»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.
lookatme :) 01:42
shinobi-cl Hi all 04:13
I read that Junctions can not be instrospected, i think they can not even be typified. Makes sense. However, i want to do something like this: 04:15
m: my Regex @rxs = (rx/3/, rx/2/); say @rxs.perl; my @L = (1, 2, 3, 10, 11, 12, 13, 21, 22, 23); my @F = grep { @rxs }, :v, @L; say @F.perl; 04:16
evalable6 Array[Regex].new(rx/3/, rx/2/)
[1, 2, 3, 10, 11, 12, 13, 21, 22, 23]
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); say @rxs.perl; my @L = (1, 2, 3, 10, 11, 12, 13, 21, 22, 23); my @F = grep { rx/2/ }, :v, @L; say @F.perl;
evalable6 Array[Regex].new(rx/3/, rx/2/)
[2, 12, 21, 22, 23]
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); say @rxs.perl; my @L = (1, 2, 3, 10, 11, 12, 13, 21, 22, 23); my @F = grep { all(@rxs) }, :v, @L; say @F.perl;
evalable6 Array[Regex].new(rx/3/, rx/2/)
[23]
shinobi-cl cool, eh? But if i wanted to have this grep inside a sub, how should i put the parameter that is going to be inside the grep later? 04:17
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); say @rxs.perl; my @L = (1, 2, 3, 10, 11, 12, 13, 21, 22, 23); my @F = grep { any(@rxs) }, :v, @L; say @F.perl; 04:21
evalable6 Array[Regex].new(rx/3/, rx/2/)
[2, 3, 12, 13, 21, 22, 23]
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); say @rxs.perl; my @L = (1 .. 23); sub search(Regex $r) { return grep { $r }, :v, @L}; say search( any(@rxs) ); 04:27
evalable6 Array[Regex].new(rx/3/, rx/2/)
(3 13 23)
(2 12 20 21 22 23)
shinobi-cl the idea,, be able to pass either a single regex to the method, or a Junction, but this junction has to be made of Regexes 04:32
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( any(@rxs) ); 04:46
evalable6 (2 3 12 13 20 21 22 23)
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( /3/ );
evalable6 (3 13 23)
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( any 42 );
evalable6 (exit code 1) Constraint type check failed in binding to parameter '$r'; expected anonymou…
Zoffix, Full output: gist.github.com/7cabce0fa5d69f1f84...c25fc59056
Zoffix Mu stops Junction autothreading and smartmatching with `Regex` typechecks both Junction and non-Junction args to Regex types 04:47
shinobi-cl my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( any(@rxs) ); 05:02
m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( any(@rxs) );
evalable6 (2 3 12 13 20 21 22 23)
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( all(@rxs) );
evalable6 (23)
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( none(@rxs) );
evalable6 (exit code 1) Died with X::TypeCheck::Binding::Parameter
in sub search at /tmp/642Q3CxtaJ line 1
in block <unit> at /tmp/642Q3CxtaJ line 1
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 23); sub search(Mu $r where Regex) { grep $r, :v, @L }; say search( one(@rxs) ); 05:07
evalable6 (exit code 1) Constraint type check failed in binding to parameter '$r'; expected anonymou…
shinobi-cl, Full output: gist.github.com/290ed9fd294916e8ab...98a01ba81f
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.all; 05:23
evalable6 (23 32)
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.one;
evalable6 (2 3 12 13 20 21 22 24 25 26 27 28 29 30 31 33)
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.any;
evalable6 (2 3 12 13 20 21 22 23 24 25 26 27 28 29 30 31 32 33)
Zoffix m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search any 42; 05:24
evalable6 (exit code 1) Type check failed in binding to parameter '<anon>'; expected Regex but got Int (42)
in block at /tmp/jGch8FWsy4 line 1
in sub search at /tmp/jGch8FWsy4 line 1
in block <unit> at /tmp/jGch8FWsy4 line 1
Zoffix Basically threads it through an empty block constrained to Regex, which will explode if anything ain't a regex (giving failed typecheck) or pass and give True from the `where` 05:25
Xliff m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.none; 05:31
m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.none;
evalable6 (1 4 5 6 7 8 9 10 11 14 15 16 17 18 19)
Xliff m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search one(@rxs.all, @rxs.none); 05:32
evalable6 (1 4 5 6 7 8 9 10 11 14 15 16 17 18 19 23 32)
shinobi-cl m: my Regex @rxs = (rx/3/, rx/2/); my @L = (1 .. 33); sub search(Mu $r where { -> Regex {}($_); True }) { grep $r, :v, @L }; say search @rxs.one;
evalable6 (2 3 12 13 20 21 22 24 25 26 27 28 29 30 31 33)
Xliff Zoffix++ # Working with combined conjunctions. 05:33
shinobi-cl hmm, i see... i created a ticket fo this... Do you think is worth of a ticket? or it was me, that was just using the wrong signature? 05:34
Xliff shinobi-cl: Which ticket??
Zoffix shinobi-cl: I don't see any tickets. What was it about? 05:35
shinobi-cl it was about this specific issue, one and none junctions throwing an error 05:41
it was sent via email, [email@hidden.address]
Zoffix shinobi-cl: ah, ok, those take some time to show up on the site. FWIW we also have this alternative Issue tracker: github.com/rakudo/rakudo/issues
shinobi-cl: and those worked as expected, I just didn't think of different types of junctions when I suggested it. `Mu $x where Regex` with a `one` Junction is basically equivalent to `$x ~~ Regex` or "where EXACTLY ONE thing is of type Regex" and you had two regex objects inside, so it was failing the typecheck 05:43
m: say <a b c>.all ~~ Str 05:44
evalable6 True
Zoffix m: say <a b c>.one ~~ Str
evalable6 False
Zoffix m: say <a b c 42>.one ~~ Int
evalable6 True
Zoffix m: say <a b c 42 42>.one ~~ Int
evalable6 False
tyil can I send a raw IRC message using IRC::Client? I want to send an NS command (instead of using PRIVMSG NickServ) 05:51
Zoffix I think so... 05:58
tyil in the source I see a send-msg that gets a raw command as arg, but it also gets a $server, so I'm not sure if I can use it as `method irc-connected($e) { $e.send-msg: 'NS identify ...' } 06:00
Zoffix tyil: it's not documented, but try to use `send-cmd` method send-cmd ($cmd, *@args is copy, :$prefix = '', :$server)
Ah
You'd call it on IRC::Client object, not on $e (which is IRC::Client::Message) 06:01
$e.irc.send-msg: 'NS blah blah'
and don't provide the server
tyil ah
that makes sense
lemme try that :>
Zoffix FWIW on Freenode you can identify with NS by using server password (:password arg on instantiation of IRC::Client object) 06:02
tyil heh
but now I already took the effort to start a module for it, so I want to see if I can make it work :p 06:03
in the least it could be good for the perl 6 module count I guess
Zoffix And I see send-cmd was meant to be documented, but was just missed, so it's safe to use it. github.com/zoffixznet/perl6-IRC-Cl.../issues/30 06:09
tyil sweet
psch o/ 06:15
there was a different VM than hack for irc clients on p6c.org, wasn't there..? 06:16
cause my screens get nuked every two days, which i can deal with wrt coding, but with irssi it's a bit inconvenient 06:17
Zoffix irc.p6c.org IIRC
shinobi-cl m: my @a = (1,2,Nil,4); say @a.grep(rx/2/, :v);
evalable6 Use of uninitialized value of type Any in string context.
Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.
(2)
in block <unit> at /tmp/Bkkc3xiDLb line 1
psch Zoffix: i actually tried that, my connection's getting refused
Zoffix m: my @a = (1,2,Nil,4); say @a.grep({defined and /2/}, :v); 06:18
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/7SD_RyDJN5
Undeclared routine:
and used at line 1. Did you mean 'rand', 'any', 'end'?
Zoffix m: my @a = (1,2,Nil,4); say @a.grep({.defined and /2/}, :v);
evalable6 (2)
psch m: my @a = (1,2,Nil,4); say quietly @a.grep(rx/2/, :v);
evalable6 Use of uninitialized value of type Any in string context.
Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.
(2)
in block <unit> at /tmp/ABR8BkAH_o line 1
shinobi-cl m: my @a = (1,2,Nil,4); say @a.grep( {.defined and rx/2/}, :p);
evalable6 (1 => 2)
shinobi-cl m: my @a = (1,2,Nil,4); say @a.grep( {.defined and rx/4/}, :p); 06:19
evalable6 (3 => 4)
psch m: my @a = (1,2,Nil,4); quietly say @a.grep(rx/2/, :v);
evalable6 (2)
psch there :s
shinobi-cl thanks :)
Zoffix m: my @a = (1,2,Nil,4); say @a.grep: * andthen /2/; 06:20
evalable6 (exit code 1) (1 2 (Any) 4)
No such method '!cursor_start' for invocant of type 'Bool'
in block <unit> at /tmp/yZppJE0uVA line 1
Zoffix no curry for andthen :(
psch m: my @a = (1,2,Nil,4); say @a.grep: { $_ andthen /2/ } 06:21
evalable6 (exit code 1) No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at /tmp/C4YTH6huhY line 1
psch m: my @a = (1,2,Nil,4); say @a.grep: { $^a andthen /2/ }
evalable6 (exit code 1) No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at /tmp/Nw8BcqCZoX line 1
psch uhm
Zoffix Regex is Callable
psch but grep does take a Block..?
oh the RHS isn't curried then, i see
Zoffix Yeah 06:22
m: /2/(4)
evalable6 (exit code 1) No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at /tmp/sUPpJus1Sf line 1
lookatme m: rx/2/(4) 06:24
evalable6 (exit code 1) No such method '!cursor_start' for invocant of type 'Int'
in block <unit> at /tmp/CxmRSt11_A line 1
lookatme m: rx/2/ .(4)
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/QtqpfOfXbB
Malformed postfix call (only alphabetic methods may be detached)
at /tmp/QtqpfOfXbB:1
------> 03rx/2/ .08⏏04(4)
psch m: /2/(Match.new("2")) # uh huh 06:25
evalable6 (exit code 1) Default constructor for 'Match' only takes named arguments
in block <unit> at /tmp/rPqRxJm3KU line 1
psch m: /2/(Match.new(orig => "2")) # uh huh
evalable6
psch yeah probably don't go there :)
Zoffix psch: dunno, talk to [Coke] about irc.p6c.org. I can see the box is running in the hypervisor, but I don't have any logins for it and too get the "connection refused" thing. 06:30
Looks like there might be a problem actually. Based on logs, masak uses it. [Coke] probably too, but neither are connected right now 06:32
psch Zoffix: oh, i see, thanks for checking 06:38
tyil hmm 06:50
No such method 'send-msg' for invocant of type 'IRC::Client' in method irc-n376 06:51
psch tyil: Zoffix mentioned send-cmd? 06:52
tyil yes, Im trying to figure out how to use it properly in the current context 06:53
psch oh, i was just confused 'cause your error mentions send-msg instead
tyil oh
OH
thanks psch ;~; 06:54
psch :)
tyil it all works as intended now
guess I should get some breakfast before going on to the next project
buggable New CPAN upload: IRC-Client-Plugin-NickServ-0.1.0.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.0.tar.gz 07:05
buggable New CPAN upload: IRC-Client-Plugin-NickServ-0.1.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.1.tar.gz 07:35
tyil hype
psch \o/ 07:47
psch hrm, why does bluetoothctl tell me it's waiting for bluetoothd if the service is just called bluetooth.service :/ 07:52
psch & 08:24
scimon One thing the recent contretemps highlighted where a few beliefs about Perl6 that seem persistent but maybe be untrue. 10:52
For instance "Perl6 can't do https requests" which AFAIK should be "libssl isn't threadsafe"
But another I saw is that Perl doesn't do DB stuff well. Now I know we don't have a go to ORM (but as I'm not a huge ORM fan this doesn't bother me hugely) but what is DBIish missing that people think it's not usable? 10:54
Anyone?
lizmat it's not DBI
sorry, am a bit cranky 10:55
afaik, Cro does HTTPS 10:56
scimon Oh I know.
lizmat *and* HTTP/2.0 for that matter
scimon And there are other modules that do HTTPS as long as you don't thread them. 10:56
(Because they are using libssl with NativeCall under the hood)
lizmat but Cro does them threaded, no ?
ah, ok 10:57
scimon I'm mostly trying to put together a list of myths that need examined.
lizmat ++scimon
scimon Frankly DBIish is pretty close to DBI
lizmat yeah, considering taking it as the base of a CPAN Butterfly Plan DBI 10:58
but not today
scimon :)
scimon I'll say it again you're doing amazing work with that. 10:59
El_Che sends some rainbows, unicorns and puppies-and-turtles-can-be-friends-video to lizmat for het crankiness 11:02
lizmat: to be honest, the name implies subpar
maybe time to change it or alias it?
lizmat well, that would be my point :-) 11:03
anyways, I'm preparing my last Perl 6 Weekly
before I have some R&R :-)
El_Che last Perl 6? nowadays ships in the carabbean are safe ;) 11:04
lizmat I'm more worried about things before getting onto the ship and after getting off the ship again :-) 11:05
El_Che hehe
DrForr Unless you take Destination:Truth for...well, truth - I watched an episode over Xmas where spooky things. 11:06
* happened;
jnthn scimon: Cro has no problem with concurrent HTTPS requests 11:07
It uses IO::Socket::Async::SSL, which does the required locking 11:08
El_Che jnthn: maybe the cro sugar could be separated to a generic module?
jnthn IO::Socket::Async::SSL already is a separate module 11:09
And the Cro::HTTP module doesn't contain the tools, UI, etc. It's just the HTTP stuff.
El_Che jnthn: I haven't used it, but the socket part in the name make it sounds low level
jnthn I'm not really sure what you're asking for. 11:10
El_Che I am not really sure what I am asking :)
jnthn :P
I don't think installing Cro::HTTP in order to have a concurrent HTTP client is really a problem, though. 11:11
Sure, there's server stuff in there too, but the server and client use the very same request/response objects and body parsing stuff.
scimon My point was that there's a myth that Perl6 can't handle SSL. Wheras the truth is rather more complex.
El_Che jnthn: I had this in mind as a slightly less higher abstraction than socket: golang.org/pkg/net/http/ 11:12
scimon I was thinking of doing a bit of mythbusting in some form (blog or some such).
tadzik I guess what matters to some people is that the truth is complex rather than being a straight "yes"
see also: "is perl6 fast?"
El_Che jnthn: if Cro::HTTP fits the bill, we need to advertise it
jnthn El_Che: Looks like it to me, glancing through that 11:13
DrForr There are all sorts of Perl 6 myths that need busting. Where are Adam and Jamie when you need them? 11:14
El_Che jnthn++
jnthn The only "sugar" atop of that which is found in Cro::HTTP is Cro::HTTP::Router.
Which accounts for probably 5% of the size of the distribution.# 11:15
El_Che my point is that many people will stay away from frameworks or thinking that they don't need microservices and don't look at Cro, while it solves the http(s) problem
jnthn: kind of what Mojo does 11:16
jnthn I've never called it a framework :-)
El_Che at the moment, it's THE framework
:)
jnthn But yeah, point taken, we can try to advertise "we have a nice HTTP client" better 11:17
scimon So. Any thoughts on what problems people might have with DBIish?
jnthn Perhaps on the homepage could make HTTP Server and HTTP Client separate entries so it's obvious there are both 11:18
El_Che yeah, having a good enough http client built-in on Go (and used a a based for more sugary implementations) helped that lang's popularity
jnthn That'd let us show off the HTTP client with some examles too.
El_Che a very well know module as the module-to-use is a good alternative
jnthn: exactly
jnthn If I do it after this week's release I can even put in an example of working with HTTP/2 push promises on the client side :) 11:19
jkramer Good news everyone! I fixed the ö/Ö talking script so it doesn't crash anymore and also made the code a bit nicer to look at :) gist.github.com/jkramer/2fa0e1d3fc...773c692a46
El_Che jnthn++
jnthn El_Che++ for feedback
El_Che jnthn: played just a little with it. Looking forward to write my first real project with it. I reserved some time for it 11:21
DrForr I should have something interesting Cro-related to show at FOSDEM :) 11:22
El_Che DrForr: nice! Looking forward to it 11:26
DrForr I'd love to show it here but unfortunately the modules are too large to fit in the margin of my server :) 11:26
El_Che margin of your server? 11:27
you rented a server in 1994?
DrForr 4Gb and I'm at 89Mb remaining, most of it's OS stuff. 11:28
And the problem children are node.js
El_Che wow
lizmat wow indeed
I mean, my notebook has more RAM :-)
El_Che if you're paying around 5€ for that, I'll give you a friendly buy firm tick on the head 11:29
:)
DrForr It's a Linode server - I think I actually can get the drive upgraded if I can get the rest of the cr*p off it. 11:30
119M node_modules/
El_Che DrForr: get a free (1 year) aws vm to show your perl stuff? 11:31
DrForr Yeah, probably will today after I get one or two problems worked out. 11:32
Not trying to be a tease, trying to get stuff straightened out. 11:33
El_Che no worries
just for the record, digital ocean smallest machines has a 25gb ssd disk. I probably have a $10 referal code in case you want to try it (but I need to check) 11:36
jnthn bbl
DrForr Sounds like linode is soaking me for a bit re: prices but I'll check later. I do want to keep this on a machine that doesn't have personal stuff on it though, so I can nuke it without any personal felings. 11:38
*feelings
El_Che DrForr: actually, the only cloud machine I have at the moment is like that. Just an OpenVPN machine 11:39
moved it around a few times: aws, digital ocean, joyent. Whereever I had credit 11:40
DrForr Aaand I just panicked MoarVM again - "use of invalid eventloop work item index -1" 11:48
Which can't be repeated... 11:49
ZzZombo m: my @a;for 1,2,3 { @a.push: class { has $.i is rw='a';submethod BUILD(:$!i='asd') {.say} } };@a[0].new.say 11:50
evalable6 (Any)
<anon|93934972633264>.new(i => "asd")
ZzZombo star: my @a;for 1,2,3 { @a.push: class { has $.i is rw='a';submethod BUILD(:$!i='asd') {.say} } };@a[0].new.say 11:52
moritz huh, camelia is missing again? 11:54
ZzZombo star: my @a;for 1,2,3 { @a.push: class { has $.i is rw='a' } };@a[0].new.say
m: my @a;for 1,2,3 { @a.push: class { has $.i is rw='a' } };@a[0].new.say
evalable6 <anon|93855688941552>.new(i => "a")
ZzZombo m: my @a;for 1,2,3 { @a.push: class { has $.i is rw=$_ } };@a[0].new.say 11:55
evalable6 <anon|94034808729584>.new(i => Any)
psch moritz: kinda "still," she fluttered off some 6+ hours ago i think
ZzZombo m: my @a;for 1,2,3 -> $v { @a.push: class { has $.i is rw=$v } };@a[0].new.say
evalable6 <anon|94070905553584>.new(i => 3)
moritz nine: the VM on which camelia resides seem to be having network trouble again :( 11:56
moritz camelia:~ # ping -c 5 perl6.org 11:57
PING perl6.org (213.95.82.53) 56(84) bytes of data.
--- perl6.org ping statistics ---
5 packets transmitted, 0 received, 100% packet loss, time 4032ms
ZzZombo m: my @a;for 1,2,3 { @a.push: class { has $.i is rw=$_ } };for @a {.new.say} 11:59
evalable6 <anon|94682981474320>.new(i => Any)
<anon|94682981474320>.new(i => Any)
<anon|94682981474320>.new(i => Any)
ZzZombo m: my @a;for 1,2,3 -> $v { @a.push: class { has $.i is rw=$v } };for @a {.new.say} 12:00
evalable6 <anon|94252978598592>.new(i => 3)
<anon|94252978598592>.new(i => 3)
<anon|94252978598592>.new(i => 3)
ZzZombo why all the classes are the same?
moritz ZzZombo: because there is only one variable $v 12:03
and an "is rw" attribute binds to the container, not the value
psch m: for ^3 { class { }.WHICH.say } 12:04
evalable6 <anon|94187547973360>|U94187527440800
<anon|94187547973360>|U94187527440800
<anon|94187547973360>|U94187527440800
psch moritz: isn't it that the class decl is compile time and thus there's only one? 12:05
moritz: $.i being the same value is the container binding though
moritz psch: erm, you're right
tobs m: ("a".."g").rotor((1, 2, Inf), :partial)».join 12:08
evalable6
tobs huh, it produces (a bc) here
psch tobs: ENOSAY
tobs m: say ("a".."g").rotor((1, 2, Inf), :partial)».join
evalable6 (a bc)
tobs my question was supposed to be if I can stop the rotor from rotoring, but Inf doesn't work 12:09
as in, (a bc defg) is what I'm after 12:10
psch m: say ("a".."g").rotor((1, 2, *), :partial)».join
evalable6 (a bc)
tyil jnthn: your fix works for me 12:11
psch m: say ("a".."g").rotor((1, 2, Int), :partial)».join 12:12
evalable6 (exit code 1) Invocant of method 'Bridge' must be an object instance of type 'Int', not a type object of type 'Int'. Did you forget a '.new'?
in block <unit> at /tmp/h77LNqYdpd line 1
psch pff
tobs: seems to me like either Inf or Whatever should do what you want, but i'm not that deep into roto 12:13
+r
tobs m: say ("a".."g").rotor((1, 2, .1), :partial)».join 12:14
evalable6 (exit code 1) Rotorizing sublist length is out of range. Is: 0.1, should be in 1..^Inf;
Did you mean to specify a Pair with => 0.1?
in block <unit> at /tmp/xg8GMwnNU8 line 1
psch tobs: you could supply the Int-coerced invocant as last argument i suppose...
tobs psch: I think so too, but ^ part of it knew that ^Inf is the limit
psch tobs: true, and that's a range excludiing Inf which means Inf itself could mean "all of the rest" 12:15
tobs welp, lunch break is over. I have to go. 12:19
lizmat has started on the P6W and it's going to be a monster one 12:41
stmuk lizmat: I'm slacking a bit but hope to get R* out in the next few hours 12:41
El_Che repeat with me: p6w is not an open letter
El_Che :P 12:41
lizmat El_Che: no, it is not 12:42
the "monster" was referring to the size
timotimo no scary posts on the p6w ;)
PrincipiaMa How do I run many shell commands right after one another? 12:43
timotimo PrincipiaMa: do you just need to start one when a previous one is finished, or do you want to shovel data in between the commands?
tadzik what about multiple shell() calls, one after another? :)
PrincipiaMa timotimo: I want it to start right after the last one is finished, so like go into a directory, then do something there 12:44
timotimo using a shell call to change directories is not going to have the effect you want
you're starting a shell to call cd, which changes the shell's directory, then the shell is finished and exits and no effect will be felt
lizmat stmuk: looking forward to announce Rakudo *
PrincipiaMa timotimo: Oh, okay. Would there be a better way to do what I'm trying to do? 12:45
timotimo and since shell calls the shell, you can use the regular multiple-commands feature that a shell offers, for example by putting ; between the commands
timotimo also, shell accepts a "cwd" named argument that you can use instead of calling cd in the shell itself 12:46
PrincipiaMa timotimo: Using ";" between commands made it work now :D Thanks.
El_Che "&&" is the better choice most of the tim 12:52
e
timotimo oh, dang, that's true
i'm a fish user, so i always do that with "; and"
El_Che I thought that shell users were lazy and wanted to minimize the typed chars :) 12:53
jnthn tyil: Thanks for testing 13:08
tyil np 13:13
timotimo you know that feeling when you should be packing boxes for the move but instead you're disassembling an lcd monitor? 13:18
jnthn I...I can't say I do! :P 13:20
I'm sure I've procrastinated packing boxing in other ways, though :)
timotimo anyway, yeah, here's a few capacitors that really wanted more freedom for their innards 13:26
El_Che when you'll move you'll have a working lCD 13:30
(and probably no clothing)
timotimo the move starts in earnest tomorrow and i haven't really started packing boxes yet. which isn't a problem because i didn't unpack much since the last move (which was only 9 months ago; 6 months ago the new landlord told us we had to move out now so their kid could get the apartment) 13:31
lizmat wishes timotimo strength the coming weeks 13:32
El_Che timotimo: fuck that
timotimo ᕦ(° Д°)ᕤ
fuck that indeed. unfortunately, i engaged 100% push-over unnecessarily-polite mode the day they came to tell us 13:33
El_Che replace the landlord with a small perl6 script 13:34
timotimo he's an urologist; i'm not sure i'm sufficiently educated about the subject to implement his work in any programming language
probably a bit like react { whenever $practice.accept -> $client { $client.say("here's some antibiotics"); $client.kick-out }; whenever Supply.interval(180) { $snacks.pop.eat } } 13:36
El_Che heh
psch timotimo: in my experience urologists always want you to pee in a cup first
timotimo: also re super-polite mode, i don't think being angry changes anything about Eigenbedarf, which is a valid reason to terminate a lease 13:37
timotimo i should have consulted an attorney, because i'm pretty sure terminating a lease that early is at least not guaranteed to work 13:38
Xliff timotimo: react { whenever $practice.accept -> $client { $client.pee; $client.say("here's some antibiotics"); $client.kick-out }; whenever Supply.interval(180) { $snacks.pop.eat } } # Fix from psch 13:50
timotimo more like my $cup = await $client.pee :)
Xliff LOL
timotimo: my $quack = Landlord.new; react { whenever $practice.accept -> $client { my $cup = await $client.pee; $quack.say("here's some antibiotics"); $client.kick-out }; whenever Supply.interval(180) { $snacks.pop.eat } } # More fixes. ;) 13:51
colomon do we have a good idiom for returning (say) two hashes from a sub? 13:54
timotimo i'd literally just return %foo, %bar 13:55
colomon timotimo: …. how do you recieve those values on the other end?
the point of the call, I mean
colomon is apparently doing it wrong 13:56
psch m: sub f { { a => 1 }, { b => 2 } }; my (%a, %b) = f; say %a<a>, %b<b>
evalable6 1(Any)
psch something something GLR was just a few months ago >_>
colomon yes, I believe that’s what I’m seeing (though not quite how I did it) 13:56
psch m: sub f { { a => 1 }, { b => 2 } }; my ($a, $b) = f; say %a<a>, %b<b>
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/npw_Ng65kT
Variabl…
psch, Full output: gist.github.com/e9b30d093468695629...6b46dfde22
psch m: sub f { { a => 1 }, { b => 2 } }; my ($a, $b) = f; say $a<a>, $b<b>
evalable6 12
colomon psch++ timotimo++ 13:57
psch m: sub f { $({ a => 1 }), $({ b => 2 }) }; my (%a, %b) = f; say %a<a>, %b<b> # probably not...
evalable6 (Any)(Any)
psch colomon: yeah, gotta itemize the assignment LHS it looks like
moritz m: sub f { { a => 1 }, { b => 2 } }; my (%a, %b) := f; say %b.perl
evalable6 {:b(2)}
psch ooh or bind, yes
moritz++
moritz
.oO( in the list to bind them )
 13:58
colomon moritz++ 13:59
colomon moritz: though I wouldn’t have to ask here if I’d found it in Perl 6 Fundamentals. ;) 13:59
colomon hasn’t actually read Perl 6 Fundamentals, which may have meant he didn’t know where to look in it 14:00
moritz :-)
colomon I have read about half of Parsing with Perl 6 Regexes and Grammars, and learned a bunch of useful stuff 14:02
moritz there's not yet a comprehensive/complete book out^H^H^Hon Perl 6 out there
i'm hoping for Andrew to write one :-)
moritz colomon: if you started from the beginning, the good parts are still ahead :-) 14:03
colomon moritz: I did start from the beginning, so yay! 14:04
moritz also, if you've liked it, a rating on amazon would be very helpful
colomon of course, that implies I do need to remember to read it!
will keep amazon in mind
moritz thanks
Xliff moritz: One day I will get your book. Yeeees. 14:15
</gollum>
www.youtube.com/watch?v=DOqKnInZov...pedjaffas1 14:16
stmuk lizmat: at least 2 p6 videos from LPW2017 available, www.youtube.com/watch?v=os75uYTSnXU www.youtube.com/watch?v=oUGUgt-_0DA
lizmat stmuk++
stmuk I thought there was supposed to be another but can't find it at a quick look ... maybe didn't make it up 14:17
act.yapc.eu/lpw2017/talks/tag/perl6 <=- don't see either the Proctor or Wicks ones 14:18
maybe they will get uploaded soon
lizmat hops
*hopes rather :-)
stmuk not like a frog then :) 14:19
lizmat although I must admit I have been hopping a few times the past weeks 14:25
stmuk not much hops but I'm looking forward to Rochefort 8 and 10 at FOSDEM 14:27
lizmat as long as you keep the consumption limited to Delirium :-) 14:28
stmuk :) 14:30
ZzZombo how do I use a parametrized role in my class? The help only shows examples of autopunned roles. 14:37
timotimo you would have it like class Bloop does Blah[1, "hi"] { } 14:38
buggable New CPAN upload: IO-Socket-Async-SSL-0.6.1.tar.gz by JNTHN cpan.metacpan.org/authors/id/J/JN/...6.1.tar.gz 14:45
lizmat notices that tbrowder is not in CREDITS ?? 15:05
lizmat do we actually have a "most recent uploads" section on modules.perl6.org that I've missed somehow ? 15:20
stmuk rakudo.org/2018/01/29/announce-raku...e-2018-01/ 15:39
Geth perl6.org: 2581447ccf | (Steve Mynott)++ | 2 files
star 2018.01
 15:42
pmurias rakudo.js update - blogs.perl.org/users/pawel_murias/2...fixes.html 15:52
moritz stmuk++ # star release!
stmuk brought to by the power of Club Mate! 15:53
moritz pmurias++
MasterDuke pmurias: very cool. btw, "I have finally figured out a long standing bug with roles that has for long which now work." is missing something 15:55
stmuk 94% of roast subset passing is pretty impressive
moritz PSA: irc.p6c.org seems to have a corrupt root file system :( 15:56
stmuk not btrfs then? ;) 15:58
moritz ext4
El_Che irc is a shell server for irssi? 15:59
moritz yes
pmurias stmuk: that's like ~73% of all roast tests 16:00
El_Che moritz: so besides $HOME, nothing special? 16:01
pmurias stmuk: I'll have to rerun everything instead of just running the subset ones to measure it accuratly
moritz El_Che: currently, there *is* something special about that system: Can't load '/usr/lib/x86_64-linux-gnu/perl/5.24/auto/re/re.so' for module re: /usr/lib/x86_64-linux-gnu/perl/5.24/auto/re/re.so: invalid ELF header at /usr/share/perl/5.24/XSLoader.pm line 96.
pmurias MasterDuke: thanks, fixed 16:06
El_Che moritz: looks broken all right 16:07
jkramer I'm trying to implement a module from the most-wanted list but apparently I'm not very good at it. :( My module has a class with a method, a submethod and (outside of that class) a sub that can optionally be exported as shortcut. They all have the same name, but apparently that's not ok? 16:20
moritz jkramer: submethod and method of the same name makes no sense 16:21
methods and submethods live in the same namespace
AlexDani` stmuk++
timotimo moritz: it kind of sort of does, though? you'd provide the method to be used by any derived classes and the submethod by that class itself? 16:21
jkramer moritz: I want to provide Foo.do-it as shortcut for Foo.new.do-it
AlexDaniel m: say 42 16:22
evalable6 42
jkramer But maybe my understanding of what a submethod actually is is incorrect
timotimo what you want is not a submethod vs a method, but two multis with one candidate having a signature like (Foo:U: $args)
moritz jkramer: sounds like you want multi methods instead
timotimo and the other would have (Foo:D: $args)
moritz m: class A { multi method f(A:D:) { say "instance" }; multi method f(A:U:) { say "static" }}; A.f; A.new.f 16:23
evalable6 static
instance
moritz jkramer: ^^
jkramer Oh ok, I'll try that :) Thanks!
timotimo u: ⋄ 16:36
unicodable6 timotimo, U+22C4 DIAMOND OPERATOR [Sm] (⋄)
jkramer m: class Foo { has Str @.optional; multi method bar(Foo:D: Str $text?) { say $text }; multi method bar(Foo:U: Str $text?, Str :@optional) { Foo.new(:@optional).bar("lol") } } ; Foo.bar("asd") 16:42
evalable6 (exit code 1) Cannot resolve caller bar(Foo: Str); none of these signatures match:
(Foo:D $: Str $text?, *%_)
(Foo:U $: Str $text?, Str :@optional, *%_)
in block <unit> at /tmp/QT04bEAEbs line 1
jkramer Why does this not work? It works when I remove the Str constraint on @.optional 16:43
And in the method argument 16:44
timotimo jkramer: don't forget that unless you return the Foo.new you build from the one bar method, or else it'll disappear into the void 16:45
jkramer timotimo: Yeah this is just for testing the problem with the mismatching signatures 16:49
Also is `run` asynchronous by default? 16:52
moritz it's not 16:53
jkramer Hmm, weird stuff happens sometimes when I do run 'vi', 'some-file'; 16:54
It seems like the script is dying/exiting while vi is still open
timotimo vi probably does fun things with tty and such
jkramer Yes that too, but before that happens the script that runs vi dies in the background 16:57
Hmmmmm does keyboard into still go to the p6 process which passes it through to the Proc? 16:58
*keyboard input 16:59
It seems like C-c in vi kills the script, which then makes vi go crazy and complain about "At EOF" forever 17:00
geekosaur that's not likely to be keyboard input 17:00
the terminal driver converts control-C to a signal sent to the entire process group. process group members do not have to be reading the keyboard input to receive it 17:01
typically, when running a program "in the foreground" you ignore SIGINT and SIGQUIT in order to prevent this problem 17:02
jkramer I just tried system("vi", "foo") in P5 at it works just find, C-c doesn't do anything there 17:02
Try run('vi', 'foo'), but use a new terminal because it's gonna break it :) 17:03
geekosaur yes. problem is, in perl 6 Proc is actually Proc::Async with some automatic wait points added. those perhaps should ignore the signals I mentioned 17:04
jkramer vim on the other hand seems to work just fine O?
stmuk depending on what you are doing you may find the ex(1) version of vim easier to script 17:05
jkramer stmuk: I'm trying to implement the editor invoke module from the most-wanted list :) 17:06
stmuk ah
geekosaur if vim disables "isig" then the signal thing doesn't come up. I don;t know what "vi" you are getting, or if it behaves differently
jkramer Also Proc.exitcode: Returns the exit code of the external process, or -1 if it has not exited yet. 17:08
What if a process exited with -1?
How would I know the difference? 17:09
moritz exit codes are uints
between 0 and 255, typically
geekosaur you can "do" exit(-1) but it will be treated as unsigned, yeh
moritz $ perl -e 'exit -1'; echo $?
255
jkramer Ah ok.
geekosaur also older commercial unixes may not always treat it as unsigned... and yes, this ends up being a portability pain point 17:10
jkramer Ok so now I added signal(SIGINT).tap({}) to work around the C-c problem, but is there a way to reverse that afterwards? In P5 I'd do `local $SIG{...}`, but how can I remove that signal tap in P6? 17:30
psch jkramer: close the Tap returned by .tap i think 17:33
jkramer psch: That seems to remove the tap I created, at least my sub is no longer executed. C-c seems to be ignored now though, I guess the default/internal handlers is not put back in place or something? 17:36
psch jkramer: i'd guess this is like the default FDs -- if you mess with them, don't forget to put them back vOv
well, assuming you want the default behavior back that is :) 17:37
jkramer Well yes I'd like to restore the original behavior, I wouldn't want a module to mess with signals and then not clean up afterwards leaving me with a weird unexpected stuff :) 17:38
geekosaur part of the problem here is I'm not sure how direct access there is to signals; they're probablyu managed by libuv 17:43
also, signals being process level complicates things a bit
jkramer Hmm even after a react block with a "whenever signal(SIGINT)" inside, SIGINT behavior is not restored to normal 17:47
jkramer Hmm there isn't even a generic kill routine for making a test script :D 17:53
psch m: say Signal::INT 18:05
evalable6 (exit code 1) Could not find symbol '&INT'
in block <unit> at /tmp/XGGSJNku1G line 1
psch m: say Signal::SIGINT 18:06
evalable6 SIGINT
psch i think this should maybe be a ticket? 18:07
i mean, the inability to restore default signal response
psch assuming ++jkramer doesn't figure out how it works that is 18:07
jkramer psch: I'm already writing a ticket, just preparing a test script to reproduce 18:10
psch apparently i'm slow... :)
jkramer github.com/rakudo/rakudo/issues/1450 18:14
Not sure if I should also write a ticket for run() not ignoring SIGINT while running the external program or if that's wanted behavior 18:16
timotimo i don't think that would be a wise thing to do by default 18:17
geekosaur bad if backgrounded, good if foregrounded 18:18
jkramer Since `run` should be synchronous, I think it should ignore SIGINT. For Proc::Async it's a different thing 18:19
geekosaur that is, if Proc is pretending to not be async, it should ignore terminal signals (INT, QUIT, possibly HUP and TSTP, possibly INFO on BSD) 18:20
run can be async though
if you ask it to tap the process's stdin.stout/stderr
timotimo or feed its out or err into another process' in 18:20
jkramer Hmm I'll wait how that other SIGINT ticket works out I guess. If that's fixed I'm fine with the current way it works 18:24
comborico1611 If I just did chown -R (folder) *, did I affect other permissions outside of folder? I checked and it looks like no, but i want to make sure. 18:50
psch comborico1611: depends on wether $folder was a relative or absolute path, and the relation $CWD has to it, i' 19:05
d say
like, "chown -R foo:bar ../different/directory *" should change more than the owner of ../different/directory 19:07
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/01/29/...s-modules/ 19:24
psch lizmat++ 19:27
psch i should gather tuits for writing down some more thoughts on jvm-interop-goals.md, probably 19:31
lizmat ++psch 19:31
psch 'cause i'm kinda stuck in some spots too, mostly wrt theoretical details really... :)
the big downside is that explorative impl work is so tedious :/
psch well, "slow" really. what with 200+ second stage parse 19:32
lizmat yeah, even in Moar I find 60 second stage parse slow :-) 19:36
part of jnthn's grant work should improve that, but most likely only on MoarVM :-(
psch i'd like to make the jvm faster there, but i have no real clue how i'd due that 19:37
s/due/do 19:38
and well, heap profiling points at e.g. "20% of the heap are byte arrays," which we can't even really change
'cause that's mostly runtime generated class files 19:39
lizmat perhaps bartolin (usev6) has some pointers ?
psch one probably could kinda-sorta "shift" bits of the type hierarchy towards the jvm level -- i.e. have Cool actually be an interface on the jvm level -- but that seems slightly nuts 19:39
lizmat: maybe, but then bartolin's a fair bit less deep into the bytecode gen than i am afaik... 19:40
lizmat: i'm probably gonna try at least :)
psch but then the few memory improvements i pushed recently were also outside of what bartolin regularly works on, so idk vOv 19:41
buggable New CPAN upload: P5study-0.0.1.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.1.tar.gz 19:45
lizmat noop for the time being ^^
jnthn lizmat: Hm, I have a vague recollection that it's a no-op in contemporary Perl 5 also :) 19:52
lizmat not according to my perldoc perlfunc, but that is 5.20 on my MBP 19:53
lizmat ah, indeed 19:54
ok, full semantic equilibrium achieved :-)
buggable New CPAN upload: P5built-ins-0.0.4.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.4.tar.gz 19:55
New CPAN upload: P5ref-0.0.2.tar.gz by ELIZABETH cpan.metacpan.org/authors/id/E/EL/...0.2.tar.gz
MasterDuke lizmat: slight correction, those were jit templates, not spesh 20:09
lizmat MasterDuke++ # fixed 20:12
El_Che lizmat++ for the weekly weekly :)
MasterDuke lizmat++ 20:13
kolikov Hi o/ 20:19
psch kolikov: hello :) 20:23
kolikov I got some question concerning the routine move (IO).
psch kolikov: shoot 20:24
kolikov my $src = IO::Path.new('/home/zz.tar.gz'); my $dest = IO::Path.new('/home/temp/xx'); $src.move($dest); 20:26
And then when I run it : Failed to copy file: illegal operation on a directory
kolikov Both exist and are valid dir / file 20:27
psch kolikov: the docs for move says rename is the one that works with directories: docs.perl6.org/routine/move
kolikov: i'm not sure in how far the "if the files are on the same storage device" applies to renaming directories, but i'd assume you'd have to write your own recursive copy/delete for copying a directory including all files across storage devices 20:28
kolikov: oh, but it might be that moving a file to a directory is the actual problem here 20:29
kolikov yes
psch kolikov: as in, your destination IO is a directory, and a file cannot be move so that it turns into an (existing) directory
kolikov: yeah, that's lots simpler. change $dest to 'IO::Path.new("/home/temp/$target-filename")' and you're set :) 20:30
kolikov: that is, as your $dest points at the directory *itself* you cannot move the file there -- overwriting a directory with a file isn't allowed. you need to specify the filename *in* the destination directory 20:32
kolikov That sounds awkward ... the methods .d and .f state valid dirs and file, why not make a multy instead of extracting src.basename to concat withe a path as soon as you know it ??? 20:33
Unoix cp and mv work just fine with those cases. 20:34
AlexDaniel “Is there a key for it on my keyboard though?” 20:35
like if people use different keyboards to type in different languages?
never understood that logic really :S
DrForr kolikov: Because Perl 6 runs elsewhere than Linux. 20:36
El_Che DrForr: walks
:P
psch El_Che: rude
how 'bout "flutters," eh? :p 20:37
El_Che psch: That's a non-rude way to describe me
AlexDaniel El_Che: it kinda jogs actually
kolikov I'll make a run 'mv' , '/toto/file.gz' , '/toto/tmp'; Then .... That's sad :-(
AlexDaniel like, compared to what it was previously
El_Che well, whenever it can jog in parallel is way faster!
AlexDaniel I remember the startup time being like several seconds at some point 20:38
El_Che AlexDaniel: I remember the early days of Moose :)
DrForr kolikov: Just use 'move' and give it '/toto/file.gz' and '/toto/tmp/file.gz' as arguments and you're done, no need to resort to nonportable shell. 20:39
psch AlexDaniel: '-e1' still is at 3+ seconds on hack -- oh wait, that's r-j ha haa 20:40
AlexDaniel haha :D
kolikov DrForr : I know that's a solution, the sad thing if that i must concat $src.basename as a string to a IO Object while IO::Path already knows directories from files .... Just LTA ... no fuss :) 20:44
psch looking at IO::Path.move i don't see an immediate reason we couldn't have the suggested logic 20:48
well, except for the TODO comment maybe, checking on the VM level might be a bit more tedious
but then there was something about torture and implementers and users i think..? :)
kolikov I don't want to give you pain !
El_Che if it the addtion doesn't break stuff, add an issue? 20:49
psch "Larry: But, we recognize that that's how one person sort of suffers on behalf of someone else, vicarious suffering is the illogical term. Or, more humorously, what we say is that we torture the implementers on behalf of the users. Implementers are actually kind of proud of the fact that they are tortured on behalf of the users." 20:50
psch kolikov: the sentiment has been around a while -- that quote is from 2015 :) 20:50
kolikov But if you insist :-p 20:51
psch kolikov: i agree with El_Che++, maybe add an RFC issue on githuhb and see if someone else chimes in. you could even provide a propsitional patch :)
kolikov Houch ! patching, that's out of my (coding) scope, but I sure can fill an issue if you don't bother. 20:53
DrForr Please do. 20:54
kolikov DrForr : here github.com/perl6/ecosystem ? 20:55
DrForr I'd do it over on the main perl6 app, but psch might have a better suggestion since he's been paying more attention :) 20:57
psch uh actually i was gone for over a year until about a week ago... :)
my first instinct would have been RT, but that's kinda deprecated and annoying to use without an account 20:58
MasterDuke github.com/rakudo/rakudo/issues
psch so i *think* the right place is github... what MasterDuke++ said
kolikov Ok then, here I go, thanks a lot ( It's always welcoming here :-) ) 21:01
DrForr And I must be off to bed as well... Hook up the display, ae some preliminary content and I should be ready to rock. 21:02
*make 21:03
G'night.
kolikov Wish-Issue submited : github.com/rakudo/rakudo/issues/1453 21:33
El_Che kolikov++: 21:42
woolfy Hi folks, some time ago I used to come a bit more regularly! :-) <3 21:51
Just got a question from somebody on the Perl6 page on Facebook, regarding the Rakudo Star 2018.01 release: " could anyone update the docker image? i think this is the one hub.docker.com/r/_/rakudo-star/ "
moritz \o woolfy 21:52
woolfy So, I did my duty. :-)
Hey moritz++
woolfy Congratulations to all for releasing a brand new Rakudo Star, and for increasing the speed of Perl 6 100x since Tuxx++ started test.t, and for being awesome alltogether. I hope you don't mind saying that I am still happy with Perl 6. I am. 22:01
Xliff Hmmm.... 22:07
@aᵀ
ooh! I want!
tobs I think I fixed the issue with rotor and whatever/Inf from 10 hours ago. Initial contact with nqp. Should I pastebin a .t file in the PR message and add it to roast later?
or what's the ideal order of things?
moritz tobs: open two pull requests, mention in the roast that it should be merged after the rakudo one 22:08
tobs: and if you don't have commit access to roast yet, tell me your github username :-)
kolikov 'Night all \o/
tobs moritz: I don't have access. The account is brand new and shiny, "taboege". 22:09
psch tobs++
moritz tobs: you should have received an invitation by email. Have fun! 22:10
sleep&
Xliff [ ] only works with infix operators, yes? 22:11
psch Xliff: you're talking about reduce? then yes 22:13
Xliff: [ ] can also be used to turn a sub into an infix though 22:14
which can get a bit silly, really... 22:16
m: sub foo($, $) { say "fooedd" }; [[&foo]] 1, 2
evalable6 Potential difficulties:
Useless use of [[&foo]] in sink context
at /tmp/26jUHIMsfu:1
------> 03sub foo($, $) { say "fooedd" }; 08⏏04[[&foo]] 1, 2
fooedd
psch the sink warning seems wrong, but we got a few of those that seem somewhat wrong i think :/
m: say [[&infix:<+>]] 1, 2, 3, 4 # vOv 22:17
evalable6 10
TimToady the basic assumption is that infix operators are highly likely to not have side effects 22:20
psch m: say &infix:<=> # :o 22:21
evalable6 sub infix:<=> (Mu \a, Mu \b) { #`(Sub+{Precedence}|94770403733552) ... }
TimToady also, I actually said "vicarious suffering is the theological term", but it got mistranscribed illogically :)
we know that = is mutating though
psch yes, it is likely *the* exception
TimToady well, and any op based on the = metaop 22:22
carefully avoiding == and === and <==
there are spots where we look at 'is pure' declarations, but I'm not sure your [[&foo]] is one of them 22:23
also, xx can be control flow
so anything thunky is probably okayish 22:24
errands
psch m: [=] (my $a), (my $b), 3, 4; say $a; say $b
evalable6 (exit code 1) Cannot modify an immutable Int (3)
in block <unit> at /tmp/pHDYORjKB6 line 1
psch m: [=] (my $a), (my $b), 3; say $a; say $b
evalable6 3
3
psch what a wonderfully chained assignment :o
+convoluted 22:25
i wonder if checking that one for (lack of) purity makes sense 22:26
Xliff Hah! 22:29
m: my($a, $b) = 3 xx 2; 22:30
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/c1hLiYjfJY
Variable '$a' is not declared
at /tmp/c1hLiYjfJY:1
------> 03my(08⏏04$a, $b) = 3 xx 2;
Xliff WAT?
m: my ($a, $b) = 3 xx 2;
evalable6
Xliff m: my ($a, $b) = 3 xx 2; $a.say; $b.say;
evalable6 3
3
Xliff Less typing. ;q
psch Xliff: i am aware, but then '2 + 3' is also less typing than '[[&infix:<+>]] 2, 3' ;) 22:34
the neat thing is creating custom inline infix-reduce operator though 22:35
m: say [[&({$^a.uc ~ $^b.lc})]] "foo", "bar", "baz"
evalable6 (exit code 1) 04===SORRY!04=== Error while compiling /tmp/LaI90y4jv2
Two ter…
psch, Full output: gist.github.com/492f4a7bd40e6b9b16...6817d4e241
psch oh, but not quite like that haha
...it is still neat in a "oh wow don't do that in prod please" way :S 22:36
actually, i think that only works for infixes in the first place 22:37
psch m: say "foo" [&({$^a.uc ~ $^b.lc})] "BAR" 22:37
evalable6 FOObar
psch anyway yeah, just lotsa "the things we allow" kinda stuff
ahh 22:38
clearly "two terms in a row" *actually* means "you have to not have spaces because those make it comprehensible"
m: say [[&({($^a.uc)~($^b.lc)})]] "foo", "BAR", "baz"
evalable6 FOOBARbaz
psch i wonder if that TTIAR is actually wrong 22:41
maybe allowing pointy blocks in there works... 22:49
it is still silly though; maybe a typed exception along the lines of X::Cannot::InlineDefineReduce or something makes sense... 22:54
comborico1611 psch: it was a relative adrrss. Thanks for help! 22:59
timotimo "Timo Paulssen made some low level changes, that make entering and leaving optimized code sections faster." - i don't remember doing anything like that? damn, this moving stress is killing my brain 23:18
lizmat d6c2a3776c68780518500ce4f6 in moar 23:20
timotimo ^^^
lizmat almost 2 weeks ago 23:21
sleep& 23:22

Powered by Cro and the Raku® Programming Language.

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

Thank you!