Re: [Tiki-devel] A Unicode-aware replacement for the natcasesort() function

Volker Wysk <post-hhF2Jplw28UoZk/[email protected]> Sat, 08 Jul 2023 15:53:57 +0200

Newsgroups	gmane.comp.cms.tiki.devel
Message-ID	<[email protected]>

Hi.

Here's the next, probably final, version of the mb_natcasesort() function.
It should be placed somewhere in the Tiki libraries. I'm not sure where.

There's no mb_natsort() function, because natural languages don't order
strings that way. See the comment at the head of the code.

See the attachment.

Volker

_______________________________________________
TikiWiki-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel

sort-nat-loc.php (application/x-php, 2.3 KB)

<?php

// Do a natural, locale aware and case-insenstitive sort of an array of strings.
//
// "Natural" means that numbers inside the strings are sorted by their value, not by the digits they contain. For
// instance, "10" is ordered after "2".
//
// When the "intl" PHP extension is available, the sort is done Unicode aware. The "$lang" argument is the language
// for which to order. Accented characters, like "ä", are ordered like their non-accented counterpart, like "a".
// For the "$lang" argument, "false" can be specified. In this case, the function investigates the language of the
// current user and uses that.
//
// When intl isn't available, mb_natcasesort() function falls back to the the natcasesort() function. This one
// isn't Unicode aware. Accented characters get ordered at the end. The "$lang" argument is ignored in this case.
//
// In case of success, true is returned. In case of an error, false is returned. This are the return values of the
// collator_asort() function from the intl extension. When the "intl" extension isn't available, the result always
// is "true".
//
// The natsort() function orders all capital letters before all lower-case letters. There's no Unicode aware
// counterpart, because natural languages appear to always order case-insensitively. The Collator class is like
// this. So we can't have a mb_natsort() function.

function mb_natcasesort(string $lang, array &$array): bool
{
    if (extension_loaded("intl")) {

        if ($lang == false) {
            $tikilib  = TikiLib::lib('tiki');
            $loginlib = TikiLib::lib('login');

            $user = $loginlib->getUser();
            $lang = $tikilib->get_language($user);
        }

        $coll = collator_create($lang);
        collator_set_attribute($coll, Collator::NUMERIC_COLLATION, Collator::ON);
        collator_set_attribute($coll, Collator::CASE_FIRST, Collator::LOWER_FIRST);
        collator_set_attribute($coll, Collator::ALTERNATE_HANDLING, Collator::SHIFTED);

        return collator_asort($coll, $array);

    } else {
        natcasesort($array); // always returns true
        return true;
    }
}



// Test

$arr = array("BBB", "bbb", "ß", "abb", "AAA", "äba", "xx100xx", "xx99xx", '"abc"', '_abc');

mb_natcasesort("de_DE", $arr);
//mb_natcasesort(false, $arr);

print_r($arr);

signature.asc (application/pgp-signature, 833 B)