[CS5][JS] Update TextFrame contents with overflow??

I regret I'm having a similar problem & the answers I'm seeing aren't providing solutions.

In my case, what I have is a large volume of unformatted text dumped into a text box awaiting mass formatting via my script which relies on Grep/Text Find & Change functions. How it's suppose to work is if specific text is selected within a text frame then only that text is formatted. If one or more text frames are selected then ALL the text in those text boxes is to be formatted INCLUDING overset text but (preferably) excluding text flowed into linked text boxes (this part I've not worked out yet). If nothing is selected then all text frames on the visible layers of the document are to be mass formatted (again this hasn't been fully worked out yet).

Currently what is happening is that if text boxes are selected, ONLY the text visible within the text boxes has the formatting applied to it. Anything hidden in overset text remains untouched. Can anyone shed some light on where I'm going wrong & how to solve it?

This is the function I've created thus far :

function doFindReplace(searchTypeString /*t|g*/, doc/*obj - the document*/) {

    // check if the default or specified document is open. If no document is provided, default document is assumed.

    var theDoc = doc;

    try{

        if( !(theDoc||(theDoc = app.activeDocument)) )

            { throw new Error("No document available."); }

        }

    catch (e) {

        abort = true;

        alert( e.name + "\n" + e.message );

        return;

    }

    // for every selected textual type item do the find & replace

    var counter1 = 0;

    do {

        try{

            // perform find & replace

            if (abort == false && (app.selection[counter1] instanceof TextFrame || app.selection[counter1] instanceof Text) ) { // If nothing in the script has previously returned an error, proceed to do search & replace on selected textual type things only

                switch ( searchTypeString) {

                    case "t":

                    if (theDoc.selection[counter1]) {

                        theDoc.selection[counter1].changeText();

                        }

                    else {

                        theDoc.changeText();                  

                        }

                        break;

                    case "g":

                    if (theDoc.selection[counter1]) {

                        theDoc.selection[counter1].changeGrep();

                        }

                    else {

                        theDoc.changeGrep();

                        }

                        break;

                        }

                    }

                }

        catch (e) {

            abort = true;

            alert( e.name + "\n" + e.message );

            return;

        }

    counter1++;

    } while (counter1 < app.selection.length);

}

//

What exactly do you mean by "jumbled and formatted differently"? Is the formatting random, or is your script changing the formatting inside the textframe but fails to process the overset text? If the latter I am guessing your script targets TextFrame.texts, TextFrame.paragraphs, TextFrame.lines, TextFrame.characters or similar in which case I would try to target TextFrame.parentStory.<whatever> instead.

Does calling the .recompose() method on the text frame help any? Just a shot in the dark...

Yep -- been there, done that, seen it.

It's caused by the way you add text and formatting to your text frame. Things like "line" don't work in an overset frame (because there is no frame to get the text width from). It seems other stuff get, well, not really "messed up", it's more like ID cannot properly access stuff that doesn't actually exist yet.

I settled for the lo-tech solution of make the text frame way larger than absolutely necessary