$Ideographic = '(?:\xE3\x80[\x87\xA1-\xA9]|\xE4(?:[\xB8-\xBF][\x80-\xB
+F])|\xE5(?:[\x80-\xBF][\x80-\xBF])|\xE6(?:[\x80-\xBF][\x80-\xBF])|\xE
+7(?:[\x80-\xBF][\x80-\xBF])|\xE8(?:[\x80-\xBF][\x80-\xBF])|\xE9(?:[\x
+80-\xBD][\x80-\xBF]|\xBE[\x80-\xA5]))';
[download]

However, my test data are normal ASCII range characters, and that doesn't succeed either, though it starts $BaseChar = '(?:[a-zA-Z]|\xC3[\x80-\x9.... So I don't know everything that's wrong with it, but it doesn't work at all (see below) if utf8 is used.

use strict;
use warnings;
use utf8;  # comment this line out and it matches
use XML::RegExp;

my $name= 'timestamp';  # contains plain ASCII letters only!
my $result= $name =~ /^$XML::RegExp::Name$/o;
print "result is $result\n";
[download]

This looks weird.

XML::RegExp was released way before 5.6.0 though, so it is not that surprising that it does not deal that well with the utf8 pragma. Do you really need it around if you only deal with US ASCII characters?

Incidently the definition of Name includes regular characters:

$Letter    =     "(?:$BaseChar|$Ideographic)";
$NameChar    = "(?:[-._:]|$Letter|$Digit|$CombiningChar|$Extender)";
$Name        = "(?:(?:[:_]|$Letter)$NameChar*)";
[download]

I know the $BaseChar includes regular characters. So the byte/char mode thing doesn't explain why it doesn't work at all, when the name to match is pure ASCII.

I'm not dealing only with US ASCII. I'm reading UTF-8. This example was a simple case showing it fail.

—John