dies with unicode bare key name

[plicease]

First reported with to Perl::Critic, since I wasn't sure if the problem was there or here. However, it seems to be reproducible with ppidump (see Perl-Critic/Perl-Critic#948 (comment) ), so I believe this is indeed a PPI issue.

code:

use utf8;
my %x = (Привет => 1);

my($key) = keys %x;
die unless $key eq 'Привет';

perlcritic:

% perlcritic .
Problem while critiquing "foo.pl": Can't parse code: Encountered unexpected character '208'

It does work when I change it to 'Привет' => 1)

I admit to not being 100% certain that you can use a bare unicode key name like this, but the code does seem to work as intended. If so a more useful diagnostic would indicate where the character is located, if that is possible.

[oalders]

Some notes on unicode support: https://metacpan.org/pod/PPI#Internationalisation

[ernstki]

I'm seeing the same behavior (at least I think I'm seeing the same behavior) with something like:

warn "Converted user name to submission ID → $subid\n";

I think this leaves me with a choice of not to use perlcritic or not to use Unicode characters, which is kind of a bummer either way.

Update: Aha, no it was a parsing error where the missing closing quote ran up to a Unicode ellipsis character (…) before finding another matching quote, not the "→" I mentioned above.

I wouldn't know offhand what to do with Encountered unexpected character '226'. I suppose now I will recognize that as 0xE2, the start of the Unicode representation for "…", but it's not something that will turn up in od -a or hexdump -C, unless I'm just using the tools wrong.

[oalders]

If you think we can provide better error messaging, a PR would be most welcome.