Question

I have cgi script pulling search terms from an HTML form through CGI::Ajax. When running Encode::Detect::Detector over the resulting string in my script, I get, instead of just UTF-8, a variety of encodings depending on the characters entered in the form: Greek chars turn up as UTF-8, umlauts as windows-1252 and no output at all for ASCII characters. I am fairly sure the problem lies with whatever CGI::Ajax does to the string it passes to Perl. It uses decodeURI() somewhere in its code, I have tried URI::Escape and Encode in all possible permutations, but none of my attempts to normalise the different strings to a single encoding were successful. As it is, one or the other set of non-ASCII chars (either umlauts or the Greek) will always be garbled. How do I tell Ajax to keep it Unicode?

Answer 1

Solved: CGI::Ajax apparently uses Javascript’s escape function, which does not handle Unicode correctly. The function has been superseded by encodeURI and encodeURIComponent, which can be set as the default escaping functions for an CGI::Ajax object $pjx like so: $pjx->js_encode_function('encodeURIComponent');. Phew.