PHP: Convert any string to UTF-8 without knowing the original character set, or at least try

Question

What you're asking for is extremely hard. If possible, getting the user to specify the encoding is the best. Preventing an attack shouldn't be much easier or harder that way.

However, you could try doing this:

iconv(mb_detect_encoding($text, mb_detect_order(), true), "UTF-8", $text);

Setting it to strict might help you get a better result.

In motherland Russia we have 4 popular encodings, so your question is in great demand here.

Only by char codes of symbols you can not detect encoding, because code pages intersect. Some codepages in different languages have even full intersection. So, we need another approach.

The only way to work with unknown encodings is working with probabilities. So, we do not want to answer the question "what is encoding of this text?", we are trying to understand "what is most likely encoding of this text?".

One guy here in popular Russian tech blog invented this approach:

Build the probability range of char codes in every encoding you want to support. You can build it using some big texts in your language (e.g. some fiction, use Shakespeare for english and Tolstoy for russian, lol ). You will get smth like this:

    encoding_1:
    190 => 0.095249209893009,
    222 => 0.095249209893009,
    ...
    encoding_2:
    239 => 0.095249209893009,
    207 => 0.095249209893009,
    ...
    encoding_N:
    charcode => probabilty

Next. You take text in unknown encoding and for every encoding in your "probability dictionary" you search for frequency of every symbol in unknown-encoded text. Sum probabilities of symbols. Encoding with bigger rating is likely the winner. Better results for bigger texts.

If you are interested, I can gladly help you with this task. We can greatly increase the accuracy by building two-charcodes probabilty list.

Btw. mb_detect_encoding certanly does not work. Yes, at all. Please, take a look of mb_detect_encoding source code in "ext/mbstring/libmbfl/mbfl/mbfl_ident.c".

You've probably tried this to but why not just use the mb_convert_encoding function? It will attempt to auto-detect char set of the text provided or you can pass it a list.

Also, I tried to run:

$text = "fiancée";
echo mb_convert_encoding($text, "UTF-8");
echo "<br/><br/>";
echo iconv(mb_detect_encoding($text), "UTF-8", $text);

and the results are the same for both. How do you see that your text is truncated to 'fianc'? is it in the DB or in a browser?

There is no way to identify the charset of a string that is completely accurate. There are ways to try to guess the charset. One of these ways, and probably/currently the best in PHP, is mb_detect_encoding(). This will scan your string and look for occurrences of stuff unique to certain charsets. Depending on your string, there may not be such distinguishable occurrences.

Take the ISO-8859-1 charset vs ISO-8859-15 ( http://en.wikipedia.org/wiki/ISO/IEC_8859-15#Changes_from_ISO-8859-1 )

There's only a handful of different characters, and to make it worse, they're represented by the same bytes. There is no way to detect, being given a string without knowing it's encoding, whether byte 0xA4 is supposed to signify ¤ or € in your string, so there is no way to know it's exact charset.

(Note: you could add a human factor, or an even more advanced scanning technique (e.g. what Oroboros102 suggests), to try to figure out based upon the surrounding context, if the character should be ¤ or €, though this seems like a bridge too far)

There are more distinguishable differences between e.g. UTF-8 and ISO-8859-1, so it's still worth trying to figure it out when you're unsure, though you can and should never rely on it being correct.

Interesting read: http://kore-nordmann.de/blog/php_charset_encoding_FAQ.html#how-do-i-determine-the-charset-encoding-of-a-string

There are other ways of ensuring the correct charset though. Concerning forms, try to enforce UTF-8 as much as possible (check out snowman to make sure yout submission will be UTF-8 in every browser: http://intertwingly.net/blog/2010/07/29/Rails-and-Snowmen ) That being done, at least you're can be sure that every text submitted through your forms is utf_8. Concerning uploaded files, try running the unix 'file -i' command on it through e.g. exec() (if possible on your server) to aid the detection (using the document's BOM.) Concerning scraping data, you could read the HTTP headers, that usually specify the charset. When parsing XML files, see if the XML meta-data contain a charset definition.

Rather than trying to automagically guess the charset, you should first try to ensure a certain charset yourself where possible, or trying to grab a definition from the source you're getting it from (if applicable) before resorting to detection.

There are some really good answers and attempts to answer your question here. I am not an encoding master, but I understand your desire to have a pure UTF-8 stack all the way through to your database. I have been using MySQL's utf8mb4 encoding for tables, fields, and connections.

My situation boiled down to "I just want my sanitizers, validators, business logic, and prepared statements to deal with UTF-8 when data comes from HTML forms, or e-mail registration links." So, in my simple way, I started off with this idea:

1. Attempt to detect encoding: $encodings = ['UTF-8', 'ISO-8859-1', 'ASCII'];
2. If encoding cannot be detected, throw new RuntimeException
3. If input is UTF-8, carry on.

4. Else, if it is ISO-8859-1 or ASCII

   a. Attempt conversion to UTF-8 (wait, not finished)

   b. Detect the encoding of the converted value

   c. If the reported encoding and converted value are both UTF-8, carry on.

   d. Else, throw new RuntimeException

From my abstract class Sanitizer

    private function isUTF8($encoding, $value)
    {
        return (($encoding === 'UTF-8') && (utf8_encode(utf8_decode($value)) === $value));
    }

    private function utf8tify(&$value)
    {
        $encodings = ['UTF-8', 'ISO-8859-1', 'ASCII'];

        mb_internal_encoding('UTF-8');
        mb_substitute_character(0xfffd); //REPLACEMENT CHARACTER
        mb_detect_order($encodings);

        $stringEncoding = mb_detect_encoding($value, $encodings, true);

        if (!$stringEncoding) {
            $value = null;
            throw new \RuntimeException("Unable to identify character encoding in sanitizer.");
        }

        if ($this->isUTF8($stringEncoding, $value)) {
            return;
        } else {
            $value = mb_convert_encoding($value, 'UTF-8', $stringEncoding);
            $stringEncoding = mb_detect_encoding($value, $encodings, true);

            if ($this->isUTF8($stringEncoding, $value)) {
                return;
            } else {
                $value = null;
                throw new \RuntimeException("Unable to convert character encoding from ISO-8859-1, or ASCII, to UTF-8 in Sanitizer.");
            }
        }

        return;
    }

One could make an argument that I should separate encoding concerns from my abstract Sanitizer class and simply inject an Encoder object into a concrete child instance of Sanitizer. However, the main problem with my approach is that, without more knowledge, I simply reject encoding types that I do not want (and I am relying on PHP mb_* functions). Without further study, I cannot know if that hurts some populations or not (or, if I am losing out on important information). So, I need to learn more. I found this article.

What every programmer absolutely, positively needs to know about encodings and character sets to work with text

Moreover, what happens when encrypted data is added to my email registration links (using OpenSSL or mcrypt)? Could this interfere with decoding? What about Windows-1252? What about security implications? The use of utf8_decode() and utf8_encode() in Sanitizer::isUTF8 are dubious.

People have pointed out short-comings in the PHP mb_* functions. I never took time to investigate iconv, but if it works better than mb_*functions, let me know.

The main problem for me is that I don't know what encoding the source of any string is going to be - it could be from a text box (using is only useful if the user is actually submitted the form), or it could be from an uploaded text file, so I really have no control over the input.

I don't think it's a problem. An application knows the source of the input. If it's from a form, use UTF-8 encoding in your case. That works. Just verify the data provided is correctly encoded (validation). Keep in mind that not all databases support UTF-8 in it's full range.

If it's a file you won't save it UTF-8 encoded into the database but in binary form. When you output the file again, use binary output as well, then this is totally transparent.

Your idea is nice that a user can tell the encoding, be he/she can tell anyway after downloading the file, as it's binary.

So I must admit I don't see a specific issue you raise with your question. But maybe you can add some more details what your problem is.