At the windows command line ...

php -r "echo ord('š');"

I get 154.

If the file is encoded Windows 1252 I get 154.

Latin small letter s with caron. 

If the file is encoded UTF-8, I get 197 as ord() is working on binary strings and š has a binary value of 0xC5 0xA1 an 0xC5 = 197. (This translates to U+0161 according to Charmap).

In PHP 5, ord() currently works on binary strings. So, depending upon the encoding used, the binary string may start whatever is appropriate.

There is a user note on ord() with regards to getting a value of a unicode "character".

This returns a value of 353 which, as if by magic, has a value of 0x0161.

That's exactly my point. The PHP documentation (http://us.php.net/manual/en/function.ord.php) says: "Returns the ASCII value of the first character of string". This is not true, as we have both shown, and should be updated appropriately.

The only thing that is potentially missing is that you will get a different response depending upon the encoding of the source file.

If you use UTF-8, the first byte (which in PHP5 and lower is synonymous with character) has a value of 197.

If you are using Windows-1252, then the character has a value of 154.

Both are correct as they are in effect different strings.

If you are needing UTF-8 string processing, then the functions mentioned in the user notes as well as other multi-byte string libraries are available.

I'm not talking about the user notes or about the accuracy of the function. The function works absolutely fine. In fact, I'm glad it works the way it does. I'm just saying that the documentation is miss leading. The ASCII character set only goes from 0 to 127. When I read the documentation and it tells me that ord returns the ASCII value that tells me to only expect a return value between 0 and 127 (or 32 to 126 for printable characters).

I have no problems with the way the ord function works. Only the documentation does not reflect what the function does.