From @finanalyst

See the following (svn revision 34915)
$ perl6

regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<digit>/ and
say 'got digit';$y~~/<!digit>/ and say 'not got digit';
got digit
not got digit
regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<!digit>/ and
say 'got digit';$y~~/<!digit>/ and say 'not got digit';
got digit
not got digit
regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<!digit>/ and
say 'got digit';$y~~/<digit>/ and say 'not got digit';
got digit

The assumption was <!digit> is the logical inverse of <digit>

Either way, /<!digit>/ is returning true with and without a digit ( $x
test in first and third lines both give 'got digit').

From @pmichaud

There seems to be a misunderstanding about what <!digit> does here --
it's a zero-width assertion that means "doesn't match a digit at this
point". Since your regexes are all unanchored, this assertion can be
true at many points (including the end of the string). To look at the
specific examples you give​:

On Tue Jan 06 05​:06​:46 2009, richardh wrote​:

See the following (svn revision 34915)
$ perl6

regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<digit>/ and
say 'got digit';$y~~/<!digit>/ and say 'not got digit';
got digit
not got digit

The $y match (the one calling <!digit>) succeeds at position 0, as it
should. The $x match succeeds at position 3.

regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<!digit>/ and
say 'got digit';$y~~/<!digit>/ and say 'not got digit';
got digit
not got digit

The $x match succeeds at position 0; the $y match succeeds at position 0.

regex digit { [0..9] }; my $x='abc9';my $y='def';$x~~/<!digit>/ and
say 'got digit';$y~~/<digit>/ and say 'not got digit';
got digit

The $x match succeeds at position 0. (The $y match fails.)

The assumption was <!digit> is the logical inverse of <digit>

This assumption is false -- <!digit> is the inverse of <?digit> .

I suspect what you're looking for is <-digit>, which means "match any
character that doesn't match <digit>". You'll still want to anchor the
pattern somehow, though, because "<-digit>" would've matched each of the
above cases also.

Finally, note that

  regex digit { [0..9] }

means "match a '0' followed by two characters followed by a '9'".
I suspect you meant to write

  regex digit { <[0..9]> }

So we end up with​:

regex digit { <[0..9]> }; say ("abc" ~~ /^ <-digit>* $/) ?? 'matched'
!! 'not matched'
matched
regex digit { <[0..9]> }; say ("abc9" ~~ /^ <-digit>* $/) ??
'matched' !! 'not matched'
not matched

Closing ticket,

Pm

The RT System itself - Status changed from 'new' to 'open'

@pmichaud - Status changed from 'open' to 'rejected'