# New Ticket Created by Lars Dɪᴇᴄᴋᴏᴡ 迪拉斯
# Please include the string: [perl #132004]
# in the subject line of all future correspondence about this issue.
# <URL: https://rt.perl.org/Ticket/Display.html?id=132004 >
› perl6 -v
This is Rakudo version 2017.07 built on MoarVM version 2017.07
implementing Perl 6.c.
› cat flail.pl
use v6;
grammar Flail {
rule TOP { <B> }
rule B { <A> 'x' 'y' | <C> }
rule A { '' | 'x' 'z' }
rule C { <C> 'w' | 'v' }
}
Flail.parse('x z x y').say;
Flail.parse('v w w w w w w').say;
› perl6 flail.pl
｢x z x y｣
B => ｢x z x y｣
A => ｢x z ｣
^C
The execution simply goes into infinite loop with the second input.
This is the least awesome failure mode.
Compare with Pegex and Regexp::Grammars, which emit a recursion
warning/error. Compare with Antlr4, where the grammar is effectively a
compile time error. Compare with Eyapp and Marpa, which can consume
this grammar and set of inputs without any problem.
I realise that the parser cannot be simply changed since it also parses
Perl6 source code. But since it is now shown that the parser is not a
general parser able to consume arbitrary valid CFG, then at least the
documentation must mention somewhere
* what the name/classification of the underlying technology/algorithm
of the parser is and
* what its shortcomings are (with examples)
so that a user can make an informed decision.
