#!/usr/bin/perl
use warnings;
use strict;

use Unicode::Normalize qw{ normalize };

my $char = "\N{LATIN SMALL LETTER U WITH DIAERESIS}";
binmode *STDOUT, ':encoding(UTF-8)';
print normalize($_, $char), ' ' for qw( D C );
[download]

Running the output through xxd:

00000000: 75cc 8820 c3bc 20                        u.. ..
[download]

But my expectation is that those modifiers don't count as character and have length 0, i.e. "Hütte" should have length 5 in both incarnations.

> how length() is implemented?

I may be wrong tho...

#!/usr/bin/perl
use strict;
use feature qw{ say };
use warnings;

use Unicode::Normalize qw{ normalize };
use Unicode::GCString;

my $char = "\N{LATIN SMALL LETTER U WITH DIAERESIS}";
binmode *STDOUT, ':encoding(UTF-8)';
for (qw( D C )) {
    my $n = normalize($_, $char);
    my $gcs = 'Unicode::GCString'->new($n);
    say join ' ', length($n),
                  $n =~ s/(\X)/$1/g,
                  $1,
                  $gcs->chars,
                  $gcs->columns,
                  $gcs->length;
}
[download]

2 1 ü 2 1 1
1 1 ü 1 1 1

Update: Added the output.

map{substr$_->[0],$_->[1]||0,1}[\*||{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^*ARGV,3]

Interesting, looks like code.

I might even be able to install those modules and try to understand the output you didn't provide (yet)!

;-P

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

> I may be wrong tho...

I certainly am...

#!/usr/bin/perl
use v5.12;
use strict;
use utf8;
use Devel::Peek;

my $trema = "\N{COMBINING DIAERESIS}";
binmode *STDOUT, ':encoding(UTF-8)';
my $huette = "Hu${trema}tte";

Dump $huette;

say "$huette\'s length: ". length($huette);
[download]

SV = PV(0x25f4a58) at 0x25266b8
  REFCNT = 1
  FLAGS = (POK,pPOK,UTF8)
  PV = 0x28da368 "Hu\314\210tte"\0 [UTF8 "Hu\x{308}tte"]
  CUR = 7
  LEN = 10
Hütte's length: 6
[download]

That's how it looks like without codetags:

Hütte's length: 6

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}

Interestingly it's possible to combine ü + ¨ to pile up tremas

Hü̈tte Hü̈̈tte

I see huts with smoking chimneys ;-)

There's also ű in Hungarian (called "double acute" in Unicode).
map{substr$_->[0],$_->[1]||0,1}[\*||{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^*ARGV,3]

Using two bars instead of dots is acceptable in German, there are fonts which realize the umlauts ä,ö,ü this way. (see also this )

That's because of the weird form of Kurrent's small e being written above the vowels.

Apparently this is also connected to the history of the Czech letter Ů ů with an o superscript.

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}