← All programs · Grammars and match trees

Grammars and match trees #13

The program

The grammar and its inputs come first. rm-dump, the same in every program of this topic, prints one line for every field of every node of the resulting Match — its text, position, .orig, .prematch, .made, its captures — so a difference in the shape of the match shows as a difference in output.

my token r0 { \w }
my token r1 { <r0> \w }
my rule r2 { \w+ }
sub rm-q($v) { $v.defined ?? $v.raku !! 'Nil' }
sub rm-dump($m, Str $p) {
    unless $m.defined { say "$p kind Nil"; return }
    if $m ~~ Match {
        say "$p kind Match";
        say "$p str {$m.Str.raku}";
        say "$p from {$m.from}";
        say "$p to {$m.to}";
        say "$p orig {rm-q($m.orig)}";
        say "$p target {rm-q($m.target)}";
        say "$p pre {rm-q($m.prematch)}";
        say "$p post {rm-q($m.postmatch)}";
        say "$p made {rm-q($m.made)}";
        for $m.list.kv -> $i, $c { rm-dump($c, $p ~ '[' ~ $i ~ ']') }
        for $m.hash.keys.sort -> $k { rm-dump($m.hash{$k}, $p ~ '<' ~ $k ~ '>') }
        # Captures that start at the same position have no specified order
        # (rakupp's varies from run to run), so ties are ordered by name.
        say "$p caps {$m.caps.sort({ .value.from, ~.key }).map({ .key ~ '=' ~ .value.Str.raku }).join(' ')}";
    }
    elsif $m ~~ Positional {
        say "$p kind List({$m.elems})";
        for $m.list.kv -> $i, $c { rm-dump($c, $p ~ '.' ~ $i) }
    }
    else { say "$p kind {$m.^name}" }
}
my @inputs = "b", ")223bc>", "";
for @inputs.kv -> $i, $in {
    say "#$i input {$in.raku}";
    if $in ~~ / 'b' | ( <r1>+? ) | [ ')' <&r0> || ')' <r1> || ')' <r1> ] \d? <r0> ** 2 / { rm-dump($/, '@' ~ $i) }
    else { say '@' ~ $i ~ ' kind Nil' }
}

Run executes the program in your browser, with the Raku++ build this site ships. Edit it and try variations.

What each compiler printed

grammars-v1 · seed 1002 · recorded 2026-09-23 with Rakudo 2026.08 and Raku++ 4.0.1-132-g3e242220-modified · Source on GitHub ↗