From @smls

When a quantified capture group matched zero times but the overall regex matches, then the List returned from `$/.list` stops iterating before the corresponding element​:

  "ac" ~~ / (a) (b)? (c) /;
  say ($0, $1, $2); # (｢a｣ Nil ｢c｣)
  say $/.list; # (｢a｣)
  say $/.list.map({1}); # (1)
  say $/.list.elems; # 3
  say $/.list[0, 1, 2, 3]; # (｢a｣ (Mu) ｢c｣ Nil

Expected behavior would be for `$/.list;` to return `(｢a｣ Nil ｢c｣)`.

mst++ suggested that maybe an InterationEnd is ending up where it shouldn't be.

From @lizmat

I’m looking into this. This appears to be an artefact of .gist on the list, caused by the say. Which also happens for .perl on the list.

say $/.list.Str; # a c

Suspecting the gistseen/perlseen machinery now. Investigating...

On 18 Sep 2016, at 19​:30, Sam S. (via RT) <perl6-bugs-followup@​perl.org> wrote​:

# New Ticket Created by Sam S.
# Please include the string​: [perl #​129299]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=129299 >

When a quantified capture group matched zero times but the overall regex matches, then the List returned from `$/.list` stops iterating before the corresponding element​:

"ac" ~~ / (a) (b)? (c) /;
say ($0, $1, $2); # (｢a｣ Nil ｢c｣)
say $/.list; # (｢a｣)
say $/.list.map({1}); # (1)
say $/.list.elems; # 3
say $/.list[0, 1, 2, 3]; # (｢a｣ (Mu) ｢c｣ Nil

Expected behavior would be for `$/.list;` to return `(｢a｣ Nil ｢c｣)`.

mst++ suggested that maybe an InterationEnd is ending up where it shouldn't be.

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

From @lizmat

Ah, found it​:

The List iterator assumes that a List cannot have “holes” in it. Apparently it can :-(

Going to sleep on how to fix this.

On 18 Sep 2016, at 19​:30, Sam S. (via RT) <perl6-bugs-followup@​perl.org> wrote​:

# New Ticket Created by Sam S.
# Please include the string​: [perl #​129299]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=129299 >

When a quantified capture group matched zero times but the overall regex matches, then the List returned from `$/.list` stops iterating before the corresponding element​:

"ac" ~~ / (a) (b)? (c) /;
say ($0, $1, $2); # (｢a｣ Nil ｢c｣)
say $/.list; # (｢a｣)
say $/.list.map({1}); # (1)
say $/.list.elems; # 3
say $/.list[0, 1, 2, 3]; # (｢a｣ (Mu) ｢c｣ Nil

Expected behavior would be for `$/.list;` to return `(｢a｣ Nil ｢c｣)`.

mst++ suggested that maybe an InterationEnd is ending up where it shouldn't be.

From @lizmat

Fixed (for now at least) with 9b6f2eb5434e71f07bc649 . Tests still needed.

On 18 Sep 2016, at 19​:30, Sam S. (via RT) <perl6-bugs-followup@​perl.org> wrote​:

# New Ticket Created by Sam S.
# Please include the string​: [perl #​129299]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=129299 >

When a quantified capture group matched zero times but the overall regex matches, then the List returned from `$/.list` stops iterating before the corresponding element​:

"ac" ~~ / (a) (b)? (c) /;
say ($0, $1, $2); # (｢a｣ Nil ｢c｣)
say $/.list; # (｢a｣)
say $/.list.map({1}); # (1)
say $/.list.elems; # 3
say $/.list[0, 1, 2, 3]; # (｢a｣ (Mu) ｢c｣ Nil

Expected behavior would be for `$/.list;` to return `(｢a｣ Nil ｢c｣)`.

mst++ suggested that maybe an InterationEnd is ending up where it shouldn't be.

From @zoffixznet

Tests added in Raku/roast@581b553b74

@zoffixznet - Status changed from 'open' to 'resolved'

From @smls

Should it really return `(｢a｣ (Mu) ｢c｣)` and not `(｢a｣ Nil ｢c｣)`, seeing as $1 is also Nil?

liz made it work on the List.iterator end by allowing it to iterate over holes in the list, but maybe Match.list should not produce a list with holes at all, but rather a list with Nil in place of missing captures?
(Maybe that's what the "for now at least" refers to.)

If so, then the `is $two.^name, 'Mu'` test added by zoffix is LTA.
Maybe it should just test `nok $two.defined` for now, until it is cleared up what type exactly it should be...

From @AlexDaniel

Should we reopen this then?

@AlexDaniel - Status changed from 'resolved' to 'open'