August 28, 2022 1 comment

While reading solutions to the PWC, I spotted a pattern. There where plenty of .map-calls that contained only simple maths. It must not be so! Task 179-2 leans itself to use vector operations.

use v6.d;

constant @sparks = "▁" .. "█";
constant \steps = +@sparks - 1;

multi MAIN(
    #| space separated list of numbers
) {
    my $scale := @data.&{ (.max - .min) / steps };
    put @sparks[((@data »-» @data.min) »/» $scale)».round].join;

multi MAIN(
    #| output an example sparkline-graph
    Bool :$test
) {
    .&MAIN for <2 4 6 8 10 12 10 8 6 4 2>, <0 1 19 20>, <0 999 4000 4999 7000 7999>;

First, I define a character range (no need for .ord here) and store the number of bar-graph steps. Then I calculate the scale like it is used in a map (those paper-things, that always work and don’t break when dropped). Now I can use that scale to shrink (or grow) values. I believe rounding is better here then to cut of to integers.

When we feed a list to a postcircumfix:<[ ]> we get a list of values in return. However, we break the method-call chain when we do so. Since operators are just Subs with a funny syntax, we can solve that issue easily.

constant &sparks = {
    constant @sparks = "▁" .. "█";
    constant \steps = +@sparks - 1;

    &postcircumfix:<[ ]>.assuming(@sparks) but role :: { has $.steps = steps }

my $scale := @data.&{ (.max - .min) / &sparks.steps };
((@data »-» @data.min) »/» $scale)».round.&sparks.join.put;

The constants are hidden within a block-scope and only steps is exposed as part of a Subs “namespace”. By declaring &sparks as constant, I move the currying to compile time.

Hypers are not getting much use in PWC-solutions. I believe another entry in raku-did-you-know is in order.

Categories: Raku

Suspicious introspection

August 22, 2022 Leave a comment

One of my routines is acting suspicious. I’m almost certain it wilfully misinterprets arguments just so it can throw an exception. There is a way to spy on it.

sub suspicious($a, :$b, *@rest, *%opts) {

    CATCH {
        put (my $sig = &?ROUTINE.signature.gist) ~ ' -> ' ~ $sig.trans(<*@ *%> => <@ %>).EVAL».raku.join(',');

suspicious(1, :2b, 3,4,5);
# OUTPUT: ($a, :$b, *@rest, *%opts) -> 1,:b(2),3 4 5,
            in sub suspicious at tmp/2021-03-08.raku line 1910
            in block <unit> at tmp/2021-03-08.raku line 1917

Please note, that the CATCH-block doesn’t do anything to the exception. It is just there to trigger the debugging code. The proper place to put that would be a macro. Alas, macro can’t inject CATCH-blocks that are actually used as exception handlers. It will have to wait for RakuAST, what would also allow me to replace the EVAL-shenanigans with proper code. Another entry in macro-ideas.txt until then.

Categories: Raku

Defeating definedness

August 18, 2022 1 comment

In his latest blogpost, p6steve went through a lot of trouble to enforce definedness. I managed to shorten it a bit.

None: Nil;
subset Option of Any where Any:D | None;

my Option $defined = 42;
my Option $nothing = None;
my Option $undefined = Mu;

.&dd for $defined, $nothing, $undefined;

# OUTPUT: Type check failed in assignment to $undefined; expected Option but got Mu (Mu)

I would really like to see the content of $nothing, and while I’m on it show why Steve’s hard labour was in vain.

try {
    my Option $defined = 42;
    my Option $nothing = None;
    my Option $undefined = Mu;

    .&dd for $defined, $nothing, $undefined;

    CATCH { default { note .message; .resume } }

# OUTPUT: Type check failed in assignment to $undefined; expected Option but got Mu (Mu)
#         Int $defined = 42
#         Label $nothing = Label.new(name => "None", file => "tmp/2021-03-08.raku", line => 1880)
#         Any $undefined = Any

Raku is a dynamic, late bound language. There are language constructs, like resumable exceptions and NativeCall, that can defeat any type-check. Trying to use Raku like Rust, Haskell or any other strictly typed language will eventually fail. No amount of :Ds will substitute testing, testing and even more testing.

I hope not to have tested p6steve’s patience and try to calm the waves with a new repo.

Categories: Raku


August 14, 2022 1 comment

On IRC deoac wished to know how to print the name of a variable. This question is ambiguous. To get the name of the container is easy.

my $foo;
say $foo.VAR.name;
# OUTPUT: $foo

If binding or constants involved, things get odd or fail.

my $bar := $foo;
constant $never-changes = 'war';
# OUTPUT: $foo
          No such method 'name' for invocant of type 'Str'.  Did you mean any of
          these: 'Date', 'Num', 'are', 'none', 'note', 'take'?

The second meaning of “variable” might be “symbol” for deoac’s use case. Getting hold of a symbol name requires early compile time measures.

use experimental :macros;
macro symbol-name($symbol) {
    quasi { $symbol.Str }

say symbol-name($foo);
say symbol-name($bar);
say symbol-name($never-changes);
# OUTPUT: ($foo)

The value of $symbol is an AST object and stringifies (in this case) to a list of Str. I didn’t dig deeper, because with RakuAST things will change.

However, this once again shows that we might do better by explaining what my actually does instead of assuming that one can guess what we mean when using the term “variable”.

Categories: Raku

Swarming Sundays

July 30, 2022 1 comment

To not miss any Sundays, PWC 175 is asking us to find them. Raku helps us a great deal, because we can easily find the last day in a month, calculate it’s distance to the next Sunday and turn the clock back by that many days.

for 2022 -> $year {
    for 1..12 -> $month {
        my $last-day-in-month = Date.new($year, $month, *);
        my $Δsunday = $last-day-in-month.day-of-week % 7;
        say $last-day-in-month.earlier(:days($Δsunday));

If you follow this blog, you will have spotted my distaste for simple loops. They get in the way of using interesting language features and thus fun. Date.new will only ever return a single date. With an Array of dates I could use vector operations to calculate all last Sundays in a month in one go. So I need a way to create a swarm of Dates, depending on a pattern. Signatures are patterns (and just fancy lists), so let’s use those.

sub swarm(::Type, @sig is copy, Range $range) {
    my $var;
    my $marker-index = @sig.first([], :k);
    @sig[$marker-index] := $var;
    gather for @$range {
        $var = $_;
        take Type.new: |@sig;

my @months = swarm(Date, (2022, @, *), 1..12);
my @Δsunday = @months».day-of-week »%» 7;
.say for @months.map({ .earlier(:days(@Δsunday.shift)) });

The sub swarm takes a type, a parameter-list meant to be used on that types’ new method and a range. It finds the empty list created by @ and replaced it with a container. That container is then filled with a value taken from the range. This is as neat as it is unnecessary. Raku will do the whole shebang for us if we use a Junction. Getting hold of the results requires a module by lizmat.

use eigenstates;

my @months = Date.new(2022, (1..12).any, *).&eigenstates;
my @Δsunday = @months».day-of-week »%» 7;
.say for @months.map({ .earlier(:days(@Δsunday.shift)) });

# OUTPUT: No such method 'List' for invocant of type 'BOOTArray'.  Found 'List'
on type 'Any'
  in sub eigenstates at /usr/local/src/rakudo/install/share/perl6/site/sources/EDAA4AE0C8813633A2EA392374FB72F6B4D61047 (eigenstates) line 4

Thank you MoarVM, for creating a nice BOOTArray for us that we can’t easily turn into a List. Some bewilderment and a PR later, the result changed. You may wish to zef upgrade if you use eigenstates.


Just displaying those Sundays wont justify getting rid of the loop. If we need that in an API, returning a list of Dates without much hassle seems valuable. Please note, that the BOOTArray indicates eagerness and I don’t think it has to be. Junctions might get lazier and faster in the future.

Categories: Raku

Coercive bits

July 13, 2022 1 comment

Altreus was looking for a way to convert a list of bitmask names, provided as a string by the user, to a bitmask. He wished for BUILDARGS, as provided by Perl’s Moose but was given good advise how to solve the problem without making object instantiation even more complex. Nobody can stop me from doing just that.

With Moose, BUILDARGS allows to modify values before they are bound to attributes of a class. We can do the same by using the COERCE-protocol, not just for Routine-arguments.

role BitMask[@names] {
    has int $.mask;
    sub names-to-bits(@a) {
        my int $mask;
        for @a -> $s {
            $mask = $mask +^ (1 +< @names.first($s, :k) // fail('bad‼'))

    sub bits-to-names($mask) {
        ($mask.base(2).comb.reverse Z @names).map: -> [Int() $is-it, $name] { $is-it ?? $name !! Empty }

    multi method COERCE(Str $s) { self.COERCE: $s.split(' ').list }
    multi method COERCE(List $l) { self.new: :mask($l.&names-to-bits) }

    method raku { "BitMask[<@names[]>](<{$!mask.&bits-to-names}>)" }
    method bits(--> Str) { $.mask.fmt('%b') }

class C {
    has BitMask[<ENODOC ENOSPEC LTA SEGV>]() $.attr is required;

my $c = C.new: :attr<ENODOC ENOSPEC SEGV>;
say $c;
say $c.attr.bits;
#         1011 

Here I build a role that carries the names of bits in a bit-field and will create the bit-field when give a Str or List containing those names. Since I use a parametrised role, my bitfield is type-safe and the type is tied to the actual names of the bits. As a consequence a user of the module that exports C can extend the accepted types by using that specific role-candidate.

enum IssueTypes ( ENODOC => 0b0001, ENOSPEC => 0b0010, LTA => 0b0100, SEGV => 0b1001 );

class IssueTypesBits does BitMask[<ENODOC ENOSPEC LTA SEGV>] {
    method new(*@a where .all ~~ IssueTypes) {
        self.bless: mask => [+|] @a

sub bitmask(*@a where .all ~~ IssueTypes) {
    IssueTypesBits.new: |@a

my $c3 = C.new: attr => BEGIN bitmask(ENODOC, LTA, SEGV);
say $c3;
say $c3.attr.bits;

Here I define an enum and provide the same names in the same order as in does BitMask. With the shim bitmask I create the type-safety bridge between IssueTypes and BitMask[<ENODOC ENOSPEC LTA SEGV>].

It is very pleasing how well the new COERCE-protocol ties into the rest of the language, because we can use a coercing-type at any place in the source code which takes a normal type as well. What is not pleasing are the LTA-messages X::Coerce::Impossible is producing.

my $c4 = C.new: attr => 42;
# OUTPUT: Impossible coercion from 'Int' into 'BitMask[List]': method new returned a type object NQPMu
            in block <unit> at parameterised-attribute.raku line 60

This is surpring because we can get hold of the candidates for COERCE at runtime.

    when X::Coerce::Impossible && .target-type.HOW === Metamodel::CurriedRoleHOW {
        put "Sadly, COERCE has failed for {.from-type.^name}. Available candidates are: ", .target-type.new.^lookup('COERCE').candidates».signature».params[*;1]».type.map(*.^name)

# OUTPUT: Sadly, COERCE has failed for Int. Available candidates are: Str List

What we can’t see are the candidates of IssueTypesBits, because that is hidden behind the constructor new. I tried to use the MOP to add another multi candidate but failed. When parametrising the underlying type-object gets cloned. Any change to the multi-chain wont propagate to any specialised types.

The COERCE-protocol is quite useful indeed. Maybe it’s time to document it.

Categories: Raku

The truth is a hard problem

July 11, 2022 2 comments

In a recent article, stevied promised a detailed walk through of code. I asked him if he would be interested in a critique of his writings. He foolishly agreed.

He claims that Arrays are eager and fill all memory if given the chance. Let’s check this claim.

my @a = 1, 1, * + * … ∞;
say @a.WHAT;

# OUTPUT: (Array)

Array.STORE goes out of it’s way to maintain iterables (not all things that are iterable do Iterable) and only reify container slots that we actively ask for. It will cache lazy lists by default, assuming that when we use .AT-POS (usually indirectly with a positional subscript) we want to use the values again. Dealing with lazy lists is a hard problem and many get it wrong (as my next blog post will discuss in more detail).

In the next paragraph I found another inaccurate statement. Int:D does not check for definedness.

multi sub bar(Any:D $) { say 'defined'; }
multi sub bar(Any:U $) { say 'undefined'; }

bar(Int); # undefined

class C { method defined { False } }
my $c = C.new; # undefined
bar(C); # undefined
bar($c); # defined
say defined $c; # False
say $c.DEFINITE; # True
say $c // 'undefined'; # undefined

In Raku objects are either type objects or definite objects. The latter are the result of nqp::create. Raku and most user code expects objects to provide a basic set of methods that can be found in Mu or Any. The default type-check is against Mu. We can check if an object is definite with the macro .DEFINITE (you can overload a method of the same name, but Rakudo will ignore it) or, if we also want to know if we got an ordinary object, with $foo ~~ Mu:D. There is a clear distinction of definedness and being definite. Raku needs to make that differentiation to support “unusual” values and types like Nil and Failure. Please note, that type-objects are not definite but can be defined. Types are a hard problem.

In the same paragraph stevied writes: “defined integer object”, even though he makes a type-check against Int.

multi sub foo(Int:D $i) { }
multi sub foo($i where { $i - $i.truncate == 0 }) { say 'lolwut‽' }
multi sub foo($) { say 'not integer'; }


# OUTPUT: lolwut‽
          not integer

Int:D will fail for any value that doesn’t got (Int) in its .^mro or is a subset of Int or any of Int‘s sub-classes. Raku sports an excellent coercion protocol and there is no reason not use it.

subset PositiveInteger of Numeric:D() where { $^i > 0 && ($i - $i.truncate == 0) || fail("Expected a positive and an integer value but got $i.")}
sub get-prime(PositiveInteger $nth where * > 0) {
    say ($x.grep: *.is-prime)[$nth - 1];
# OUTPUT: 11
          Expected a positive and an integer value but got 0.5.
            in block  at 2021-03-08.raku line 1809
            in block <unit> at 2021-03-08.raku line 1288

A value check is needed in this instance because .AT-POS will coerce to Int (by truncating) what may give a reasonable answer for an unreasonable question. No matter how sharp the knife, we wont get the ½th prime number. Being positive and integer is a hard problem.

Please don’t use say unless you actually want to run .gist.

say Nil;
put Nil;
          Use of Nil in string context
            in block  at 2021-03-08.raku line 1818

When you say you may miss subtile bugs, especially when running CI-tests.

Further down we find Sequence but grep returns a Seq – in this example. I never had to make the distinction between Seq and HyperSeq but the latter is not a child of Cool so a lot of interfaces are missing. The same is true for the role Sequence. The build-in types are a hard problem.

If we must ask what exactly * is (I wouldn’t, because that’s a really hard question.), it is best to provide the correct answer.

my &b = * + *;
say .WHAT, .signature with &b;
my &b2 = { $^a + $^b };
say .WHAT, .signature with &b2;
# OUTPUT: (WhateverCode)(;; $whatevercode_arg_11 is raw, $whatevercode_arg_12 is raw)
          (Block)($a, $b)

A WhateverCode-object will have all its arguments as is raw, what means it will always have access to the containers used by the callee (There are many things we can do, but probably shouldn’t, with .VAR). It will never have control exceptions (no return or phasers, etc.), doesn’t have a proper scope (no use-statement) and a few more things. The idea is to have a subclass of Code that is so simple that the compiler can always inline it. Code-objects are a hard problem.

I would have written that code a little different.

sub get-primes(*@nth) {
    (^∞).hyper.grep(-> Int:D() $_ is raw { .is-prime })[@nth »-» 1]

.put for get-primes(5, 50, 500, 5000, 50000);

I spend a pretty penny on that Threadripper, so I better .hyper as often as I can. If sensible I try to use a slurpy to move the burden of dealing with plurals from the user to the implementer. We can ask .AT-POS for a list of values, so there is no reason not to.

There are quite a few more inaccuracies in Steve’s article. Bless him, Raku is a bitch and we are not good at explaining how the language actually works. For most programs that doesn’t matter, they will run just fine even if the programmer is sloppy. I learned most of the gritty details when tracking down bugs. Granted, I stick my arm deep into the machinery, so it’s my own bloody fault when I get pinched. Since bugs happen we need to get better at explaining how to hunt them down. Raku is a hard problem.

Categories: Raku

Sinking Errors

June 26, 2022 1 comment

I was looking for a way to output debug messages that can also carry additional values, when not output to the screen. That is easy. The tricky part is golfing the interface. After quite a bit of struggle. I ended up with the following.

use Log;

dd ERROR('foo') ~~ LogLevel::ERROR;
dd ERROR('foo') ~~ LogLevel::DEBUG;
say ERROR('foo').&{ .file, .line };
VERBOSE 'Detailed, very much so indeed.';
my $*LOGLEVEL = 2;
VERBOSE 'Detailed, very much so indeed.';

# Bool::True
# Bool::False
# (2021-03-08.raku 1661)
# 2021-03-08.raku:1664 Detailed, very much so indeed.

In sink-context, the “functions” will output to $*ERR and when not sunk return an object that provides a few useful methods. To get both, we need to implement the methods sink and CALL-ME.

# Log.rakumod

use v6.d;

my role Functor {
    has $.message;
    my $.level;
    has $.captured-loglevel;
    has $.file;
    has $.line;
    method CALL-ME(*@message) {
        my ($line, $file) = callframe(1).&{ .line, .file };
        self.new: message => "$file:$line " ~ @message.join($?NL), :captured-loglevel($*LOGLEVEL // 0), :$line, :$file;
    method sink { $*ERR.put(self.message) if ($!captured-loglevel // 0) ≥ $.level }
    method Str(::?CLASS:D:) { self.message }
    method gist(::?CLASS:D:) { self.^shortname ~ ' ' ~ self.message }

my $ERROR;                                                                                                                                                                    my $VERBOSE;
my $DEBUG;

class LogLevel {
    constant ERROR := $ERROR;
    constant VERBOSE := $VERBOSE;
    constant DEBUG := $DEBUG;

$ERROR = class :: is LogLevel does Callable does Functor { my $.level = 1; }
$VERBOSE = class :: is LogLevel does Callable does Functor { my $.level = 2; }
$DEBUG = class :: is LogLevel does Callable does Functor { my $.level = 3; }

sub EXPORT {
    Map.new: '&ERROR' => $ERROR,
             '&VERBOSE' => $VERBOSE,
             '&DEBUG' => $DEBUG,
             'LogLevel' => LogLevel,

To behave like a function a Callable must be bound to an &-sigiled symbol (or Rakudo complains when we use it as a term). As soon as Rakudo spots a class, it will stuff it into UNIT::EXPORT::ALL. When importing any &-sigiled key of sub EXPORT, with the same basename as a class with that basename, the symbol from EXPORT will be ignored. The only way to avoid that is to have anonymous classes and do the forward-declarations by hand. The latter is needed to allow ERROR('foo') ~~ LogLevel.

All that is a bit convoluted but works as expected. What I didn’t expect, is that dynvars that are user-defined are not visible in method sink. Not much of a problem in my case, as capturing the state of $*LOGLEVEL inside CALL-ME is the right thing to do anyway. What threw me off is the inconsistency with other functions and methods that are called by a time-machine. We type BEGIN, BUILD and friends in all-caps for that very reason. This may warrant a problem solving issue. It’s an ENODOC for sure.

I believe golfing APIs is a most desireable effort. Laziness isn’t just a virtue for programmers, often it’s a given.

Categories: Raku

Reducing sets

May 25, 2022 1 comment

This week´s PWC asks us for hexadecimal words and distinctly different directories.

sub term:<pwc-166-1> {
    sub less-substitutions-then ($_, $n) {
        .subst(/<-[olist]>/, :g, '').chars < $n

        .grep({.chars ≤ 8 && .&less-substitutions-then(4)})\
        .map(*.trans(<o l i s t> => <0 1 1 5 7>))\
        .grep(/^ <[0..9 a..f A..F]>+ $/)\

Once again, we write the algorithm down. Get the words, drop anything longer then 8 chars or what would need more then 4 substitutions. Then do the substitutions and grep anything that looks like a hexadecimal numeral. Sort for good measure and output the first 100 elements.

The second task provides us with a little challenge. We need to mock listing directories and working with them. Since dir returns a sequence of IO::Path I can create those by hand and mixin a role that mimics some filesystem operations. I can overload dir to provide a drop-in replacement.

sub term:<pwc-166-2> {
    sub basename(IO::Path $_) { .basename ~ (.d ?? '/' !! '') }
    sub pad(Str $_, $width, $padding = ' ') { .Str ~ $padding x ($width - .chars) }

    sub dir(Str $name) {
        sub mock-file(*@names) { @names.map({ IO::Path.new($_) but role :: { method f ( --> True ) {}; method e ( --> True ) {} } } ) }
        sub mock-dir(*@names) { @names.map({ IO::Path.new($_) but role :: { method d ( --> True ) {}; method e ( --> True) {} } }) }

        constant %dirs = dir_a => flat(mock-file(<Arial.ttf Comic_Sans.ttf Georgia.ttf Helvetica.ttf Impact.otf Verdana.ttf>), mock-dir(<Old_Fonts>)),
                         dir_b => mock-file(<Arial.ttf Comic_Sans.ttf Courier_New.ttf Helvetica.ttf Impact.otf Tahoma.ttf Verdana.ttf>),
                         dir_c => mock-file(<Arial.ttf Courier_New.ttf Helvetica.ttf Impact.otf Monaco.ttf Verdana.ttf>);


    sub dir-diff(+@dirs) {
        my @content = @dirs».&dir».&basename;
        my @relevant = (([∪] @content) ∖ [∩] @content).keys.sort;

        my @columns = @content.map(-> @col { @relevant.map({ $_ ∈ @col ?? $_ !! '' }) });
        my $col-width = [max] @columns[*;*]».chars;

        put @dirs».&pad($col-width).join(' | ');
        put (''.&pad($col-width, '-') xx 3).join('-+-');
        .put for ([Z] @columns)».&pad($col-width)».join(' | ');

    dir-diff(<dir_a dir_b dir_c>);

I’m asked to add a / to directories and do so with a custom basename. The rest is liberal application of set theory. Only names that don’t show up in all directories are relevant. Columns are created by matching the content of each directory against the relevant names. The width of columns is the longest string. The header is put on screen. To output the columns line by line, x and y are flipped with [Z].

After careful study of the solutions written in other languages, I believe it is fair to call Raku an eco-friendly language. Our keyboards are going to last at least twice a long.

Categories: Raku, Uncategorized

Writing it down

May 21, 2022 2 comments

PWC 165 refers us to mathsisfun for the algorithm to be used. Let’s write it down.

    my $input = '333,129  39,189 140,156 292,134 393,52  160,166 362,122  13,193
                341,104 320,113 109,177 203,152 343,100 225,110  23,186 282,102
                284,98  205,133 297,114 292,126 339,112 327,79  253,136  61,169
                128,176 346,72  316,103 124,162  65,181 159,137 212,116 337,86
                215,136 153,137 390,104 100,180  76,188  77,181  69,195  92,186
                275,96  250,147  34,174 213,134 186,129 189,154 361,82  363,89';

    my @points = $input.words».split(',')».Int;

    my \term:<x²> := @points[*;0]».&(*²);
    my \xy = @points[*;0] »*« @points[*;1];
    my \Σx = [+] @points[*;0];
    my \Σy = [+] @points[*;1];
    my \term:<Σx²> = [+] x²;
    my \Σxy = [+] xy;
    my \N = +@points;

    my $m = (N * Σxy - Σx * Σy) / (N * Σx² - (Σx)²);
    my $b = (Σy - $m * Σx) / N;

    say [$m, $b];

That was exceedingly simple. If the point cloud is large, this might also be exceedingly slow. There are no side effects and everything is nice and constant. Should be easy to get the CPU busy.

    sub delayed(&c) is rw {
        my $p = start { c };
        Proxy.new: STORE => method (|) {}, FETCH => method { await $p; $p.result }

    my \term:<x²> = delayed { @points[*;0]».&(*²) };
    my \xy = delayed { @points[*;0] »*« @points[*;1] };
    my \Σx = delayed { [+] @points[*;0] };
    my \Σy = delayed { [+] @points[*;1] }
    my \term:<Σx²> = delayed { [+] x² };
    my \Σxy = delayed { [+] xy };
    my \N = +@points;

    my $m = (N * Σxy - Σx * Σy) / (N * Σx² - (Σx)²);
    my $b = (Σy - $m * Σx) / N;

    say [$m, $b];

Since I use sigil-less symbols, I get binding for free. This in turn makes using Proxy easy. The await $p part is called more then once per Proxy but not in a loop so there is no real need to optimise here. On my box the Proxyed version is almost twice as fast. A nice win for such a simple change.

I somehow feel this solution might please a mathematician — as Raku should.

Categories: Raku