Replace iconv's transliteration with another transliteration library about detox-php

[dharple]

Pleased to present for your consideration:

# cat test.php
<?php

require_once 'vendor/autoload.php';

printf("%s %s\n", ICONV_IMPL, ICONV_VERSION);

$str = 'Sláinte';

for ($i = 0; $i < mb_strlen($str); $i++) {
    printf("%s -> %s\n",
        mb_substr($str, $i, 1),
        iconv('UTF-8', 'ASCII//TRANSLIT', mb_substr($str, $i, 1))
    );
}
# php7.4 -f test.php
glibc 2.31
S -> S
l -> l
á -> a
i -> i
n -> n
t -> t
e -> e
# php8.0 -f test.php
glibc 2.31
S -> S
l -> l
á -> ?
i -> i
n -> n
t -> t
e -> e
# LANG=en_US.UTF-8 php8.0 -f test.php
glibc 2.31
S -> S
l -> l
á -> ?
i -> i
n -> n
t -> t
e -> e

However:

# cat test-force-utf8.php
<?php

require_once 'vendor/autoload.php';

printf("%s %s\n", ICONV_IMPL, ICONV_VERSION);

setlocale(LC_ALL, 'en_US.UTF-8');

$str = 'Sláinte';

for ($i = 0; $i < mb_strlen($str); $i++) {
    printf("%s -> %s\n",
        mb_substr($str, $i, 1),
        iconv('UTF-8', 'ASCII//TRANSLIT', mb_substr($str, $i, 1))
    );
}
# php8.0 -f test-force-utf8.php
glibc 2.31
S -> S
l -> l
á -> a
i -> i
n -> n
t -> t
e -> e

[dharple]

Essentially - the behavior of iconv is unpredictable.

[dharple]

My solution:

# cat test-behat.php
<?php

require_once 'vendor/autoload.php';

use Behat\Transliterator\Transliterator;

$str = 'Sláinte';

for ($i = 0; $i < mb_strlen($str); $i++) {
    printf("%s -> %s vs %s\n",
        mb_substr($str, $i, 1),
        iconv('UTF-8', 'ASCII//TRANSLIT', mb_substr($str, $i, 1)),
        Transliterator::utf8toAscii(mb_substr($str, $i, 1)),
    );
}
# php7.4  -f test-behat.php
S -> S vs S
l -> l vs l
á -> a vs a
i -> i vs i
n -> n vs n
t -> t vs t
e -> e vs e
# php8.0 -f test-behat.php
S -> S vs S
l -> l vs l
á -> ? vs a
i -> i vs i
n -> n vs n
t -> t vs t
e -> e vs e