any or | or & or all will give me a junction?
moritz I guess you need to resolve the actual candidate (if it's not done already), and see if it has a return type annotation 11:12
(add one if it's missing)
kresike r: my $a = eval { say "a"; }; say $a;
p6eval rakudo 55dbc0: OUTPUT«Cannot call 'eval'; none of these signatures match:␤:(Str $code, :lang(:$lang) = { ... }, PseudoStash :context(:$context))␤␤ in block at /tmp/d3OQZ1Eizp:1␤␤»
timotimo moritz: do you mean add it to the builtin & and | infix operators? 11:13
moritz yes 11:14
GlitchMr perl6: if False { say "Hello, world!" }
p6eval rakudo 55dbc0, niecza v24-12-g8e50362: ( no output )
timotimo it seems like self.find_lexical can do most of the work for me in the optimiser
GlitchMr It's compile time ERROR in Java
moritz timotimo: note that the optimizer already resolves some multis 11:15
FROGGS someone should glue the java thingy to p6eval
timotimo in this case i don't want to simply inline, i'd like to rewrite the tree itself to move the junction outward 11:16
moritz I didn't say "inline", I said "resolve"
well, it might inline too
you can control that with the number in --optimize=1 etc.
wait, does your code run before or after the rakudo optimizer? 11:17
11:24 sisar joined
timotimo i've reached the land of a thousand tunnels. also: bad reception 11:27
that's a good question. i'm not sure. 11:28
i would think after
11:30 SamuraiJack left 11:32 SamuraiJack joined 11:38 Chillance joined 11:41 brrt left
timotimo been able to recurse using subs in subs. got a pretty-printed graph now. it will only get more complicated now :| 11:50
11:55 sisar left
moritz timotimo: have you discovered the QAST dumper already? 11:55
print($qast.dump) 11:56
timotimo oooh, that's neat 12:00
but since i have to traverse it anyway, i'll keep the visit method i "built" anyway
moritz I implemented QAST::Node.dump for debugging the sink context stuff 12:01
masak g'day, #perl6. 12:04
tadzik heey masak 12:05
moritz \o
masak++ # release
timotimo so, i can probably find the method infix:<|> or infix:<&> from the lexical scope easily, but even if i can figure out that the return type is Junction, i probably can't just call it at compile time, no?
but that would be necessary to programmatically turn 1 | 2 == 1 into 1 == 1 or 2 == 1 12:06
moritz if the arguments are know at compile time, you can
otherwise not
timotimo oh, well that's nice :)
12:06 brrt joined
moritz "known at compile time" means you have to know their values 12:07
ie won't work for my $x = 3; if $x == 42|5 { }
timotimo so if the & or | has only Wants and for each the compile time value is known (which is the case for my first example here), i can ... just call it?
moritz erm wait, in my example you can, yes
timotimo ah, so optimising $x & $y == 5 won't work like that? 12:08
moritz correct
timotimo ah, i can, well that's great :)
moritz for that, you'd have to identify the &infix:<&> as the one from the setting
timotimo is there an easy way to do that?
moritz and duplicate its functiality in the compiler by generating different code for that 12:09
well, I can think of two ways
timotimo that's what i wanted to do, yes
moritz the first is to write a special trait, apply it to the operator, and check it from the compiler
and the second is to look up the symbol in CORE::
r: say &CORE::infix:<&> 12:10
p6eval rakudo 55dbc0: OUTPUT«sub+{precedence} infix:<&>(*@values) { ... }␤»
timotimo is the code to look it up the same in nqp?
nqp: say &CORE::infix:<&>
p6eval nqp: OUTPUT«Confused at line 2, near "say &CORE:"␤current instr.: 'panic' pc 13221 (src/stage2/gen/NQPHLL.pir:4682) (src/stage2/gen/NQPHLL.pm:328)␤»
timotimo right
moritz no, you have to find out what it compiles to
timotimo nqp: say(&CORE::infix:<&>) 12:11
p6eval nqp: OUTPUT«Confused at line 2, near "say(&CORE:"␤current instr.: 'panic' pc 13221 (src/stage2/gen/NQPHLL.pir:4682) (src/stage2/gen/NQPHLL.pm:328)␤»
timotimo i suppose the trait thing would be simpler.
moritz perl6 --target=past -e '&CORE::infix:<&>'
timotimo would it be enough to just do has $!is_trivial_implementation; and expose that through a getter? or is there a mechanism that's not visible to the user of the rakudo interpreter and such? 12:12
moritz doesn't even need to be an attribute 12:14
it can just mix in the role { method IS_DEFAULT_IMPLEMENTATION { True } } or so
timotimo that seems good
and then i'd look up i the object has that method and call it to find out if it returns true? 12:15
moritz and then you write if nqp::can($code, 'IS_DEFAULT_IMPLEMENTATION') && $code.IS_DEFAULT_IMPLEMENTATION { ... }
timotimo oh, clever :)
FROGGS moritz: in order to add a despatchee to a proto I need it's code, how do I get the code from a symbol name? I'm iterating through $package.WHO... 12:23
... in Perl6/World.pm that is 12:24
12:24 brrt left
masak colomon++'s advent post for today isn't scheduled. 12:24
is that intentional?
colomon masak: I'm proof-reading it right now, will post as soon as I'm done.
masak oki nice 12:25
just checking :)
colomon posted
masak \o/ 12:26
colomon++
moritz colomon++
colomon now to make improvements to the testing script. :) 12:27
timotimo building a new rakudo/ a new setting is slow on this laptop anyway, but on battery ... oh my :|
masak anyone mind if I take gist.github.com/4341223 , make a draft out of it in Wordpress, and edit it heavily?
moritz FROGGS: I... think you have it already
masak it will be published under my name. if I could, I would publish it under lichtkind's name. but I can always make "Guest post by..." banner or something. 12:28
moritz or we can just invite lichtkind to the calendar, and have him publish it
(but please heavily edit it first anyway)
GlitchMr Rakudo Warszawa :-) 12:29
masak moritz: right. question is, should I do the heavy editing in a gist or in Wordpress? 12:30
why does "the wrong track" translate to "the wood way" in German? :) 12:31
moritz FROGGS: the .WHO is a hash. The keys are the names of the symbols, and the values are the symbols (e.g. Sub objects) themselves
masak: whatever you prefer
FROGGS hmmm
masak moritz: I think I *prefer* to do it in Wordpress, because then I have visual feedback as to how the post will end up looking. 12:32
masak makes it so
moritz masak: re Holzweg/wood way: de.wikipedia.org/wiki/Auf_dem_Holzweg_sein
masak excellent. thank you :)
FROGGS masak: it is about being on a track that is not built completely 12:33
masak my heavy editing starts with renaming the article... from "Parsing an IP" to "Day 22 - Parsing an IP address".
moritz or maybe IPv4 address, unless he covers IPv6 too
masak he doesn't. good point. 12:34
I'm confused by the reference "...relevant even in Redmond" in the beginning of the post. does this refer to any particular Microsoft product famously incorporating Perl 5 regexes? 12:35
timotimo i think i may have made "make" run an endless loop when trying to compile my change to the junction class :|
moritz what did you change? just the trait? 12:36
timotimo yes, but i think i did it wrong 12:37
moritz did you test it outside the setting first?
that's how I go about such stuff
i.e. write the trait in a normal .pl file, and apply it to a new operator in there 12:38
timotimo i'm doing that now. i went to the toilet in the mean time 12:39
FROGGS moar details please^^
:P
moritz r: multi sub trait_mod:<is>(Routine $r, :$DEFAULT_IMPLEMENTATION!) { $r does role { method IS_DEFAULT_IMPLEMENTATION() { True } } }; multi f() is DEFAULT_IMPLEMENTATION { }; say &f.?IS_DEFAULT_IMPLEMENTATION
p6eval rakudo 55dbc0: OUTPUT«Nil␤»
masak I see that C# has regular expressions. maybe that's what lichtkind is referring to. 12:40
does VB have regular expression support? :)
moritz r: multi sub trait_mod:<is>(Routine:D $r, :$DEFAULT_IMPLEMENTATION!) { say "applying trait"; $r does role { method IS_DEFAULT_IMPLEMENTATION() { True } }; 0 }; multi f() is DEFAULT_IMPLEMENTATION { }; say &f.?IS_DEFAULT_IMPLEMENTATION
p6eval rakudo 55dbc0: OUTPUT«applying trait␤Nil␤»
tadzik I think it does
FROGGS masak: it has
tadzik (VB)
moritz did I misspell it?
r: multi sub trait_mod:<is>(Routine:D $r, :$DEFAULT_IMPLEMENTATION!) { say "applying trait"; $r does role { method IS_DEFAULT_IMPLEMENTATION() { True } }; 0 }; multi f() is DEFAULT_IMPLEMENTATION { }; say &f.?IS_DEFAULT_IMPLEMENTATION 12:41
p6eval rakudo 55dbc0: OUTPUT«applying trait␤Nil␤»
moritz r: multi sub trait_mod:<is>(Routine:D $r, :$DEFAULT_IMPLEMENTATION!) { say "applying trait"; $r does role { method IS_DEFAULT_IMPLEMENTATION() { True } }; 0 }; sub f() is DEFAULT_IMPLEMENTATION { }; say &f.?IS_DEFAULT_IMPLEMENTATION
p6eval rakudo 55dbc0: OUTPUT«applying trait␤True␤»
moritz ah
when it's a multi, the role is only mixed into the candidate, but &f is the proto 12:42
r: multi sub trait_mod:<is>(Routine:D $r, :$DEFAULT_IMPLEMENTATION!) { say "applying trait"; $r does role { method IS_DEFAULT_IMPLEMENTATION() { True } }; 0 }; multi f() is DEFAULT_IMPLEMENTATION { }; say &f.candidates>>.?IS_DEFAULT_IMPLEMENTATION
p6eval rakudo 55dbc0: OUTPUT«applying trait␤True␤»
moritz there you go.
r: proto f(Int $x) {*}; multi f($) { say 'default' }; f 'foo' 12:43
p6eval rakudo 55dbc0: OUTPUT«default␤»
moritz bug!
masak submits rakudobug
moritz it should check the proto signature, but doesn't 12:44
masak right. 12:45
draft is in place.
saved but not scheduled yet.
will start heavy editing now.
colomon masak++ 12:48
colomon just finally got around to running the (improved!) testing script on the Collatz scripts under Niecza, and discovered one of them didn't actually work there. All fixed now, with an updated timings table. 12:49
12:58 kaleem left
masak did Perl 5 bring regexes to regular programming to any significant degree? surely it was in sed and ack way before that. 12:59
I guess Ruby can be said to have regex support *because* Perl 5 has it.
probably Java too, actually.
any other languages you know that have regex support because Perl 5 has it? 13:00
oh! JavaScript.
it even has them with the same slash-delimiter syntax.
moritz aye 13:01
GlitchMr Almost any language has regexpes because of Perl 5. 13:02
rindolf masak: I think in awk regexes are quoted as strings and used by functions (except in case of the line/rule-matchine).
GlitchMr Perl 5 made regexpes popular.
masak GlitchMr: yes, but what evidence supports that claim?
GlitchMr JavaScript specification
rindolf Java didn't get built-in regex support until JDK-1.4.0 13:03
GlitchMr es5.github.com/#x15.10
masak GlitchMr: I already gave JavaScript as a data point above.
I'm looking for more data points.
timotimo moritz, i did the trait completely diffeently, just adding a role. i guess ill better copy your code from above 13:04
GlitchMr docs.python.org/2/library/re.html
"This module provides regular expression matching operations similar to those found in Perl."
timotimo can the trait go into junction.pl?
GlitchMr Besides, almost any language with regular expressions has lookaheads (?= ) and (?! )
Those were first in Perl 5. 13:05
Similarly, (?: )
masak lichtkind: try to avoid phrasings like "that's how it's spelled now", which induces a temporal sequencing between Perl 5 and Perl 6 that does not accord with observable reality. ;)
13:05 Chillance left, Celelibi joined
masak GlitchMr: yes, but which languages? 13:05
moritz masak: the term PCRE attests to Perl's influence on regexes
masak Python too, good.
13:05 Chillance joined
GlitchMr Practically anything that was written after Perl had regular expressions. 13:06
masak I'm looking for indirect influences, too. which languages/technologies use PCRE, then?
FROGGS PHP
masak ooh, yes.
FROGGS++
GlitchMr PCRE is used in PHP and R.
moritz UNIX
man 3 pcre
GlitchMr That was late addition.
UNIX doesn't use Perl regular expressions. 13:07
Besides, man 3 pcre isn't even part of UNIX
It's part of PCRE library
13:07 rjbs left 13:10 MayDaniel joined
GlitchMr If you are going to tell me that grep can use --perl-regexp option, this is GNU option, not UNIX. 13:10
masak and GNU is not UNIX ;) 13:11
moritz $ $ apt-cache rdepends libpcre3|wc -l
243
masak moritz: wow :)
moritz that's just dependencies on one particular implementation
aaand I think just direct dependencies, not recursive oes 13:12
GlitchMr glitchmr@strawberry ~/Pobrane> apt-cache rdepends libpcre3 | wc -l
moritz *ones
masak lichtkind says that regex syntax is "less special", compared to the rest of the language? what does he mean? does he mean that regex syntax is more general than the main slang? because I don't think it is.
GlitchMr 246
Interesting
masak takes a short walk
GlitchMr apt-cache rdepends seems to include the programs I don't have installed... makes sense. 13:13
moritz I just tried aptitude purge libpcre3
it wants to deintall 722 packages
masak :D
moritz *deinstall
and that's just from the ones I have installed on my desktop 13:14
masak I think what lichtkind wanted to say was either "less special syntax compared to Perl 5's regexes" or "less special compared to all the other changes that happened between Perl 5 and Perl 6".
I don't know which one. 13:15
moritz I guess "more regular than the Perl 5 regexes"
masak aye.
going with that one.
GlitchMr gist.github.com/4352748 13:16
Remove the system!
That's like almost every package in my system 13:17
masak ok, so PCRE is important. I get it :)
so far I have Ruby, Python, Java, JavaScript, VB, and PHP among the languages. 13:18
still not so many.
13:18 am0c_ joined
GlitchMr Does m:P5// in Perl 6 count? 13:19
masak "Before you get scared too much..." -- so, a little bit scared is what we're aiming for here? :P
GlitchMr: oh, you're implying Perl 6 was inspired by Perl 5 regexes? :P
GlitchMr Also, I think that .NET should be on this list. Both VB.NET and C#.
masak yes, I guess both m:P5 and the Perl 6 regexes count, in a way.
masak takes a short walk
moritz masak: exim, apache, ngrep, lighttpd, i3-wm, drizzle, nmap, grep, zsh # just a few popular programs that depend on libpcre3 13:20
GlitchMr GNU grep 13:21
The UNIX grep doesn't use PCRE
GNU grep depends on PCRE in only one specific case - when using --perl-regexp option.
moritz what is "The UNIX grep" exactly?
GlitchMr It's grep in UNIX 13:22
tadzik what's "UNIX"?
GlitchMr duckduckgo.com/?q=UNIX
moritz in which UNIX?
GlitchMr That's interesting question...
tadzik UNIX is a trademark of AT&T or something, I don't think there even is "The UNIX" which is used anywhere 13:23
moritz just like there's no "the C compiler" or "the shell" or "the food"
GlitchMr But whatever UNIX it would be, Linux is not UNIX. 13:24
moritz why not? because there's no UNIX in the name?
tadzik no, because that's what the name says :P
moritz I'd say linux is a UNIX flavor, just as gcc is a flavor of C compiler
cedrvint thought Linux was short for "Le Unix" :)
tadzik :D 13:25
FROGGS UNIX is a brand, Unix is not
GlitchMr Linux isn't POSIX compatible.
moritz it's not?
tadzik news
13:26 am0c_ is now known as am0c
moritz en.wikipedia.org/wiki/POSIX lists GNU/Linux als "Mostly compliant" 13:26
FROGGS BSD's and Mac OS X are Unixes, linux is not
it's just System V compatible
moritz so, what does a system need to be to be a UNIX?
erm sorry, a Unix 13:27
tadzik I think an appropriate anwser is "who cares?" :P
FROGGS it needs to have parts of the original kernel
cedrvint tadzik++
tadzik does it have cat and grep? Then it's unix enough for me
GlitchMr The problem is that Linux doesn't have cat or grep. 13:28
cedrvint no cat but a penguin
moritz thinks we are far into trolling territory right now
tadzik yep
soon someone will start enforcing the GNU/Linux meme
GlitchMr GNU/Linux does have cat or grep. Linux doesn't. 13:29
BinGOs Gun-Oo
flussence
.oO( surely you mean GNU/Linux/Xorg/Mozilla/Google/Qt/Oracle/... )
 13:30
GlitchMr I don't have Google or Oracle software on my computer.
flussence I don't have GNU on my router, but it still runs Linux :) 13:31
tadzik "I see that you're using glibc? Yeah! Start calling your system differently from now on!"
cedrvint GlitchMr: as far as I know Oracle contributes to the Linux kernel (FS).
GlitchMr Or perhaps you mean Google Linux. You know, en.wikipedia.org/wiki/Android_(operating_system)
Except Google Linux doesn't have grep. But as far I can see, it does have cat. 13:32
FROGGS I have pics of my cat on my android too
FROGGS .oO( and fwiw my cat can grab ) 13:33
13:37 arafatkamaal joined
GlitchMr Also, www.johndcook.com/regex.html 13:38
Perl regexpes in Microsoft PowerShell
FROGGS all .Net things can use regexes
since you just have to load a specific dll, like in VB
moritz even niecza :-) 13:39
GlitchMr How can I use .NET in Niecza?
tadzik look at examples
there's Gtk tetris and other stuff there
GlitchMr Just wondering, what is the difference between MAIN.exe and Niecza.exe? 13:40
Woodi_ hey, hey :)
colomon GlitchMr: basically you just import the symbols and Niecza handles the rest. 13:41
colomon did not know there was a MAIN.exe
GlitchMr Let's say I want to use System.Text.RegularExpressions.Regex
Is it possible in Niecza?
Woodi_ looks afternoon is nice after sto^H^H^Hworld-end :) 13:42
advocating Perl influence in grep is quite fun :) 13:44
colomon GlitchMr: seriously, look at the examples
it will probably be faster for you to try it out than it would be to explain it. 13:45
GlitchMr Where can I see some examples?
colomon niecza/examples/gtk-*
GlitchMr github.com/sorear/niecza/blob/mast...watcher.pl 13:47
Oh, that...
add_ is for += overloading?
colomon I don't know, that's moderately magical and it's been a long time since i looked at it. 13:48
but I think it's a .NET thing, not Niecza-specific. 13:49
13:49 avar joined, avar left, avar joined 13:50 rindolf left
colomon yeah, .add_Changed is the same as .Changed += 13:51
Woodi_ Plan9 probably forked 'regexp' meaning or at least is trying to make step forward from *nix world... swtch.com/~rsc/regexp/regexp1.html but mayve this link is so 2007's ? 13:52
tadzik plan9 even has its own C standard librar
GlitchMr niecza> CLR::<System.Text.RegularExpressions.Regex,System,PublicKeyToken=b77a5c561934e089>.new('ab?c').Match('<ac>')
ac
Seems to work.
But why doing this has to be so ugly? PublicKeyToken? 13:53
FROGGS dunno, mabye its a microsoft policy
(doing ugly stuff) 13:54
colomon GlitchMr: if you look at, say, examples/gtk-clock.pl, you'll see how to factor the code so that all the ugliness is contained and the main code looks beautiful. 13:55
13:55 SamuraiJack left
GlitchMr Also, how can I use static methods, like Regex.Escape()? 13:56
13:57 shlomif joined 14:04 shlomif is now known as rindolf
colomon GlitchMr: I'm not sure. 14:05
might be as simple as <System.Text.RegularExpressions.Regex,System,PublicKeyToken=b77a5c561934e089>.Escape 14:06
GlitchMr Thanks. 14:07
It worked.
niecza> CLR::<System.Text.RegularExpressions.Regex,System,PublicKeyToken=b77a5c561934e089>.Escape('[](){}')
\[]\(\)\{}
Interesting. Regex.Escape() doesn't escape ] or }. 14:08
14:09 Kharec left
moritz not needed in normal regexes; you only need to escape ] inside charclasses 14:09
14:10 Kharec joined
GlitchMr Too bad. 14:11
If it would escape ] and -, it would be useful for character classes.
But perhaps it's good. preg_quote() in PHP also escapes characters like =, !, <, >, :. And doesn't escape -. 14:12
Not surprised. It's PHP. Everything is broken.
14:16 PacoAir joined
[Coke] enjoys a subtitled film on his day off. 14:23
(french zombie movie) 14:26
Ulti outside of the core language is there a list of modules that are wanted for perl6? 14:28
like a list of priority ports from perl5 or something?
moritz there was an old wiki page with such a list, but I don't know where it is, and neither whether it's up to date 14:29
masak is back, having stocked his coffers with turkey, chocolate, and stout
Ulti I'll take a google then :)
masak this should last me through Christmas :)
Ulti there is a list on github for most wanted modules 14:30
probably not going to touch SSL though ;P
14:30 dayangkun joined
masak Woodi_: I don't understand -- are you saying swtch.com/~rsc/regexp/regexp1.html is related to Plan9 somehow? 14:33
by the way, that URL has a long and proud history on this channel :) 14:34
arafatkamaal Hi masak 14:35
what exactly is the difference between print and .print?
GlitchMr print is function 14:36
.print is method
timotimo moritz: you were suggesting to create DEFAULT_IMPLEMENTATION, but i see there's already a role "default" which has a method "default()" that just returns true
should i just use that one instead?
arafatkamaal perl6 -e 'my @a=1...10;for @a {print}'
perl6 -e 'my @a=1...10;for @a {.print}'
give two different outputs
FROGGS perl6: my @a=1...10;for @a {.print} 14:37
GlitchMr Unlike Perl 5, print doesn't have default implicit argument.
p6eval rakudo 55dbc0, niecza v24-12-g8e50362: OUTPUT«12345678910»
GlitchMr Use .print for that
FROGGS perl6: my @a=1...10;for @a {print}
p6eval niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Unsupported use of bare 'print'; in Perl 6 please use .print if you meant $_, or use an explicit invocant or argument at /tmp/jRsfWwoO5T line 1:␤------> my @a=1...10;for @a {print⏏}␤␤Unhandled exception: C…
..rakudo 55dbc0: ( no output )
FROGGS perl6: my @a=1...10;for @a { print($_) }
arafatkamaal rakudo allows print
p6eval rakudo 55dbc0, niecza v24-12-g8e50362: OUTPUT«12345678910»
arafatkamaal but prints nothing
GlitchMr arafatkamaal: NYI
arafatkamaal actually my question was something else 14:38
GlitchMr arafatkamaal: This error is Not Yet Implemented.
arafatkamaal ok
actually I have a lot of Perl 5 habits
tadzik print() is probably equivalent to print('') at the moment
GlitchMr Errors are supposed to find every possible Perl 5 habit. 14:39
But sadly, most of those aren't yet implemented in Rakudo.
arafatkamaal I was expecting while(<FILEHANDLE>){chomp;print} to somehow work
GlitchMr Nope
<FILEHANDLE> is one element list.
Containing 'FILEHANDLE' 14:40
<> is new syntax for qw// (qw still works, by the way).
masak it's `for $fh.lines { .say }` nowadays :)
arafatkamaal: hi, btw :)
arafatkamaal :)
masak r: for lines { my $firstword = .words[0]; say $firstword.uc }
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Missing block␤at /tmp/tC3vGA8Ui_:1␤------> rstword = .words[0]; say $firstword.uc }⏏<EOL>␤ expecting any of:␤ method arguments␤ postfix␤ infix or meta-infix␤ infix stopper␤ stat…
GlitchMr std: while <FILEHANDLE> { chomp; print }
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Unsupported use of bare 'chomp'; in Perl 6 please use .chomp if you meant $_, or use an explicit invocant or argument at /tmp/jQq0nUonVB line 1:␤------> while <FILEHANDLE> { chomp⏏; print }␤Unsupported use of bare 'pr…
masak oh right.
r: for lines() { my $firstword = .words[0]; say $firstword.uc } 14:41
p6eval rakudo 55dbc0: OUTPUT«LAND␤LAND␤LAND␤HEIMAT␤VOLK,␤VIELGERÜHMTES␤VIELGERÜHMTES␤␤HEISS␤LIEGST␤EINEM␤HAST␤HOHER␤VIELGEPRÜFTES␤VIELGEPRÜFTES␤␤MUTIG␤FREI␤ARBEITSFROH␤EINIG␤VATERLAND,␤VIELGELIEBTES␤VIELGELIEBTES␤»
masak (the input is the Austrian national hymn)
r: for lines() { my $firstword = .words[* - 1]; say $firstword.lc }
p6eval rakudo 55dbc0: OUTPUT«strome,␤dome,␤zukunftsreich!␤söhne,␤schöne,␤österreich,␤österreich!␤Cannot use negative index -1 on List␤ in method <anon> at src/gen/CORE.setting:10307␤ in at src/gen/Metamodel.pm:2404␤ in any find_method_fallback at src/gen/Metamodel.pm:2392␤ in any find_met…
masak heh.
GlitchMr Cannot use negative index -1 on List
I guess it's empty line
arafatkamaal Any reason why implicit printing of $_ was removed from Perl 6? Eg: perl6 -e 'my @a=1...10;for @a {say}' - Doesn't really print $_ automagically? 14:42
masak r: for lines() { next unless .words; my $firstword = .words[* - 1]; say $firstword.lc }
GlitchMr Because of "␤VIELGERÜHMTES␤␤HEISS"
p6eval rakudo 55dbc0: OUTPUT«strome,␤dome,␤zukunftsreich!␤söhne,␤schöne,␤österreich,␤österreich!␤umstritten␤inmitten,␤gleich.␤ahnentagen␤getragen,␤österreich,␤österreich!␤zeiten,␤schreiten,␤hoffnungsreich.␤brüderchören,␤schwören,␤österreich,␤österreich!␤»…
GlitchMr arafatkamaal: Because it was too clever.
masak arafatkamaal: yes, there's a reason.
GlitchMr But you can use .say;
It's close enough.
masak arafatkamaal: we try to eliminate the cases where you have to learn a lot of "special" behavior by heart.
arafatkamaal ok 14:43
masak arafatkamaal: and some subroutines defaulting to $_ counts as that kind of special behavior.
as GlitchMr points out, we have the more general mechanism of doing .say instead of $_.say now.
GlitchMr Still, regular expressions default to $_ .
masak yes, but they're not subs :)
GlitchMr So you can still use .say if /word/; or something like that. 14:44
masak for loops and given statements also default to $_.
arafatkamaal perl6 -e 'my @a=1...10;for @a{.say}' doesn't work but...
hoelzro what exactly is CORE.setting? the default environment that Perl 6 programs run in?
arafatkamaal perl6 -e 'my @a=1...10;for @a {.say}' does work
masak arafatkamaal: yes, that's the next thing :)
GlitchMr arafatkamaal: Whitespace
masak arafatkamaal: we're a bit more touchy about whitespace in Perl 6.
GlitchMr $a{'b'} means key 'b' in hash $a.
masak arafatkamaal: and there are reasons for that as well.
arafatkamaal :) obviously im missing years of discussions 14:45
masak :)
GlitchMr Without doing that, omitting parenthesis would be impossible in if.
masak arafatkamaal: [] alone means array, but [] directly after something means positional indexing.
arafatkamaal: {} alone means block or hash, but {} directly after something means named indexing.
arafatkamaal: <> alone means qw list, but <> directly after something means autoquoted named indexing. 14:46
after a while, a pattern emerges: we care a lot about whitespace directly after terms.
GlitchMr if $a{'a'}{.say} \n { say 42 } # is .say hash element or not...
masak so Perl 6 simply cares about all of that kind of whitespace.
arafatkamaal: () alone means an isolated statement (not expression, surprisingly), but () dreictly after something means a function call with a parameter list. 14:47
FROGGS so if( 123 ) is a function call
masak right.
since we care about the whitespace after terms, we could even define infix operators that would otherwise collide with postfix operators, but don't. 14:48
arafatkamaal It was quite surprising for me to learn for @a -> $b is also a function call!
masak r: sub infix:<++>($l, $r) { $l + $r + 1 }; say 4 ++ 5
p6eval rakudo 55dbc0: OUTPUT«10␤»
masak arafatkamaal: in a sense, yes.
we lean quite heavily on blocks. see S04 :)
arafatkamaal is there a .for?
masak yes.
it's called .map :P
arafatkamaal !!!
masak literally, 'for' and '.map' use the same mechanism in Perl 6. 14:49
GlitchMr gist.github.com/4353237
arafatkamaal So basically it all comes down to reading specs and working out how they all make sense together
GlitchMr arafatkamaal: consider code I've pasted
masak it's just that 'for' tends to be eager by default, since it's often used in sink (=void) context. whereas .map tends to be lazy by default, since it often isn't.
GlitchMr This is one of examples why Perl 6 needs to use whitespace
masak arafatkamaal: yes, read specs, but keep coming back here and talking about it. it helps.
arafatkamaal: also, keep writing code to try stuff. that really helps. 14:50
arafatkamaal ok
timotimo what is "is default" used for inside rakudo? can i use it to mark infix:<&> and infix:<|> as the "default implementation"?
masak arafatkamaal: we're always very grateful when people come back with feedback. so don't be afraid you're bothering the people here with questions. we love it :)
GlitchMr timotimo: "is default" is used in case more than one multi-dispatch rule matches.
When used, "is default" is most important method. 14:51
Usually it isn't needed, as Perl 6 tries to be clever about that.
masak arafatkamaal: oh! and because we care about whitespace after terms, you can do a few cool new things inside qq strings. here, let me demonstrate.
r: my @array = 1, 2, 3, 4; say "The third element is @array[2], you know."
p6eval rakudo 55dbc0: OUTPUT«The third element is 3, you know.␤»
masak r: my %hash = foo => 1, bar => 2; say "Go to the bar and get %hash<bar> :)"
p6eval rakudo 55dbc0: OUTPUT«Go to the bar and get 2 :)␤»
GlitchMr Actually, this is one of those places where Perl 5 cared about whitespace.
masak r: say "You can even call &uc('functions')!" 14:52
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unable to parse expression in double quotes; couldn't find final '"'␤at /tmp/tX0em_ouwH:1␤------> ay "You can even call &uc('functions')!"⏏<EOL>␤ expecting any of:␤ postfix␤ double quotes␤»…
14:52 stopbit joined
timotimo in that case, i shall create a trait DEFAULT_IMPLEMENTATION (shouted to convey this is internal magic, of course) 14:52
GlitchMr eval: my @a = 1; print "Hello, $a[0].";
buubot_backup GlitchMr: Hello, 1.1
GlitchMr eval: my @a = 1; "Hello, $a[0].";
buubot_backup GlitchMr: Hello, 1.
GlitchMr eval: my @a = 1; "Hello, $a [0].";
buubot_backup GlitchMr: Hello, [0].
masak hm, what happened there?
r: say "You can even call &uc('functions')"
p6eval rakudo 55dbc0: OUTPUT«You can even call FUNCTIONS␤»
arafatkamaal yes
masak submits rakudobug
GlitchMr I usually use {uc 'functions'} for that, it's less ugly. 14:54
masak sometimes, discussing things uncovers bugs in the implementations ;)
GlitchMr r: say "You can even call {uc 'functions
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unable to parse expression in single quotes; couldn't find final "'"␤at /tmp/v5N6Zk_hCZ:1␤------> say "You can even call {uc 'functions⏏<EOL>␤ expecting any of:␤ argument list␤ prefix or term␤ …
GlitchMr rn: say "You can even call {uc 'functions'}!"
p6eval rakudo 55dbc0, niecza v24-12-g8e50362: OUTPUT«You can even call FUNCTIONS!␤»
masak not as often as before, though.
GlitchMr: yes, whatever floats your goat. actually, I would probably do that too, in most cases.
r: my @a = 1, 2, 3; say "@a[1]!" 14:55
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unable to parse expression in double quotes; couldn't find final '"'␤at /tmp/tseZolTZEq:1␤------> my @a = 1, 2, 3; say "@a[1]!"⏏<EOL>␤ expecting any of:␤ postfix␤ double quotes␤»
masak r: my @a = 1, 2, 3; say "@a!"
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unable to parse expression in double quotes; couldn't find final '"'␤at /tmp/eXWz0iwRYh:1␤------> my @a = 1, 2, 3; say "@a!"⏏<EOL>␤ expecting any of:␤ postfix␤ double quotes␤»
timotimo i don't quite know what i'm doing wrong - i added a multi trait_mod:<is>(Routine:D $r, :$DEFAULT_IMPLEMENTATION!) to src/core/traits.pm and put sub infix:<|>(**@values) is DEFAULT_IMPLEMENTATION into src/core/Junction.pm, but i get Could not find sub &infix:<does> when trying to compile it :| 14:59
masak timotimo: could you nopaste your patch? 15:00
timotimo sure 15:01
sprunge.us/NgOc
r: say True + True + False
p6eval rakudo 55dbc0: OUTPUT«2␤»
timotimo oh, that's neat, i *am* going to be able to turn junctions involving ^ into simpler logical formulas 15:02
arafatkamaal This one is from the spec
perl6 -e 'my @primesquares = (do $_ if .is-prime for 1..100) »**» 2;say @primesquares'
15:03 FROGGS left
arafatkamaal what exactly is »**» , how do you even type that on a normal keyboard? 15:03
timotimo you can also type >>**>> if you want
it applies ** to all list items on the left and the right (kind of like a zip would), but will make the list on the right as long as needed 15:04
(because on the right side there's only one 2, but you want as many 2s as you have primes on the left)
arafatkamaal wasn't 'do' supposed to execute any block in a loop only once?
timotimo isn't that what FIRST is for? 15:05
r: { FIRST { say "first!"; } say "blubb" } for ^5;
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/XA_ir6Qago:1␤------> { FIRST { say "first!"; } ⏏say "blubb" } for ^5;␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ …
timotimo r: { FIRST { say "first!"; }; say "blubb" } for ^5;
p6eval rakudo 55dbc0: OUTPUT«first!␤blubb␤blubb␤blubb␤blubb␤blubb␤»
timotimo hm, i wonder ... 15:07
arafatkamaal One last question till I get further practice
timotimo r: { say "a"; FIRST { say "first!"; };} for ^5;
p6eval rakudo 55dbc0: OUTPUT«first!␤a␤a␤a␤a␤a␤»
arafatkamaal where exactly can I read about all the .functions eg: .say, .is-prime ? 15:08
timotimo if you want to, you can look into the rakudo source, because rakudo is for the most part implemented in perl6 15:09
other than that, see the synopses i guess? perlcabal.org/syn/
arafatkamaal thanks 15:10
timotimo you're welcome. and have fun with perl6! :)
masak arafatkamaal: in Perl 6, 'do' simply turns a statement into an expression. 15:11
nothing to do with loops.
arafatkamaal: if you're using vim, you cn type » as Ctrl+K > >
arafatkamaal: if you're on Emacs, there's a wonderful plugin called LEIM with all sorts of nifty input methods. 15:12
(if you're on some other editor, get a better editor) :P
arafatkamaal: also check out doc.perl6.org/ -- it's not complete, but what's there is really nice.
moritz++ 15:13
timotimo or a better keyboard layout. otherwise, most linuces come with proper Xcompose mappings
masak yeah.
in general, look beyond what keys you have on your keyboard, and let software remap things for you.
timotimo masak: did you catch my diff? i can't make heads nor tails of it 15:16
masak timotimo: url?
timotimo sprunge.us/NgOc
masak looks 15:17
$r does role { method DEFAULT_IMPLEMENTATION() { True } }
trivially, this is where the error happens.
but I also don't see why.
timotimo yeah, in the other is infixes it's done the same way 15:18
i don't have to stub anything in junction.pm or something?
masak maybe it's a circularity saw problem.
maybe &infix:<does> isn't defined at the time you try to use it. 15:19
jnthn or pmichaud would know, but they're not here right now.
beyond that, I'm a little dismayed at putting a new traid into the setting like that. it looks like a kind of super-default, is that right? 15:20
I'm assuming the spec doesn't mention it?
is the intention to start putting DEFAULT_IMPLEMENTATION on anything in the setting from now on?
(because I don't think that's a very good idea)
what's the scope of your change? :)
timotimo i was hoping there would be a way to make it only visible internally inside rakudo 15:21
in order to rewrite simple-ish junctions in the AST into logic formulas (think $x & $y == -1|0|1) i need to make sure those & and | are the "default implementation"
masak sounds like you want AST manipulation on the level of macros, not to change the setting. YMMV. 15:22
timotimo moritz suggested an alternative, which is to lookup if the infix:<&> or what have you comes from ::CORE, but i didn't see how to do that at all 15:23
masak I like that better.
timotimo the AST would have been manipulated in the optimiser, the trait is just for the analysis
15:26 bot48 joined 15:32 kaleem joined
timotimo not having a lot of success here :P - i guess i should try doing something simpler instead. like add more typed exceptions or close more old bugs =_= 15:36
arafatkamaal Is there a way to know all methods inside a class? 15:39
like say perl6 -e 'my %h,say %h.allmethods' 15:40
should print all methods callable on %h
timotimo r: say {a=>1,b=>2).^methods 15:41
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unable to parse expression in block; couldn't find final '}'␤at /tmp/eqJJZKh649:1␤------> say {a=>1,b=>2⏏).^methods␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ sta…
timotimo r: say {a=>1, b=>2}.^methods
p6eval rakudo 55dbc0: OUTPUT«new bind_key STORE_AT_KEY STORE delete push PARAMETERIZE_TYPE at_key perl elems exists iterator list keys kv values pairs invert at_key STORE_AT_KEY Capture FLATTENABLE_LIST FLATTENABLE_HASH fmt hash of Bool ACCEPTS perl elems infinite item fmt Int Num chrs Numeric…
timotimo .^ is a call to the metaclass 15:42
arafatkamaal ok 15:43
15:44 arlinius joined 15:45 skids joined
timotimo r: Hash.new.HOW.^methods 15:49
p6eval rakudo 55dbc0: ( no output )
timotimo r: say Hash.new.HOW.^methods
p6eval rakudo 55dbc0: OUTPUT«No such method 'gist' for invocant of type 'NQPRoutine'␤ in method gist at src/gen/CORE.setting:5038␤ in sub say at src/gen/CORE.setting:7488␤ in block at /tmp/21u807fF0G:1␤␤»
timotimo i wonder if anyone would be mad at me for implementing this ... 15:50
15:50 kaleem left
moritz nobody would be mad at you for implementing stuff that's missing 15:52
timotimo: re infix does and circularity: that's a problem because infix:<does> is defined rather alte in the setting 15:56
timotimo: but you can take a look at how infix:<does> does the mixing in, and duplicate that
stuff like that is often done in the setting
15:59 am0c left 16:00 kaleem joined
masak today's autopun: "How many SEO experts do you need to exchange a light bulb, lightbulb, light bulb, energy saving light bulb, blowjob, xxx, free porn" 16:00
(courtesy of mjd, and mst's quote collection) 16:01
16:01 FROGGS joined
moritz :-) 16:02
16:02 dpk joined, mtk joined, dpk left
Woodi_ masak: I am pretty sure that url is related to plan9 becouse Thompson algo is very related to p9 QED... I didn't know that uri have history here, however algo description looks similiar to recent jnthn++ work (which is why I posted link...) 16:02
arafatkamaal perl6 -e 'my @a=0...Inf;for @a {.say if .is-prime}' 16:03
how would you rewrite this using map?
not I have used Inf, so don't run evalbot on it
moritz evalbot has a protection against stuff like that 16:04
anway, I'd write it as
kresike bye folks
arafatkamaal typo: note
16:04 kresike left
moritz r: .say for (0..*).grep: *.is-prime 16:04
p6eval rakudo 55dbc0: OUTPUT«(timeout)»
Woodi_ btw. just found Ken Thompson patented Thompson algo what probably explains why it was so *un*adopted in last decades...
arafatkamaal but thats using grep 16:05
16:07 kaleem left 16:10 odoacre joined
rindolf Woodi_: which one? 16:12
masak: heh, nice joke. 16:13
masak Woodi_: there's an interesting relationship between that URL and Perl 6's LTM mechanism. 16:16
r: (0..Inf).map: { .say if .is-prime } 16:17
p6eval rakudo 55dbc0: ( no output )
[Coke] I am not a fan of borat at all, and even I laughed at this: twitter.com/DEVOPS_BORAT/status/28...6682757120
masak r: eager (0..Inf).map: { .say if .is-prime }
arafatkamaal: like that.
p6eval rakudo 55dbc0: OUTPUT«(timeout)2␤3␤5␤7␤11␤13␤17␤19␤23␤29␤31␤37␤41␤43␤47␤53␤59␤61␤67␤71␤73␤79␤83␤89␤97␤101␤103␤107␤109␤113␤127␤131␤137␤139␤149␤151␤157␤163␤167␤173␤179␤181␤191␤193␤197␤199␤211␤223␤227␤229␤233␤239␤241␤251␤257␤263␤269␤271␤277␤281␤283␤293␤307␤311␤313␤317␤331␤337␤347␤349␤353􏿽xE2􏿽x90
masak [Coke]: it's because it's funny-true :) 16:18
16:21 silug joined 16:23 thou joined
arafatkamaal Good night folks 16:23
masak arafatkamaal: 'night
16:23 arafatkamaal left
[Coke] bueños sueños. 16:24
[Coke] over ñ'd that.
masak .oO( spamusement.com/index.php/comics/view/20 ) 16:26
that website still has the ability to incapacitate me with silly laughter. :) 16:27
hoelzro waves goodbye to productivity 16:33
masak hahaha spamusement.com/index.php/comics/view/117 16:37
16:40 thou left 16:44 thou joined 16:45 benabik left 16:48 isBEKaml joined
masak tadzik would like this one: spamusement.com/index.php/comics/view/185 :) 16:49
isBEKaml hola, #perl6! 16:51
masak: (backloggin') I guess the Redmond reference in lichtkind's post was about powershell. They were said to be heavily inspired by perl scripting. 16:52
masak: and you're right about Java and ruby bringing Perl regexes to mainstream. PCRE libs played their part too. 16:53
timotimo if i want to use the nqp i just changed (but have not committed) for building my rakudo, do i have to do Configure.pl at all? or is it enough to make in nqp/ and the make in rakudo? 16:54
tadzik :D
isBEKaml I have now taken to the habit of backlogging and adding stuff. I'm not even sure if it's a good thing. :D 16:55
masak isBEKaml: it's a good thing. thanks for adding stuff.
isBEKaml tadzik: You're our meme repository! :P
tadzik am I? :)
timotimo so sad that spamusement won't make new content :( 16:56
isBEKaml tadzik: Ja.
masak: Danke.
masak tadzik: pandas, zubrs, rage faces. yes, you are :) 16:57
hm, meant żubry, of course :)
tadzik (:
17:11 xinming joined
timotimo i was going to implement a .gist for NQPRoutine, so that you can .HOW.^methods, but i have no idea what i'm doing and now i don't know how to do what i guess is correct ... which would be to return something generated with nqp::repr_box_str. 17:13
17:13 Kharec left 17:16 dakkar left
masak "so you don't move your eyeballs back and forth before you can understand it" -- heh, I understand what lichtkind means. something like "you don't need to read to the end of the regex before you know the conditions under which it should be read". :) 17:16
masak rewrites
17:16 kaare_ joined
masak r: my $str = "1 23 456 78.9"; $str ~~ m:g/(\d+)/; say $0.^name 17:18
p6eval rakudo 55dbc0: OUTPUT«Any␤»
masak r: my $str = "1 23 456 78.9"; $str ~~ m:g/(\d+)/; say $/.^name
p6eval rakudo 55dbc0: OUTPUT«Any␤»
masak I'm... not sure I want to include this example in the blog post. it doesn't seem to work.
I wouldn't do m:g anyway, I'd do .comb 17:19
masak rewrites to use .comb instead.
17:23 buubot_backup joined
masak oh, .comb shows up a few paragraphs earlier. 17:25
is there support in the spec that $str ~~ m:g/(\d+)/ means the same as $str.comb(/\d+/) ? 17:26
moritz r: $_ = '1 23 456 78.9'; .say for m:g/(\d+)/
p6eval rakudo 55dbc0: OUTPUT«｢1｣␤ 0 => ｢1｣␤␤｢23｣␤ 0 => ｢23｣␤␤｢456｣␤ 0 => ｢456｣␤␤｢78｣␤ 0 => ｢78｣␤␤｢9｣␤ 0 => ｢9｣␤␤»
moritz you can use that as an example
masak moritz++ 17:27
r: $_ = '1 23 456 78.9'; .Str.say for m:g/(\d+)/
p6eval rakudo 55dbc0: OUTPUT«1␤23␤456␤78␤9␤»
masak r: $_ = '1 23 456 78.9'; say .Str for m:g/(\d+)/
p6eval rakudo 55dbc0: OUTPUT«1␤23␤456␤78␤9␤»
moritz there are two possiblities for m:g// 17:28
the first is to return a List
then $str ~~ m:g/.../ matches the string against the List, and (nearly) always fails the smart-match
or it returns a Match 17:29
which might or might not flatten, but $/ can never flatten, because it starts with a $
masak each time we get back to ~~ and m//, it seems to me that we've painted us into an unfortunate design corner there.
moritz r: .say for 'abc' ~~ /(.)(.)(.)/ 17:30
p6eval rakudo 55dbc0: OUTPUT«｢abc｣␤ 0 => ｢a｣␤ 1 => ｢b｣␤ 2 => ｢c｣␤␤»
masak where we have some expectations, and the spec ends up saying "no, that simply can't work". I wonder why.
moritz r: say .Str for 'abc' ~~ /(.)(.)(.)/
p6eval rakudo 55dbc0: OUTPUT«abc␤»
moritz masak: I've blogged about that :-)
lemme look up the link for you
masak: perlgeek.de/blog-en/perl-6/three-fu...match.html
TimToady++ wants to fix it with his OKness proposal 17:31
masak nod, and nod.
moritz but I still think orthogonalizing matching and topicalization is the way to go 17:32
though I can't provide a good solution either
masak the OKness proposal still rubs me the wrong way, for some reason. it's like we're putting dualvars on *everything*.
and calling the top type "Mu" feels like a small break from tradition compared to the OKness proposal. 17:33
moritz speaking of which
now that we don't smart-match against Mu all the time, can we rename it to Object again?
the name Mu is an FAQ, and it's wrongly huffmanized 17:34
masak \o/ 17:35
I'm all for it.
moritz
.oO( you think this is Mu today )
 17:36
is that an autopun? 17:37
masak r: /(\d{1,3})\.(\d{1,3})\.(\d{1,3})\.(\d{1,3})/
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Unsupported use of {N,M} as general quantifier; in Perl 6 please use ** N..M (or ** N..*)␤at /tmp/PMGekASt8e:1␤------> /(\d{1,3}⏏)\.(\d{1,3})\.(\d{1,3})\.(\d{1,3})/␤»
masak moritz: doesn't feel so autopunny... :)
moritz masak: it doesn't, but it kinda uses the thing it mentions (or fails to mention?) 17:38
masak just a calque, I think.
who are Bob and Tom in lichtkind's article? someone I should know about? 17:43
timotimo isn't it normally tim and ... who was that other guy? 17:44
tim and al.
masak oh good, so it isn't just me.
masak rewrites
timotimo (from the tv show 'home improvement')
masak aye. I've seen it. 17:46
not excellent TV, but kinda cute and wholesome.
timotimo i agree. it was more than good enough for me back then, but i was quite young, too 17:47
17:47 Kharec joined
masak (for a moment, I was wondering if "Bob" was another lichtkind reference to Redmond. but I don't believe there was ever a Microsoft Tom) :P 17:48
timotimo :D 17:49
17:51 GlitchMr left
moritz Bob the Visual Studio Builder assistent? 17:52
17:52 GlitchMr joined
masak :P 17:54
r: say .Str for "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32." ~~ m/ (\d ** 1..3) ** 4 % '.' / 17:56
p6eval rakudo 55dbc0: OUTPUT«127.0.0.1␤»
masak r: say .Str for "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32." ~~ m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«False␤»
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say .Str for m:g/ (\d ** 1..3) ** 4 % '.' / 17:57
p6eval rakudo 55dbc0: OUTPUT«127.0.0.1␤173.194.32.32␤»
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say @().perl for m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«().list␤().list␤»
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say @($/).perl for m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«(Any,).list␤(Any,).list␤»
masak hm.
timotimo so is regex matching in perl6 surprisingly non-cool?
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say $/.list.perl for m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«(Any,).list␤(Any,).list␤»
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say .list.perl for m:g/ (\d ** 1..3) ** 4 % '.' / 17:58
p6eval rakudo 55dbc0: OUTPUT«((Match.new(orig => "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32.", from => 19, to => 22, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.…
masak ah, yes.
r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say .list>>.Str.perl for m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«("127", "0", "0", "1")␤("173", "194", "32", "32")␤»
isBEKaml timotimo: Regex is not Cool.
timotimo that would explain it!
masak r: $_ = "So, I told him, go 127.0.0.1! But he got lost and knocked on 173.194.32.32."; say .>>Str.perl for m:g/ (\d ** 1..3) ** 4 % '.' /
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Confused␤at /tmp/7u2EVzmqCa:1␤------> st and knocked on 173.194.32.32."; say .⏏>>Str.perl for m:g/ (\d ** 1..3) ** 4 % ␤ expecting any of:␤ dotty method or postfix␤»
timotimo why was Regex not specced to be Cool?
masak because Cool is all about the interconvertability between string, numeric and boolean values from Perl 5. 18:01
and a Regex isn't a string, a number, or a boolean. 18:02
isBEKaml yeah, what masak said. Cool is a mixed bag for all operations, it contains numeric, string and other stuff. 18:04
now that I look at Regex.pm, I see some pir stuff. I thought we got rid of them in favour of nqp? (flaky memory again, I guess :) 18:05
masak lichtkind post reviewed, rewritten, and scheduled for tomorrow. 18:08
feel free to log into Wordpress and review the result :)
"people making apocalypse jokes like there's no tomorrow" -- this is quite likely to be an autopun, however :) 18:10
oh dear; at this rate I'll have material for a second lightning talk... :P
isBEKaml that's going to get old real quick! :P
masak isBEKaml: autopuns have an amazing ability to renew themselves. they don't wear out so much. 18:11
which is odd because they all build on *exactly* the same principle.
18:11 SamuraiJack joined
masak blames our brains 18:12
masak heads out into the winter for the second time today
isBEKaml masak: well, I was just talking about how we pull jokes around events and all the fun is lost after they're past.
masak: winter in .se? must be night-long days! :D 18:13
masak isBEKaml: I swear, I haven't seen the sun for weeks. I hope it's doing some good for the rest of planet. :) 18:14
isBEKaml masak: Well, I can't say it's any good if the earth is tilted so far back. :P 18:17
isBEKaml should stop making bad jokes 18:18
18:37 kaleem joined 18:45 isBEKaml left 18:50 SamuraiJack left 18:52 Tabrenus joined
masak greetings, Tabrenus. 18:55
Tabrenus hello masak
masak are you new here? I don't recall seeing you around. 18:56
if so, welcome. :)
Tabrenus thanks 18:57
18:58 wooden_ left
masak nr: say "the implementations say 'welcome, {my $name = 'Tabrenus'}', too :)" 19:03
p6eval niecza v24-12-g8e50362: OUTPUT«Potential difficulties:␤ $name is declared but not used at /tmp/iyrhu2OB10 line 1:␤------> "the implementations say 'welcome, {my ⏏$name = 'Tabrenus'}', too :)"␤␤the implementations say 'welcome, Tabrenus', too :)␤»
..rakudo 55dbc0: OUTPUT«the implementations say 'welcome, Tabrenus', too :)␤»
masak is not so fond of the 'declared but not used' warning :/
diakopter neither 19:04
hoelzro is that niecza-specific?
masak hoelzro: no, it's from STD and only Niecza implements it. 19:05
FROGGS I like the C way there, emitting these warnings only if I do -Wunused
hoelzro can someone come up with a snippet of code that would result in a "Confused"
19:05 wooden joined
hoelzro FROGGS: agreed. 19:05
19:05 wooden left, wooden joined, Tabrenus left
flussence r: sub infix:<_> { * }; 5_() 19:06
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Confused␤at /tmp/0TCJFtVs7d:1␤------> sub infix:<_> { * }; 5⏏_()␤ expecting any of:␤ whitespace␤»
flussence useless, but does what you asked for :)
19:07 rhr joined
hoelzro hmm 19:07
thanks!
19:17 popl joined
sahadev r: my @a = 1..5; my @b = 'a' .. 'e'; my @c = \@a, \@b; my @d = [\@a, \@b]; say "\@c[0] = @c[0], \@d[0] = @d[0]" 19:43
p6eval rakudo 55dbc0: OUTPUT«@c[0] = 1 2 3 4 5, @d[0] = 1 2 3 4 5 a b c d e␤»
19:44 fgomez left
tomgj Kind of a back to front question coming up. Suppose I am specifying the API of a Perl 6 module. A class contains "has Str $.foo is rw". Is this best documented in the API doc as "has Str $.foo is rw"? If one wanted to document the same API in terms of method(s), what would they be? 19:46
19:47 REPLeffect joined
tomgj "method foo() is rw returns Str" ? 19:47
19:56 fgomez joined
masak tomgj: in a sense, $.foo is a method. 19:58
slavik1 getter/setter? 20:00
20:03 SmokeMachine left
timotimo so, there must be some kind of wrapping that i would need to do if i want to make a method of an NQPRoutine not only callable from perl6 code (which already seems to work), but also wrap the string I want to return so that it is a sensible object for perl6/rakudo. what could that be? repr_box_str perhaps? from looking at its signature, i couldn't figure out how to correctly call it :( 20:04
moritz nqp::p6box_s($nqp_str)
timotimo well, that was simpler than i thought :)
masak (it's as if NQP was made for this kind of thing) :P 20:06
20:07 fgomez left
FROGGS coincidence! 20:07
tomgj slavik1: AIUI the method created by "has" as written above returns an lvalue, so it doesn't really fit what I understand by getter/setter 20:08
masak: okay, but I'm really trying to ask about best practices for api documentation. I was wondering whether it would be expected generally that the consumer of an api, reading the documentation, would understand the "has ..." one as meaning the "method ..." one. But actually "method foo() is rw returns Str" is starting to look less messy now anyway, so I think I'll just stick with that. 20:12
slavik1 tomgj: Perl6 automatically generates getter/setter methods, in java you would have to write them manually. If you call $object.foo("hello"), this will set the value of $.foo to "hello", likewise, $object.foo() returns the value stored by $.foo 20:13
has Str $.foo is rw < this talks about an attribute/member, not method. method gets generated automagically. that is my udnerstanding of how Perl6 (and Moose in Perl5) functions 20:14
masak tomgj: yes, the consensus seems to be to put things in code when you can and in documentation only if you must.
slavik1 listen to masak, his IQ has as many digits as the value of my IQ ... >.> 20:15
TimToady no, that's not how it works; you'd have to assign to $object.foo
slavik1 TimToady: so it's $object.foo = "hello";?
wasn't away that it was changed ... 20:16
TimToady it was never the other way
slavik1 oh
diakopter rn: class A { has $.foo = $!foo; method boo { $!foo } }; say A.new.boo
p6eval rakudo 55dbc0, niecza v24-12-g8e50362: OUTPUT«Any()␤»
TimToady arguments to a method are only used to help establish the identity of the lvalue
so the signature of the "setter" is identical to the "getter" 20:17
that is, we don't distinguish setter from getter
slavik1 right
TimToady a thing either returns an lvalue or it doesn't 20:18
masak rn: class A { has $!x; has $!y; method foo($_) is rw { when 'x' { $!x }; $!y } }; given A.new { .foo('x') = 5; say .foo('x') }
p6eval rakudo 55dbc0: OUTPUT«Cannot assign to a non-container␤ in block at /tmp/QuevMOdD2B:1␤␤»
..niecza v24-12-g8e50362: OUTPUT«5␤»
masak niecza++
slavik1 I recall reading somewhere that the idea was that if you call $object.foo() without any args, it acts as a getter, if you give it stuff, it acts as a setter ...
TimToady in the RFCs maybe, never after that
tomgj TimToady: good. Would you be able to confirm for me that I am correct in saying that the signature of the method generated when I write "has Str $.foo" is "method foo () returns Str is rw"
20:19 sjohnson joined
timotimo so, uh, i've tried calling p6box_s, but i'm getting "could not find sub &p6box_s". perhaps i'm working at the wrong level? i just added method gist() { p6box_s($!do); } to my knowhow NQPRoutine { ... } in src/core/NQPRoutine.pm. 20:19
TimToady there is no 'is rw' unless you put one on the "has"
20:20 fgomez joined
tomgj oh dear, I thought declaring a method rw meant it returned an lvalue. Back to experiments / reading for a while then.. 20:20
timotimo return-rw is there for you
20:20 skids left
TimToady you asked if it was equivalent to the "has", and it isn't 20:20
it would be equivalent if you put "is rw" on the "has" 20:21
tomgj ah, whoops, I missed off "is rw" from the "has" clause the most recent time I wrote it. Thanks for confirming that with it, it is equivalent. 20:23
20:24 telex left
tomgj masak: regarding code vs documentation, the comment seems to rather gloss over the distinction between interfaces, which should be specified properly, and implementation. 20:25
masak still no takers on perl6dvent slot #24?
20:25 Kharec_ joined
slavik1 huh? 20:25
what is that? 20:26
masak tomgj: well, I'd consider signature, return type, and names to be part of the interface.
tomgj: and anything inside of the method body to be part of the implementation.
20:26 telex joined
masak slavik1: it's the remaining empty slot in the Perl 6 Advent Calendar this year. 20:26
slavik1 to do what, though? 20:27
tomgj masak: sure - part of.
slavik1 I have no idea what that calendar is for ...
tomgj I think I understand what you meant better now
slavik1 :(
masak slavik1: the calendar is here: perl6advent.wordpress.com/ 20:28
20:28 Kharec left
masak slavik1: it's the fourth year we're doing it, so I'm a little surprised you've missed it until now. :) 20:28
slavik1 yeah
I haven't been followng Perl6 much in the last 4 years due to wr0k
masak *nod* 20:29
slavik1 I am not smart enough to have something cool about Perl6 :(
20:30 colomon joined 20:32 jfried joined
masak the 24th slot is usually something special. in 2009, pmichaud++ wrote about STD.pm6. in 2010, TimToady++ wrote about positive trolling. 20:34
timotimo moritz: did you read my little line about p6box_s above? this is the diff: sprunge.us/VgfE 20:36
20:42 kaleem left
moritz timotimo: nqp::p6box_s is only availble in rakudo, not in nqp 20:42
and it's wrong for NQP to return Perl 6 types anyway 20:44
20:47 sahadev1 joined
timotimo oh. so i'll have to do magic in rakudo, not in nqp 20:48
moritz www.piumarta.com/software/maru/ # interesting 20:49
20:49 Kharec_ left, sahadev1 left 20:50 sahadev left, sahadev1 joined 20:51 cognominal joined, fgomez left 20:52 snearch joined
masak I believe I've seen Maru once before. a really nice proof-of-concept. 20:53
20:57 cognominal left 20:59 fgomez joined
timotimo grepping through all of perl6+nqp+parrot shows only appearances of NQPRoutine in nqp and nothing in src/Perl6/Metamodel stands out as something that would explain how NQPRoutine gets into rakudo. where should i look instead? 20:59
21:01 isBEKaml joined 21:02 sahadev1 left
moritz nqp: sub f() { }; say(f.name) 21:05
p6eval nqp: OUTPUT«Method 'name' not found for invocant of class 'NQPMu'␤current instr.: '' pc 93 ((file unknown):58) (/tmp/tBaLSZy_xE:1)␤»
moritz nqp: sub &f() { }; say(&f)
p6eval nqp: OUTPUT«Use of undeclared variable '&f' at line 2, near ")"␤current instr.: 'panic' pc 13221 (src/stage2/gen/NQPHLL.pir:4682) (src/stage2/gen/NQPHLL.pm:328)␤»
moritz timotimo: I think that subroutines and/or methods in NQP are compiled to NQPRoutine 21:06
timotimo: and since the metamodel is written in NQP, you get that once you go a level more meta than you usually do
timotimo in that case i don't know how to cross the metas sufficiently to return a perl6 string from the gist method a NQPRoutine has (or rather: doesn't have yet) 21:08
21:08 GlitchMr left
moritz timotimo: the answer is probably rather magical, in terms of MOP metacircularity or so 21:14
and not easily figured out
moritz tests the qast-sink-1 branch one more time, and hopes to merge it Very Soon[tm] 21:16
timotimo OK, then i'll look for another thing to break my teeth off with 21:17
besides fixing more old bugs that are "closable with tests", what comparably simple things could i try? 21:18
moritz timotimo: I'm not sure how easy it is, but I'd love to see better recognition of illegal Null Regexes in Rakudo 21:19
STD does it very well
std: /a | | b/
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/fmNpJ9Tfmg line 1:␤------> /a | ⏏| b/␤Parse failed␤FAILED 00:00 41m␤»
moritz r: / a | | b /
p6eval rakudo 55dbc0: ( no output )
moritz nr: / a | | b /
p6eval rakudo 55dbc0: ( no output )
..niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Null pattern not allowed at /tmp/EjTpknugZS line 1:␤------> / a | ⏏| b /␤␤Parse failed␤␤»
timotimo std: / a | ( | ) | b / 21:20
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/ukHS6BTAoN line 1:␤------> / a | ( | ⏏) | b /␤Parse failed␤FAILED 00:00 41m␤»
timotimo this is something that would be done in Actions.pm?
don't think it could be done in Grammar already 21:21
moritz it's done in the grammar 21:22
note that grammars have the same <term> <infix> <term> structures as mainline code 21:23
and two terms in a row and two operators in a row are also detected in the grammar
timotimo a | | b / <- this would be two operators in a row, right? 21:24
moritz you can "just" search STD.pm6 for "Null pattern" and "Null regex", and try to port the error detection to rakudo. Such things are often, but not always, not too hard
correct
timotimo i'll see what i can do. i'll be focusing on src/Perl6/Grammar.pm::Perl6::RegexGrammar, is that right? 21:26
moritz right 21:27
and possibly also on the regex grammar in NQP
timotimo opens QRegex::P6Regex::Grammar in a split, too
moritz from which Perl6::RegexGrammar inherits
21:27 ifim joined
timotimo it seems like nqp/src/stage1/gen/NQPP6Regex.pm is the one i want to look at, is that right? 21:30
or maybe stage2?
moritz no, everything in directories with 'gen' in the name is generated
src/QRegex/P6Regex/Grammar.pm iirc
and src/stage* is also generated (though the .pm files are often just copies)
timotimo ah, yes 21:31
moritz .nqp, not .pm
timotimo that's about 1000 lines less than the generated one. but nothing important is missing?
moritz I guess the generated one also contains the regex compiler and actions 21:32
21:32 isBEKaml left
timotimo do you wish to get a structured exception class to go with it? (in case i get anything done in the first place ;) ) 21:32
moritz there's already one for it
timotimo perfect!
moritz just not triggered often enough 21:33
timotimo er, there's only 3 relevant occurences of "Null" in STD.pm6. am i missing something? 21:35
dalek Heuristic branch merge: pushed 35 commits to rakudo/nom by moritz 21:36
flussence yay
timotimo great, moritz++ :)
flussence moritz++ # kitchen sink
21:38 Tabrenus joined
timotimo er, will i be allowed to add "invalid null pattern" errors to nqp as well? intuition says: yes 21:38
moritz yes 21:39
21:42 stevan_ joined
FROGGS boys and girls: I almost-fixed the merge-import bug! 21:50
moritz \o/ 21:51
timotimo is there no alternative spelling of $¢?
FROGGS phenny: tell jnthn I have a patch that allows one to import multis from two different modules (like the trait_mod:<is>), but one thing remains open: the dump of the list of dispatchers is incomplete, even if all dispatchers work
phenny FROGGS: I'll pass that on when jnthn is around.
FROGGS thanks phenny 21:52
moritz timotimo: $/.CUSRSOR
FROGGS: incomplete how?
are the ones from the setting missing?
FROGGS I have Foo.pm that exports multi trait_mod:<is> ( Routine $r, :$foo!), and Bar.pm that does that with :$bar! 21:53
timotimo r: / /
p6eval rakudo 55dbc0: OUTPUT«===SORRY!===␤Null regex not allowed␤at /tmp/JQ1osiSPjh:1␤------> / /⏏<EOL>␤»
timotimo at least that still works in my local copy :)
FROGGS you can use both traits, but if you use a unkown trait it will just print the core ones plus bar, but not foo 21:54
I would like to push that to a branch, and it would be just awesome if I had commit bit 21:56
FROGGS makes puppy eyes
dalek kudo/nom: 0199241 | jnthn++ | src/ (2 files):
Make "my $a; { $a; my $a } an error.

This commit doesn't break any tests and actually makes a TODO test pass. However, it's a bit close to release to include it now, so it goes in a branch for now and can be merged afterwards.		 22:03
kudo/nom: 3c3f780 | moritz++ | src/ (2 files):
Merge remote branch 'origin/outer-bound' into nom
22:03 tcpr1347 joined 22:04 skids joined
dalek kudo/nom: c8de2eb | moritz++ | docs/ChangeLog:
update ChangeLog for outer-bound branch merge
 22:04
timotimo the next release is already going to be nice :P 22:05
moritz phenny: tell jnthn YOU MAKED BRANCH outer-bound BUT I MERGED IT KXTHXBY
phenny moritz: I'll pass that on when jnthn is around.
22:09 Tabrenus left 22:10 colomon left 22:11 whiteknight joined
timotimo i'll go to bed and see if i can get anything working tomorrow. i've copied over a rule from STD to nqp, but it doesn't seem to have changed anything. i'll need to look more closely and understand more, i s'pose. 22:16
22:16 cognominal joined 22:17 colomon joined
masak 'night, #perl6 22:38
colomon \o 22:40
22:42 snearch left 22:52 fgomez left 23:00 cognominal left
rindolf masak: good night. 23:03
23:12 kaare_ left 23:15 popl left 23:19 cognominal joined 23:26 fgomez joined 23:29 stopbit left 23:33 xinming left, rummik left, rummik joined 23:35 xinming joined, cognominal left 23:40 cognominal joined 23:41 rurban joined 23:45 rurban left 23:47 rurban_ joined 23:49 rurban_ is now known as rurban, cognominal 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!