We definitely need an RFC for that. But besides that, it looks good.

I'm happy to write one. My understanding was for simple functions like this it wasn't necessary.

about the function name exposed , I think it will be better to be consistent with is_string, is_int etc.
http://cn2.php.net/manual/en/function.is-string.php

so, maybe is_ascii? os is_ascii_string?

string_is_ascii probably not the best name. We don't have consistent prefixes for string functions, unfortunately, but we have none starting with string_. We have str_something, we have strsomething, we have is_something, but no string_something. I think adding yet another prefix is not a good idea.

Otherwise, the code looks good.

No, the comparison with is_long, is_string, ... Doesn't apply as they compare with types. This function checks the actual value.

Good point about the name. I've renamed the user facing function to str_is_ascii. I left the internal function untouched because there are a few functions that start with string_ in the same file (e.g. string_natural_compare_function) I gust got RFC karma, I'll try to write it up tonight.

"mb_check_encoding($str, 'ASCII')" and "ctype_print($str) || ctype_cntrl($str)" are enough solutions. Symfony 2 spefify ctype functions for the requirements. Defining the fallback function for mb_check_encoding is easy since htmlspecialchars has an ability for validating various encodings (https://github.com/php/php-src/blob/master/ext/standard/html.c#L105). Considering PHP's history and existing projects, standardizing mb_check_encoding seems to be a good topic. RFC: Alternative implementation of mbstring using ICU is good stuff for that purpose.

if(!function_exists('mb_check_encoding')) {

    function mb_check_encoding($str, $encoding = 'UTF-8')
    {
        if ('ASCII' === $encoding) {
            return ctype_print($str) || ctype_cntrl($str);
        }

        $list = ['BIG-5' => 'BIG5', 'CP936' => 'GB2312'];

        if (array_key_exists($encoding, $list)) {

            $encoding = $list[$encoding];   

        }

        $str === htmlspecialchars_decode(htmlspecialchars($str, ENT_COMPAT, $encoding));
    }
}

function is_ascii($str)
{
    return mb_check_encoding($str, 'ASCII');
}

function is_utf8($str)
{
    return mb_check_encoding($str, 'UTF-8');
}

What's the state of this? As @masakielastic already pointed out str_is_ascii($str) is equivalent to mb_check_encoding($str, 'ASCII'). Does this mean that this PR can be closed or are there additional reasons to support this function?

Thanks for this PR. However, I think we don't need a specialized function as mb_ can already handle this. Therefore I am closing this PR.