UTF8-C8 string changes under concatenation

[samcv]

use Test;
sub decode_utf8c {
    my @ints = 103, 248, 111, 217, 210, 97;
    my $b = Buf.new(@ints);
    my Str $u=$b.decode("utf8-c8");
    $u.=subst("a","b");
}
#?rakudo.jvm todo "Unknown encoding 'utf8-c8' RT #127878"
lives-ok &decode_utf8c, 'RT #127878: Can decode and work with interesting byte sequences';

This is a test in t/spec/integration/weird-errors.t (roast).

Or more simply, replacing a's with a's so the string should not be changed:

my @ints = 103, 248, 111, 217, 210, 97; my $b = Buf.new(@ints); my Str $u=$b.decode("utf8-c8"); say $u.chars; $u.=subst("a","a"); say $u.chars
#→  OUTPUT: «6␤9␤»

When re_nfg() in src/strings/ops.c gets called, it iterates by codepoint into the normalizer, losing the UTF8-c8 representation.

I discovered this by setting MVM_DEBUG_NFG

[samcv]

This has been closed with this commit 2f71945d

Closing :-)