»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.
Geth perl6-most-wanted: 81c4d9bf63 | cbk++ (committed using GitHub Web editor) | most-wanted/modules.md
Updated modules.md to list API::USNavalObservatory as a work in progress (WIP)
 00:06
Geth doc: c1175753eb | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
Fix typo
 04:49
fatguy i can't seem to use p6doc on any function or anything on help section. i'm using rakudo 2017.04.3 binary for mac os x. i already run index build, can someone tell me how to make it work ? 04:50
[Coke] can you post the command you're running and the error you're getting? 04:56
(gist.github.com works for multi-line snippets) 04:57
Geth ecosystem: 326210508b | loren++ (committed by Zoffix Znet) | META.list
Add Terminal::Table version 0.0.1 to the ecosystem (#340)
 04:58
doc/master: 4 commits pushed by (Will "Coke" Coleda)++
fatguy why i need to put ';' at then of the block with this one liner: -e 'my $count; for "banana".comb -> $letter { ++$count if $letter eq "a" }; say $count 05:18
missed ' at the end
geekosaur because a semicolon is only inferred after a close brace if it is at the end of a line 05:19
geekosaur keep in mind that a Block is an object; it can be passed as a parameter, and can have other parameters after it 05:19
but in this case for is a builtin and doesn't expect anything after the Block so the compiler grumps 05:20
fatguy geekosaur: thanks 05:22
geekosaur also perl 6 is trying to avoid the "mass of oddly interacting special cases everywhere" behavior of prl 5, so only simple rules like the line-final } rule instead of complex ones based on specific keywords 05:23
especially since you can add keywords by means of slangs 05:24
araraloren I'm more used to add smicolon after every `}` :) 05:29
Geth doc: 25bce38c1b | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.succ
 05:32
Geth doc: 82339608d2 | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Improve IO::Handle.t docs
 05:49
doc: 3274322f0f | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
Add :D to return type
 05:50
doc: b4006a2f5a | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Be explicit what IO::Handle.opened returns

for unopened handles.		 05:53
Geth doc: c4f27a7424 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.pred
 05:55
Geth doc: 860333f578 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Remove entirely-invented "File test operators"

Remind people of Pair.ACCEPTS from file test method docs, which is what these "operators" really are.		 06:02
Geth doc: ab0bd7abad | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.Numeric/.Int

IMO these really shouldn't exist, but they made it into 6.c language.		 06:30
araraloren Anybody know how to include multi version in **META6** file ? 06:36
Geth doc: 396d5032e1 | (Zoffix Znet)++ | doc/Type/Str.pod6
Document Str.Numeric/.Int
 06:41
Zoffix araraloren: multi version of what? Dist? Isn't the META6 file specifically for a single version? 06:42
What are you trying to do? 06:43
araraloren such like versions : [ "v0.0.1", "v0.0.2" ], Zoffix 06:54
Zoffix araraloren: I don't think something like that is currently supported 06:55
nine: lizmat: ^ if you're short of PTS topics to discuss 06:56
araraloren Zoffix, oh 06:57
Zoffix araraloren: though my question would be, why package two versions in the same dist? Release them separately. If someone wants earlier version, they'll just install that? 06:58
araraloren Zoffix, yeah
That's what i mean
Zoffix Reminds me of the mess with Perl 5's Date::Manip. Giant, dense docs describing all the different "interface versions". It's like... pick one 06:59
And I'm unsure how well releasing different versions of the same dists works with our ecosystem. I know some did it like this; adding more than one entry to meta file: github.com/perl6/ecosystem/blob/ma...#L726-L727 07:01
araraloren I remember Perl6 has a `class Fo:ver(0.0.1)`, but when I can declare one class with different version
m: class F:ver<1.10.1> { }; class F:ver<1.0.1> { };
camelia 5===SORRY!5=== Error while compiling <tmp>
Redeclaration of symbol 'F'
at <tmp>:1
------> 3ss F:ver<1.10.1> { }; class F:ver<1.0.1>7⏏5 { };
expecting any of:
generic role
Zoffix with our current ecosystem I mean; I know CPAN supports it just fine and IIRC you can tell zef to install from CPAN
araraloren class F:ver<1.10.1> { };
m: class F:ver<1.10.1> { };
camelia ( no output )
araraloren Zoffix, yeah, that a way to do multi version, but it's not convenient 07:02
When I release one version, I need to create a PR to **ecosystem** 07:03
Geth doc: 4f81f08210 | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Improve IO::Handle.get docs
Zoffix araraloren: but why are old version still relevant?
Feels it'd be a rareish case, for popular software, to keep old code working when you make non-back-compat change in a new version. 07:04
araraloren Em, If I changed API in new version, the module use old version would not work.. 07:05
Zoffix Sure, but ARE you changing the API in new version?
Even with our github ecosystem you can tag previous releases and let people install those via a link to a zip 07:06
araraloren Is zef or panda support this ? 07:07
Check tag, and install from that source 07:08
Zoffix araraloren: panda isn't relevant; it's dead. And yes, zef supports it. E.g.: zef install github.com/zoffixznet/perl6-IRC-Cl...001002.zip
And I got that URL from github.com/zoffixznet/perl6-IRC-Cl...r=3.003003
And it supports installing from CPAN too, as I recall
araraloren Zoffix, but when I use zef, how I know what version the module released ? 07:09
Zoffix araraloren: also, I sent you an invite to perl6 org. You should have access to most of the org's repos, so you don't need to PR to ecosystem, you can just commit directly by editing from github.com/perl6/ecosystem/edit/ma.../META.list
araraloren Oh, thanks 07:10
Zoffix araraloren: I don't understand the question. How you know what version you installed?
is that the question? 07:11
There's `zef info Some::Module` command
araraloren No..
It's zef list
It only list one version info
Zoffix .... which appears to be busted up. I installed two versions of IRC::Client, but it only reports 1
Zoffix And `list` don't tell me anything. Even zef list --installed IRC::Client shows nothing, despite my having 2 IRC::Client versions. 07:13
But anyway. My advice: don't over think this. It's a waste of time to fret over how to release multiple versions with our temporary ecosystem when you don't have any real need to release multiple versions.
\o
araraloren Oh, The problem turn back to ecosystem again, forget it .. 07:13
Ok 07:14
Geth ecosystem: bcabd9ab2e | loren++ (committed using GitHub Web editor) | META.list
Replace with **version** uri
 07:18
Geth doc: c45d389d98 | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Finish off IO::Handle.getc/&getc docs
 07:22
doc: ca7f7c921c | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
Document &getc defaults to $*ARGFILES

ATM it's NIY in Rakudo and will error out with poor error, but it'll be implemented in less than a week.		 07:25
Geth doc: e0645be5db | (Zoffix Znet)++ | doc/Type/Mu.pod6
Fix up &gist

  - Use correct signature
  - Mention say on Str types doesn't use gist		 07:39
Geth doc: a4012e0237 | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Document IO::Handle.gist

Rakudo impl: github.com/rakudo/rakudo/commit/276d4a7e7e Tests: github.com/perl6/roast/blob/a3410d....t#L29-L36		 08:10
Geth doc: d15b0c7003 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.gist

Tests:
  github.com/perl6/roast/commit/31e39932d8
  github.com/perl6/roast/commit/e2a3a25751		 08:35
modules.perl6.org/problems: eae812be19 | (Tadeusz “tadzik” Sośnierz)++ | site-tips.txt
Add todo routes to site tips
Geth modules.perl6.org/problems: 9d69c873e3 | (Tadeusz “tadzik” Sośnierz)++ | 2 files
Remove leftover debug info
 10:32
modules.perl6.org/problems: b4c484e0a7 | (Tadeusz “tadzik” Sośnierz)++ | 5 files
Have problems have varying severity. Indicate it in the UI
AlexDaniel does this trap docs.perl6.org/language/traps#Cons...mpile_Time also apply to sigilless variables? 10:58
jnthn No 11:00
Geth modules.perl6.org/problems: f1af625ae4 | (Tadeusz “tadzik” Sośnierz)++ | 2 files
Mark the lack of MANIFEST file as a problem

As per discussions on PTS, this'd be pretty much a requirement for CPAN usage. It's not a hard requirement now, so the severity is not cranked up to 11 just yet.		 11:02
modules.perl6.org/problems: 1e55111306 | (Tadeusz “tadzik” Sośnierz)++ | 2 files
Improve severity visuals
masak interesting: insearchofsecrets.com/2014/08/04/l...-symmetry/ -- captures something I've felt sometimes while implementing Perl 6 macros: that the MOP is drinking at least *some* of macros' milkshake 12:26
Geth modules.perl6.org/problems: 02ab412d2c | (Tadeusz “tadzik” Sośnierz)++ | 3 files
Make Todo controller lookup authors, not full URLs

Required guessing author when it's not specified, and gave us an excuse to store a Problem when an author's not set.
This also fixes a bug with _builder being cleared if something's stored there too early: we don't need to declare it anywhere in particular, autoviv does the work for us.		 13:23
modules.perl6.org/problems: 68df45f230 | (Tadeusz “tadzik” Sośnierz)++ | lib/ModulesPerl6/Controller/Todo.pm
Don't show dists that do everything right on the /todo

  (How do I ask DBIC to do that?)
tadzik I'd welcome feedback on github.com/perl6/modules.perl6.org...6:problems It's meant as a place to track sort of a CPAN-readiness of our modules, both as a statistic, as a tool for module authors (“let's see which of my modules do something completely horrible”) and potential contributors, too look for LHF in the ecosystem 13:28
it looks somewhat like this in practice: tadzik.net/pub/SZ25liL5bD.png
Geth specs: dfbaac8a53 | (Stefan Seifert)++ | S22-package-format.pod
Deconfuse auth in the META data.
 13:38
specs: b48be2c62d | (Stefan Seifert)++ | S22-package-format.pod
Unoutdate S22 a bit

CompUnitRepo has long been renamed to CompUnit::Repository.
  %?RESOURCE was implemented as %?RESOURCES and is just a plain array.
moritz if it's an array, why the % ? 13:43
ugexe because the items get referenced 13:49
$original => $some_sha1_string
e.g. %?RESOURCES<$original>.Str eq $some_sha1_string
ugexe but its declared as "resources" : [ "original" ] 13:52
Geth specs: 7000e04e42 | (Stefan Seifert)++ | S22-package-format.pod
Introduce versioning of the META6.json spec for real

Use a more descriptive name of the version field and mark it as required. Tools will provide some backwards compatibility, but this document describes the state as it _should_ be at some point in the future. And there we'll want to have a mandatory meta-version field.		 13:53
eater how does `try sink await $async-proc.start` work? 13:56
is sink a magic keyword?
AlexDaniel eater: you mean this? docs.perl6.org/language/traps#Sunk_Proc 13:57
eater ye 13:58
is it like run("blah").sink ? 13:59
AlexDaniel m: class Foo { method sink { say ‘woosh’ } }; my $x = Foo.new; say $x 14:00
camelia Foo.new
AlexDaniel m: class Foo { method sink { say ‘woosh’ } }; my $x = Foo.new; $x
camelia WARNINGS for <tmp>:
Useless use of $x in sink context (line 1)
AlexDaniel hm, interesting
eater m: say lc "HELP"
camelia help
AlexDaniel m: class Foo { method sink { say ‘woosh’ } }; Foo.new
camelia woosh
eater I guess it's doing that w/ sink 14:01
also your results are also intresting :')
AlexDaniel eater: well, yes 14:02
or not…
AlexDaniel is confused 14:03
eater :')
Geth whateverable: dce3462b15 | (Aleks-Daniel Jakimenko-Aleksejev)++ | build.p6
Make build.p6 more universal

We've been using an almost identical script for MoarVM, it's time to bring them back together.
In theory, now it should be easy to throw in some other backends.
Somehow I don't like the code here. Refactoring is welcome!		 14:14
AlexDaniel hm, it probably does not work 14:15
we will see :) 14:16
eater is there an apt repo w/ up to date perl6 packages? 14:35
Zoffix eater: yes, it's like run(...).sink; it sinks the proc inside the try, instead of having it get sunk outside of it. 14:40
Zoffix In `m: class Foo { method sink { say ‘woosh’ } }; my $x = Foo.new; $x` you're not sinking Foo; you're sinking the variable, so you get the warning and no `woosh` 14:40
In `class Foo { method sink { say ‘woosh’ } }; Foo.new` you're actually sinking Foo, so you get the `woosh` 14:41
eater: I know there are built packages; don't know if there's a repo
huggable: deb
huggable Zoffix, CentOS, Debian, Fedora and Ubuntu Rakudo packages: github.com/nxadm/rakudo-pkg/releases
eater thanks Zoffix! 14:42
Zoffix m: say "foo/\x[308]bar" ~~ /<[\/ \\]>/ 15:02
camelia Nil
Zoffix Any trick of making a *regex* that would match such a slash and just the slash, without the combiner?
m: dd split /<[\/ \\]>/, "foo/\x[308]bar"
camelia ("foo/̈bar",).Seq 15:03
Zoffix So that would be ("foo", "\x[308]bar").Seq
Zoffix m: use nqp; dd split /.<?{my str $slash = ~$/; (nqp::iseq_i(nqp::ord($slash), 92) || nqp::iseq_i(nqp::ord($slash), 47)) }>/, "foo/bar/\x[308]ber" 15:07
camelia ("foo", "bar", "ber").Seq
Zoffix I also need the \x[308] to not be removed
eater so the result must be foo, bar, \x[308]ber 15:13
?
Zoffix Yeah, like «("foo", "bar", "\x[308]ber").Seq␤»
m: ("foo", "bar", "\x[308]ber").Seq.perl.say 15:14
camelia ("foo", "bar", "\x[308]ber").Seq
eater but also on
Zoffix weird
eater foo\bar\\x[308]ber
Zoffix m: ("foo", "bar", "x\x[308]ber").Seq.perl.say 15:14
camelia ("foo", "bar", "ẍber").Seq
Zoffix Ah, Str.perl needs a thing to combine on...
eater: yeah, if that's split with the regex, it'd also be ("foo", "bar", "\x[308]ber").Seq 15:15
And it has to be a regex. I can easily do it without one
Zoffix tries messing with $/ 15:16
eater m: dd split /[ \/ || \\ <!before x>]/, 'foo/bar/\x[308]ber' 15:20
camelia ("foo", "bar", "\\x[308]ber").Seq
eater there is that
Zoffix lol
eater I think the problem is that we have a unicode regex parser 15:24
it's applying \x[308] to /
so it's not a / anymore
Zoffix Yup
eater and I don't see any flags that will tell it to not combine any characters 15:25
Zoffix There aren't any yet.
Zoffix m: use nqp; my $str = "foo/bar/\x[308]ber"; nqp::stmts((my $p := nqp::encode(nqp::unbox_s($str), 'utf8-c8', buf8.new)), (my int $els = nqp::elems($p)), (my int $i = -1), nqp::while( nqp::isne_i($els, $i = nqp::add_i($i, 1)), nqp::if(nqp::iseq_i(nqp::atpos_i($p, $i), 47) || nqp::iseq_i(nqp::atpos_i($p, $i), 92), nqp::atposref_i($p, $i) = 0)), $str := nqp::decode($p, 'utf8-c8')); dd split /\0/, $str 15:29
camelia ("foo", "bar", "\x[308]ber").Seq
Zoffix Well, that's one (slow) way of doing it, I guess, if being careful to change nuls back to slashes in all the relevant bits. 15:31
eater :')
Zoffix This is for IO::Win32::Spec; I guess I'll try to rewrite stuff without regex where possible
Thanks \o
tadzik m: "mojo".succ.say 15:37
camelia mojp
tadzik there 15:38
m: "Mojolicious".succ.say 15:40
camelia Mojoliciout
eater m: "one".succ.say 15:41
camelia onf
Geth doc: 04355a4f80 | (Zoffix Znet)++ | doc/Language/traps.pod6
Include example without `try sink`

Since normally you don't need any trys		 15:59
eater m: role A { method b { ... } }; class B is A {} 16:22
camelia 5===SORRY!5=== Error while compiling <tmp>
Class B can't pun role A because it has required methods: b. Did you mean to use "does" instead?
at <tmp>:1
eater well that's a better error then "Method 'b' must be implemented by A because it is required by roles: A." 16:23
lucs m: "exampleXX".say 16:31
camelia exampleXX
lucs m: "example".say
camelia example
lucs m: "example".print
camelia example
araraloren eater, no, **is** do inheritance, it need a class 16:41
m: role A { method b { ... } }; class B is A { method b() { } }
camelia 5===SORRY!5=== Error while compiling <tmp>
Class B can't pun role A because it has required methods: b. Did you mean to use "does" instead?
at <tmp>:1
timotimo araraloren: heyo 16:42
araraloren Em ?
timotimo araraloren: will you give your table module a section in the readme for "how is it different from the other table modules"?
araraloren timotimo, other module ? 16:43
timotimo yeah, there's like 6 different modules with "table" in their name
huggable: eco table
huggable timotimo, nothing found
timotimo buggable: eco table
buggable timotimo, Found 11 results: Text::Table::List, MIME::QuotedPrint, Text::Abbrev, File::Which, Image::PNG::Portable. See modules.perl6.org/#q=table
timotimo well, only one of these is interesting 16:44
Terminal::Table, Term::TablePrint, Text::Table::List, Text::Table::Simple
araraloren I see, but what you mean different ? 16:45
timotimo how does a user know what module to use 16:46
yours shows what the table looks like, that's good
araraloren OH
timotimo Term::TablePrint has no output example :( 16:46
i'll open an issue for that
araraloren I will update some example output
timotimo, sorry for that 16:47
timotimo no problem :) 16:48
you already have one example output, that's good
araraloren yeah, I have some example
eater how do I pipe to a proc with Proc::Async? 16:51
timotimo you need to open it with :in 16:52
then you'll have wrie, print, say, and close-stdin
alternatively, you can pass stuff to the :in(...) 16:53
wait, it looks like the arguments are called :r and :w?
eater yes
that's why I ask :p
I now just use a tap
eater but that means I need to build a buffer somewhere 16:54
timotimo i was confusing the async and non-async version of proc 16:55
you can call .Channel on a supply
if you want to use it like polling or something
gotta go, food calls! 16:56
araraloren timotimo, I add some output in readme, is that ok? github.com/araraloren/perl6-terminal-table 17:14
Geth doc: 1cf69327aa | (Zoffix Znet)++ | 3 files
[io grant] Document IO::Spec::*.is-absolute
 17:15
araraloren ~~ night 17:16
Geth doc: 4e88b84b32 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Finish up IO::Path.is-absolute

  - Include generic example
  - Point out "/foo" is absolute even on Windows
  - Include Windows-specific examples		 17:21
mcafee Hey, so, if I have an infinite list, is there any way to drop its first element and get an infinite list of the rest of the original list? 17:33
Zoffix mcafee: yeah, .skip 17:36
m: my @foo = (1...*).skip: 5; say @foo[^10 17:37
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in subscript; couldn't find final ']'
at <tmp>:1
------> 3my @foo = (1...*).skip: 5; say @foo[^107⏏5<EOL>
expecting any of:
statement end
st…
Zoffix m: my @foo = (1...*).skip: 5; say @foo[^10]
camelia (6 7 8 9 10 11 12 13 14 15)
Zoffix m: my @foo = (1...*)[4..*]; say @foo[^10] # also this
Zoffix hah 17:37
robocide? :)
camelia (timeout)
Zoffix really 17:38
I guess that doesn't propage .is-lazy
m: say (1...*)[4..*].is-lazy
camelia (timeout)
Zoffix why does it reify?
mcafee I get "No such method 'skip' for invocant of type 'Seq'". Is it pretty new? 17:39
Zoffix
.oO( ♪ oh tell me whyyy... do we build stuff that reify ♪ )
 17:40
Zoffix mcafee: newish, yeah 17:40
mcafee Okey dokey, time to upgrade then I suppose.
Zoffix m: my @a = 1..*; say @a.tail(*-5)[^10] 17:41
camelia (6 7 8 9 10 11 12 13 14 15)
Zoffix star: my @a = 1..*; say @a.tail(*-5)[^10]
camelia Cannot resolve caller tail(Array: WhateverCode); none of these signatures match:
(Any:D $: *%_)
(Any:D $: Cool $n, *%_)
in block <unit> at <tmp> line 1
Zoffix also works, but also new
AlexDaniel
.oO( did YOU upgrade your rakudo? Whateverables just did )
 17:42
Geth whateverable: edcb6687b1 | (Aleks-Daniel Jakimenko-Aleksejev)++ | Whateverable.pm6
Fix rakudo repo path after dce3462b153

We keep several rakudo repos (which is excessive but safe), so everything will need to be a bit more specific. For now, let's default to rakudo-moar directory.
Zoffix Ah, I see my mistake.
lazy index list is about stopping when the array ends, not about keeping it all lazy 17:43
m: my @a = 1…*; @a.shift xx 5; say @a[^10]
camelia (6 7 8 9 10 11 12 13 14 15)
Zoffix star: my @a = 1…*; @a.shift xx 5; say @a[^10]
camelia (6 7 8 9 10 11 12 13 14 15)
Zoffix :) cheatsy-doodle
m: my @a = 1…*; @a.splice: 0, 5; say @a[^10] 17:44
camelia (6 7 8 9 10 11 12 13 14 15)
Geth doc: 497e7f7027 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Finish off IO::Path.is-relative
 17:49
vetmaster m: say "test"; 18:32
camelia test
vetmaster p6: while True { say "AAA" } 18:34
camelia (timeout)AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA
AAA…		 18:34
vetmaster p6: foreach (^70) { say "vonni??" } 18:35
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of 'foreach'; in Perl 6 please use 'for'
at <tmp>:1
------> 3foreach7⏏5 (^70) { say "vonni??" }
vetmaster for ^70 { say "vonni??" } 18:36
AlexDaniel m: for ^70 { say "vonni??" } 18:36
camelia vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonni??
vonn…
vetmaster thnx bro
:)
AlexDaniel vetmaster: if you have any other questions, feel free to ask :) 18:37
vetmaster okay, I've just installed rakudo. and I'm impressed..
eater :D 18:38
AlexDaniel vetmaster: may I ask what exactly impressed you? :)
Geth doc: e05dc4665e | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
s/new line/value of .nl-in/
 18:39
vetmaster i.e. grammars
I often work with regex on Python 18:40
AlexDaniel oh yeah :) 18:41
vetmaster I also use Perl5, but the 6th version looks more modern
AlexDaniel well, in perl6 there are ways to write things cleaner, that's for sure 18:42
vetmaster hmm.. it's intuitive! 18:44
are there any libraries for natural language processing? (for Perl6) 18:46
thanks
vetmaster like ntlk in Python3, for example? 18:47
Geth doc: 3eaf7f2df8 | (Jan-Olof Hendig)++ | doc/Type/Str.pod6
Added doc for Str.indices. eater++
 18:49
AlexDaniel vetmaster: I don't know anything about this. But you can probably just use ntlk using Inline::Python github.com/niner/Inline-Python 18:52
vetmaster thank you, it would be a good solution 18:55
vim Perl 5 & 6 support github.com/vim-perl/vim-perl 18:56
vetmaster could you please suggest some good AI library? 19:08
timotimo .tell araraloren fantastic! i didn't realize you could have tables in triangle shape or similar! that's really cool and probably sets your module apart from all others 19:22
yoleaux timotimo: I'll pass your message to araraloren.
eater timotimo: w/o context that looks like such an sarcastic and denouncing statement :') 19:27
timotimo :( 19:38
ugexe you should be able to upload dists to cpan normally now and have them discovered with zef wi(using the --cpan option) 19:58
timotimo wow
we've come so far!
that ought to make ti easier to do multiple versions, too, right? 19:59
ugexe zef also has a branch with cpan testers matrix integration
its easier in that my cpan indexer will auto populate the source-url 20:00
but either way you are submitting a new dist to cpan or submitting a new dist to the ecosystem list
timotimo what is "a dist" in cpanverse? 20:01
ugexe i mean a tarball containing a unique name + version
timotimo OK
b2gills vetmaster: One of the things I really like about grammars are that they are just classes, and regexes/tokens/rules are just methods; so you can compose them using roles. 20:08
eater IO::Path.add is added recently? 20:31
timotimo bisectable6: IO::Path.new("/").add("foo") 20:34
bisectable6 timotimo, On both starting points (old=2015.12 new=95e49dc) the exit code is 1 and the output is identical as well
timotimo, Output on both points: «IO::Path is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting line 1␤ in method new at src/RESTRICTED.setting line 32␤ in block <unit> at /tmp/t6dPk9TO1F line 1␤»
timotimo m( of course
eater :") 20:36
timotimo: it was added a month ago 20:41
timotimo yeah, zoffix was able to bisect it
Geth whateverable: 39cde1e3e9 | (Aleks-Daniel Jakimenko-Aleksejev)++ | build.p6
Told you, I'll never remember to avoid this trap

Similarly to 96ccddccd804b33e0b9c5ed6580c1c23a91a5737		 20:57
eater is there a reason you can't retrieve the Proc object from the Proc::Async object later on? 21:01
Geth doc: f7e75c1bec | (Zoffix Znet)++ | doc/Type/IO/Handle.pod6
[io grant] Document IO::Handle.nl-in
 21:03
eater oh, it magically builds a fake Proc object 21:06
Geth whateverable: c63ed89d4e | (Daniel Green)++ | Benchable.p6
Close the pipe between pzstd and tar

Otherwise sockets get left open and it will die once it hits the 1024 limit.		 21:30
timotimo a-ha! 21:31
tyil Im working on a module, but its not yet completely done. the things I so far implemented work, and they dont depend on things not yet implemented, and they have tests available. Is it wise to keep it out of the ecosystem until its "complete", or should I add it in advance to let people know someone's working on it? 21:41
moritz_ tyil: add it to the ecosystem now 21:42
tyil right this moment? :p 21:42
Geth whateverable: 09c46d5915 | (Aleks-Daniel Jakimenko-Aleksejev)++ | build.p6
Build some of the dangling branches

The title of this commit does not express how cool this is.		 21:43
moritz_ tyil: sure!
timotimo faster!
the moment is almost over! 21:44
tyil D:
AlexDaniel timotimo: if your “ah-ha” was refering to $proc.out.close, then keep in mind that this is a third commit fixing this… 21:45
Geth ecosystem: Tyil++ created pull request #342:
Add MPD::Client		 21:46
AlexDaniel timotimo: and I don't know what amound of documentation is going to resolve this… github.com/perl6/doc/issues/1304
amount*
timotimo yeah, damn
tyil timotimo: was I still in time
timotimo tyil: really close! ;)
tyil I delayed rolling my joint for this!
but its worth it
timotimo AlexDaniel: was somethign keeping the objects around? or can we perhaps put a warning into DESTROY like we have for Failure? 21:47
AlexDaniel timotimo: I don't think so. $proc should be thrown away relatively quickly 21:47
tyil eater: wanna do a p6 coding weekend again sometime? 21:47
timotimo wanna try if we can check the pipes for being closed before the DESTROY gets called? 21:48
might be interesting if that triggers
eater tyil: when I have time and money again :p
Geth ecosystem: 0ea4a4ccbd | (Patrick Spek)++ (committed by Moritz Lenz) | META.list
Add MPD::Client
tyil thanks moritz_ :>
eater: cool
eater github.com/PostCocoon/P6-SystemPac...in.pm6#L69 21:49
tyil I have some plans for perl programs once this module is done
eater is there a better way to do this?
wrapping start and then sleeping till it's started feel really dirty
tyil also, I'd like to see how far I can get with an init system in perl 6
eater s/feel/feels
AlexDaniel timotimo: well, I don't think I have any time for this now 21:50
timotimo: but the code to reproduce this issue is just these two lines right there
timotimo eater: isn't started a promise?
eater timotimo: that resolves when it's exited 21:51
AlexDaniel timotimo: or even this example from our docs :P docs.perl6.org/type/Proc
eater not really the time to insert stuff into the STDIN :p
timotimo when it *exits*?
eater yeah that
timotimo ah, not a promise 21:52
eater "Returns a promise that will be kept with a Proc object once the external program exits"
docs.perl6.org/type/Proc::Async#method_start
timotimo ah, i see
why do you even have to wait? 21:53
eater because if I put data in before it's started
it will tell me that it's not started
AlexDaniel timotimo: to see how many sockets were left behind you can do something like ｢lsof | grep socket｣
timotimo huh. 21:54
eater and it's not immeditially started after it ran .start
timotimo of course not, it's async after all :)
AlexDaniel maybe narrow it down by pid, but this should be pretty much enough :)
timotimo eater: i think wrapping the "start" method of a single instance of Proc::Async isn't possible. you're probably poisoning every single Proc::Async in existence 21:55
eater timotimo: whooo 21:56
it's not needed if I do the until sleep loop anyway
timotimo it's strange that we don't have a method that'd let us wait for a program to start 21:58
eater Im currently looking up the MoarVM callback for it
timotimo like, the docs have an example where it just "await $proc.write(Buf.new(12, 42))" 21:59
without doing any waiting for the proc to start or anything
eater MoarVM has the "ready" callback
eater tbh I expected .started to be a Promise 22:00
timotimo me, too
that's an api change we can't easily make, though
how do i find a process that takes long enough to start so i can get the error you were talking about? 22:01
timotimo i can just start and immediately write to cat :P 22:01
but cat is small
eater bash -c 'sleep 5; cat;'
oh wait 22:02
I get what you mean
timotimo even with perl6 -e ... i can't get it to start up slow enough 22:03
are you sure you aren't somehow accidentally calling write before you called .start?
eater well... if you saw the code, I wrapped start, ran callsame and then did the thing 22:04
oH!
I see the problem :')
because it applies that wrap to all procs
it ran when I started the first process
timotimo ah, yeah 22:05
that would totally do it
i'd suggest you force the consumer of your piper class to have a method to call after you do .new, and one to call after you .start
or something
eater why no ready promise? :') 22:06
eater well time for sleep now, see ya and thanks for the help timotimo :) 22:07
timotimo i don't think there's any point in time where it wouldn't immediately be kept 22:07
so $proc.ready.then would be a no-op 22:08

Powered by Cro and the Raku® Programming Language.

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

Thank you!