Samemark breaks regional indicator symbols, should it be this way?

[AlexDaniel]

Code:

say ‘🇦🇬’.samemark(‘a’)

Output:

🇦

I don't think samemark functionality should apply to flags, this looks like a bug to me. But maybe it's not, considering that often people use samemark to get a string with single separate codepoints without prepend/append characters.

[samcv]

.samemark is currently broken. it assumes the first codepoint of a grapheme is the "base", and anything occuring afterward is a "mark".
(below the * is used in the same meaning as regex, and things in brackets are single codepoint tokens)

How samemark assumes the world exists
[base] [mark]*

The real world it is actually more like
[mark]*[base][mark]*

But then you have emoji, flags etc.
[Regional Indicator][Regional Indicator]
or
[Base Emoji]([ZWJ][Emoji])*

What I think is the most reasonable solution is to treat Grapheme_Cluster_Break=Extend and Grapheme_Cluster_Break=Prepend as “mark”’s. This would cause it to treat flags or emoji sequences as the "base", since it's not really separable the same way accent marks are separable.

This does not address the fact that "\c[Canada]".samemark("é") would give you a Canadian flag with an accent mark on it, and why anybody would desire this. As this is how samemark has always worked it doesn't make sense to change this.

[samcv]

Here is an example of what I am proposing:

use Test;
is 'hi'.samemark('é'), 'h́í', 'plain base';
is 'ēi'.samemark('é'), 'éí', 'mark nonmark base';
is ‘🇦🇬’.samemark('é').ords, '127462 127468 769';

# Prepend + Other will add the extend
is "\c[arabic number sign]a".samemark('é').NFD.list, '97 769', 'prepend test';
is "a".samemark("\c[arabic number sign]é").NFD.list,
  "\c[arabic number sign]".ord ~ ' 97 769', 'prepend test';
done-testing;