»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
lollercopter m: use nqp; for ^10000 { my $v = "Hostname=This is a-test"; if $v.starts-with: 'Hostname=' { my $x = nqp::join('_', nqp::split('-', nqp::join('_', nqp::split(' ', nqp::substr($v, 9))))) } }; say now - INIT now 00:00
camelia 0.03849455
lollercopter m: use nqp; my $v = "Hostname=This is a-test"; for ^10000 { nqp::index($v, 'Hostname=') || nqp::join('_', nqp::split('-', nqp::join('_', nqp::split(' ', nqp::substr($v, 9))))) }; say now - INIT now 00:01
camelia 0.0294379
TEttinger faster as in speed or faster as in amount of code?
lollercopter m: use nqp; my $v = "Hostname=This is a-test"; nqp::index($v, 'Hostname=') || nqp::join('_', nqp::split('-', nqp::join('_', nqp::split(' ', nqp::substr($v, 9))))).say 00:02
camelia This_is_a_test
TEttinger I mean, those are definitely better on speed
lollercopter m: use nqp; my $v = "Hostname=This is a-test"; for ^10000 { }; say now - INIT now
camelia 0.0029200
00:02 Cabanossi left
lollercopter m: say 2.114/.159 00:02
camelia 13.295597
lollercopter m: say 2.114/.027 00:03
camelia 78.296296
lollercopter Voldenet: so there. 13x faster if you want it pure perl 6 and 78x faster if you don't mind using nqp
00:04 Cabanossi joined, zacts joined
lollercopter Voldenet: getting rid of blocks and regexes are generally a way to speed stuff up 00:04
Well, I mean a sure-fire way 00:05
00:06 mcmillhj joined
lollercopter crawls back into the hole 00:07
00:07 lollercopter left, araujo joined 00:10 cdg_ joined 00:11 mcmillhj left 00:14 cdg left, mcmillhj joined 00:15 cdg_ left 00:19 mcmillhj left
Voldenet Nice! :D 00:20
00:25 mcmillhj joined 00:31 mcmillhj left 00:40 cdg joined, mcmillhj joined 00:43 mr-foobar left 00:44 mr-foobar joined 00:45 cdg left, mcmillhj left 00:51 mcmillhj joined 00:56 zacts left 00:57 mcmillhj left 01:02 Cabanossi left, mcmillhj joined 01:04 Cabanossi joined 01:05 zacts joined 01:13 mr-foobar left 01:15 mr-foobar joined 01:19 LeCamarade left 01:29 mcmillhj left, troys is now known as troys_ 01:43 mcmillhj joined, mr-foobar left 01:45 mr-foobar joined 01:48 mcmillhj left 01:56 troys_ is now known as troys 01:58 mcmillhj joined 02:03 mcmillhj left 02:10 mcmillhj joined 02:11 AlexDaniel left 02:13 mr-foobar left 02:15 mcmillhj left, cyphase left 02:16 mr-foobar joined 02:20 ctilmes left, cyphase joined 02:22 noganex_ joined
awwaiid It was nice seeing some of you at TPC-2017-DC (aka YAPC::NA)! 02:23
02:24 noganex left 02:27 mcmillhj joined 02:31 troys is now known as troys_, mcmillhj left 02:43 mr-foobar left 02:47 mr-foobar joined 02:51 mcmillhj joined 02:56 mcmillhj left 03:00 troys_ is now known as troys 03:02 mcmillhj joined
BenGoldberg So I've got a hard problem, and it doesn't involve cache invalidation or an off by one error ;) 03:02
03:02 Cabanossi left 03:04 Cabanossi joined
BenGoldberg I'm working on some magic so that you can have class Foo with a repr('CStruct'), with attributes in it which can contain subroutines, which will be properly serialized. 03:05
I'm leaning towards a syntax of: has $foo is fptr :(int32 $arg --> int32); 03:06
But, I'm unhappy with the "fptr" name :P 03:07
03:07 mcmillhj left
BenGoldberg Could someone suggest a better name? 03:08
samcv can anybody see any problems with this implementation of recursive directory finder? sub get-dirs (IO::Path:D $path) { 03:12
my @paths = $path.dir.grep(*.d);
if @paths {
@paths.append: @paths».&get-dirs($_);
}
return @paths;
}
crap didn't mean to paste that here... ugh
meant to paste the link...
03:13 Cabanossi left, Cabanossi joined
samcv here it is: gist.github.com/samcv/7dade9fb0a93...1d864d21c7 03:13
i wanted to make it as simple as possible
03:14 mr-foobar left 03:16 mr-foobar joined 03:18 mcmillhj joined 03:23 mcmillhj left 03:29 araraloren joined 03:31 troys is now known as troys_ 03:34 mcmillhj joined 03:40 mcmillhj left
llfourn samcv: you passed $_ to it when you've already passed it implicitly with .&get-dirs 03:42
(I didn't run the code but I'd hope that's a compile time error) 03:43
samcv it's not
llfourn :(
samcv i mean you can pass variables into it. i'm actually going to add more arguments so i'm going to need to make it explicit yes?
though i guess this is different @dirs.map({&get-dirs($_, :$d)}) 03:44
llfourn but it only takes one arg and you've passed $_ twice to it
samcv but with brackets around it
ah
that passes it twice?
llfourn yeah
BenGoldberg This isn't perl5, you don't need the & there.
llfourn .&foo syntax passed the invocant in as the first argument to the sub
samcv BenGoldberg, even as a method?
BenGoldberg Well, maybe as a method, but not in {&get-dirs(...)} 03:45
samcv well yeah. i just altered it from a method
and didn't delete it.
nobody was harmed :)
llfourn @dirs.map(*.&get-dirs(:$d)) # should do it 03:47
03:51 mcmillhj joined
samcv ok turns out it wasn't passing in :$d 03:55
doing it like that...
or something @dirs.map({ get-dirs($_; :$d) } ) # Unexpected named argument 'd' passed wtf
llfourn ^ err was that smicolon meant to be there? 03:56
03:56 mcmillhj left
samcv yes 03:56
llfourn why?
samcv sorry uploading new version
ok check again
because i'm using named arguments 03:57
llfourn hmm
I seem to remember something vaguely about semicolons and named arguments
samcv oh crap i used a semicolon 03:58
sorry i misread semicolon as colon
llfourn lol
samcv haha
i'm going blind!
llfourn that is an LTA error though
but there is some kind of semicolon in pararmeter list syntax
samcv yeah
03:58 mr-foobar left 04:00 mcmillhj joined
samcv well it seems to work 04:00
04:01 mr-foobar joined
samcv i get 32326 directories with find and 32414 with perl 6 04:01
not sure why it's different 04:02
at least find returns one '.' but that would make find have one more directory
llfourn maybe sort and diff the output 04:03
samcv this liessss 04:05
docs.perl6.org/routine/flatmap#(Li...od_flatmap the docs page for flatmap
"invokes uc four times." untrue!
m: say ((1, 2), <a b>).flatmap({ $a++; uc($_)}).perl; say $a 04:06
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$a' is not declared
at <tmp>:1
------> 3say ((1, 2), <a b>).flatmap({ 7⏏5$a++; uc($_)}).perl; say $a
samcv m: my $a = 0; say ((1, 2), <a b>).flatmap({ $a++; uc($_)}).perl; say $a
camelia ("1 2", "A B").Seq
2
llfourn yep that's wrong
samcv map runs 4 times 04:07
flatmap doesn't
also both have the same output
llfourn flatmap just does .map then .flat
samcv so it's the worst example ever
llfourn yeah .flatmap is considered harmful
samcv oh wait nvm
i forgot to reset the value of $a. both of them still do it twice
llfourn samcv: lol, github.com/perl6/doc/issues/851
samcv and both have the same output. map or flatmap 04:08
for the love of god someone do something!
llfourn haha
samcv panics and runs around screaming
Voldenet > sub get-dirs (IO::Path:D $path) { flat gather for $path.dir.grep(*.d) { take $_, get-dirs($_) }}
it's ugly, but it works
samcv smls opened this issue on Aug 21, 2016
not that ugly 04:09
i wonder how it is speedwise
04:09 mcmillhj left
llfourn I've never actually gather take 04:09
always do map/grep
actually used* 04:10
samcv it is kinda nice if you want a sequence
and not all at once
Voldenet gather take is supposed to be lighter on memory if I reason correctly 04:11
04:11 Cabanossi left
llfourn doesn't map grep produce a lazy seq anyway? 04:11
or do I have to put lazy in front 04:12
04:12 kyan joined 04:13 Cabanossi joined 04:19 khw left
Voldenet Hm, I believe map grep do produce lazy seq when you give them lazy seq 04:23
04:27 wamba joined
BenGoldberg m: dd map *, grep {1}, ^10; 04:28
camelia Cannot resolve caller map(Whatever, Seq); none of these signatures match:
(&code, + is raw)
in block <unit> at <tmp> line 1
llfourn m: my $a = (^10).map: { .say }; # only runs if it's sunk
camelia ( no output )
BenGoldberg m: dd map {$_}, grep {1}, ^10;
camelia (0, 1, 2, 3, 4, 5, 6, 7, 8, 9).Seq
BenGoldberg m: dd map {say}, grep {1}, ^10; 04:29
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of bare "say"; in Perl 6 please use .say if you meant to call it as a method on $_, or use an explicit invocant or argument, or use &say to refer to the function as a noun
at <tmp>:1
--…
BenGoldberg m: dd map {.say}, grep {1}, ^10;
camelia 0
1
2
3
4
5
6
7
8
9
(Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True).Seq
llfourn m: my @a = (^10).map: { .say }; # if it's assigned it probably runs too
camelia 0
1
2
3
4
5
6
7
8
9
BenGoldberg m: my $nosink := map {.say}, grep {1}, ^10; say 'hmm';
camelia hmm
BenGoldberg m: my $nosink := map {.say}, grep {1}, ^10; say 'hmm'; say $nosink.perl
camelia hmm
0
1
2
3
4
5
6
7
8
9
(Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True, Bool::True).Seq
llfourn m: my @a = lazy (^10).map: { .say }; # but if you put lazy it shouldn't
camelia ( no output )
Voldenet my $nosink = map {.say}, grep {1}, ^10; say $nosink.take(5) 04:32
m: my $nosink = map {.say}, grep {1}, ^10; say $nosink.take(5)
camelia Too many positionals passed; expected 1 argument but got 2
in block <unit> at <tmp> line 1
Voldenet erm 04:33
m: my $nosink = map {.say}, grep {1}, ^10; say $nosink.head(5);
camelia 0
1
2
3
4
(True True True True True)
04:36 mcmillhj joined
Voldenet in fact 04:38
sub get-dirs (IO::Path:D $path) { $path.dir.grep(*.d).deepmap: { $_, get-dirs($_) } }
that would be even better way to do this :D
llfourn isn't deepmap pointless because dir is only going to give you a an array of paths? 04:44
04:44 mcmillhj left 04:47 mcmillhj joined 04:48 raiph joined
Voldenet hm, get-dirs gives you a lazy sequence of a dir, so deepmap would give you a sequence representing a directories tree 04:48
m: sub get-dirs (IO::Path:D $path) { $path.dir.grep(*.d).deepmap({$_, get-dirs($_)}).flat }; say get-dirs(".".IO).elems 04:50
camelia Type check failed in binding to parameter '$path'; expected IO::Path but got IO::Path.new(".", :SP...
in sub get-dirs at <tmp> line 1
in block <unit> at <tmp> line 1
04:51 mcmillhj left
Voldenet It works on my box, I guess "no fun allowed" anymore :P 04:53
04:53 troys_ is now known as troys 04:57 Cabanossi left 04:58 Cabanossi joined
samcv ok i'm fi xing flatmap docs right now 05:00
Geth doc: b90d2f82c6 | (Samantha McVey)++ | doc/Type/List.pod6
Provide correct documentation for .flatmap

The previous example's explanation was wrong, and the example had the same result on both .map and .flatmap, making it an even more confusing function.
The documentation now strongly recommends use of .map.flat and recommends to instead using the confusing .flatmap
Fixes issue #851.		 05:02
samcv that bug was so old :(
05:03 mcmillhj joined
samcv sad! 05:03
05:03 troys left
llfourn samcv++ 05:05
samcv i just solved world hunger
ok maybe nothing that important
but i feel happy about it
ugexe m: my $files = -> $p = $*CWD { $p.f ?? $p !! dir($p)>>.&?BLOCK.Slip }; say $files().map(*.parent).unique;
camelia Failed to get the directory contents of '/home/camelia/.local/share/systemd/user': Failed to open dir: 2
in block <unit> at <tmp> line 1
05:06 wamba left 05:07 mcmillhj left
Voldenet uh, getting a list of files and then doing .map(*parent).unique is a very unique way of doing this :) 05:13
05:17 pilne left 05:18 roguelazer joined, mcmillhj joined
Geth doc: 286e135bac | (Samantha McVey)++ | doc/Type/Hash.pod6
Be more clear about %() being preferred to { } for hash creation

Work on fixing #1380
In addition to more highly recommending %(), we also use less text and make the examples of the pitfalls more clear.		 05:19
samcv this also makes me happier too
BenGoldberg I would have, instead of 'To assign to a variable without the C<%> sigil', 'To assign a hash to a variable which does not have the C<%> sigil' 05:23
05:23 mcmillhj left
samcv isn't that the same thing? 05:24
BenGoldberg Also, s/Instead, you should use/This would have been avoided if you had used/.
samcv which does not vs without
does not have vs without
BenGoldberg Well, it's almost the same thing, except we're still using a %, just on the right side of the = instead of the left. 05:25
Voldenet ugexe: apparently calling unique on IO doesn't work all that well on my system, but calling it on *.parent.Str works
samcv Voldenet, did you sort them with eqv? 05:26
.sort(&[eqv]).unique 05:27
05:29 mr-foobar left 05:31 mr-foobar joined
Voldenet .unique(:with(&[eqv])) kind of works, but it's slow 05:31
samcv kind of works? or works but slowly
samcv wonders if we have an eqv candidate for IO::Path 05:32
Voldenet It works, but it's incredibly slow
this takes a blink of an eye > .map(*.parent.Str).unique; 05:33
this takes longer than 10s > .map(*.parent).unique(:with(&[eqv]))
samcv yeah i know it is slow :P
Voldenet Ah, /eqv/ might not be the operator I was wanting to use 05:34
samcv and does .sort not work?
Voldenet Nope, see for yourself: my $files = -> $p = $*CWD { $p.f ?? $p !! dir($p)>>.&?BLOCK.Slip }; .say for $files().map(*.parent).sort.unique;
05:34 mcmillhj joined
samcv oh you're running unique without sort? does th't work? 05:34
Voldenet The set is pretty much sorted anyway, so there's no need to sort it again :) 05:35
samcv that doesn't work for me
well i don't think unique even works if thigs are not sorted.. 05:36
ah maybe i'm wrong
that doesn't even run that is 05:37
Voldenet It seems that .unique works right on Str
samcv Failed to get the directory contents of '/home/samantha/git/rakudo/.nav-marker-rules': Failed to open dir: 2 05:38
timotimo i'd expect unique to work with .WHICH
samcv why?
well it does call WHICH 05:39
ah yeah cause it literally is checking if they're the same object period
but it doesn't work if you have two thinsg which are different objects but have the same contents 05:40
unless it's a Str
and prolly an Int
or Rat
other things no
05:40 mcmillhj left
samcv m: say dir[0].WHICH 05:40
camelia IO::Path|60080000
timotimo you could of course implement "unique" based on eqv 05:41
samcv m: say dir[0].WHICH; say dir[0].WHICH
camelia IO::Path|43442208
IO::Path|43442288
timotimo but that'll blow up run time quadratically
samcv er maybe you want unique with cmp? 05:42
or uh. would eqv work? since uh. it only gives True or False
usually you implement unique by sorting and then removing duplicates next to each other
timotimo you'd have to eqv the new element with every existing element
samcv yeah
Voldenet with IO::Path .unique(:as(*.Str)) would work reasonably well
timotimo that'd be faster, yeah 05:43
though not everything you can eqv you can cmp
samcv yep
Voldenet, that won't work if the CWD is different i would guess right?
Voldenet Yeah.
timotimo yeah, it'd have to be :as(*.absolute)
05:44 xtreak joined
Voldenet well, I'd still rather use: sub get-dirs (IO::Path:D $path) { $path.dir.grep(*.d).deepmap({$_, get-dirs($_)}).flat } 05:44
samcv well that won't find unique paths on/ly ones that refer to the same thing
since paths store more info than that
that cuold cause weird bugs? depending?
not sure
maybe if you use .Str ever 05:45
i've never used deepmap hm
how does it differ from map
Voldenet it's invoked on subsequences as well 05:46
and since get-dirs returns a sequence, the map would invoke it recursively 05:47
samcv map works for sub arrays but not subsequences or what? i've never tried to go very deep 05:48
Voldenet m: (1, (2, 3), 4).map({ .say; *+1 }) 05:50
camelia 1
(2 3)
4
Voldenet m: (1, (2, 3), 4).map({ .say; $_ + 1 }).say
camelia 1
(2 3)
4
(2 3 5)
Voldenet not what I'd expect
m: (1, (2, 3), 4).deepmap({ .say; $_ + 1 }).say
camelia 1
2
3
4
(2 (3 4) 5)
Voldenet better :)
m: (1, ^5, (2, 3), 4).deepmap({ $_ + 1 }).say 05:51
camelia (2 (1 2 3 4 5) (3 4) 5)
05:53 travis-ci joined
travis-ci Doc build errored. Samantha McVey 'Provide correct documentation for .flatmap 05:53
travis-ci.org/perl6/doc/builds/246458765 github.com/perl6/doc/compare/7d9e8...0d2f82c6ff
05:53 travis-ci left 05:56 mr-fooba_ joined, Cabanossi left 05:57 mr-foobar left 05:59 Cabanossi joined, xtreak_ joined 06:02 xtreak left
ugexe m: (1, (^5), ((4,5),), 6).tree(*.self, *.reverse, *.sum).say 06:04
camelia (1 (4 3 2 1 0) (9) 6)
Voldenet what does .tree do? 06:05
ugexe m: (1,(2,3),4).tree(*.self, *.reverse).say; (1,(2,3),4).tree(*.reverse, *.self).say 06:06
camelia (1 (3 2) 4)
(4 (2 3) 1)
ugexe m: (1,(2,3),(4,(5,6))).tree(*.self, *.self, *.reverse).say; 06:07
camelia (1 (2 3) (4 (6 5)))
06:09 travis-ci joined
travis-ci Doc build errored. Samantha McVey 'Be more clear about %() being preferred to { } for hash creation 06:09
travis-ci.org/perl6/doc/builds/246461571 github.com/perl6/doc/compare/b90d2...6e135bac9d
06:09 travis-ci left, nbrown joined
samcv ahh go away travis 06:09
travis is a big meanie. won't even finish building 06:11
though! the cache did work
we need one build that does highlighting and one that doesn't i think 06:13
but i don't have time for that today
06:15 domidumont joined 06:19 tony-o joined 06:21 domidumont left, domidumont joined 06:24 mr-fooba_ left 06:25 parv joined 06:28 MasterDuke left 06:32 domidumont left 06:33 domidumont joined 06:38 mr-foobar joined 06:42 jamesaxl joined
jamesaxl hi 06:42
mst: do you use perl 6?
TEttinger from what I understand, mst uses some of every language ever comprised. ever heard of Kernel? 06:43
jamesaxl TEttinger: I think that he does not use anything 06:45
TEttinger: how about you? 06:46
TEttinger I do not use perl 6. 06:47
jamesaxl TEttinger: only perl5?
TEttinger I can write a teensy bit of very simple perl6. my perl5 is limited to regexes. 06:48
I'm interested in how perl6's grammars work
I think they are an elegant way of constructing DSLs
jamesaxl TEttinger: do you think that perl6 can be good in the future? 06:50
TEttinger yes.
I think there are speed issues holding it back currently, but that's addressable by MOAR VM improvements 06:51
I think the language itself is phenomenally complex and this makes it hard to attract users 06:52
that said, C++ has tons of users and is a very complex language, probably just as much or more so than perl6
things like junctions are brilliant for describing what you mean rather than building complex if-else statements 06:53
jamesaxl TEttinger: and how about modules, I see that perl6 lacks to very important modules.
TEttinger yeah it lacks tons, but Inline::Perl5 makes it apparently very easy to call perl5 code 06:54
06:55 domidumont left
jamesaxl TEttinger: And how can I share modules of perl6 (I meant my modules)? 06:57
TEttinger I do not use perl 6. but I think CPAN is in the process of changing to accept perl6 as well as perl5 06:58
nine TEttinger: it already does
TEttinger I yield the floor to nine 06:59
jamesaxl nine: can I upload it use pause? 07:00
nine jamesaxl: indeed, you can :)
You can use mi6 (App::Mi6) for added convenience 07:01
jamesaxl that is nice 07:03
nine zef will find the distro about half an hour or an hour later. The only bit that's missing to my knowledge is listing the module on modules.perl6.org
BenGoldberg m: package Foo { our proto bar(|) {*}; our multi bar(Int) { 'int' } }; dd Foo::bar(42); 07:08
camelia "int"
BenGoldberg wonders if that's considered a bug.
nine Why would it? 07:09
BenGoldberg m: package Foo { our proto bar(|) {*}; my multi bar(Int) { 'int' } }; dd Foo::bar(42);
camelia "int"
BenGoldberg pasted the wrong version.
The 'my' is basically ignored. 07:10
nine BenGoldberg: it's not: 07:13
m: package Foo { our proto bar(|) {*}; { my multi bar(Int) { 'int' } } }; dd Foo::bar(42);
camelia Cannot resolve caller bar(Int); Routine does not have any candidates. Is only the proto defined?
in block <unit> at <tmp> line 1
nine Foo::bar(42) conceptually calls the proto bar(|) which then calls the appropriate candidate which it does find in it's lexical scope 07:14
07:16 mr-foobar left 07:23 bwisti left, mr-foobar joined 07:24 llfourn_ joined, llfourn left 07:26 BenGoldberg left, unclechu left, xui_nya[m] left 07:29 xui_nya[m] joined, unclechu joined 07:32 domidumont joined 07:34 zapwai left 07:39 skids left 07:44 domidumont left 07:45 setty1 joined 07:56 xtreak joined 07:57 xtreak__ joined 07:59 xtreak_ left 08:01 xtreak left 08:08 rindolf joined 08:17 parv left 08:26 darutoko joined 08:27 setty1 left
Geth ecosystem: d5e94b9211 | (Martin Barth)++ (committed using GitHub Web editor) | META.list
Added HTTP-Server-Orgre

Added HTTP-Server-Ogre - just another p6w http server		 08:30
08:30 ufobat joined 08:36 parv joined 08:42 LeCamarade joined 09:06 xtreak joined 09:10 xtreak__ left 09:11 xtreak_ joined 09:14 parv left, xtreak left 09:16 ufobat left 09:25 ufobat joined 09:30 lizmat left 09:33 seatek joined 09:37 xtreak_ left, lizmat joined 09:38 ufobat left 09:41 xtreak joined 09:49 parv joined 09:55 TEttinger left, parv left 09:56 xtreak left 10:12 MARTIMM joined 10:30 jamesaxl left 10:32 jamesaxl joined 10:36 MARTIMM left 10:45 Sense8 joined 10:47 Sense8 left 11:00 lizmat left 11:05 nhywyll joined, nhywyll left 11:06 lizmat joined 11:13 setty1 joined 11:17 nhywyll joined, nhywyll left 11:28 nhywyll joined, nhywyll left 11:41 nhywyll joined, nhywyll left 11:46 MasterDuke joined 11:48 AlexDaniel joined 11:49 grondilu left 12:24 robertle joined 12:28 domidumont joined 12:29 kaare_ left 12:40 cpage_ left 12:42 Cabanossi left 12:45 ufobat joined 12:46 Cabanossi joined 12:48 ufobat left 12:57 domidumont left, nbrown left 13:00 lucasb joined 13:12 domidumont joined, Cabanossi left 13:16 Cabanossi joined 13:24 user3 joined
user3 why do I have to put a 'my' here 'my Str sub f() {}' while it's not needed in this one 'sub f() of Str {}' 13:24
13:27 pilne joined, wamba joined
masak user3: that question does not seem to make sense out of the context you're in but have not stated 13:31
oh wait, maybe it does
user3: I think the simple answer is "because if you want to specify a type like 'Str', you have to front with a declarator keyword" 13:32
whether you make that keyword `my`, or `our`, or `has` or something else, is up to you
masak loses today's Oxford Comma nomination 13:33
user3 k
masak user3: the same answer, but phrased differently, would be "a type like `Str` cannot front a declaration statement like `sub`" 13:34
user3 yes, i understand the logic of it
thx
masak pzh
moritz or from a different point of view: If it's not part of a declaration, a type name like Str is a term, so the parser expects an operator after it
masak oh, right 13:38
which makes for a nice contrast with e.g. Java, where type names and ordinary terms live in different namespaces 13:39
so in Perl 6, the "cost" of the tradeoff is that you need to do `my Str` or `our Str` etc. at the start of declarations
user3 its a good bargain
masak whereas in Java, it's that you need to do `MyType.class` (to bring a type term from type land to term land) instead of just `MyType` 13:40
and it also makes `instanceof` a slightly magical operator :) 13:41
(because it has to tell the parser, "hey, the rhs is a type, not a term")
13:44 domidumont left 13:53 cpage_ joined 13:55 user3 left 14:02 BenGoldberg joined 14:04 bwisti joined 14:10 itaylor57 left 14:19 mscha joined
mscha m: my $p = start { for 1..5 { .say; sleep 1; } }; sleep 2; $p.break; 14:19
camelia 1
2
3
Access denied to keep/break this Promise; already vowed
in block <unit> at <tmp> line 1
mscha How do I break a promise (from the “outside”)?
mst POSIX::_exit(255); # all promises are now broken 14:26
TimToady mst: you can't break a promise unless the inside grants you access to the "vow"
mscha But what if I don't want to exit, just kill that promise?
TimToady that's by design 14:27
mscha So in my case, once that promise is running, there's no way to stop it (except by exiting)?
TimToady I suppose it depends on where the inside put its vow?
if it's in an object attribute you can get at it sneakily... 14:28
mscha I tried $p.vow.break, but it wouldn't let me either.
m: my $p = start { for 1..5 { .say; sleep 1; } }; sleep 2; $p.vow.break;
camelia 1
2
3
Access denied to keep/break this Promise; already vowed
in block <unit> at <tmp> line 1
 14:29
14:29 khw joined
mst TimToady: how does the inside of the start block gets its vow? 14:30
TimToady doesn't remember
TimToady doesn't break many vows... 14:31
the tests use Promise.new rather than start, so maybe start is doing something weird 14:35
MasterDuke docs.perl6.org/type/Promise#method_break mentions that using e.g., start, means you can't break it, so use Promise.new if you want to 14:37
mscha But if you don't use start, how do you (asynchronously) run code? 14:39
TimToady apparently start keeps its vow private, but if you can get the code to throw an exception, it'll break 14:40
you can always Promise.new yourself
look at method start in src/core/Promise.pm 14:41
TimToady heads for DCA, LAX, SJC... 14:43
mscha Hmm, okay, so looks like I can't do that ­- breaking a promise from the “outside” when the code in the promise is already running. 14:46
14:46 Ven joined 14:47 Ven is now known as Guest80266 14:48 Guest80266 left 14:51 nhywyll joined 14:54 skids joined 14:56 lucasb left 15:04 wamba left 15:10 davido_ joined 15:14 reino4477 joined
reino4477 hey, has anybody created a web app in Perl 6 recently? 15:15
15:15 kyan left
reino4477 i'm trying to understand if its eco system is ready more or less so I can use it for building web apps for myself 15:16
15:16 davido_ left
AlexDaniel reino4477: I think it depends 15:17
15:18 davido_ joined
reino4477 www.reddit.com/r/perl6/comments/6j...ment_what/ 15:19
MasterDuke reino4477: Gabor Szabo is working on a book about it www.indiegogo.com/projects/book-we...bsite--2#/ 15:22
jnthn fwiw, start and Promise are only related in that start blocks evaluates to a Promise. The start implementation looks something like sub start(&foo) { my $p = Promise.new; my $v = $p.vow; $*SCHEDULER.cue: { $v.keep(foo()); }, :catch({ $v.break($_) }); return $p }
15:23 reino4477 left
jnthn Thus the vow is taken by start and kept with the result or broken if it throws an exception 15:24
15:24 davido_ left, domidumont joined
jnthn I don't really get why you'd want to break the Promise from the outside, though. It's not like doing that would cancel the code runing in the Promise, or even prevent it from being schedulered if it wasn't already running. 15:25
It'd just result in it exploding with a "already kept/broke this Promise" at the end
So I guess "what are you trying to achieve" is the righter question :)
15:27 davido_ joined 15:28 kyan joined 15:44 skids left 15:48 Ven joined 15:49 k-man left, Ven is now known as Guest44526 15:51 kaare_ joined 15:52 nhywyll left, abraxxa joined
abraxxa I just built 2017.06 and now use v6c; isn't working any more 15:53
when I change it to use v6.c; it works again, did I miss a change there?
15:55 lollercopter joined
lollercopter abraxxa: it worked only because of a bug. The right thing is v6.c 15:56
15:56 k-man joined
abraxxa lollercopter: I guessed so, thanks! 15:56
lollercopter c: 2017.05 my str $x = '6c'; my num $y = $x; say $y
committable6 lollercopter, ¦2017.05: «6»
lollercopter c: 2017.06 my str $x = '6c'; my num $y = $x; say $y
committable6 lollercopter, ¦2017.06: «Can't convert '6c' to num: trailing characters␤ in block <unit> at /tmp/inMsBbqJbY line 1␤ «exit code = 1»»
lollercopter ^ that's the bug that got fixed 15:57
abraxxa do the release notes mention that somewhere? I couldn't find it
15:58 davido_ left
lollercopter hm 15:58
m: dd v6c
camelia v6.c
lollercopter abraxxa: OK, nevermind, it's a new bug :D since v6c works a version literal for v6.c, it should work in `use` too, I think 15:59
abraxxa: no, release notes don't mention it
16:00 kaare_ left
abraxxa ok 16:00
lollercopter I can fix it... 16:01
16:02 kaare_ joined
abraxxa lollercopter++ 16:04
ugexe i couldn't think `use v6c;` should work... 16:22
wouldnt^
lollercopter ugexe: why not? 16:23
16:24 dotness joined
ugexe m: say v6c cmp v6.c; # i guess it depends on how you fe