»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_log/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 July 2018.
timotimo greetings 00:00
please state the nature of your lepidopterological emergency
tbrowder__ i want to assign a function with multple args to a scalar, and then use it, but how is that done? the docs show a weird example for 2 args but it’s not clear how to practically use it. here is what i would like to do: 00:04
tbrowder__ timotimo: hi 00:04
tbrowder__ “my $sub = &do-something; $sub($a, $b, $c); 00:05
timotimo that's exactly how you do that 00:06
m: sub do-something($one, $two, $three) { dd $one, $two, $three }; my $sub = &do-something; $sub(1, 2, 3) 00:07
camelia 1
2
3
TimToady and you can use .assuming if you want to pre-supply some of those args
tbrowder__ ok, thanks. i need to go back and make sure i’m not making a typo 00:09
tbrowder__ hm, do the subs have to be declared before the reference is taken? 00:35
the subs i’m trying to reference are at the end of the program 00:36
geekosaur probably yes. otherwise you'd get a type object 00:37
tbrowder__ calling them directly is no problem
geekosaur there are ways to trip this with variables due to some things happening at compile time and others at run time, and compilation being one-pass 00:38
timotimo m: my $foo = testsub; sub testsub() { say "yes" }; $foo() 00:39
camelia yes
No such method 'CALL-ME' for invocant of type 'Bool'
in block <unit> at <tmp> line 1
timotimo m: my $foo = &testsub; sub testsub() { say "yes" }; $foo()
camelia yes
timotimo that apparently ought to work
tbrowder__ let’s see 00:45
m: my $f = &bar; $f(); sub bar() { say “yes”} 00:46
camelia yes
tbrowder__ m: my $f = &bar; $f(1, 2); sub bar($a, $b) { say “$a, $b”} 00:48
camelia 1, 2
TimToady subs are the one spot where we worked really hard to allow forward refs
tbrowder__ hm, i must have a typo, this all looks great and as expected (wanted)
TimToady because mutual recursion
tbrowder__ roger! 00:49
thanks TimToady and timotimo!
TimToady it does have to be resolved by the end of the file
tbrowder__ yes!
tbrowder__ sub refs working great after i found err elsewhere! 02:33
joy_ hello, while compiling perl6 from source on a Centos 7 machine (for which im not root, so i'm installing locally), i got an erorr 04:17
Can't locate Digest/SHA.pm in @INC (@INC contains: /usr/local/lib64/perl5 /usr/local/share/perl5 /usr/lib64/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib64/perl5 /usr/share/perl5 .) at tools/build/gen-version.pl line 9.
i guess i could install Digest/SHA.pm locally (how?) and if so, how can i tell the perl6 make setup to modify @INC to look in my local (i.e. my home folder) Perl library path? 04:18
geekosaur export PERL5LIB=whateverdir 04:30
joy_ geekosaur: thanks much. One last thing, how to install Digest/SHA.pm locally (i.e. into whateverdir) 04:31
geekosaur (centos installs a lobotomized perl 5 by default) 04:33
perl -MCPAN -e shell
install File::Slurp
except here it'd be Digest::SHA1, and you rpoably need to do something to overridethe install directory 04:34
joy_ let me try 04:36
geekosaur except watch the broken perl5 centos installs ntohave theCPAN module :( 04:39
joy_ using perlbrew ... 04:55
geekosaur that may be your best bet 05:12
geekosaur mention centos to anyone with respec tto a perl 5 question and they mostly won't talk to you unless you use perlbrew or install perl-core 05:15
Geth doc: 00b7b00a91 | (JJ Merelo)++ | 3 files
Rewrites Test and Enumeration to index it properly
 05:51
EdwardIII15 how come * under bash can both glob and work as a literal arg for some program, but not for zsh? eg. ls somedir/Downloa* && find somedir -type d -name Downloa* will both work under bash, but for zsh the latter asterisk attempts to glob and messes up the whole command, returning zsh: no matches found: Downloa* 06:57
.c .010817 * 23
yoleaux I don't know
lookatme_q I remember this is an option about that in zsh 07:03
tyil they're different shells, doing different things 07:09
zsh != bash != csh and one could go on and list every shell there 07:10
that's like saying "why does X work in Python but not in Ruby"
lookatme_q I think that guy not here now 07:25
There is a option can disable this behavior
Any14Coffee Morning all 07:56
masak autopun spotting: twitter.com/shadowcat_mst/status/1...4667912192 (via diakopter++) 08:02
Any14Coffee I can use $ARGV to capture my Cmd line arguments however is there a variable for me to capture what I’m stunning my script against? I.e. ./blah.pl <example> 08:03
masak m: sub foo { return 1; LEAVE { return 2 } }; say foo() 08:06
camelia Attempt to return outside of any Routine
in sub foo at <tmp> line 1
in block <unit> at <tmp> line 1
masak oh, that's one way to handle that, I guess ;)
masak I was thinking of Java's (and other languages') ability to return a second time in a `finally` clause, overriding the first return value 08:07
Any14Coffee Ah ok, will give it a try 08:08
Thanks
masak the LEAVE above, is it truly "outside of any Routine"? 08:25
masak I mean, it's clearly textually inside `foo`, but how about dynamically? 08:25
I think I had always assumed that LEAVE executes *before* we leave a surrounding block
masak ...is there a mechanism in Perl 6 similar to Java et al.'s "return another value from a `finally` block"? 08:40
(I don't have a use case for it, but I'm curious about nailing down the semantics)
my guess is that there isn't, because of how `return` works in Perl 6 as opposed to Java 08:43
domm not sure if this was discussed here, but a Perl 6 solution is missing in this comparison: github.com/drujensen/fib 08:46
Ulti is there a way in the ecosystem to be able to flag operating system compatibility 09:22
Ulti Im noticing a lot of things fail on macOS that have a native component but quite often its something like UUID as a dep a million miles from what Im installing, it would be quite useful to flag that just one dep wont install on your operating system to save the ten minute wait for failure 09:23
Ulti also can I test the ecosystem and feed that back somewhere as something helpful? 09:24
can you already give binary deps like assumed libraries in the META file? 09:26
pmurias skiplang.com/blog/2017/11/20/fixing...rrier.html - a blog most relevant to the p5ism errors discussion 10:13
yoleaux 27 Sep 2018 00:16Z <Zoffix> pmurias: thanks. I reverted that propspec in github.com/perl6/roast/commit/7e0b791fa3
lizmat weekly: skiplang.com/blog/2017/11/20/fixing...rrier.html 10:16
notable6 lizmat, Noted!
tbrowder__ hi, folks 10:25
tbrowder__ got bit for a while by a left curly brace at the end of a string (in ascii “ quites). i had to escape it, but the error msgs weren’t very useful. let me see what happens here... 10:28
m: say “a left curly {“
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in curly double quotes; couldn't find final '”' (corresponding starter was at line 1)
at <tmp>:1
------> 3say “a left curly {“7⏏5<EOL>
expecting any of:…
tbrowder__ thats msg looks good to me. now i have to find out why i didn’t get it in my code (could be because was writing a code generator). 10:30
tbrowder__ hm, my code seems to be ok otherwise...still checking. 10:33
masak m: foo: say foo.^name 10:42
camelia Label
masak m: foo: say foo.^methods
camelia (new name goto leave Int next redo last gist BUILDALL)
masak moritz: on the topic of "should an object of that type really hold that method?", what about Label objects having a .goto method? :) 10:43
I think it's comparable to integer objects having a .times method -- it's not that it's super-wrong exactly, but it's kind of ignoring a surrounding "context" that's really doing the heavy lifting 10:44
in both the case of .times and .goto, there's an implicit runtime somewhere that does the appropriate control flow
m: foo: say "HELLO WORLD"; foo.goto() 10:45
camelia HELLO WORLD
Label.goto() not yet implemented. Sorry.
in block <unit> at <tmp> line 1
masak haha :D
masak "Sorry." -- somehow that "sorry" felt a lot less genuine than the one you get when there's a syntax error. this one is more like "we haven't had the time to implement this feature, because we're *busy*, you know" 10:46
lizmat SORRY! 10:49
:-)
moritz masak: that's an interesting question. But at least Label is already a control flow object, whereas 5.times(callback) seems like a category error (number vs. control flow) 10:55
masak moritz: not arguing for the justness of 5.times here, but -- from a Peano/Church interpretation, it's actually not completely crazy that natural numbers represent a kind of iteration/repetition 11:00
moritz there's a reason we use integers to count iterations, yes :-) 11:01
sena_kun `5.times` seems more like a stealing from smalltalk, where we have something like `1 to: 20 do: [:x | x printNl ]`. and in smalltalk there are no real categories between things, as everything is an actor && actor can do only calls.
moritz the inductive defintion of integers very closely iterations
sena_kun: I know it from ruby, but I guess ruby might've stolen it from smalltalk :) 11:02
sena_kun moritz, yes, I think they might. :)
masak yes, they did 11:03
my understanding of Smalltalk is that Smalltalk is very un-Liskovian :)
and the criterion is more like "does it feel cute today to send 1 a to:do: message? then let's do that"
this is not a criticism so much as trying to capture what it is Smalltalk is aiming for 11:04
inheritance in Smalltalk is not used for some neat/pure subset relation, but for pragmatic re-use 11:05
pmurias masak: re Int having a times method isn't that more a philosophical question as it seems to depend on what we assume numbers are? 11:12
masak pmurias: it's that too, but I was approaching it as an object-oriented design question 11:13
masak as in "where does this method go in a 'good' design?" 11:13
masak like I said elsewhere, I don't consider 5.times to be either more or less wrong than "OH HAI".say 11:14
I think if we could've done Perl 6 from scratch, then `.&say for @values;` would've been enough, and no need for a method 11:15
(not bashing methods in general; just that one) :) 11:16
jnthn If you were to assume that 5 is represented as a Church numeral, wouldn't passing a function to apply 5 times kinda makes sense? :P
*make
masak yes, like I wrote above :)
harrison m: say 1.^methods; 11:17
camelia (new Capture Int Num Rat FatRat abs Bridge chr sqrt base polymod expmod is-prime floor ceiling round lsb msb narrow Range succ conj cos cosh acosec sign cotanh log asech acosech pred log10 tanh roots sin acos cosec atan2 sinh cotan isNaN rand tan sech…
jnthn Oh, I dind't read backlog :)
masak <+masak> moritz: not arguing for the justness of 5.times here, but -- from a Peano/Church interpretation, it's actually not completely crazy that natural numbers represent a kind of iteration/repetition
jnthn haha
.oO( Sufficiently warped minds think alike )
masak 'fraid so
similarly, a semicolon is really just an operator that passes the rest of the statement list into the current statement as a continuation 11:18
harrison Wouldn't it be easier if *.^methods return the list of methods ordered in alphabetical order, for easier scanning?
I mean, a person scanning with their eyes.
masak m: say 1.^methods.sort(*.name) 11:19
camelia (ACCEPTS Bool Bridge Capture Complex DUMP FatRat Int Num Numeric Range Rat Real Str WHICH abs acos acosec acosech acosh acotan acotanh asec asech asin asinh atan atan2 atanh base ceiling chr cis conj cos cosec cosech cosh cotan cotanh exp expmod floor…
moritz harrison: the MOP is optimized for human aye scanning
*NOT
m: say 1.^methods.sort
camelia Method object coerced to string (please use .gist or .perl to do that)
in block <unit> at <tmp> line 1
Method object coerced to string (please use .gist or .perl to do that)
in block <unit> at <tmp> line 1
Method object coerced to string (…
masak .oO( "should I come in again?" ) :P
moritz m: say 1.^methods».name.sort
camelia (ACCEPTS Bool Bridge Capture Complex DUMP FatRat Int Num Numeric Range Rat Real Str WHICH abs acos acosec acosech acosh acotan acotanh asec asech asin asinh atan atan2 atanh base ceiling chr cis conj cos cosec cosech cosh cotan cotanh exp expmod floor…
harrison moritz: Oh. 11:20
masak .oO( TODO: make a MOP that's optimized for human eye scanning )
jnthn Well, the MOP is at least partly there for people to introspect stuff, but it's just as much there for meta-programming, and making every programmatic use of the data pay for a sort it almost certainly doesn't need would be a bit unfortunate 11:21
masak jnthn++ has a tendency to express things exactly the way I want to say them :> 11:23
harrison jnthn: makes sense.
How hard would it be to make something like distillery for Perl6. github.com/bitwalker/distillery. A quote from the documentation: "Distillery is a tool for packaging Elixir applications for deployment using OTP releases. In a nutshell, Distillery produces an artifact, a tarball, which contains your application and everything needed to run it." 11:27
lizmat harrison: isn't that called Docker nowadays > 11:30
?
tbrowder__ hi, folks, got a question on contributing to roast: 11:35
pmurias "OH HAI".say doesn't seem to be a good design at all 11:36
tbrowder__ i need an exhaustive bunch of pod test for the implementation of ‘#’ :numbered alias and am writing a test generator. should that generator script be in roast, too, mabe somewhere under packages? 11:38
harrison lizmat: Yes, but I actually mean, something a little bit simpler to adopt. Like I run `some-command-line --param1 --param2 --param3` and it generates a tar file with moarvm inside, modules I am using and my application. So all I have to do is upload on my server and run it. No tinkering with things and installing things on the server. 11:38
masak pmurias: that's what I'm saying. who gave strings the responsibility to print themselves? 11:40
El_Che harrison: as long rakudo is not relocatable, it would be very useful 11:41
lizmat m: "this way".put 11:42
camelia this way
lizmat m: "this way".note
camelia this way
lizmat m: 42.note
camelia 42
lizmat masak: I think it's Cool things that we gave the resposibility 11:43
masak yes, you're right 11:44
and Cool is just a manifestation of the p5 scalar concept, that some types are interconvertible
but it doesn't make much more sense for Cool to own the capacity to print itself than for Str to do so 11:45
lizmat perhaps, to me it feels like TIMTOWTDI
specifically when testing things
masak anyway, I'm not angry about this, or proposing a Grand Change -- I'm fully aware Perl 6 has never been about that kind of ideological purity
harrison El_Che: Yes, it can even be easier to choose a version of rakudo to use so that you can easily generate runnable units of the same app but for different rakudo versions.
lizmat so you're writing this long expression to test things, and than you realize you forgot the "say" 11:46
masak lizmat: aye, it is very pragmatic. my point was that `.&say` (invented later) would've done just as well, and would've made the method unnecessary
lizmat then it's nice to be able to do just '.say'
masak: fwiw, I find the .&foo syntax an abomination from the P5 days 11:47
where you didn't know whether something was intended to be used as a sub or a method
tbrowder__ never mind, i think i see some examples
masak lizmat: interesting. I kind of love it, and I don't mind at all the fluency that exists between subs and methods 11:48
in fact the sub <-> method "isomorphism" goes very deep in Perl 6
lizmat as it goes even deeper in Perl 5 :-)
yet, most attempts at MOP's try to make a difference between sub / method use 11:49
El_Che harrison: yes, but that's not an issue if rakudo is part of the package (the best situation). Now you'll need root/admin to write to a generic situation like /opt/rakudo or c:\rakudo because the rakudo path is hardcoded. 11:50
masak m: my method foo(Str $s:) { say $s }; sub bar($s) { say $s }; "this works".&foo; "so does this".&bar
camelia this works
so does this
El_Che harrison: you won't be able to run it from e.g. your home (or / if a container)
masak lizmat: Perl 6 is the only language I know where you can have both subs *and* methods inside a class. (and there's a distinction, and it kinda makes sense)
harrison El_Che: I see, I hadn't thought of it like that. 11:51
masak lizmat: but the distinction disappears if you go deep enough into the semantics, because (after all) methods are just routines that you called "on an invocant", and so not so different from subs
lizmat: I dunno, I think I've just made peace with that ;)
jnthn masak: subs in a class pretty much fill the same niche as static methods, no? :) 11:52
masak I remember being frustrated at some point that error messages considered the invocant to be one of the parameters, and so it would kind of report the parameters "off-by-one"... but I've come to accept that as a necessary sacrifice when there's no fundamental distinction between subs and methods
jnthn: yeah, guess so. except that's not how I think about it in Perl 6, but I guess it makes sense 11:53
Geth ecosystem: 230a11d9ee | Altai-man++ (committed using GitHub Web editor) | META.list
Add ECMA262Regex module
 12:16
lizmat weekly: niner.name/blog/house_cleaning_with...index.html 12:48
notable6 lizmat, Noted!
MasterDuke domm: github.com/drujensen/fib/pull/44 13:35
domm MasterDuke++ 14:00
lizmat MasterDuke: that bench would be a lot faster as: $n < 1 ?? 1 !! fib($n - 1) + fib($n - 2) 14:03
while still being idiomatic Perl 6, I would think ? 14:04
MasterDuke lizmat: heh, github.com/drujensen/fib/pull/45
MasterDuke and github.com/drujensen/fib/pull/50 14:04
lizmat :-) 14:05
sena_kun >$n <= 1 ?? 1 !! fib($n - 1) + f(ib$n - 2); 14:06
>f(ib$n)
oops
MasterDuke, ^
also s/f(46)/fib(46)/ 14:07
MasterDuke sena_kun: ?
sena_kun isn't code at github.com/drujensen/fib/pull/50/files has thinko? 14:08
MasterDuke sena_kun++
sena_kun I may be mis-understanding something here, sorry if so.
sena_kun but kind of fun how fast memoized version is compared to other benches. I wonder if it's considered as cheating in this repo. 14:09
MasterDuke++
MasterDuke the perl6 version without return is faster than perl5 for fib(35) 14:11
lizmat yeah, it takes a little while for Perl 6 to heat up 14:13
timotimo yeah, return is *really* slow :) 14:18
lizmat hmmm... perhaps we could statically determine in the optimizer that we can lose any 'return' at the end of a sub / method ? 14:20
timotimo i think we already have an opt that tries to do that
it must have bitrotted
jnthn Hm, I wonder if return control exceptions are rewritten into goto yet 14:21
(Also, I think that `return` once wasn't so obviously slow. I suspect lots of other things have sped up around it.) 14:22
timotimo surely possible :) 14:23
Zoffix jnthn: can't we simply statically throw out `return` if it's in the last statement? 14:48
jnthn Zoffix: Yes, though it'd be nice to make the other cases faster too
Zoffix I had a bot reminder to check that, but I kinda ignored it on the assumption that something so obvious would've been done already if it weren't already
'cause one thing I noticed is lots of Perl people stick an explicit `return` for "better style" practices, without realizing it has a perf impact in p6 14:49
jnthn Well, it won't once we optimize it :) 14:50
awwaiid AlexDaniel`: Ahh!! no wonder I haven't been getting so much email, I let my domain name expire. Fixing...
Zoffix And some people stick it in just 'cus: github.com/koorchik/formula-evalua....pl#L8-L33
Zoffix
.oO( wonder if it's detrimental to throw out `&return` but leave in `RETURN -> - QAST::Op(lastexpayload)` in the tree
 14:58
)
AlexDaniel awwaiid: o yey it's back! 15:03
awwaiid: thank you!
AlexDaniel what's the proper name for `@foo.bar: baz` syntax (as opposed to `@foo.bar(baz)`)? 15:07
Zoffix Don't immediatelly see any logic in the optimizer for telling "this is a routine body and not a block/thunk body" so giving up on return elision right now (need to do 6.d stuff), if anyone wants to take a crack at it. 15:08
AlexDaniel: colon call? :) 15:09
AlexDaniel: FWIW, I saw some term used in the docs for it. 15:10
precedence lowering or something (didn't make much sense to me TBH)
timotimo yeah, listop precedence i think?
it's like the indirect object notation, except the object isn't indirect 15:11
so it should be called "notation"?
Zoffix
.oO( colonotascopy )
AlexDaniel on thelackthereof.org/Perl6_Colons it is called “Precedence dropper” 15:23
but our docs never say that 15:24
awwaiid (site will be back eventually) 15:26
I think I got that language from TimToady 15:27
AlexDaniel awwaiid: it's already back 15:28
AlexDaniel here at least 15:28
awwaiid reloads harder 15:32
Scorpion thanks 15:59
I want to contribute to the Perl6 project. 16:00
sena_kun Scorpion, sure you can! you can start with reading "Getting involved" section at perl6.org/community/ 16:03
Zoffix Scorpion: awesome! What sort of tasks were you looking for? 16:04
BTW, do we have Hacktoberfest labels set up?
don't look like it 16:05
MasterDuke Zoffix: i think jmerelo and AlexDaniel were chatting about that recently, but i don't remember the outcome
Scorpion Thanks guys. However I can help. I use perl5 daily in network/system engineering. Perhaps I can start testing code
MasterDuke Scorpion: github.com/rakudo/rakudo/labels/testneeded 16:06
Zoffix Scorpion: there are some already-fixed bugs that need tests written for them: github.com/rakudo/rakudo/issues?q=...testneeded
Scorpion Thank you Zoffix
Zoffix Scorpion: there's also a Brief Newcomer's Guide: rakudo.party/post/Newcomer-Guide-t...ore-Perl-6
Zoffix As well as this channel and #perl6-dev :) 16:06
Scorpion MasterDuke Thank you 16:07
Zoffix I'm adding "Hacktoberfest" label to all "good first issue" rakudo/rakudo issues and sticking rakudo.party/post/Newcomer-Guide-t...ore-Perl-6 as first line in the issue 16:11
MasterDuke Zoffix++ 16:12
Zoffix Scorpion: what's your github username? 16:17
jmerelo Hi! 16:30
AlexDaniel MasterDuke: yes but hacktoberfest is not that related to squashathons so the label needs to be added anyway 16:31
jmerelo: so the upcoming squashathon needs a doc saying that people who want their modules to be worked on have to add a webhook to whateverable.6lang.org:4243/ with Content type: `application/json` and `Send me everything` 16:34
jmerelo: and then mark issues with `hacktoberfest` label 16:35
the bot will then announce here that the webhook was added, and I guess that's it?
we also need a list somewhere of all repos
and by doing that, yes, you invite hacktoberfesters and squashathoners at the same time 16:36
sena_kun hmm, the module was properly added in github.com/ugexe/Perl6-ecosystems/...49033b0557 but then removed in github.com/ugexe/Perl6-ecosystems/...849021050, what can the reason be? 16:37
AlexDaniel also there will be a winner who gets a plush camelia :)
jmerelo: can you write something like that but properly? :)
MasterDuke heh github.com/drujensen/fib/issues/51 17:03
jmerelo AlexDaniel: OK 17:04
I guess you mean something like the two sentences above, but in a place and extension that can make it more understandable, right?
I mean, a howto so that anyone can do that.
kephra_lk cheers jmerelo 17:05
jmerelo Hi, kephra_lk !
squashable6: status
squashable6 jmerelo, ⚠🍕 Next SQUASHathon in 6 days and ≈16 hours (2018-10-06 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
kephra_lk jou know how if there is still crosswith method to overload X metaop
jmerelo OK, I'll try to do that over the weekend AlexDaniel 17:06
kephra_lk: you can't overload X directly?
kephra_lk jmerelo i can but i still wonder what happened to crosswith 17:07
jmerelo Looking over the source, it looks like it disappeared sometime after 2010... github.com/rakudo/rakudo/commit/2f...57b68ad54d 17:09
kephra_lk jmerelo thank you very much, it case you wonder im the matrix guy from glasgow just i travel now 17:10
kephra_lk jmerelo could you show me the signature of the X metop ? 17:12
tbrowder__ m: my $s=“1\n2”;say $s 17:18
camelia 1
2
tbrowder__ m: my $s=“1\n2”;say $s eq “1\n2”; 17:24
camelia True
AlexDaniel jmerelo: yes, like a markdown page or something like that 17:34
kephra_lk can anyone help me a little with overloading the X metaop ?
AlexDaniel jmerelo: similarly to the beautiful page you did for the unbitrot squashathon
jmerelo AlexDaniel: OK
AlexDaniel jmerelo: perhaps a list of modules that are participating on the same page
jmerelo AlexDaniel: OK, will try. Feel free to prod me back in a couple of days if I don't. 17:35
Xliff \o 18:08
What does it mean when I get the following error message?
cbwood@cbwood-VirtualBox2:~/Projects/p6-GtkPlus$ perl6 -Ilib t/10-menu.t
Default constructor for 'GTK::MenuItem' only takes named arguments
in block <unit> at t/10-menu.t line 22
Because I am trying to use both positional AND mixed parameters at t/10-menu.t line 22 18:09
Zoffix Xliff: it's hitting Mu.new. You got custom multies? Your signatures aren't matching 18:11
If you want to block off Mu.new from being considered add a `proto method new` in your class
Xliff Zoffix: Thanks. Here's the code. 18:12
github.com/Xliff/p6-GtkPlus/blob/m.../10-menu.t
Zoffix Xliff: what's with nqp stuff? 18:15
Xliff nqp? 18:16
What are you looking at?
Zoffix use nqp;
my $o = nqp::create(self)
in MenuItem.pm6
Xliff Oh! I am overriding method bless there.
Got that bit from timotimo
Zoffix nqp is not user-supported. You shouldn't be using it. 18:17
Xliff: blocking off Mu.new gives a better error: gist.github.com/zoffixznet/b33a036...424c001f36
Xliff I am well aware of that.
I have signed the nqp disclaimers. It works. That's all I care about.
Zoffix Xliff: but that's exactly the point. It works *right now*. It can change at any moment and not work anymore. 18:18
Xliff I am aware of that. 18:19
Do you have a better method of overriding method bless?
Zoffix method bless(*%attrinit) { 18:20
my $o := self.Mu::bless: |%attrinit;
$o.setType('GTK::MenuItem');
$o;
}
Xliff Zoffix++ !! 18:21
moritz might be better use |c captures 18:22
more efficient, that is
not sure though
Zoffix
.oO( do `.setType('GTK::MenuItem'); later in the food chain—.new, .BUILD, .TWEAK—and leave .bless alone)
 18:23
moritz too simple
Xliff Zoffix: Tried that. BUILD does NOT work becuase of inheritance chain. setType() is only supposed to run ONCE. 18:24
same with TWEAK
Zoffix m: class Foo {}; sub (Foo() :$z) {}() 18:31
camelia No such method 'Foo' for invocant of type 'Any'
in sub at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix m: class Foo {}; multi z (Foo() :$z) {}; z 18:31
camelia Cannot resolve caller z(...); none of these signatures match:
(:$z)
in block <unit> at <tmp> line 1
 18:32
Zoffix Xliff: it's this coercer: github.com/Xliff/p6-GtkPlus/blob/m...em.pm6#L62
Xliff: you're not giving it a default, so it's trying to coerce the default default value of that parameter (`Any`) to that type, and since it can't that multi gets regarded as non-matching
Xliff Zoffix: Thanks! That's helpful, since I need those. It's the only time I've tried that using a named-arg. 18:33
So I guess a better way to write that would be -- "GtkWidget() $submenu = GtkWidget" 18:34
Zoffix prolly
Xliff Because I need to use a "with" construct later in the code... which should actually be in BUILD.
Thanks a lot!
Zoffix Xliff: this potentially is also a bug. You're giving `$_` to the subs, yet it doesn't appear to be set to anything: github.com/Xliff/p6-GtkPlus/blob/m...m6#L64-L67 18:35
Xliff Whoops! That was originally in a with block. Thanks! 18:37
lizmat weekly: perlmonks.org/index.pl?node_id=1223228 18:39
notable6 lizmat, Noted!
SmokeMachine is there a way to colorize perl6 code on wordpress? 20:00
moritz you can let vim or other tools generate colored HTML, and add that to WP 20:02
Zoffix .tell skids just a minor comment: it's slightly annoying to come across several dozen merges during spec review because it takes a second to realize which changes you reviewed already in earlier commits. Unless it can't be merged to master yet or it's somethign controvercial, just commit directly instead of going through PRs. The commits are bot-announced and the spec is reviewed pre-release, so we don't 20:08
yoleaux Zoffix: I'll pass your message to skids.
Zoffix really need any FYI-type notifications about new tests.
.tell skids really need any FYI-type notifications about new tests.
yoleaux Zoffix: I'll pass your message to skids.
SmokeMachine moritz: yes, thanks... 20:09
kensanata I'm a bit confused about documenting my code. I start with =head1, then =begin pod, some general stuff, =end pod. Then I add #|{...} for functions and classes. The result doesn't look very nice using p6doc. Pod in #|{...} is not rendered. And the text has all line breaks removed but no word wrap added. This doesn't mix well. What do you do in this situation? 21:22
Example alexschroeder.ch/cgit/oddmuse6/tre...e/Diff.pm6
timotimo i think you're mixing up a few different things 21:53
i myself would put =begin pod and =end pod around things like =head1 21:54
and i didn't actually know you could put parenthesis (any kind) directly after the | or =
if it's just one line, you can of course leave it out 21:55
tyil timotimo: he's not online anymore 21:58
I sent him an email with a patch to solve it, and explain why I changed things around
[Coke] tyil++ 21:59
timotimo damn 22:04
that's real nice
timotimo yes, tyil++ indeed 22:07
Kaiepi is getting the terminal width and height something that'd be better off left as a library or implemented natively? 22:18
[Coke] library 22:20
wait.
japhb Kaiepi: Terminal::Print does this using tput to emit the terminal-specific sequences for this.
[Coke] I read that as 'library vs. core' not 'library vs. native'. my bad
Kaiepi oh 22:21
japhb Terminal size is a mess that is implemented by library on all posix-like OSen.
It's such a pain that we wrapped the command line tool, instead of picking a library
japhb ... and it's still an ugly mess (see github.com/ab5tract/Terminal-Print...mands.pm6) 22:22
japhb wonders if the Rakudo bug referenced in line 50 is fixed .... 22:24
Kaiepi i do it like this in Net::Telnet hastebin.com/upetelofew.pl
geekosaur you do somewhat less; telnet isn't trying to figure out whether the user's terminal emulator is antisocial 22:27
Kaiepi wdym? 22:28
geekosaur (like, when you output a character in the last clumn, where is the cursor afterward? telnet doesn't care. Termina::Print ants to know if it's a terminal that leaves the cursor in thelast column instead of the first column of the next line) 22:28
Kaiepi oh 22:29
geekosaur this stuff gets so very _fun_… not 22:30
tbrowder__ tyil: ref the pod problem, could you please let me see what you sent kensanata! 22:49
s/!/?/ 22:50
allowing some formatting in declarator blocks is on my todo list. 22:51
tyil tbrowder__: sure 22:52
tbrowder__: p.tyil.nl/bee6 22:55
tbrowder__ tyil: thanks! 22:57
tyil :> 22:58
tbrowder__ just added new rakudo pod issue GH# 2332 23:13
synopsebot GH#2332 [open]: github.com/rakudo/rakudo/issues/2332 [POD] allow some formatting in pod declarator blocks
tbrowder__ note that has been a TODO note in current code 23:14
Geth doc: threadless-screw++ created pull request #2335:
Rewrite of Identifier-section		 23:41

Powered by Cro and the Raku® Programming Language.

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

Thank you!