CODEWITHSUNDEEP
javascriptfirefoxdominnerhtml
user123444555621
user123444555621

Reputation: 153184

How to use innerHTML on plain txt files in Firefox?

If you open a text file (.txt, .js, .css, ...) in your browser, it will get wrapped up in a nice DOM tree.

For example, open this .txt file and enter

javascript:alert(document.documentElement.innerHTML);

into your address bar. Nice... every major browser supports DOM manipulation on this wrapped text files, which is a great thing for writing powerful bookmarklets or user scripts.

However, Firefox fails on assigning any element's innerHTML. For example,

javascript: document.body.innerHTML = document.body.innerHTML.replace(/(\d+\s+\w+(?=\s+\d+))/g, '<span style="color:red">$1</span>'); void 0;

will work in every browser but Firefox.

Is there a trick to work around this issue?

(No, I don't want to parse the innerHTML string manually, and no, it doesn't work with jQuery either.)

Upvotes: 1

Views: 5336

Answers (4)

user123444555621
user123444555621

Reputation: 153184

I think I found a working solution. First of all, let me give some more details on the question.

The problem is: Firefox creates something like

[some wrapper]
+---document
    +---<html>[=documentElement]
        +---<body>
            +---<head/>
            +---<pre>
                +---[actual plain text contents]

but the wrapped document object does not support setting innerHTML properly. So, the basic idea is, create a new document object with full innerHTML support. Here's how it works:

var setInnerHTML = function(el, string) {
    if (typeof window.supportsInnerHTML == 'undefined') {
        var testParent = document.createElement('div');
        testParent.innerHTML = '<br/>';
        window.supportsInnerHTML = (testParent.firstChild.nodeType == 1);
    }
    if (window.supportsInnerHTML) {
        el.innerHTML = string;
    } else {
        if (!window.cleanDocumentObject) {
            /* this is where we get a 'clean' document object */
            var f = document.createElement('iframe');
            f.style.setProperty('display', 'none', 'important');
            f.src = 'data:text/html,<!DOCTYPE html><html><title></title></html>';
            document.body.appendChild(f); /* <- this is where FF creates f.contentDocument */
            window.cleanDocumentObject = f.contentDocument;
            document.body.removeChild(f);
        }

        /* let browser do the parsing */
        var div = window.cleanDocumentObject.createElement('div');
        div.innerHTML = string; /* this does work */

        /* copy childNodes */
        while(el.firstChild) {
            el.removeChild(el.firstChild); /* cleanup */
        }
        for (var i = 0; i < div.childNodes.length; i++) {
            el.appendChild(div.childNodes[i].cloneNode(true));
        }
        delete div;
    }
}

edit:

This version is better and faster; using XSLTProcessor instead of iFrame.

var setInnerHTML = function(el, string) {
    // element.innerHTML does not work on plain text files in FF; this restriction is similar to
    // http://groups.google.com/group/mozilla.dev.extensions/t/55662db3ea44a198
    var self = arguments.callee;
    if (typeof self.supportsInnerHTML == 'undefined') {
        var testParent = document.createElement('div');
        testParent.innerHTML = '<p/>';
        self.supportsInnerHTML = (testParent.firstChild.nodeType == 1);
    }
    if (self.supportsInnerHTML) {
        el.innerHTML = string;
        return el;
    } else if (typeof XSLTProcessor == 'undefined') {
        return undefined;
    } else {
        if (typeof self.cleanDocument == 'undefined')
            self.cleanDocument = createHTMLDocument();

        if (el.parentNode) {
            var cleanEl = self.cleanDocument.importNode(el, false);
            cleanEl.innerHTML = string;
            el.parentNode.replaceChild(document.adoptNode(cleanEl), el);
        } else {
            var cleanEl = self.cleanDocument.adoptNode(el);
            cleanEl.innerHTML = string;
            el = document.adoptNode(cleanEl);
        }

        return el;
    }

    function createHTMLDocument() {
        // Firefox does not support document.implementation.createHTMLDocument()
        // cf. http://www.quirksmode.org/dom/w3c_html.html#t12
        // the following is taken from http://gist.github.com/49453
        var xmlDoc = document.implementation.createDocument('', 'fooblar', null);
        var templ = '<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">'
                + '<xsl:output method="html"/><xsl:template match="/">'
                + '<html><title/><body/></html>'
                + '</xsl:template></xsl:stylesheet>';
        var proc = new XSLTProcessor();
        proc.importStylesheet(new DOMParser().parseFromString(templ,'text/xml'));
        return proc.transformToDocument(xmlDoc);
    }
};

Upvotes: 1

bobince
bobince

Reputation: 536615

It seems on a text document in Firefox 3, assigning the innerHTML of any node acts as if you are assigning to innerText (with “<html><body><pre>” prepended).

(Since DOM scripting on a non-XML/HTML document is completely undefined, it's certainly within Firefox's rights to do this; it appears to be a quick hack to display text files in an HTML page.)

So you can't use innerHTML on Firefox, but other DOM methods work:

var span= createElement('span');
span.style.color= 'red';
span.appendChild(document.createTextNode(match));

Upvotes: 0

Andrew Hare
Andrew Hare

Reputation: 351616

It is failing because there is no body - even the file you linked is just a text file without a body (perhaps you are looking at it in firebug?).

The best thing to do would be a regex replace since you are working with text.

Upvotes: 1

Itay Moav -Malimovka
Itay Moav -Malimovka

Reputation: 53607

use GreaseMonkey

Upvotes: 0

Related Questions