This channel is intended for people just starting with the Raku Programming Language (raku.org). Logs are available at irclogs.raku.org/raku-beginner/live.html
Set by lizmat on 8 June 2022.
01:11 beanbrain joined, beanbrain left, beanbrain joined 08:10 dakkar joined 08:12 teatwo left 08:13 teatwo joined 08:20 teatwo left, teatwo joined 09:03 teatwo left 09:04 teatwo joined 09:44 beanbrain left 11:21 Chanakan left, Chanakan joined 11:35 discord-raku-bot left, discord-raku-bot joined 11:36 avuserow_ left 11:37 nicole left 11:44 avuserow_ joined, nicole joined 12:15 avuserow_ left, nicole left 12:17 avuserow_ joined, nicole joined
librasteve @gfldex ... I am smiling since this is a beginner channel that likes to go deep ... and argue the finer points (too late to go back now ;-)) [with apologies to all beginners who are wondering wtf - just ignore our ping pong) 14:02
so thanks for your patient explanation so far, my interpretation is this: a regular parameter will (a) implicitly "declare" its name in the scope of the block, (b) Bind the name to a thing and (c) store the thing's type in T... 14:10
m: sub f(::T $x){say T; say $x}; f(42)
Raku eval (Int) 42
librasteve so, this is analagous to this linear code: 14:11
m: my Int $x = 42; dd $x; 14:12
Raku eval Int $x = 42
librasteve based on that, an optional parameter with an absent argument will (a) implicitly "declare" its name in the scope of the block, errr that's it... 14:13
m: sub f(::T $x?){say T; say $x}; f()
Raku eval (Any) (Any)
librasteve so, this is analagous to this linear code:
m: my Int $x; dd $x; 14:14
Raku eval Int $x = Int
librasteve oops, sorry, I meant:
my $x; dd $x;
m: my $x; dd $x;
Raku eval Any $x = Any
librasteve so far I think that this is a good design - and if I switch $ to @ I get the same analogy 14:15
m: sub f(::T @a?){say T; say @a}; f() 14:17
Raku eval (Array) []
librasteve m: my @a; say @a.^name; say @a; 14:18
Raku eval Array []
librasteve so I am inclined to think that so far the easy things are easy to do and easy to remember since they are consistent 14:19
btw in this view, I your is raw example is no different - the raku spec says the implicitly declared / bound internals vars are read-only unless the is raw trait is used so from a coder pov that's a restriction on how the arg is bound ... and the optional form gives a sensible error message in the absence of an arg... 14:26
m: sub bar($x? is raw) {dd $x; $x=3}; my $y = 42; bar($y) 14:27
Raku eval Int $y = 42
librasteve m: sub bar($x? is raw) {dd $x; $x=3}; bar()
Raku eval Exit code: 1 Any assign requires a concrete object (got a Any type object instead) in sub bar at main.raku line 1 in block <unit> at main.raku line 1
librasteve I think we agree to differ on how best to answer the OP question - for me "don't test an array for defined, test fo Empty (since in raku even Empty Arrays are defined)" is the easy things are (fairly) easy and "you have to use a Capture with a sub-signature and introspect the Capture" is the hard things possible ;-) 14:30
15:39 teatwo left 15:40 teatwo joined 16:20 sjn left 16:21 sjn joined 16:41 dakkar left
gfldex I don't agree on your analog with my Int $x = 42; dd $x; The whole point of a type-capture is that you don't know the type at compile time but what to do fancy stuff with it at runtime (at least in a Signature, roles are specialised at compile-time). And a type-capture on @a? is outright odd. If no argument is provided by the caller it will be Array, if one is provided it will do Positional. If I would spot 20:54
that anywhere in the wild, I would start hunting for bugs, right away.
is raw in a Signature allows to get hold of a container the caller is using. So should ::T be Scalar if a container is supplied? 20:56
Also, if we have Nil to indicate the absence of a value, why not use it, if no value is supplied? 20:58
21:03 teatwo left 21:04 teatwo joined
m: sub impossible(Any:D $a?) { } impossible(); 21:11
Raku eval Exit code: 1 Parameter '$a' of routine 'impossible' must be an object instance of type 'Any', not a type object of type 'Any'. Did you forget a '.new'? in sub impossible at main.raku line 1 in block <unit> at main.raku line 2
gfldex :-|
ab5tract IMO this is a bit overboard for a beginner oriented chatroom, but I do agree it's worth discussing 21:28
that said, I don't understand where the confusion lies with your latest example gfldex 21:29
you've declared that the argument needs to be an Any:D, and you didn't provide a defined argument 21:30
it's marked optional, of course, but I don't know if I agree that this should supercede the type declaration 21:31
maybe it's an LTA though 21:32
21:54 lizmat_ joined 21:58 lizmat left 22:00 lizmat_ left 22:01 lizmat joined
pelevesque Guys, sorry I didn't really contribute to the discussion trying to answer my questions. 22:58
Everyone, sorry I didn't contribute to the discussion on my question. I read everything, but it's a bit over my head. I don't know the Raku internals well enough to contribute. Even some of the code posted as examples go over my head. sub-signatures... I feel like I entered a new galazy. 23:00
I think I am more than a beginner. I am a beginner beginning to begin. 23:02
antononcube @pelevesque You can also say: > a sub-beginner who sub-signs in a new sub-galaxy. 23:14
@pelevesque I think asking the question is enough of a contribution. The discussion is over my head too.
gfldex I'm a Raku beginner since 2008. :) 23:19
pelevesque It would be nice to have some universal way to know if optional args are set... Like $a? @a? \a? or whatever all behaving the same... like a $a.wasSet or something.
gfldex Sadly, that is not easily done becaue of is raw. We could have a method on `.VAR´ but that requires the container to be created by the binder, not the caller. 23:21
pelevesque Seems that if there is a universal way to make an argument optional (the ? char), there should also be a way a universal way to check if it was set. Maybe put in the docs that it can't be done for x reason? 23:26
gfldex It can be done, but only with Capture trickery. 23:27
However, you should not need that because you can just have a multi that does the same in a cleaner way. There might be edge cases, where you can't get around a Capture but I'm not sure if that is in scope of the docs. Feels more like it should be part of the Cook Book (that we don't have right now). 23:29
pelevesque Ah right, so multi would be the right choice for this. Hmmm 23:30
gfldex Likely, but as you didn't tell us what you acutally need that for, it's a little tricky to give advice.
As is common with questions that spawn lengthy Raku-discussions. :) 23:31
m: sub nice($a? = Nil) { dd $a; } nice(42); nice(); 23:32
Raku eval 42 Nil
gfldex If you don't want a multi (there can be good reason not to), that might be a nice way.
pelevesque I have a super ugly sub that I made work in the end. 23:33
# -------------------------------------------------------------------- sub get_frames ($score-file, $frame-numbers?) { my $grabbing = False; my @frame; my @frames; for $score-file.IO.lines -> $l { my $line = $l.trim; # Ignore empty lines, comments ｢-｣, and sections ｢@｣. if $line.chars && $line !~~ /^ <[-@]> / { # Create a new frame and
set grabbing. if $line.starts-with: '#' { if @frame.elems { @frames.push: @frame.clone; @frame = Empty; } if ( ! $frame-numbers.defined || $line.comb(/\d/).join.Int ~~ any @$frame-numbers ) { $grabbing = True; } else {
$grabbing = False; } } elsif $grabbing { # Save a frame's row. my @row = $line.comb; @frame.push: @row; } } } if @frame.elems { @frames.push: @frame.clone; } return @frames; } I was able to use $ for a list... then use @$ to smart match against it, so it worked in the
end.
I'm now have a way better sub than this using only a few lines with a REGEX, so this is abandonned, but it's why the question in the first place. 23:34
I wanted undefined to mean all frame-numbers, [] none, and [1, 24, ...| some 23:35
gfldex m: sub nice($a? = *) { given $a { when Whatever { say '*' } when () { say '()' } when .elems > 0 { say 'many' } } } nice(); nice([42]): nice([}); 23:37
Raku eval Exit code: 1 ===SORRY!=== Error while compiling /home/glot/main.raku Confused at /home/glot/main.raku:9 ------> nice([42]):⏏<EOL> expecting any of: colon pair
gfldex Without the typos, that would work. 23:39
pelevesque Thanks. I will take note. 23:40
With $ I can just do .defined though... @a? = * :: gonna test if that would work. 23:41
gfldex m: constant NaA = Array.new; sub nice(@a? = NaA) { say @a =:= NaA } nice(); nice([]); 23:54
Raku eval True False
gfldex That works for @-sigiled bindings.
m: constant NaA = Positional.new; sub nice(@a? = NaA) { say @a =:= NaA } nice(); nice([]); 23:55
Raku eval True False
gfldex That might be even better.

Powered by Cro and the Raku® Programming Language.

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

Thank you!