»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.
timotimo wouldn't a zen-slice[] copy a list? 00:00
i must be misunderstanding coroutines :| 00:05
or maybe binding or lazyness or something entirely different. 00:06
00:07 bot48 joined 00:15 bot48 left
timotimo hm, i would have liked to gist this and feed it to p6eval, but could as well just sprunge it 00:16
sprunge.us/AOFJ
the only surprise i can't get rid of is that each of the combinations is apparently a parcel with a list as second argument, but i can't really turn it into a single parcel, because that just flattens the outermost list into utter unhelpfulnes 00:17
also it's very interesting how the number of elements you take out of the outermost lazy list causes the first element to be longer, that seems kind of unnatural :D
00:18 bot48 joined
timotimo sprunge.us/DbKJ - sample output for [^30], sample output for [^10]: sprunge.us/NIbe 00:19
00:25 rurban1 left 00:26 MayDaniel left
timotimo and i can't run the perl6-debugger because it's saying "Missing or wrong version of dependency 'src/gen/Metamodel.pm'" and i can't install it via panda because github is down :| 00:27
00:44 slayer101_ joined 00:53 ServerSage joined 01:21 rurban joined 01:22 slayer101_ left 01:23 bowtie left 01:29 pmurias left 01:31 rurban left 01:33 rummik left, rummik joined 01:38 Celelibi joined 01:40 PacoAir left 01:44 cognominal joined 02:00 mj41 joined 02:07 thou_ joined 02:09 slayer101_ joined
diakopter urp. github still down. 02:11
scheduled or unscheduled?
sorear diakopter: failures during scheduled maintance, Chernobyl style 02:12
diakopter: status.github.com/messages
diakopter if only this were distributed.. oh wait. 02:13
02:15 rurban joined 02:22 skids joined 02:28 mj41 left 02:34 MikeFair_ left 02:36 rurban left
cognominal unlike chernobyl, the fallout reaches France. 02:38
02:49 mj41 joined 02:54 mj41 left, FROGGS_ joined 02:58 FROGGS left
[Coke] No such method 'length' for invocant of type 'Str' 03:00
03:03 Chillance left
diakopter No such method 'resurrect' for invocant of type 'github' 03:22
sorear [Coke]: length is gone, sorry. did you mean "chars"? 03:23
03:26 cognominal left, mj41 joined, cognominal joined 03:31 mj41 left 03:46 mj41 joined 04:09 slayer101_ left 04:14 cognominal left, raiph left 04:19 cognominal joined 04:23 slayer101_ joined 04:25 cognominal_ joined 04:26 cognominal left, snarkyboojum_ is now known as snarkyboojum 04:30 cognominal_ left 04:34 slayer101_ left 04:35 hash_table left 04:36 slayer101_ joined 04:39 chubbyLady joined 04:40 chubbyLady left 05:11 telex left 05:18 telex joined 05:21 aindilis joined 05:29 lolage left 05:31 telex left 05:37 telex joined 05:40 slayer101_ left 05:42 mj41 left 05:58 mj41 joined 06:06 rurban joined 06:10 rurban left 06:27 noggle left 06:28 noggle joined 06:37 rurban joined 06:39 cognominal joined 06:42 rurban left 06:53 cognominal left 07:01 cognominal joined 07:05 noggle left 07:06 mj41 left 07:09 rurban joined 07:11 noggle joined 07:16 rurban left 07:21 mj41 joined 07:28 mj41 left 07:34 cognominal left 07:35 cognominal joined 07:37 FROGGS_ left
moritz timotimo: yes, perl-6.de is lots out of date 07:43
07:44 mj41 joined
moritz I guess I should update it, but so far I have too many other things on Perl 6 TODO heap 07:48
I'm putting the source code on github.com/moritz/perl-6.de 07:58
if anybody wants to help, just ask for a commit bit
08:07 cognominal left, cognominal joined 08:10 rurban joined 08:11 SamuraiJack joined 08:15 cognominal left 08:16 gaussblurinc joined, cognominal joined 08:18 rurban left 08:23 cognominal left
moritz timotimo: are you still looking for stuff to hack on? 08:24
timotimo: rakudo's substr-rw doesn't understand *-1
n: my $a = '1234'; say substr($a, 1, 1, '42) 08:25
p6eval niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Unable to parse argument list at /tmp/EBV219gS1w line 1:␤------> my $a = '1234'; say substr(⏏$a, 1, 1, '42)␤Couldn't find final ')'; gave up at /tmp/EBV219gS1w line 1:␤------> my $a = '1234'; say subs…
08:25 cognominal joined
moritz n: my $a = '1234'; say substr($a, 1, 1, '42') 08:25
p6eval niecza v24-12-g8e50362: OUTPUT«4␤»
moritz erm, what?
r: my $a = '1234'; say substr($a, 1, 1, '42') 08:26
p6eval rakudo c8de2e: OUTPUT«Too many positional parameters passed; got 4 but expected between 2 and 3␤ in sub substr at src/gen/CORE.setting:2339␤ in block at /tmp/eT3nbR6ttb:1␤␤»
08:27 donghongbo joined 08:30 cognominal left 08:34 cognominal joined 08:39 xinming left, kaare_ joined 08:41 cognominal left, GlitchMr joined 08:43 xinming joined 08:44 cognominal joined 08:45 rurban joined 08:54 popl joined, rurban left 08:55 spider-mario joined 08:56 mj41 left 09:01 mj41 joined
moritz timotimo: I've updated perl-6.de a bit; still needs lots of updating 09:03
09:07 cognominal left, mj41 left
moritz n: say do for ^4 { $_ ** 2 } 09:09
p6eval niecza v24-12-g8e50362: OUTPUT«Nil␤»
moritz n: say Date.today 09:10
p6eval niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'Date' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /home/p…
09:11 cognominal joined 09:16 cognominal_ joined
GlitchMr I've lately moved to Arch. I like how repositories have Rakudo Star 2012.11. 09:16
popl I considered moving to Arch Linux.
09:16 cognominal left
popl my laptop is on the way out anyways, though. 09:17
So I have Slackware and a broken Parrot
and no Perl6
oh well
OH, that was not the topic of conversation. My apologies. 09:18
moritz reminds masak++ to publish a p6advent post today 09:19
09:21 FROGGS_ joined
FROGGS_ morning 09:23
moritz \o FROGGS_
FROGGS_ moritz: is github/rakudo an org now? 09:24
09:25 popl left
FROGGS_ I'm asking because of the commit bit I'd like to have 09:25
bbkr wordpress on advent calendar destroys code thinking it's HTML :( I've posted comment under masak IPv4 post about usefulnes of <?{}> 09:26
rakudo: $_ = "Go 127.666.0.1, I said! He went to 173.194.32.32."; say .Str for m:g/ ( (\d ** 1..3) <?{ 0 < $0 < 255 }> ) ** 4 % "." /;
p6eval rakudo c8de2e: OUTPUT«173.194.32.32␤»
moritz FROGGS_: it is, but I still can't hand them out on a whim
FROGGS_ moritz: k 09:27
moritz FROGGS_: you need to have submitted a CLA, and the majority of the developers need to agree, and none object
FROGGS_ I sent the CLA right before 2012.11 release 09:28
moritz: what can I do to get the devs feedback? making a gist so that they can reply there? or send a mail to some address? 09:31
09:40 tcpr1347 left 09:41 Psyche^ joined 09:42 Patterner left, Psyche^ is now known as Patterner 09:45 mj41 joined, rurban joined 09:52 rurban left 09:57 cognominal_ left 10:02 cognominal_ joined 10:04 rindolf joined
moritz FROGGS_: talk to jnthn and pmichaud 10:06
FROGGS_: you have my +1 10:07
10:07 ivan`` joined 10:15 ps joined, ps is now known as Guest90247
FROGGS_ moritz: thanks, and thanks, will do :o) 10:15
10:16 cognominal_ left
GlitchMr I wanted to put problem 1 solution 10:16
It looked like this:
say [+] grep * %% (3|5), ^1000;
The problem is that somebody already has done that... 10:17
Exactly like that
moritz GlitchMr: what are you talking about?
GlitchMr github.com/perl6/perl6-examples/bl...rondilu.pl
moritz so, what's the problem? 10:18
GlitchMr But well, the problem is probably so simple that it's hard to think of other Perl6ish solution.
rindolf What is %%?
Is it numeric reminder? 10:19
GlitchMr Divides by operator
rindolf remainder sorry.
Ah.
Why not a simple "%"?
GlitchMr Because % is modulo
3 % 2 == 1, but 3 %% 2 == False.
rindolf Ah, hah.
GlitchMr But opposite of modulo would work too
!($a % $b) is just like ($a %% $b) 10:20
moritz rindolf: you can look up operators on doc.perl6.org/language/operators
GlitchMr %% is actually useful operator :-).
The $a % $b == 0 I often see in other programming languages annoys me. 10:21
rindolf doc.perl6.org/language/operators#infix+%25%25 - this link does not work in Firefox with JS blacklisted.
GlitchMr !($a % $b) is even more annoying.
10:21 cognominal_ joined
moritz rindolf: I know there's some trouble with the anchor links 10:22
the browsers don't quite agree on what needs escaping and what not
and I haven't figured it out yet
rindolf moritz: OK.
I tend to avoid doing that and use custom id="..." 10:23
GlitchMr print(sum(i for i in range(1000) if i % 3 == 0 or i % 5 == 0)) 10:25
I believe it would look like this in Python
Now for Rubyish way 10:30
puts (1..999).select {|i| i % 3 == 0 or i % 5 == 0 }.inject(:+)
You can write really short programs in Perl 6.
Of course, you can write Ruby in Perl 6 10:33
say (1..999).grep(-> $i { $i % 3 == 0 or $i % 5 == 0 }).reduce(&[+])
Close enough
10:39 jokar left 10:42 Guest90247 left 10:44 cognominal_ left 10:46 rurban joined 10:49 cognominal joined 10:51 cognominal left 10:53 pmurias joined 10:55 PacoAir joined 10:57 rurban left 11:00 Kharec joined 11:03 PacoAir left, PacoAir joined
masak hola, #perl6 11:17
sorry I haven't prepared a slot 23 post yet.
I completely forgot...
if you have the patience for it, I'll have one ready in a couple hours.
moritz what would be the alternative? you prepare it for tomorrow, and somebody else does a post fo today? 11:19
masak I guess.
up to you folks and bots. 11:20
GlitchMr Illegally post-declared types 11:23
Forward goto is disallowed in Niecza? 11:24
rn: gist.github.com/4363016 11:32
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/BFHqh1Q8Y1:13␤------> zero⏏:␤ expecting any of:␤ argument list␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ statement modifier␤…
..niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Illegally post-declared types:␤ 'no' used at line 34␤ 'yes' used at line 23␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 11…
GlitchMr std: gist.github.com/4363016
felher masak: I think we can wait a couple of hours :)
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Illegally post-declared types:␤ 'no' used at line 34␤ 'yes' used at line 23␤Check failed␤FAILED 00:00 43m␤»
GlitchMr So, I cannot have forward labels?
"must enclose the label in quotes" 11:33
oh, ok
11:34 rummik left 11:35 rummik joined
moritz masak: I agree with felher 11:36
11:38 donghongbo left
FROGGS_ IMO quality is much more important than beeing on time 11:39
masak: (and your posts tend to be good) 11:40
jnthn morning o/ 11:43
phenny jnthn: 21 Dec 21:51Z <FROGGS> 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
jnthn: 21 Dec 22:05Z <moritz> tell jnthn YOU MAKED BRANCH outer-bound BUT I MERGED IT KXTHXBY
jnthn kthx by...whom? :P
jnthn will be about for a little bit :)
FROGGS_: What does "list of discpatchers" mean? Maybe it's easiest if I can see a patch... 11:44
moritz by CEILING CAT
jnthn: I think FROGGS_ talks about the list of signatures in the error message
FROGGS_ moritz: correct
jnthn ah, those'd be dispatch*ee*s :P
11:45 Kharec left
jnthn Anyway, I can probably spot any issues by looking at the patch 11:45
FROGGS_ jnthn: I can't show you right now
jnthn ok
I'm gonna be fairly tied up with family stuff, but may find the odd moment in the next few days
FROGGS_ currently I'm traveling
pmurias jnthn: what do uninitialised lexicals contain in nqp? NQPMu? 11:46
jnthn Best bet is to gist it and phenny me it.
pmurias: For scalars, yes
FROGGS_ jnthn: I will be too, no hurry
jnthn @ and % are automatically initialized with an empty array/hash
FROGGS_ jnthn: k
11:46 rurban joined
moritz now I wanted to remove the sink ROADMAP item, and found that we had none :-) 11:46
jnthn moritz: aww! 11:47
moritz: Noticed it got merged. Yay. \o/
moritz++
FROGGS_ cool
how much is the parse time now?
moritz not much changed 11:48
FROGGS_ k
moritz still between 1 and 4 minutes, depending on the machine I run it on
FROGGS_ one minute? wow
moritz jnthn: any idea for medium-sized rakudo hacking projects? 11:49
FROGGS_: I've got a new workstation at $work
it managed the parsing stage in 62s or so
FROGGS_ had 80s right before the slow-down patch, now got 240s
moritz erm, what?
the slow-down part of the patch has been reverted 11:50
FROGGS_ the patch that added set operators
moritz FROGGS_: are you on up-to-date nom branch?
jnthn has just kicked off a build
FROGGS_ ohh, then I should pull :o)
no
moritz yes, you should :-)
FROGGS_ ya
jnthn moritz: Hm, lemme glance the ROADMAP for things :) 11:51
11:51 pmurias left
moritz one that sounds like fun ins NFG strings... but I think it's way too much work 11:51
jnthn moritz: Where did the logical cascades work get to?
moritz jnthn: it mostly works, but there were some trouble setting $_ and/or $! in the thunks 11:53
jnthn ah, ok
orelse too?
There could be some things in regexes
11:53 pmurias joined
moritz yes, orelese too 11:54
though I'm not sure of the corner cases there
jnthn The array variable interpolation is...difficult.
moritz hm, might still be a TODO
jnthn But OTOH not a huge amount of code to write.
11:55 rurban left
jnthn The problem is that it involves some fairly dynamic NFA stuff 11:55
Doing ~~ inside regexes may well be more accessible. 11:56
moritz what's the "missing bits of enums" in the Language Features section?
jnthn I'm not sure
Do non-numeric enums work?
I'm sure I saw a comment the other day in Actions that suggested not.
moritz r: my enum A a => 'b', c => 'd'; say c
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Malformed my␤at /tmp/Y3RWkIw78A:1␤------> my enum A ⏏a => 'b', c => 'd'; say c␤ expecting any of:␤ scoped declarator␤»
moritz r: enum A (a => 'b', c => 'd'); say c 11:57
p6eval rakudo c8de2e: OUTPUT«c␤»
jnthn The sigspace changes are also an option.
pmurias jnthn: are there any specific method names I should use for conversions to nqp numbers,ints and strings? (I currently use to_n,to_i,to_s)
jnthn .Num, .Int, .Str :)
moritz r: enum A (a => 'b', c => 1); say c
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Type error in enum. Got 'Int' Expected: 'Str'␤at /tmp/1khutqr76M:1␤------> enum A (a => 'b', c => 1)⏏; say c␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement …
moritz are those supposed to work?
jnthn r: enum A (a => 'b', 'c'); say c 11:58
p6eval rakudo c8de2e: OUTPUT«c␤»
moritz well, maybe I'm doing the rest of p6cc first :-)
jnthn r: enum A (a => 'b', 'c'); say c.perl
p6eval rakudo c8de2e: OUTPUT«A::c␤»
jnthn r: enum A (a => 'b', 'c'); say c.value
p6eval rakudo c8de2e: OUTPUT«c␤»
jnthn Hm, seems they do work
moritz r: enum A (a => 'x', 'c'); say c.value
pmurias jnthn: isn't Int a bigint rather then something the op add_i wants?
p6eval rakudo c8de2e: OUTPUT«y␤»
jnthn Maybe consider the spectest
moritz ok
jnthn pmurias: In Perl 6, but NQP doesn't use bigints natively
pmurias: In reality, all the numeric ops in NQP are floating point. 11:59
The only way to get an int is with the native types stuff.
Well, plus literals I guess
This is to avoid having to multi-dispatch all the operators. 12:00
moritz: I think STD has got a decent solution on the sigspace things.
pmurias and nqp::add_i?
jnthn pmurias: Yes, if you do that it's integer addition of course
pmurias: I meant what + does
It always (in Actions) spits out nqp::add_n 12:01
12:01 gattschardo joined
jnthn My plan was to change it to add_i in the optimizer when both sides are integers. 12:01
moritz: Don't take the :D/:U/:T task as I've plans for that (which will involve some notable changes) 12:02
pmurias that would acutally hurt performance in js
as javascript has no integers
jnthn moritz: If you want to do something REALLY evil...masak's macro grant doesn't cover textual macros, and now that I did that refactor that puts the target/orig off in a struct that all cursors share throughout the parse, they're actually kinda doable ;) 12:03
pmurias: Optimizer doesn't have to do the same thing for all backends ;) 12:04
moritz: Another small task that may be fun and isn't a huge amount of work: look up $*LAST_NIBBLE in STD, and port the related bits to Rakudo. 12:05
moritz: It does runaway string detection.
It's a nice error improvement for a probably quite modest amount of work. :) 12:06
Other possibility there: implement suppose. I didn't get around to it yet, and we could use it in a few places.
12:06 fgomez joined
jnthn (today we get bogus TTIAR errors sometimes, when the next thing actually isn't a term...) 12:07
moritz: Hope that lot provides some ideas ;)
12:08 gaussblurinc left
moritz jnthn: it does, takk 12:08
jnthn var saa god :) 12:09
gattschardo is there a simple way to load (or reload) a .pl file in rakudo when using it as a shell? 12:13
12:13 LlamaRider left, LlamaRider joined
FROGGS_ slurp and eval? 12:14
jnthn Heh, could do that. 12:15
"use" won't look for it.
r: evalfile
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Undeclared routine:␤ evalfile used at line 1␤␤»
jnthn hm, probably misremembering...I thought there was a builtin that did slurp/eval for you
but yeah, eval slurp 'file.pl' 12:16
gattschardo well, slurp/eval works, thanks
12:16 am0c joined
gattschardo just thought there might be something built in 12:16
jnthn And if you repeat it, it'll slurp it afresh
jnthn realizes that this is probably a cute debugger prompt idiom too...
Since eval slurp 'foo.pl' at the debugger prompt would let you debug foo.pl :) 12:17
GlitchMr In Perl 5, do is enough 12:18
do 'foo.pl'
But not like eval slurped files is good idea 12:19
12:23 MayDaniel joined 12:25 rindolf left 12:26 rindolf joined 12:27 noggle left 12:28 noggle joined 12:35 GlitchMr left 12:37 GlitchMr joined 12:38 LlamaRider left 12:40 GlitchMr_ joined 12:42 GlitchMr left, GlitchMr_ left 12:44 mj41 left 12:47 mj41 joined, rurban joined, jokar joined 12:50 GlitchMr joined 12:51 rurban left 12:52 pmurias left, pmurias joined, mj41 left 12:55 GlitchMr left 13:04 FROGGS_ left 13:05 FROGGS_ joined
jnthn OK, enough for today :) o/ 13:14
moritz bye 13:15
13:17 GlitchMr joined 13:20 mj41 joined 13:24 donghongbo joined 13:27 isBEKaml joined 13:30 GlitchMr left
isBEKaml hola, #perl6! 13:33
13:35 SamuraiJack left 13:37 Chillance joined 13:40 not_gerd joined
not_gerd hello, #perl6 13:40
isBEKaml rn: sub froo($xx, $yy?) { say "x is $xx" ~ $yy.defined ?? " with yy as $yy." !! ""; }; froo(24); 13:41
p6eval rakudo c8de2e: OUTPUT«use of uninitialized value of type Any in string context in sub froo at /tmp/y0LWn_zjkK:1␤␤ with yy as .␤»
..niecza v24-12-g8e50362: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/lib/CORE.setting line 1287 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 266 (Mu.Str @ 15) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/uZzmxBKJTo line 1 (froo @ 4) ␤…
not_gerd moritz: one way to make html ids from arbitrary strings: gist.github.com/4363369 13:42
horribly slow on Rakudo, though :(
isBEKaml rn: sub froo($xx, $yy?) { say "x is $xx" ~ $yy.defined ?? " with yy as $yy." !! ""; }; froo(24); froo(24, 42)
p6eval niecza v24-12-g8e50362: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/lib/CORE.setting line 1287 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 266 (Mu.Str @ 15) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/PKLdLjr8FA line 1 (froo @ 4) ␤…
..rakudo c8de2e: OUTPUT«use of uninitialized value of type Any in string context in sub froo at /tmp/UjkC3z7vvW:1␤␤ with yy as .␤ with yy as 42.␤»
isBEKaml is that correct? This was out of p6book. :) 13:43
rn: sub froo($xx, $yy?) { say "x is $xx" ~ ($yy.defined ?? " with yy as $yy." !! ""); }; froo(24); froo(24, 42) 13:45
p6eval rakudo c8de2e, niecza v24-12-g8e50362: OUTPUT«x is 24␤x is 24 with yy as 42.␤»
isBEKaml oh, parens. how I love you. :|
13:46 SamuraiJack joined 13:47 rurban joined
pmurias jnthn: why do we have BUILDPLAN? 13:50
jnthn: why do we construct and execute the plan instead of just doing the work?
moritz because doing the work takes time 13:51
pmurias so it's supposed to avoid doing the traversal every time? 13:52
moritz yes
pmurias would be nice to have that mentioned in comments 13:54
not_gerd rn: constant @x = []; constant @y = @x; say @x =:= @y; my @u = []; my @v = @u; say @u =:= @v
p6eval rakudo c8de2e, niecza v24-12-g8e50362: OUTPUT«True␤False␤»
not_gerd doesn't that violate least surprise?
(assuming it's not a bug, of course) 13:55
moritz IMHO it's a bug
assigning to a constant
13:55 rurban left
isBEKaml OOC, what does =:= mean? 13:56
equality both by type and values?
13:57 SamuraiJack left
not_gerd isBEKaml: container identity, ie both variables are bound to the same thing 13:57
isBEKaml not_gerd: ah, thanks. 13:58
13:58 mikemol joined
not_gerd just got a 10x speedup by replacing constant with my 14:00
isBEKaml not_gerd: the difference is 0.5 sec on my machine (between constant and my) 14:02
"my" being the faster one. It took 5 secs to just go through encoding and decoding two strings. 14:03
14:03 donghongbo left
not_gerd isBEKaml: my @foo = 0..5, 10..20 creates a new array, whereas constant @foo = ... just binds the right-hand side 14:04
==> fast iteration in case of my, slow iteration in case of constant
14:04 lolage joined
moritz you can work around it by slapping .list on the RHS 14:04
but it's really a bug in rakudo 14:05
it should do proper list assingment
14:05 FROGGS_ left
isBEKaml not_gerd: I can see that. Also, s/:=/=/g gives me 2.5 sec running time. :) 14:06
14:06 SamuraiJack joined
moritz nr: my @a = 1..5; say @a.WHAT 14:08
p6eval rakudo c8de2e, niecza v24-12-g8e50362: OUTPUT«Array()␤»
isBEKaml moritz: do we know if variable bindings are slow on rakudo? 14:09
14:09 am0c left
isBEKaml moritz: that is, $var := $another-var or $var := literal-stuff 14:09
moritz isBEKaml: they aren't slow 14:12
isBEKaml: they are only slow if they lead to multiple evaluations in later code
14:14 bowtie joined 14:15 GlitchMr joined 14:22 rurban joined
isBEKaml moritz: sorry if I sound a bit confused - do "multiple evaluations" also include lookups? The way I see it, they shouldn't (ideally) slow things down, right? 14:22
moritz isBEKaml: I'm talking about multiple evaluations of lazy lists and stuff like that 14:23
isBEKaml: not lookups
not_gerd rn: my $n = 42; /a ** $n/ 14:25
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Quantifier quantifies nothing␤at /tmp/4yt24UZp5x:1␤------> my $n = 42; /a ** ⏏$n/␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ statement modif…
..niecza v24-12-g8e50362: OUTPUT«Potential difficulties:␤ Unsupported use of atom ** $n as separator; nowadays please use atom+ % $n at /tmp/lhcsayHHjd line 1:␤------> my $n = 42; /a ** $n⏏/␤␤»
isBEKaml moritz: wheww... okay. :)
14:27 rurban left, FROGGS joined
isBEKaml moritz: where can I find the full index of all functions documented in doc.perl6.org? 14:29
moritz: or is that what we have right on the first page?
moritz doc.perl6.org/#Routines 14:30
though that doesn't list the operators from language/operators
isBEKaml moritz: okay, I guess I'll start with some string functions. .subst, .trans and stuff like that. :)
14:35 cibs joined 14:40 spider-mario left 14:48 FROGGS left 14:55 cibs left 14:57 rurban joined 14:58 rurban left 15:00 mikemol left
masak ok, starting to write the macros advent post now. 15:01
15:03 cibs joined
felher masak: btw: here is my Perl6 and Haskell solution. The Haskell solution is quite long and not very efficient. But since I have little to no experience with Haskell and am quite busy, I was happy to get it working at all and then left it there :) 15:05
masak: gist.github.com/4363849 15:06
15:07 isBEKaml left
masak looks 15:07
felher: nice. 15:08
your solution is more... composable than mine.
felher: looks like you can get rid of 'list' almost anywhere.
felher: because of currying, when the left side and right side end in the same terms, you can just remove them, like in an equation. 15:09
felher masak: oh, right. Point free ftw :)
masak felher: your Perl 6 solution looks like what I wanted to write all along.
kudos.
filterWith, nice. 15:10
felher masak: thanks. I updated the gist to make the haskell version more point-free :) 15:12
masak \o/ 15:13
masak .oO( resistance is point-free )
felher :D
masak: btw. Your haskell version is much shorter and much more efficient I guess. I just was too lazy to look up Set and MultiSet :) 15:15
masak I got #haskell to help me with those. 15:17
my Haskell port is a straight port of my Perl 6, which was a straight port of the Python. 15:18
I essentially innovated nothing along the way.
felher masak: nothing wrong with that as long as it gives you a nice solution :) 15:21
15:23 mikemol joined 15:24 SamuraiJack left
masak r: macro checkpoint { state $n = 0; my $i = ++$n; quasi { say "CHECKPOINT $i" } }; checkpoint; say "foo"; checkpoint; say "bar"; checkpoint 15:26
p6eval rakudo c8de2e: OUTPUT«CHECKPOINT 1␤foo␤CHECKPOINT 2␤bar␤CHECKPOINT 3␤»
masak \o/
15:29 donghongbo joined
not_gerd bye, #perl6 15:31
15:31 not_gerd left, donghongbo left 15:32 rummik left
masak what's the biggest difference between C's textual macros and Perl 5's source filters? 15:32
15:32 lorn left
felher r: macro checkpoint { state $n = 0; quasi { say "CHECKPOINT {++$n}" } }; checkpoint for ^3; say "foo"; checkpoint; say "bar"; checkpoint 15:32
p6eval rakudo c8de2e: OUTPUT«CHECKPOINT 1␤CHECKPOINT 2␤CHECKPOINT 3␤foo␤CHECKPOINT 4␤bar␤CHECKPOINT 5␤»
15:32 rummik joined
felher r: macro checkpoint { state $n = 0; ++$n; quasi { say "CHECKPOINT {$n}" } }; checkpoint for ^3; say "foo"; checkpoint; say "bar"; checkpoint 15:33
p6eval rakudo c8de2e: OUTPUT«CHECKPOINT 3␤CHECKPOINT 3␤CHECKPOINT 3␤foo␤CHECKPOINT 3␤bar␤CHECKPOINT 3␤»
15:33 lorn joined
felher Ah, i think i got why there is a $n and a $i :) 15:34
masak closure semantics.
felher yep :) sweet :)
masak r: macro checkpoint { my $i = ++(state $n); quasi { say "CHECKPOINT $i" } }; checkpoint; say "foo"; checkpoint; say "bar"; checkpoint 15:35
p6eval rakudo c8de2e: OUTPUT«CHECKPOINT 1␤foo␤CHECKPOINT 2␤bar␤CHECKPOINT 3␤»
diakopter masak: lol point-free 15:36
masak felher: it's sweet, yes. it's like the code you're dealing with is alive, and smiling at you. it's not just text, it's *behavior*. :)
felher masak: indeed :)
masak felher: the more I program, the more I realize that we're all Dr Frankenstein. we're putting behavior into dead matter. and we're slowly getting increasingly good at it.
the future, barring some major cataclysm, will be awesome. 15:37
felher yeah :) 15:38
masak is on a roll with this advent post 15:44
15:51 isBEKaml joined 15:59 hash_table joined
masak r; macro checkpoint { my $i = ++(state $n); quasi { say "CHECKPOINT $i" } }; checkpoint; for ^5 { checkpoint }; checkpoint 16:00
r: macro checkpoint { my $i = ++(state $n); quasi { say "CHECKPOINT $i" } }; checkpoint; for ^5 { checkpoint }; checkpoint 16:01
p6eval rakudo c8de2e: OUTPUT«CHECKPOINT 1␤CHECKPOINT 2␤CHECKPOINT 2␤CHECKPOINT 2␤CHECKPOINT 2␤CHECKPOINT 2␤CHECKPOINT 3␤»
masak this nicely illustrates a difference between macros and subs.
16:02 GlitchMr left 16:04 erkan joined, erkan left, erkan joined
timotimo is there a problem regarding the 23rd advent post? 16:07
masak working on it. 16:08
(I meant to write it yesterday, but forgot)
timotimo that's fine :) 16:11
16:12 birdwindupbird joined 16:17 GlitchMr joined 16:26 Vlavv left 16:33 Vlavv_ joined 16:42 david joined 16:44 david left
isBEKaml rn: my $foo = "boo"; say $foo.subst(rx:m/boo/, "far", :mm); #NYI? 16:44
p6eval niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Unrecognized adverb :m(1) at /tmp/IYfl8z2IpH line 1:␤------> my $foo = "boo"; say $foo.subst(rx⏏:m/boo/, "far", :mm); #NYI?␤␤Regex modifier m not valid on regex literal at /tmp/IYfl8z2IpH line 1:␤------> …
..rakudo c8de2e: OUTPUT«===SORRY!===␤Adverb m not allowed on rx␤at /tmp/MuMFtXSjoB:1␤------> y $foo = "boo"; say $foo.subst(rx:m/boo/⏏, "far", :mm); #NYI?␤»
masak I don't think rx takes adverbs at all. 16:45
by spec, I mean.
isBEKaml masak: well then, the spec needs change. :)
S05
GlitchMr unlink Perl6 || true
aur.archlinux.org/packages/ra/rakudo/PKGBUILD
Interesting use of || true
16:55 mj41 left 16:58 birdwindupbird left 17:01 mj41 joined
masak advent post published: perl6advent.wordpress.com/2012/12/2...23-macros/ 17:07
tadzik yay
masak++
17:10 mj41 left, isBEKaml left
felher <3 "A Song of Ice and Fire" reference 17:15
masak "A textual macro is a sword without a hilt. There's no safe way to grasp it." 17:18
:)
felher masak++ # nice blog post about macros that makes you wish the end of the next year is already here :) 17:25
masak I don't promise all of those things will be in place by the end of next year. ;) 17:27
two of the points are within the current grant.
17:34 SamuraiJack joined 17:45 azawawi joined
felher masak: oh, there is a solution on Wikipedia for python. Is that the one you translated to Perl 6 and then to Haskell? 17:52
masak felher: it is. 17:58
I renamed all the variables and eliminated some of the Python boilerplate that Perl 6 doesn't need. (such as checking whether a hash entry exists before incrementing it)
felher masak: I see :) 18:01
azawawi hi 18:04
phenny azawawi: 02 Dec 18:44Z <japhb> ask azawawi Would you mind using a different pastebin than the one at paste.debian.net (or configuring it to keep the pastes for longer), because they are expiring quite quickly (sometimes faster than I can backlog if I've been AFK for a couple days)?
azawawi: 03 Dec 04:53Z <japhb> tell azawawi your 1 == any(range) benchmark is now in perl6-bench, with two range limits to show scaling; but note that sorear says it may be somewhat problematic against smart implementations (but at least it notices the difference between smart and not)
18:07 Abcxyz joined, tcpr1347 joined
azawawi feather.perl6.nl:3030/ # Farabi (with a Perl backend / JS Frontend), Web-based REPL, Run, Open file autocomplete... simple Perl 6 syntax highlighting... etc... etc :) 18:08
18:09 spider-mario joined
azawawi nap & 18:11
18:11 azawawi left
GlitchMr What macros are useful for? 18:26
Making gotos? 18:27
diakopter r: macro foo { quasi { {{{ macro boo { 1 } }}} } }; foo 18:28
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Error while compiling block : Error while compiling op call: Error while compiling block : Error while compiling block : Unknown QAST node type NQPMu␤»
18:28 mj41 joined
diakopter r: macro foo { quasi { say {{{ macro boo { quasi { {{{ 3 }}} } } }}} } }; foo 18:30
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Method 'flat' not found for invocant of class 'NQPMu'␤»
18:36 mtk left 18:38 mtk joined 18:40 xilo_ joined 18:50 Kharec joined
hoelzro hooray, heredoc hilighting! \o/ 18:56
GlitchMr :-)
timotimo masak: that was a good post :) 19:02
19:04 aindilis left 19:05 aindilis joined
masak timotimo: thanks. it did feel like it had something going for it. :) 19:12
GlitchMr I've done Niecza Arch module. I guess I can upload it now? 19:13
It looks like this: github.com/GlitchMr/arch-niecza 19:14
masak GlitchMr: what are macros useful for? I like the quote in the latest advent post as an answer to that question. 19:18
GlitchMr: I don't think you need anyone's permission to upload a Niecza Arch module. even if you do, forgiveness > permission. 19:19
GlitchMr well, ok 19:20
aur.archlinux.org/packages/niecza/
Arch already had nice Rakudo Perl package, so why not Niecza? 19:28
masak GlitchMr++ 19:34
Abcxyz (1..5 ==> reverse) ==> [\+] 19:35
r: (1..5 ==> reverse) ==> [\+]
p6eval rakudo c8de2e: ( no output )
GlitchMr "bash" as dependency is there just for fun (it's meaningless).
Abcxyz r: say (1..5 ==> reverse) ==> [\+]
p6eval rakudo c8de2e: OUTPUT«5 4 3 2 1␤»
GlitchMr (but it's used in /usr/bin/niecza)
Abcxyz r: say 1..5 ==> reverse ==> [\+] 19:36
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Preceding context expects a term, but found infix ==> instead␤at /tmp/quldCkupyL:1␤------> say 1..5 ==> reverse ==>⏏ [\+]␤»
sorear good * #perl6 19:38
Abcxyz I think the parenthesis should be redundant in this, right?
GlitchMr They aren't because of bug
PerlJam Greetings
GlitchMr std: say 1..5 ==> reverse ==> [\+]
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Preceding context expects a term, but found infix ==> instead at /tmp/CPBOrQQm3a line 1:␤------> say 1..5 ==> reverse ⏏==> [\+]␤Parse failed␤FAILED 00:00 42m␤»
tadzik in case you missed it: an open-source geeky board game:github.com/TableTopSecurity/d0x3d-the-game 19:39
Abcxyz Is operator: ==> right associative or left associative?
Where can I find these detailed information?
GlitchMr doc.perl6.org/language/operators says it's neither 19:40
masak it's a statement terminator.
hoelzro masak++ # GoT reference
Abcxyz OK... @_@ 19:41
tadzik heh, masak puts GoT references in his blogposts
you could say that...
( •_•)>⌐■-■
it is known 19:42
(⌐■_■)
YEEEEEAH
masak ;) 19:45
Abcxyz What is GoT reference? 19:47
tadzik GoT is Game of Thrones, a book 19:48
the beginning to a long series of books :)
hoelzro a very good series of books if you've got the time =) 19:49
tadzik yeah, you need plenty of time. And winder is coming
hoelzro indeed
Abcxyz I see... 19:52
I have some for Perl6 some bug reports and lots of new ideas.
Is this the right place to discuss these? 19:53
tadzik sure
19:53 FROGGS joined
Abcxyz 1. 19:54
r: print$_
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/1nZf4GbVMG:1␤------> print⏏$_␤ expecting any of:␤ argument list␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ statement modifier…
Abcxyz :r say 3x$_ 19:55
FROGGS r: say 3x$_
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Confused␤at /tmp/EXdrOZgj7K:1␤------> say 3⏏x$_␤ expecting any of:␤ whitespace␤»
FROGGS hi Abcxyz 19:56
flussence that one seems self-explanatory
Abcxyz Yes
Are all the white spaces mandatory in Perl6 now?
In Perl5 all these work fine.
FROGGS I think so, yes
tadzik not all of them, surely
but some of them, yes 19:57
FROGGS std: print$_
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Unsupported use of bare 'print'; in Perl 6 please use .print if you meant $_, or use an explicit invocant or argument at /tmp/VZePAZqOpi line 1:␤------> print⏏$_␤Two terms in a row (listop with args requires whitespac…
flussence the ones between [<alnum>|<punct>] ** 2 usually are
Abcxyz Will this improve in the future? Since most of them are so obvious according to context: there is no ambiguity at all.
19:59 hash_table left
FROGGS if withspaces around x arn't mandatory, you don't know if x in 3x$_ is suffix, infix or prefix 19:59
Abcxyz I see. 20:00
That's why Perl5 accepts them, but Perl6 rejects.
Since Perl5 doesn't have to distinguish these?
FROGGS Perl 5 only has infixes, right? 20:01
Abcxyz I guess so.
FROGGS except [] and () and stuff
huf and +
FROGGS but these can't appear as infix
right
huf and a ton of others :)
Abcxyz Yes.
FROGGS I'm not a language pro, just a "user"
Abcxyz Fine. 20:02
Usually I include whtespace in my one liner character countings.
That's why I checked these...
masak FROGGS: it's "whitespace", singular, as a mass noun. 20:03
Abcxyz Yes.
FROGGS ahh, see, everyday I learn something ;o)
masak Abcxyz: even where there is no ambiguity, someone could define new operators and there would be ambiguity.
Abcxyz Right. 20:04
That is what is different in Perl6 from Perl5 now.
So many possibilities of new operators.
doy where are the rules for where whitespace is required documented?
Abcxyz I don't think it is fully documented. 20:05
But Perl5 uses heuristics to parse.
So usually it understands you and does the right thing.
Just like when you talk to a people face to face, you don't have to be grammatically correct, but they understand you.
That's human language, right? 20:06
And a lot of time Perl5 is parsing in that way.
doy perl 5 only uses heuristics in very specific places, and i can't think of any that involve whitespace (except for a few annoying things that are newline-sensitive)
Abcxyz I guess I would define it this way: 20:08
When there is ambiguity, one of them makes the most possible sense, then interpret it this way.
For example 20:09
The defined-or operator: //
and the regexp matching operator: //
If there is ambiguity, Perl5 interprets it as defined-or 20:10
doy "one of them makes the most sense" is only usually true
Abcxyz Since you can always write the other in the form: m//
masak doy: if you count STD.pm6 as documentation, it's all in there.
huf Abcxyz: can you even have defined-or in a position to confuse it with //?
masak doy: otherwise, the rule is quite simple: the operator after a term is infix if there's whitespace in-between or post(circum)fix if not. 20:11
Abcxyz Maybe not, but the parser is not smart enough to distinguish that.
Yet.
So it uses heuristics directly in this case.
doy masak: thanks, that was what i was curious about
huf: yes, you can 20:12
Abcxyz And another case, but we almost don't realize:
1..10
masak Abcxyz: in Perl 6, an empty match // is disallowed.
Abcxyz 1. . 0
This is ambiguity.
But we almost always mean 1 .. 10
masak Abcxyz: and in the case of 1..10, Perl 6 disallows '1.' as a number :)
you see, we put great thought into this. :) 20:13
Abcxyz So the parse treats it that way, the most possible one, being understanding.
masak ...whereas in Perl 6, we try to make the underlying parsing sane rather than DWIMmy.
will probably scale better to derived dialects of Perl 6.
Abcxyz What does it mean by: disallowing '1.' as a number?
masak std: 1.
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Decimal point must be followed by digit at /tmp/rEFlzRG2zo line 1:␤------> 1.⏏<EOL>␤Unsupported use of . to concatenate strings; in Perl 6 please use ~ at /tmp/rEFlzRG2zo line 1 (EOF):␤------> 1.⏏<EOL>…
masak means it's not allowed. 20:14
Abcxyz ...
That sounds scary to me.
masak boo!
20:14 pmurias left
Abcxyz Since it means if I am Copy and Pasting the numerical results from other program's output into my perl script, I have to modify these every time? 20:15
Before, perl is always the most flexible and versatile language in these kind of things...
timotimo quote it and parse it with a simple regex? 20:16
doy if you're copying and pasting numbers that use , as a decimal point, you have to modify them every time too
there are all kinds of number formats, it's a bit unreasonable to expect perl to parse all of them
Abcxyz I don't have to.
masak also, it's *output*. when was the last time you saw "1." or similar in output?
Abcxyz use pdl [ STRING ]
if there is comma: ,
use qw{ STRING } if there is whitespace 20:17
So easy.
Many language does this: in C: double d = 1.; 20:18
Although you don't have to. 20:19
masak r: say "1. 2. 3.".subst(/<?after \d> \./, "", :g)
p6eval rakudo c8de2e: OUTPUT«1 2 3␤»
Abcxyz But some language enforces the decimal point.
masak so easy.
you have a *programming language* at your fingertips.
slightly wrongly formatted output shouldn't hinder you.
r: say "1. 2.0 3".subst(/<?after \d> \. <!before \d>/, "", :g) 20:20
p6eval rakudo c8de2e: OUTPUT«1 2.0 3␤»
Abcxyz I know, there can always be work arounds.
masak this isn't even that.
it's just adapting output before you paste it as code. :)
Abcxyz :r say 1 20:21
:r say 1.0
masak r: say 1.0
p6eval rakudo c8de2e: OUTPUT«1␤»
Abcxyz :r say 1.
:r say 0.1
:r say .1
masak r: say 1.
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Confused␤at /tmp/TfkY83FcTT:1␤------> say 1.⏏<EOL>␤ expecting any of:␤ dotty method or postfix␤»
20:21 pmurias joined
masak r: say .1 20:21
p6eval rakudo c8de2e: OUTPUT«0.1␤»
Abcxyz In all these forms, only "1." is rejected. 20:22
masak aye.
that's the form that is causing trouble with things like infix:<..> and method dot. 20:23
Abcxyz Just make the regexp greedy then...
timotimo r: sub postfix:<.>($i) { $i }; say 1.;
p6eval rakudo c8de2e: OUTPUT«1␤»
timotimo i win.
masak timotimo: :)
Abcxyz As in 1..10, we always mean 1 .. 10 without even notice there is such underlying ambiguity.
FROGGS masak: how do you make multiline commit messages in git with a subject? do you add \n's to -m="" ? 20:24
masak Abcxyz: yes, but the parser is already in the 'number' rule, parsing the term.
Abcxyz: it's won't back out if it already saw '1.', and then it won't understand '.10'
Abcxyz: "Just make the regexp greedy" is more like the problem, not the solution.
FROGGS: I tend to write them in $EDITOR... 20:25
FROGGS: but in a Unix environment, you can just hit Enter until you terminate the quote.
FROGGS: (in Windows, not so much)
FROGGS okay, thanks
Abcxyz Yes, I agree it is a implementation difficulty. 20:26
timotimo i believe you can type \ before hitting enter in DOS.
Abcxyz I just mean that in real life the meaning is almost certain, as my previous definition for heuristics, the most possible interpretation is clearly there.
masak Abcxyz: indeed. 20:27
doy r: sub postfix:<.> (Int $a) { $a }; say 1.
masak Abcxyz: and Perl 5 takes that road.
p6eval rakudo c8de2e: OUTPUT«1␤»
Abcxyz Right.
masak Abcxyz: Perl 6 doesn't. it lets every rule do its thing.
which leads naturally to '1.' being disallowed. 20:28
doy r: sub postfix:<.> (Int $a) { $a }; say 1. + 2.
p6eval rakudo c8de2e: OUTPUT«3␤»
Abcxyz I can imagine it is because the practical way we implements it.
masak r: sub postfix:<.> (Int $a) { $a }; say 1..10
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/_qp7pQyZa0:1␤------> sub postfix:<.> (Int $a) { $a }; say 1..⏏10␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ state…
masak Abcxyz: no, it's because it's *consistent*.
Abcxyz: I explained the reasons to you. did you understand them when I did? 20:29
Abcxyz I guess partly.
20:29 SamuraiJack left
masak I agree that it is a fair bit different from how Perl 5 does it. 20:29
Abcxyz Yes.
masak that's because the guy who wrote the Perl 5 parser is involved in this project too, and he chose what he percieves to be a better way this time. 20:30
Abcxyz He has changed. :)
masak his code sure has :)
20:30 stevan_ left
timotimo it would be a real shame if perl6 were just perl5 reheated 20:31
20:31 stevan_ joined
Abcxyz It is not impossible to solve it in the grammar - rule system. But probably that will bring in a lot extra complexity to the grammar just to solve this one small problem, and it might be simpler just to make the rules easier to the computer. 20:32
lue hello world o/
timotimo not using the knowledge you gained from the previous version to create not only a "new" version , but also an "improved version" ... that would be sad
doy masak: why doesn't ltm fix that? 20:34
shouldn't .. be a priority over .?
Abcxyz It is not easy. 20:35
It takes a lot of work.
And it only solve a small problem, so it probably won't be a priority when there are other things to improve now. 20:36
Unless for perfectionism
FROGGS phenny: tell jnthn that this is my questionable patch: github.com/FROGGS/rakudo/commit/eb...213da2b43d 20:37
phenny FROGGS: I'll pass that on when jnthn is around.
masak doy: because LTM decides which alternation wins. if the parser is already inside a term, parsing it, the decision between a decimal dot and the '..' isn't up to LTM, the decimal dot automatically wins as being part of the term. 20:39
Abcxyz: if you're building a language that's meant to be extended, it makes a whole lot of sense to keep exceptions to a minimum.
lue! \o/
doy so effectively, it's just because .. is infix and . is postfix? 20:40
lue
.oO(If C<1.> was allowed, what would happen in C<1...10> ?)
masak doy: oh! in your case with the postfix, it's because there's no whitespace. 20:41
as I said above, that's how infix is distinguished from post(circum)fix.
20:41 stevan_ left
masak lue: also a good point. 20:41
doy masak: hmmm 20:42
lue also personally never liked how "1." looked visually.
doy so infix operators are only disambiguated by whitespace then
it's only relevant if there's an ambiguity?
postfix operators just always win if there is no whitespace, but postfix operators fail to even be considered if there is intervening whitespace 20:43
is that it?
masak if there's no whitespace, and both a postfix operator and an infix operator (not necessaarily of the same length) can match, the postfix wins. 20:44
if there's whitespace, only infix operators are considered.
doy okay 20:46
that makes sense
20:47 slayer101_ joined
masak what's more, it keeps making sense as new terms and operators are added in derived Perl 6 dialects. :) 20:48
doy yeah
perl 6's whitespace rules have just always been something that seemed a bit odd and confusing to me 20:49
but i feel like that's probably because i just don't know what they are
in perl 5, you can get by almost entirely with just "word boundaries are equivalent to whitespace", or something along those lines, perl 6 does things a bit differently 20:50
20:51 hash_table joined
moritz masak++ # macro advent post 20:52
doy i think some amount of documentation/spec that explains exactly what those rules are would be helpful
20:52 sid_ joined
masak doy: I agree. 20:53
it's information that just sits in STD.pm6, the compilers, and the brains of old regulars like me :P
should be documented somewhere. 20:54
Abcxyz Is it possible to define an operator "...^*" other than "...^" ?
moritz calls for a braindump
Abcxyz: yes, if I understood the question correctly
the longest operator wins, in case of doubt 20:55
timotimo r: sub infix:<...^*>($a, $b) { say "i have $a and $b"; $a...^$b }; say 5...^*10;
p6eval rakudo c8de2e: OUTPUT«i have 5 and 10␤5 6 7 8 9␤»
timotimo easiest to just try
20:55 cognominal joined
Abcxyz So after I've defined that, will this work without whitespace? "1, *+1, ...^*>10" 20:56
Sorry, "1,*+1...^*>10"
moritz Abcxyz: why don't you just try it?
Abcxyz :r 1,*+1...^*>10 20:57
masak it's 'r' and then ':'
Abcxyz r: 1,*+1...^*>10
p6eval rakudo c8de2e: ( no output )
masak and you need to 'say' your output.
Abcxyz r: say 1,*+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
20:58 sid_ left
Abcxyz r: sub infix:<< ...^*> >>($a, $b){}; say 1,*+1...^*>10 20:59
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
20:59 stevan_ joined
masak WhateverCode.new is what '*+1' comes out looking like. 20:59
Abcxyz r: say 1,*+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
masak er, '*+(1...^*>10)', I mean. 21:00
Abcxyz Interesting.
masak actually, I'm not 100% sure where the '>' ends up in the AST. 21:01
Abcxyz So it interprets it as: ( 1, ( * + ( 1 infix:<< ...^*> >> 10 ) ) ) 21:02
21:02 GlitchMr left 21:05 hash_table left
Abcxyz :r 1, * + Nil 21:05
masak 'r', then ':'
Abcxyz ...
masak it's short for 'rakudo:' 21:06
moritz :wq!
hugme hugs moritz, good vi(m) user!
Abcxyz r: 1, * + Nil
p6eval rakudo c8de2e: ( no output )
LoRe do not use :wq, use :x instead
masak hugs moritz, too, even though he probably did it intentionally ;)
LoRe: nono, use ZZ :)
21:06 FROGGS left
timotimo no, pull the plug and do everything with swap files 21:06
Abcxyz r: say 1, * + Nil
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
LoRe or that, but :wq is harmful 21:07
diakopter rn: :wq
Abcxyz I fogot the 'say' again
p6eval rakudo c8de2e, niecza v24-12-g8e50362: ( no output )
Abcxyz r: say 1, * + Nil
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
Abcxyz r: sub infix:<< ...^*> >>($a, $b){}; say 1,*+1...^*>10 21:08
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
Abcxyz It's the same thing. 21:09
tadzik LoRe: how is wq harmful?
LoRe tadzik: it always writes the file, even if it is unchanged, :x only writes if modified
21:09 FROGGS_ joined
moritz r: sub infix:<< ...^*> >>($a, $b){ 'blubb'}; say 1,*+1...^*>10 21:10
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
moritz r: sub infix:«...^*>»($a, $b){ 'blubb'}; say 1,*+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
LoRe ok, it's not that harmful, but changes timestamps
timotimo you need to call it
Abcxyz r: say 1, * + 'blubb'
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
timotimo the * + is doing the conversion
tadzik yeah, :x seems saner
Abcxyz That' how it reads
moritz r: sub infix:«...^*>»($a, $b){ 'blubb'}; say 1,+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1blubb␤»
moritz timotimo++
21:11 bruges joined
masak nice. paulgraham.com/fix.html # via @sheeshee++ 21:14
moritz I disagree with some of the details, but in general, yes, nice 21:15
masak there's certainly room for disagreement. 21:16
diakopter Perl 6: No other language has enough of anything
moritz (including insanity) 21:17
masak or "Perl 6: Perl 5 is a kludge." 21:18
guess that one's almost too obvious to state.
Abcxyz So even the Perl group consider this as true?
diakopter Perl 6: Where else will you find the kitchen sink?
masak Abcxyz: everyone is entitled to their own opinions. everyone isn't entitled to their own facts. ;)
timotimo everyone is unentitled? 21:19
masak Perl 5 *is* a kludge. one that makes bazillions out in the industry each day, but still a kludge.
Abcxyz I think Perl5 is really nice, except for so many historical vestiges that cannot be thrown away for compatibility sake. 21:20
masak Perl 6 is very much sitting down and saying "there are *good* things in there. let's put them on a sane foundation and see what language we get."
lue
.oO(Perl 6: There's too big a leap required to get to P7)
Abcxyz The regexp is a big surprise to me, since before I was using C++ to treat streams...
masak (this was what my 24th slot post was gonna be about, btw) :)
lue masak: what specifically? 21:21
masak lue: "You should think of Perl 6 more as Perl 16" :)
lue: nothing specifically. the whole *language* is a rewrite according to that kind of thinking.
diakopter Perl 6: No ho-hum. Just tra-la-la. 21:23
lue (I'll assume you didn't mean P16 < P6)
diakopter Perl 6: 100 years ahead, but 12 years behind, so 88 years ahead. 21:24
masak diakopter: *lol*
flussence 88MPH?!? 21:25
something I was thinking about recently, a generic HTTP test suite would be useful to have. Trying to figure out how it'd be automated is pretty hard though... 21:30
masak mocks.
if you can mock time, surely you can mock HTTP ;)
lue perl6: "because I" ==> (say so $_);
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Sorry, do not know how to handle this case of a feed operator yet.␤at /tmp/ETZ8Sg20EY:1␤------> "because I" ==> (say so $_)⏏;␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤»…
..niecza v24-12-g8e50362: OUTPUT«Unhandled exception: Feed ops NYI␤ at /home/p6eval/niecza/lib/CORE.setting line 1435 (die @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3353 (infix:<==>> @ 4) ␤ at /tmp/xQb4gHHohG line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting li…
masak flussence: en.wikipedia.org/wiki/Mock_object
flussence: github.com/jnthn/test-mock 21:31
flussence ooh, pre-existing stuff
masak :) 21:32
21:32 stevan_ left
diakopter Perl 6: Because no one blithely loves recursive Death Marches enough 21:33
cognominal that's perl 6..* the developpers have almost finished their first take to reify the sequence beginning to gather the first version number. 21:34
flussence masak: I have a sort-of plan for doing it already, it'd be a runtime-agnostic list of instructions for what requests and responses are supposed to happen, then I write a bailador app (or whatever) to implement those. It wouldn't necessarily have to be perl6-specific either :)
cognominal I suppose that 6..** with hyperwhatever is infinity and beyond 21:35
diakopter Perl 6: Because no one had taken Second System Syndrome far enough yet 21:36
masak diakopter: note that's pg's original one's weren't overly bitter. 21:37
ones*
lue Perl 6: No-one had yet created their language with -Ofun
masak I'm pretty sure that's not strictly true :)
diakopter masak: truly one can state those entirely non-bitterly; blithely even 21:38
some people love death marches
timotimo well, if laughs are equivalent to fun, LOLCODE would certainly be a contestant.
diakopter masak: embrace the fail; it loves you anyway 21:40
masak I... guess.
diakopter I mean, one could talk about redefining success.. 21:41
or just leaving it uselessly ambiguous 21:42
Abcxyz r: say sub infix:«...^*>»($a, $b) is looser(&infix:<+>){ 'blubb' } 21:46
p6eval rakudo c8de2e: OUTPUT«sub+{precedence} infix:<...^*>>($a, $b) { ... }␤»
Abcxyz r: say 1,*+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
timotimo that .perling is misleading 21:47
Abcxyz r: sub infix:«...^*>»($a, $b) is looser(&infix:<+>){ 'blubb' }; say 1,*+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1WhateverCode.new()␤»
Abcxyz r: sub infix:«...^*>»($a, $b) is looser(&infix:<+>){ 'blubb' }; say 1,+1...^*>10
p6eval rakudo c8de2e: OUTPUT«1blubb␤»
21:49 mj41 left 21:57 slayer101_ left 21:59 pmurias left 22:01 pmurias joined, kaare_ left 22:02 Kharec left
Abcxyz r: say given 9 { 0 when $_ %% 3; 1 } 22:03
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Two terms in a row␤at /tmp/bsaYNIOHpY:1␤------> say given 9 ⏏{ 0 when $_ %% 3; 1 }␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤»
Abcxyz r: sub s () { given 9 { 0 when $_ %% 3; 1 } }; say s; 22:04
p6eval rakudo c8de2e: OUTPUT«1␤»
Abcxyz Why this does not return 0 ?
masak Abcxyz: because statement-mod 'when' doesn't leave the contextualizer block. 22:06
the ordinary 'when' does.
Abcxyz so $_ is Nil? 22:09
I am translating my previous Perl5 ProjectEuler code into Perl6. 22:10
So all these are working code in Perl5.
It's good way to find bugs also. 22:11
masak in Perl 5, statement-mod 'when' does leave the contextualizer block.
but that's a bug in Perl 5, not in the Perl 6 implementations ;)
(IMNSHO)
Abcxyz What does it mean by leaving the contextulizer block? How does that effect $_?
masak r: sub s () { given 9 { when $_ %% 3 { 0 }; 1 } }; say s; 22:12
p6eval rakudo c8de2e: OUTPUT«0␤»
masak does that answer your question?
Abcxyz Yes,
but what is the meaning of your previous sentence?
Is there a link I can read on this? 22:13
22:14 mj41 joined
masak S04, I guess. 22:15
22:21 stevan_ joined
doy r: sub s () { given 9 { return 0 when $_ %% 3; 1 } }; say s 22:22
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Null regex not allowed␤at /tmp/L1ou2HrobS:1␤------> 9 { return 0 when $_ %% 3; 1 } }; say s⏏<EOL>␤ expecting any of:␤ argument list␤ prefix or term␤ prefix or meta-prefix␤»
doy r: sub s () { given 9 { return 0 when $_ %% 3; 1 } }; say s;
p6eval rakudo c8de2e: OUTPUT«0␤»
masak that first one is arguably a bug. 22:24
r: sub s {}; say s
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Null regex not allowed␤at /tmp/LJm1RVH5Du:1␤------> sub s {}; say s⏏<EOL>␤ expecting any of:␤ argument list␤ prefix or term␤ prefix or meta-prefix␤»
masak submits rakudobug 22:25
std: sub s {}; say s 22:26
p6eval std a8bc48f: OUTPUT«===SORRY!===␤No delimiter found at /tmp/0ivzQQ5JQE line 1 (EOF):␤------> sub s {}; say s⏏<EOL>␤Parse failed␤FAILED 00:00 42m␤»
masak std: sub s {}; say s;
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Regex not terminated at /tmp/kOLlJYIE3C line 1 (EOF):␤------> sub s {}; say s;⏏<EOL>␤Parse failed␤FAILED 00:00 43m␤»
masak right.
it's the 's;' case that is wrongly parsed.
'night, #perl6 22:28
22:31 mcglk joined
pmurias jnthn: ping 22:35
masak: 'night 22:36
Abcxyz night
gattschardo std: sub s {}; say s(); 22:38
p6eval std a8bc48f: OUTPUT«ok 00:00 42m␤»
22:41 sftp joined 22:43 cognominal left
Abcxyz std: sub s {}; say s(); 22:50
p6eval std a8bc48f: OUTPUT«ok 00:00 42m␤»
Abcxyz std: sub s {}; say s()
p6eval std a8bc48f: OUTPUT«ok 00:00 43m␤»
Abcxyz std: sub s {}; say s
p6eval std a8bc48f: OUTPUT«===SORRY!===␤No delimiter found at /tmp/WZYxiQM5mo line 1 (EOF):␤------> sub s {}; say s⏏<EOL>␤Parse failed␤FAILED 00:00 42m␤»
Abcxyz std: sub s {}; say s;
p6eval std a8bc48f: OUTPUT«===SORRY!===␤Regex not terminated at /tmp/xKF_rk8RnY line 1 (EOF):␤------> sub s {}; say s;⏏<EOL>␤Parse failed␤FAILED 00:00 43m␤»
Abcxyz It seems that Perl6 does not understand bareword functions well. 22:51
diakopter r: sub s { 55 }; say s; 22:52
p6eval rakudo c8de2e: OUTPUT«55␤»
diakopter r: sub s { say 55 }; s
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Null regex not allowed␤at /tmp/fATj0WLr68:1␤------> sub s { say 55 }; s⏏<EOL>␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ statement modifier␤…
diakopter r: sub s { say 55 }; s;
p6eval rakudo c8de2e: OUTPUT«55␤»
diakopter r: sub s { say 55 }; s␤
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Null regex not allowed␤at /tmp/O5CgWGj9zh:2␤------> <BOL>⏏<EOL>␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ statement modifier␤ statem…
Abcxyz In: "for lines() { ... }", the parenthesis cannot be omitted neither. 22:53
I guess this should be a bug for now.
diakopter rn: for (lines) { .say }
p6eval rakudo c8de2e: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmi…
..niecza v24-12-g8e50362: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdtei…
Abcxyz What happens if you remove the "()" 22:54
diakopter rn: for lines { .say }
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Missing block␤at /tmp/DDJpTN5vRR:1␤------> for lines { .say }⏏<EOL>␤ expecting any of:␤ method arguments␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ stat…
..niecza v24-12-g8e50362: OUTPUT«===SORRY!===␤␤Missing block at /tmp/5Ve7vISWzO line 1 (EOF):␤------> for lines { .say }⏏<EOL>␤␤Parse failed␤␤»
Abcxyz r: for lines { .say } 22:55
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Missing block␤at /tmp/cf3jsPDorE:1␤------> for lines { .say }⏏<EOL>␤ expecting any of:␤ method arguments␤ postfix␤ infix or meta-infix␤ infix stopper␤ statement end␤ stat…
Abcxyz r: for (lines) { .say }
p6eval rakudo c8de2e: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmi…
Abcxyz 123
r: for lines() { .say }
p6eval rakudo c8de2e: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmi…
timotimo r: say glob
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Undeclared routine:␤ glob used at line 1␤␤»
timotimo r: say qp/./.^methods 22:56
p6eval rakudo c8de2e: OUTPUT«===SORRY!===␤Confused␤at /tmp/piWNoX8QHs:1␤------> say qp/.⏏/.^methods␤ expecting any of:␤ dotty method or postfix␤»
22:57 REPLeffect left, mj41 left, REPLeffect joined 23:00 spider-mario left 23:06 xenoterracide joined 23:09 cognominal joined 23:17 REPLeffect left, REPLeffect joined 23:25 PacoAir left 23:28 Tim-Tom joined
Abcxyz How do you express this idea? 23:33
my @list = 5..10; say 0 ...^ * -> $idx { @list[$idx] > 8 }
23:35 rindolf left, pmurias left
Tim-Tom Abcxyz: I'm not sure what you're attempting. Do you want to report all elements of list that are greater than 8, create a list of true false values for items greater than 8? 23:35
23:37 mj41 joined
Tim-Tom Abcxyz: As it is, what you've written would report false until it gets to the 4th element, then report true for 3 elements, then false for all eternity (perhaps with warnings about undef). Which is probably not what you intended 23:39
Abcxyz I wanted to get all the index for @list that contains the elems le 8 23:40
my @list = 5..10; say @list[ 0 ...^ * -> $idx { @list[$idx] > 8 } ]
With this I hope to print: 5 6 7 8
23:42 xinming left
Abcxyz Because: "my @list = 5..10; say @list[ 0 ...^ @list[*] > 8 ]" Definite does not do what I want, the * in @list[*] is used as @list.elems instead. 23:42
Tim-Tom Well if your end result is supposed to be 5,6,7,8, you probably want "my @list = 5..10; say @list.grep: * <= 8" 23:43
23:44 xinming joined
Abcxyz Yeah, maybe I should do that. 23:44
I just wanted to make the * really functional.
In a sense that they may be more interesting sequences that it can express directly. 23:45
But I can't foresee that yet.
Tim-Tom But if you want to go via indexes, you really want 0 ..^ @list
or the equivalent just ^@list
Abcxyz I want to make the "* > 8" in "0, *+1 ...^ * > 8" a more powerful function. 23:48
Tim-Tom Sorry, that should apparently be ^+@list
Abcxyz I know.
I wanted to know how to express something beyond the * > 8.
Like @list[*] > 8 23:49
Or rank(*) > 8 23:50
Tim-Tom Ah. The simple answer is to just use the subroutine form. so 0, *+1 ...^ { @list[$^a] <= 8 }
Or -> $a { @list[$a] <= 8} if you don't like the automatic parameters 23:51
Abcxyz Yes!
This is exactly what I want!
Thanks!
Tim-Tom Sure, glad I could help :)
Abcxyz r: my @list = 5..10; 0, *+1 ...^ -> $a { @list[$a] > 8 } 23:52
p6eval rakudo c8de2e: ( no output )
Abcxyz r: my @list = 5..10; say 0, *+1 ...^ -> $a { @list[$a] > 8 }
p6eval rakudo c8de2e: OUTPUT«0 1 2 3␤»
Abcxyz r: my @list = 5..10; say @list[ 0, *+1 ...^ -> $a { @list[$a] > 8 } ]
p6eval rakudo c8de2e: OUTPUT«5 6 7 8␤»
Abcxyz Thanks 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!