»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_log/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 July 2018.
timotimo i made an animated 00:08
timotimo not sure if imgur is liking the file 00:08
if not, it'll go up on hack
timotimo hack.p6c.org/~timo/flashy_ca_video.mp4 - potentially bad for people with epilepsy or motion sensitivity 00:11
(also, it's sped up a little)
Herby_ hmm 00:12
in firefox nightly, getting a "video can't be played because the file is corrupt"
timotimo :o 00:15
MasterDuke same in ff 61 00:20
timotimo can you download it and play it with any old media player? 00:20
MasterDuke works in vlc. and man, you weren't kidding about the flashing 00:22
timotimo my deflicker mode didn't do much :D
Herby_ could be my browser, wont let me download it either 00:24
timotimo o_O
i've got a different one that's less flashing 00:25
timotimo but man, i should really render it differently than this :d 00:25
MasterDuke what is it? i hope not a visual virus meant to infect the human brain 00:28
timotimo just a simple cellular automaton
SmokeMachine How difficult would it be to some one port Rakudo to iOS, as some one did with python? 00:46
SmokeMachine That is an app I’d love to buy! 00:49
Herby_ timotimo: what text editors or ide do you use for your p6 coding? 01:45
MasterDuke Herby_: over in #moarvm he said he was going to bed about an hour ago 01:59
Herby__ ahh ok
MasterDuke: what do you use?
I'm pingponging between vscode and vim
MasterDuke vim 02:00
Herby__ hmmm 02:01
rindolf hi all! Please help with the error in github.com/shlomif/perl6-CI-Gen/bl...CI/Gen.pm6 08:00
rindolf see www.shlomifish.org/Files/files/text/f.txt 08:02
buggable New CPAN upload: Trait-Env-0.5.1.tar.gz by SCIMON modules.perl6.org/dist/Trait::Env:cpan:SCIMON 08:03
rindolf some of the rakudo errors are not too helpful
rindolf anyone? 08:06
scimon Just looking. 08:12
rindolf how do i pass an array ref to a subroutine? 08:14
scimon xml-g-write-bash is calling write-bash with named parameters ( param-name=>$param-name etc) but wr-tebash is expecting positional parameters.
Perl6 doesn't have Array Ref's it's got array objects. 08:15
Xliff \o
scimon That you can just pass.
Xliff What's the best way to convert a signed integer to uint8/16/32/64?
Xliff I was doing $int +& 0xffff, but that's only good for unsigned. 08:16
scimon Easiest fix for rindolf is check write-bash to have this signature : (:$param-name, :@pkgs )
scimon Makes the arguments named. :) 08:16
Not sure about the signed thing. Bit shifting isn't my forte.
Xliff ($int +& 0xffff) * ($int < 0) ?? -1 !! 1 08:17
Or.., maybe ($int +& 0x7fff) * ($int < 0) ?? -1 !! 1
rindolf scimon: thanks 08:18
rindolf scimon: thanks, fixed now 08:25
Xliff: perhaps use $int.abs
Xliff rindolf: Good point. Thanks! 08:26
rindolf Xliff: you're welcome 08:27
lizmat weekly: brrt-to-the-future.blogspot.com/201...pdate.html 08:35
notable6 lizmat, Noted!
scimon rindolf: www.youtube.com/watch?v=obYlOurt-44 A little talk I did on signatures in Perl6. 08:41
(I did a longer one with Types and Multi Methods as TPCiG)
rindolf scimon: ah 08:42
scimon: thanks
scimon: are there slides? 08:43
masak woot, hygienic macros! brrt++
scimon www.slideshare.net/SimonProctor8/p...signatures 08:44
rindolf scimon: thanks
scimon github.com/ugexe/Perl6-ecosystems : So this says is updates every 1-2 hours but it's also saying it last updated 11 hours ago? It that because it only pushes changes if there's been releases? 08:51
scimon rindolf: No problems. I hope it helps :) 08:52
rindolf scimon: it does 08:55
scimon Cool... I'll wait :) 09:05
(Problem is the 0.5.0 release of Trait::Env I did last night kinda breaks. I rolled out a fix but as I'm using it in work I've sort of borked a release pipeline. Ho hum). 09:07
On the bright side my @path is env; is now a thing and does what you expect so that's kinda cool.
pmurias the irc log seems to be down :( 10:44
timotimo isup.me confirms that it seems to be down 11:07
timotimo hack.p6c.org/~timo/wobbly_ca_anim.mp4 - new animation that's not nearly as flashy as the previous one! 12:02
pmurias timotimo: that's footage of a Perl 6 game of life? 12:03
timotimo it's different from game of life in many respects, but yeah. also, that's not a screen recording, that's frames rendered out and then assembled into a video 12:04
are you interested in the specifics? :) 12:06
ggoebel can't backlog today :-( colabti.org is inaccessible. Does anyone know what's up? 12:39
timotimo not sure, just that multiple people noticed :( 12:41
pmurias maybe the EU sent a strike team to take out our remaining irc log? 12:43
;)
timotimo i'll try to get you something
hack.p6c.org/~timo/perl6_log.html 12:47
ggoebel: check this out ^ 12:49
i swear i saw someone else ask for logs 12:52
anyway, same url format for moarvm and perl6-dev as well
lichtkind how do i say that a role relies on a attribute of the class to be present? 13:30
timotimo well, attributes are, to an outside observer, "just" methods, so you could ask for a method to be present 13:32
lichtkind yes
thank you
and you can not get a sub via a role? 13:33
timotimo no, subs are a package thing, not a class thing 13:34
lichtkind allright thatswhy they were not mentioned , thanks again 13:35
timotimo no prob :)
er
i should be more clear
subs are "my" scoped by default, so they aren't even a package thing, but totally lexical
"our" subs are a package thing, though 13:36
El_Che good safe :) 13:37
yoleaux 4 Sep 2018 19:09Z <AlexDaniel> El_Che: Not sure what's going on in M#951 but can I have arm arch configured for the canary repo? :)
synopsebot M#951 [open]: github.com/MoarVM/MoarVM/issues/951 [⚠ blocker ⚠] Problem running programs on RPi with 2018.08
yoleaux 4 Sep 2018 19:11Z <AlexDaniel> El_Che: and if it's possible, then I guess arm64 won't hurt also?
timotimo yeah, actually it'd be fair to say i was wrong the first time
lichtkind: ^- important clarification/correction
Geth doc: 0b8c7ba2f3 | (Will "Coke" Coleda)++ | doc/Type/Proc.pod6
Treat code block as signature
 13:38
synopsebot Link: doc.perl6.org/type/Proc
El_Che .tell AlexDaniel` That needs some work. Travis is x64 only, so will have to run a qemu image to run the acutal image. It can be done, but not easily 13:40
yoleaux El_Che: I'll pass your message to AlexDaniel`.
vrurg Hi! Any known workaround for bug rt.perl.org/Public/Bug/Display.html?id=130794 ? I can only think of splitting conflicting symbols in separate files and still not sure if it would work out. 14:03
sarna hey, what does * mean in a twigil like @*foo? 14:10
geekosaur normally, a dynamic variable. but in a signature it's special (dynamic variables can't occur there), means its "slurpy" 14:12
timotimo well, in signatures the * actually goes in front of the sigil 14:13
though saying *@ is a twigil is probably also not wrong 14:14
like, an argument could be made
geekosaur whoops, rigth
geekosaur anyway docs.perl6.org/language/variables#The_*_twigil 14:16
El_Che I always think of %*ENV 14:17
and I translate dynamic to global in my head 14:18
sarna thanks :) 14:19
sarna can I somehow call subs like methods so I get a nice .foo instead of foo($_) while dealing with a topic variable? 15:30
sena_kun m: sub foo($_) {.say}; 1.&foo; 15:31
camelia 1
sena_kun sarna, like this? ^
m: sub foo($_) {.say}; my $_ = 1; .&foo; 15:32
camelia Potential difficulties:
Redeclaration of symbol '$_'
at <tmp>:1
------> 3sub foo($_) {.say}; my $_7⏏5 = 1; .&foo;
1
sena_kun It gives a warning because of an explicit declaration, but I hope I understood you correctly. :) 15:32
sarna sena_kun: whoa, that's interesting. I can't name the parameter though in this case, can I? 15:34
sena_kun sarna, I think you can. I just used `$_` as parameter to make code shorter. :) 15:35
m: sub foo($foo, $bar) {say $foo + $bar}; my $_ = 2; .&foo(2);
camelia Potential difficulties:
Redeclaration of symbol '$_'
at <tmp>:1
------> 3foo($foo, $bar) {say $foo + $bar}; my $_7⏏5 = 2; .&foo(2);
4
sena_kun sarna, ^
geekosaur bare . doesn't care about $_ so much as it wants to be method instead of sub 15:36
sarna m: sub foo($x) {say $x}; given "hello" { .&foo } 15:37
camelia hello
geekosaur althogh .&... overrides that
sarna oh my!
thanks sena_kun <3
geekosaur but it's using . / metod syntax with no invocantt hat defaults to $_, not someting abiout the sub/method
geekosaur hates this keyboard 15:38
(managed to break oldone, this is a crappy backup)
sena_kun that's a very useful info, thanks. 15:39
sena_kun takes notes
sarna is there a sub like is-upper? I'm browsing the docs but can't find anything like this, only .uc 15:47
sena_kun sarna, I can come up with just doing `$foo eq $foo.uc`. 15:47
sarna sena_kun: well yeah, that's what I'll end up doing I think 15:48
sena_kun you also can check with regex, but I believe that might be not so fast.
sarna and less readable, I think 15:49
jkramer m: sub is-upper(\c) { ?c.uniprops.first('Lu') }; say 'a'.&is-upper; say 'A'.&is-upper 15:54
camelia False
True
sena_kun jkramer, didn't they mean checking a string, not a letter? But yes, with minor patching if string checking is needed, this way seems very nice. 15:55
sarna, ^
jkramer m: sub is-upper(\c) { ?c.uniprops.first('Lu') }; 'FOO'.comb.all.&is-upper.say 15:56
camelia all(True, True, True)
jkramer m: sub is-upper(\c) { ?c.uniprops.first('Lu') }; say ?'FOO'.comb.all.&is-upper
camelia True
jkramer m: sub is-upper(\string) { string.uniprops.all eq 'Lu' }; say 'FOO'.&is-upper; say 'FOo'.&is-upper 15:58
camelia all(True, True, True)
all(True, True, False)
jkramer m: sub is-upper(\string --> Bool) { string.uniprops.all eq 'Lu' }; say 'FOO'.&is-upper; say 'FOo'.&is-upper
camelia Type check failed for return value; expected Bool but got Junction (all(Bool::True, Bool:...)
in sub is-upper at <tmp> line 1
in block <unit> at <tmp> line 1
jkramer m: sub is-upper(\string) { ?(string.uniprops.all eq 'Lu') }; say 'FOO'.&is-upper; say 'FOo'.&is-upper
camelia True
False
chenyf hi, #perl6 15:59
sena_kun chenyf, o/ 16:00
telma Hi there, any one knows why this rule would not work? 16:03
rule propuesta { 'SE PROPONE' .* <?before [:i 'Palabra's?' clave:'||'Responsable's?]>
this other rule _IS working_ fine -> rule asunto { 'ASUNTO' .* <?before ['RESULT'||'CONSIDER']'ANDO QUE' > } 16:04
chenyf if it is single character, try: ord('A') ~~ 65..90 16:05
geekosaur assumng you only care abut the US 16:07
sarna sena_kun, jkramer: I wanted it to work on a string, not only a character
jkramer m: say ord('Ö') ~~ 65..90
camelia False
jkramer sarna: This should do then: sub is-upper(\string) { ?(string.uniprops.all eq 'Lu') 16:08
jkramer But I guess $foo eq $foo.uc is just as good :) 16:08
sarna jkramer: thanks! your sub has many things I have yet to learn :^) 16:09
jkramer Oh actually mine wouldn't work if there are special characters in the string
sarna special, like Ö, or special like &?
jkramer m: say 'FOO.foo'.uniprops
camelia (Lu Lu Lu Po Ll Ll Ll)
jkramer Anything that's not a letter 16:10
sarna I see
and you did \string instead of $string because we don't want to mutate it, right?
jkramer Yup and I think it's a bit faster 16:11
cog hi, what is the equivalent of C typedef in Perl 6. Say I want to use C as a synonymous of: A does B 16:12
timotimo m: constant Frob = Int; my Frob $a = 9; say $a
camelia 9
timotimo m: constant Frob = Int; my Frob $a = "9000"; say $a
camelia Type check failed in assignment to $a; expected Int but got Str ("9000")
in block <unit> at <tmp> line 1
cog thx, timotimo 16:13
jkramer m: sub is-upper(\s) { ?(s.comb.none eq 'Ll') }; say 'FOO'.&is-upper; say 'Foo'.&is-upper 16:14
camelia True
True
jkramer m: sub is-upper(\s) { ?(s.uniprops.none eq 'Ll') }; say 'FOO'.&is-upper; say 'Foo'.&is-upper
camelia True
False
SmokeMachine m: role B {}; class A {}; subset C of A where *.^does: B; my C $a = A.new 16:15
camelia Type check failed in assignment to $a; expected C but got A (A.new)
in block <unit> at <tmp> line 1
jkramer sarna: ^^ This only returns true if the string contains lowercase letters
SmokeMachine m: role B {}; class A {}; subset C of A where *.^does: B; my C $a = A.new but B
camelia ( no output )
SmokeMachine cog: ^
sarna jkramer: thanks again :^) 16:16
cog thx, SmokeMachine 16:17
cog m: constant Frob = Int but 'foo'; my Frob $a; say ~$a 16:21
camelia foo
cog m: constant Frob = Int but 'foo'; my Frob $a; say $a
camelia (Int+{<anon|1>})
cog m: constant Frob = Int where * < 0 ; my Frob $a = -1; say $a 16:22
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3constant Frob = Int7⏏5 where * < 0 ; my Frob $a = -1; say $a
expecting any of:
infix
infix stopper
postfix
…
cog m: constant Frob = subset of Int where * < 0 ; my Frob $a = -1; say $a 16:24
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3constant Frob = subset of7⏏5 Int where * < 0 ; my Frob $a = -1; say
expecting any of:
infix
infix stopper
postfix
…
cog m: subset Frob of Int where * < 0 ; my Frob $a = -1; say $a 16:25
camelia -1
cog m: subset Frob of Int where * < 0 ; my Frob $a = 0; say $a
camelia Type check failed in assignment to $a; expected Frob but got Int (0)
in block <unit> at <tmp> line 1
jmerelo Hi. Back home. (Briefly) 16:25
yoleaux 31 Aug 2018 04:57Z <AlexDaniel`> jmerelo: it's SQUASHathon time!!! :)
2 Sep 2018 14:01Z <AlexDaniel> jmerelo: So I wonder, did we run out of squashable tickets in perl6/doc repo? It does feel like the fun stuff is no longer there after so many squashathons. Should we just go with random topic every month?
AlexDaniel jmerelo: o/ 16:49
AlexDaniel hmmm I have to rethink topics for upcoming squashathons 16:50
jmerelo AlexDaniel: I would be OK if we continue with docs every other month. 16:54
jmerelo AlexDaniel: there's a lot of stuff to be done still. Easy and not so easy. 16:56
AlexDaniel: on that note: github.com/perl6/doc/issues/2295 16:57
TimToady telma: .* in a rule is almost always wrong, because rules don't backtrack by default, so the .* will always eat up the rest of the string, and never give any back 17:01
you need the 'regex' keyword instead of 'rule' or 'token' if you want backtracking by default 17:02
you can also put a ! or a ? on the .* to force it to backtrack right-to-left or left-to-right, but it won't by itself in either a rule or token 17:05
masak there's also the whole "if you're using rules it's probably because you're parsing a language, and trying to avoid multiple passes" underlying reason 17:28
TimToady: I have lots of questions, but I've currently forgotten most of them :) 17:29
TimToady: here's one: do you think of concatenation in regexes as a "zero-width infix operator"? 17:30
TimToady masak: not really; regexes just naturally match a bunch of things in a row, so it's more like an iterable recursive function 18:07
so more like "foo bar baz 42" only written the other way around
I tend to think of the prototypical infix as a symmetrical operation, but the terms in / a b / are more hierarchical 18:10
TimToady of course, not all our infixes are entirely symmetrical... 18:12
but the relationship of a and b above is more like an "andthen" than it is like a concatenation 18:13
telma TimToady thanks! 19:10
robertle I am mucking around with another benchmark to compare perl 6 performance, the aim being that the tests would be closer to what I want to run and more full-sized rather than synthetic (github.com/robertlemmen/p6bench). But I am having weird issues with older versions of rakudo. for example 2017.05 (with matching nqp and moar) fails with 19:22
/usr/bin/moar --libpath="/usr/share/nqp/lib" --libpath="/usr/share/nqp/lib" perl6.moarvm --nqp-lib=blib --setting=NULL --ll-exception --optimize=3 --target=mbc --stagestats --output=CORE.setting.moarvm gen/moar/CORE.setting
Unhandled exception: Failed to write string to stream: broken pipe
at gen/moar/stage2/NQPHLL.nqp:1713 (/usr/share/nqp/lib/NQPHLL.moarvm:command_eval)
which is odd, I am sure that I did manage to build rakudo versions early last year. also when I did have build problems in the past, the error messages were far more helpful 19:23
has anyone seen stuff like this before? or any ideas? Is my asumption that matched versions of rakudo/nqp/moar should work together incorrect? is it possible they don't build anymore beacuse other stuff on my computer (e.g. libuv or so) is too recent? 19:24
clarification for the above, it's the rakudo build that fails of course 19:25
lizmat robertle: if you're not using the libuv that comes with that version of MoarVM, that feels like a recipe for disaster
as I think Moar has been adapted to match subtle changes in libuv more than once 19:26
afk for a bit&
robertle ok, trying that. I was running with --has-stuff for everything
robertle yeah, I think that does get me a step further into the past versions! 19:36
El_Che robertle: there have been some broken rakudo + moarvm combinations 19:37
where dot releases where needed
AlexDaniel robertle: there's also this ticket github.com/MoarVM/MoarVM/issues/951 19:39
robertle: which is maybe unrelated but you should know about it
robertle yeah, where we had dot releases I am picking the latest of them (for each component) 19:45
the rpi issue is not what I have here, this is a box-standard pc, but is highly interesting anyway, that will probably affect the debian builds of 2018.08
robertle rpi2 is armhf, right? armhf has extra float alignment requirements... 19:47
but that should lead to sigbus not segv
AlexDaniel robertle: also there will be 2018.09 in 10 days if we resolve that issue 19:55
releasable6: status
releasable6 AlexDaniel, Next release in ≈9 days and ≈23 hours. 0 blockers. 0 out of 36 commits logged
AlexDaniel, Details: gist.github.com/d7c1ecdcaf741ac49e...74af8499b0
robertle 0 blockers?
bazzaar \o Perl6
robertle I guess it looks at rakudo blockers only?
AlexDaniel releasable6: status 19:56
releasable6 AlexDaniel, Next release in ≈9 days and ≈23 hours. 1 blocker. 0 out of 36 commits logged
AlexDaniel, Details: gist.github.com/cc9b0538cf50853511...9aa7d4eb78
AlexDaniel robertle: yes… fixed :)
well, “fixed”
robertle where does releasable get it's data from?
AlexDaniel robertle: github.com/perl6/whateverable/blob...p6#L30-L31 19:57
bazzaar $contents is a slurped .IO file and a loop of construct 'for $contents.lines -> Str $line {' gives an error, whereas 'for $contents.lines() -> Str $line {' does not. Is it to do with lazy code? Would appreciate some guidance on where in docs to look for an explanation. 20:01
AlexDaniel bazzaar: how can I reproduce this? I'm seeing no error in both cases 20:03
bazzaar ok, I'll create a gist, the code is simple, it slurps in a csv file (I can't provide that though), it looks for lines with fewer than the required number of fields, and pads those lines out. 20:07
lizmat bazzaar: which version of Rakudo are you using ? 20:16
lizmat just worked on Str.lines 20:16
bazzaar lizmat: I'm behind the times, "This is Rakudo Star version 2018.04.1 built on MoarVM version 2018.04.1" 20:18
lizmat bazzaar: ok, so nothing I broke recently :-) 20:19
could you gist the error ?
bazzaar gist.github.com/bazzaar/914a5c3e6f...e9d16e360b 20:23
lizmat bazzaar: you're using Slang::Tuxic, which implies you may need parens in some places ? 20:27
specifically the "$contents.lines" may need "($contents.lines)" 20:28
bazzaar lizmat: aaah, sounds like you are on the trail there. :-) 20:29
lizmat: the code also worked with 'for $contents.lines[^5] -> Str $line {' ..... I remember 20:32
lizmat bazzaar: is there a reason you need Slang::Tuxic in *your* code ?? 20:34
if not, I suggest you remove it 20:35
it will at least make your script compile a lot faster
bazzaar lizmat: I think it was in the TEXT::CSV example / docs, ... I'll do as you say 20:36
lizmat aah... 20:37
ok, lemme see if I can fix that :-)
bazzaar lizmat: when I comment out the 'use Slang::Tuxic' line, I get the two terms in a row error on the 'my $csv_line = Text::CSV.new⏏ (eol => "\n");' line 20:40
lizmat indeed: you cannot have a space there in standard Perl 6
bazzaar lizmat: aaah, thank you for that. 20:41
AlexDaniel but
my $csv_line = Text::CSV.new: eol => "\n";
lizmat that would also work, but I was going for the least change :-) 20:43
lichtkind can i export a sub from a role into its consuming class? 20:44
bazzaar AlexDaniel: Thankyou also :-), both examples work, and Slang::Tuxic is gone 20:45
lizmat bazzaar: cool! 20:49
lichtkind: no, closest thing is a private method 20:50
m: role A { method !foo { say "foo" } }; class B does A { B!foo }
camelia foo
lichtkind lizmat, thanks but not that what i had before, hugs
bazzaar perl6 is totally awesome, .. world [of programming] domination will surely happen! 20:57
masak TimToady: I see what you're saying about the hierarchic/andthen thing. after all, parsers are monads, so it makes sense :) 21:03
TimToady: there's a years-old comment in src/QRegex/P6Regex/Grammar.nqp about turning the nibbler into an EXPR 21:05
masak that's what got me thinking about zero-width infixes 21:05
masak guess it comes down to semantics; one person's "zero-width infix" is another's TTIAR... 21:07
MasterDuke looks like colabti.org is down, anyone know why?
timotimo dunno :( 21:10
MasterDuke oh well. anything interesting in the channels today? 21:11
timotimo want my logs?
timotimo hack.p6c.org/~timo/moarvm_log.html hack.p6c.org/~timo/perl6_log.html hack.p6c.org/~timo/perl6-dev_log.html 21:12
MasterDuke cool, thanks
timotimo YW 21:13
MasterDuke: mastodon.social/@timotimo/100674655508783180 - have a look at my pretty videos if you wanna 21:14
MasterDuke i like the blue ones 21:17
Herby_ \o 22:31
timotimo o/ 22:34
Herby_ timotimo: how goes it? 22:35
timotimo all right. i finally got back to playing with cellular automata. but my wrists are getting worse since yesterday :| 22:37
how's you?
Herby_ cellular automata? Can't say I'm too familiar with that 22:38
and I'm doing well, just got off work, drinking coffee, and trying my hand at programming exercises 22:39
timotimo sounds like a chill time
Herby_ I can't complain :)
whats happening with your wrists? rsi? 22:40
timotimo yeah 22:42
Herby_ bummer :(
timotimo it really is :\
timotimo want a short primer on cellular automata? 22:42
Herby_ I'm always up for learning 22:43
timotimo OK, so the most general definition is this: you have some kind of grid with individual cells. every cell can have one of N different states - think colors or numbers or animals 22:44
(that's not all of it!)
(just as much as fits into one line) 22:45
timotimo so, the whole system makes "steps", where every cell is looked at, along with its "neighborhood" - for example "the cells north, east, south, and west of it on a chess board 22:45
and depending only on the states of the cell itself and the cells in its neighborhood, does the cell change its own state 22:46
and things like what states can exist, and what the neighborhood is, and what the rules for changing state are, that's the same for every cell
that's the basic case for cellular automata. Game of Life is probably the most famous one 22:47
in game of life, the neighborhood is north, northeast, east, southeast, south, southwest, west, northwest
and the rules are expressed as "if the cell is 'alive' and there are between X and Y cells nearby, it dies (starvation/suffocation), otherwise it stays alive" 22:48
"there are between X and Y 'alive' cells nearby"
timotimo "if the cell is 'dead' and there are between F and G alive cells nearby, the cell becomes alive (reproduction/splitting/growth), otherwise it stays dead" 22:49
i don't know the "default" numbers by heart, but there are variants with different values anyway
Herby_ ok. I've often heard of the Game of Life but I never really understood what it was about. this helps 22:51
timotimo cool :)
game of life is an example of CAs with "simple" rules that can execute arbitrary calculations 22:52
in fact, there is a method to build a game of life simulator out of configurations of cells inside a game of life world
Herby_ sounds like inception. 22:54
What are you currently exploring with CA right now?
timotimo the ones i have right now are on a regular 2d grid with NESW neighborhood and just two states, but every cell has a random chance in every step to just do nothing 22:55
that chance is at 5% right now, but sometimes i change it on a whim
mastodon.social/@timotimo/100674655508783180 - i've made a thread on mastodon with a few example videos 22:56
Herby_ those are awesome. is the code posted anywhere? 22:57
timotimo not yet, still working on it :)
i just made it 40% faster by removing an "int8" from one variable %) 22:58
Herby_ i love optimizing things, even with my very limited programming skillset 23:00
speaking of making things faster, could you eyeball my code for this puzzle to see if there are any easy wins to speed it up?
gist.github.com/sylints/df1c8e2619...09d7668c09
puzzle is here: www.reddit.com/r/dailyprogrammer/c..._funnel_1/
The goal for Optional Bonus 2 is to have it run within 1 to 2 seconds. I'm currently sitting at 10 seconds 23:01
but the result is correct
timotimo since $word doesn't change in that loop inside funnel-bonus, you could move $word.chars out of the "while" statement 23:02
that'd be a miniscule win at best
substr is a bit faster if you use the form without Ranges, even though it looks less beautiful
timotimo sometimes using .IO.slurp.lines is faster than .IO.lines 23:03
oh, and you're reading the file in twice, perhaps read the words list once, and create the Set from that afterwards 23:04
it's possible that making $index and $counter native ints, i.e. "my int $counter", though with a bit of bad luck it'll end up slower 23:05
timotimo you could use my %w := funnel-bonus, which is probably faster, since it doesn't copy the hash that was created inside funnel-bonus 23:06
i.e. binding := instead of assignment =
timotimo let me know how it goes! 23:07
Herby_ thanks for the feedback! i'll tweak the code and let you know how the timing changes
probably the root cause of the run time is my poor choice of algorithm 23:08
timotimo not enough focus to look much at the algorithm and such 23:13
Herby_ no worries at all, just thinking aloud :) 23:14
MasterDuke Herby_: you check if %words.elems > 4 in funnel-bonus and in the mainline 23:23
might be able to change to something like `if my %w = funnel-bonus($line, $word_set) { $counter++; say "bonus('$line') => " ~ %w.keys; }` 23:24
timotimo or just if funnel-bonus($line, $word_set) -> %w { ... } 23:25
Herby_ MasterDuke: doh! nice catch 23:26
timotimo Herby_: "return" at the end of a sub is a little slower than just putting the result value there 23:32
vrurg Can somebody have an idea on how to get around this bug: rt.perl.org/Public/Bug/Display.html?id=130794 ? 23:33
s/have/give/
timotimo yes, you can make an inner lexical scope to do the "use" in, and return what you wanted from it into a constant, for example with a do block 23:35
m: class is::all::right { }; constant OtherNamespace = do { use Test; Test }; OtherNamespace::is "me", "me", "yay" 23:36
camelia Could not find symbol '&is'
in block <unit> at <tmp> line 1
timotimo m: class is::all::right { }; constant &is = do { use Test; &is }; OtherNamespace::is "me", "me", "yay" 23:37
camelia ok 1 - yay
timotimo wait what
that's very wrong %)
m: class is::all::right { }; constant &namespaced-is = do { use Test; &is }; namespaced-is "me", "me", "yay"
camelia ok 1 - yay
timotimo m: class is::all::right { }; constant &namespaced-is = do { use Test; &is }; namespaced-is "me", "me", "yay"; say is.WHAT
camelia ok 1 - yay
(is)
timotimo m: class is::all::right { }; constant &namespaced-is = do { use Test; &is }; namespaced-is "me", "me", "yay"; say is.^name 23:38
camelia ok 1 - yay
is
timotimo m: class is::all::right { }; constant &namespaced-is = do { use Test; &is }; namespaced-is "me", "me", "yay"; say is.HOW.^name
camelia ok 1 - yay
Perl6::Metamodel::PackageHOW
timotimo is that helpful at all? the example isn't quite 1:1 because it's a sub that i'm importing, not apackage
oh ... also "is" as a sub wouldn't conflict with "&is" anyway
timotimo let's see 23:38
timotimo m: class CArray::Foo { }; constant MyNativeCall = do { require NativeCall; NativeCall }; say MyNativeCall::.keys 23:40
camelia 5===SORRY!5=== Error while compiling <tmp>
An exception occurred while evaluating a constant
at <tmp>:1
Exception details:
Type check failed in binding to parameter '$top-existing-pkg'; expected Any but got NativeCall (NativeCall)
…
timotimo um, huh?
i don't get that error locally
vrurg timotimo: Thanks! But it turns out that there is a better solution for my task. I didn't know that a method could be run on a role itself without applying it to a class. 23:46
timotimo ah, yes indeed
vrurg But your examples are very interesting from the syntax point of view. Another thing I've just learned. :)
timotimo it'll act as if you had an empty class to compose it into
that's called "punning"
vrurg I thought punning happens on instantiation. Is FooRole.method() punning too? 23:48
timotimo yes
vrurg Ok, another things to know. Thanks a lot!
timotimo you're quite welcome :) 23:49
vrurg Trully, RTFM: "... or use it as a type object ..." I've been through that paragraph for a couple of times now. But never noticed... 23:50
timotimo yeah, that happens all the time to me, too 23:52
Herby_ MasterDuke and timotimo: I made changes based on your feedback and have dropped the runtime from roughly 9.5 seconds to 4 seconds 23:58
timotimo nice!
Herby_ using substr() instead of $word.substr() had the biggest impact by far
Herby_ using that binding idea you mentioned gave me this error: Type check failed in binding; expected Associative but got Slip (Empty) 23:59
but > 50% reduction in runtime is awesome, thanks

Powered by Cro and the Raku® Programming Language.

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

Thank you!