Well, there is indeed a bug in the JavaScript file (or in the calls) for Internet Explorer v 6.0 SP1 or less.

The JS file has a function called tinyMCE_GZ.checkCompress() which returns 0 (false) if the person is using IE 6.0 SP1 or less.  (IE 6.0 SP2 or IE7 return 1, as well as all non-IE browsers.)

If checkCompress() is 0, then '&compress=false' is added to the end of the tiny_mce_gzip.aspx (or .php) QueryString, which in turn deactivates gzip compression for the file that gets loaded.

Now, with compressioned turned off, I believe the problem is that all the normal function calls are made, as if the file is compressed, so when a bad function call is made, the JavaScript stops executing, so TinyMCE does not get instantiated.

Here's the weird part (and why I didn't spend much more time trying to figure out the root cause): the very first time IE 6.0 SP1 or less is run, the editor loads, but each time after that it fails.  (Just refresh the page to see.)

To immediately get around the error, comment out the following line in tiny_mce_gzip.js:

v += this.checkCompress() ? '' : '&compress=false';

becomes:

//v += this.checkCompress() ? '' : '&compress=false';

After the line is commented out, the problem does not seem to happen in IE6 anymore.

Of course, this is just a band-aid, and does not address whatever the root cause is.

TinyMCE doesn't support multiple languages on the same page. This is most likely to change sometime in the future so the gzip compressor has support for it so that it will function ones TinyMCE has that support.

If you are still having problems getting TinyMCE to work (tinyMCE is not defined error) - although you have put tinyMCE_GZ.init({ and tinyMCE.init({ into two seperate JavaScript tags - and you are using a foreign language, you might want to take a look here:

http://tinymce.moxiecode.com/punbb/view … 946#p16946

Regards
Leo

Interesting thanks a lot for this one. I will include it in the next release of the gzip compressor.

For the following code in the tiny_mce_gzip.js file:

        for (i=0, nl = d.getElementsByTagName('script'); i<nl.length; i++) {
            if (nl[i].src && nl[i].src.indexOf('tiny_mce_gzip') != -1) {
                sr = nl[i].src;
                sr = sr.substring(0, sr.lastIndexOf('/'));

                if (b != '' && b.indexOf('://') == -1)
                    b += sr;
                else
                    b = sr;
            }
        }

Shouldn't you cosider aborting the for/next when you find a matching SCRIPT so as to avoid over concatenating the b variable?

Last edited by jeffmowens (2006-12-05 22:15:53)

(Copied this from another thread)

It is not a good idea to comment out that part.

    // Check if it supports gzip
    if (isset($_SERVER['HTTP_ACCEPT_ENCODING']))
        $encodings = explode(',', strtolower(preg_replace("/\s+/", "", $_SERVER['HTTP_ACCEPT_ENCODING'])));

    if ((in_array('gzip', $encodings) || in_array('x-gzip', $encodings) || isset($_SERVER['---------------'])) && function_exists('ob_gzhandler') && !ini_get('zlib.output_compression')) {
        $enc = in_array('x-gzip', $encodings) ? "x-gzip" : "gzip";
        $supportsGzip = true;
    }

Add this before those lines

$encodings = array();

What kind of server are you running by the way? It seems like your server doesn't have this server variables set at all
$_SERVER['HTTP_ACCEPT_ENCODING']

Upload a php file with phpinfo(); in it, and check what the HTTP_ACCEPT_ENCODING string says (if it exists). Write it here plz.

So the new compressor works around the IE bug by basically not using compression if the browser is IE, correct?