# New Ticket Created by Stephane Payrard
# Please include the string: [perl #111288]
# in the subject line of all future correspondence about this issue.
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=111288 >
$ perl6
> (grammar { token TOP { <a>? $<b>='b' }; token a { a } }).parse: 'ab'; say
> $/<b>.keys;
a
> (grammar { token TOP { <a> $<b>='b' }; token a { a } }).parse: 'ab'; say
> $/<b>.keys;
>
In the first line $/<b>.keys is not empty contrary to expectations.
Somehow the results of the <a>?
rule leaks into $/<b>.
Note that in the second line, the question mark in <a>? is suppressed
and the bug is not triggered any more.
Maybe the bug will be fixed when the implementation of ? will be
rewritten as per spec (not returning an array).
<cognominal> nom: (grammar { token TOP { <a>? $<b>='b' }; token a {
a } }).parse: 'ab'; say $/<b>.keys;
<p6eval> nom 4130f6: OUTPUT«a»
<cognominal> nom: (grammar { token TOP { <a> $<b>='b' }; token a {
a } }).parse: 'ab'; say $/<b>.keys;
<p6eval> nom 4130f6: OUTPUT«»
<cognominal> jnthn, this does not make any sense to me that $/<b>.keys
returns anything different than a void Parcel
<cognominal> should I file a bug?
* NamelessTee has quit (Ping timeout: 244 seconds)
<moritz> nom: (grammar { token TOP { <a> $<b>='b' }; token { a }
}).parse: 'ab'; say $/<b>.keys.WHAT;
<p6eval> nom 4130f6: OUTPUT«Method 'a' not found for invocant of class
'<anon>' in regex TOP at /tmp/fZXy8AXoHU:1 in method parse at
src/gen/CORE.setting:8019 in block <anon> at /tmp/fZXy8AXoHU:1»
<moritz> nom: (grammar { token TOP { <a> $<b>='b' }; token a { a }
}).parse: 'ab'; say $/<b>.keys.WHAT;
<p6eval> nom 4130f6: OUTPUT«List()»
<moritz> cognominal: it's returns an empty List, not an empty Parcel.
Is that your problem?
<cognominal> see the first example, it prints "a\n".
<cognominal> I don't care if it is a parcel or a list. Whatever it is
should be empty.
<moritz> agreed
<cognominal> it does not happen when I delete the '?' suffix.
<moritz> though of course the ? quantifier on <a> should not affect $<b>
<cognominal> yes, that's my point.
<moritz> agreed. Bug it is.
--
cognominal stef
