Uncaught DOMException on https://losst.ru/obnovlenie-debian-9

[rNeomy]

I am getting

Uncaught DOMException: Failed to execute 'setAttribute' on 'Element': 'adsbygoogle.js"' is not a valid attribute name.
    at Readability._setNodeTag (Readability.js:477:19)
    at Readability._grabArticle (Readability.js:774:25)
    at Readability.parse (Readability.js:2062:31)
    at 

which refers to

  _setNodeTag: function (node, tag) {
    this.log("_setNodeTag", node, tag);
    if (node.__JSDOMParser__) {
      node.localName = tag.toLowerCase();
      node.tagName = tag.toUpperCase();
      return node;
    }

    var replacement = node.ownerDocument.createElement(tag);
    while (node.firstChild) {
      replacement.appendChild(node.firstChild);
    }
    node.parentNode.replaceChild(replacement, node);
    if (node.readability)
      replacement.readability = node.readability;

    for (var i = 0; i < node.attributes.length; i++) {     /*the next line*/
      replacement.setAttribute(node.attributes[i].name, node.attributes[i].value);
    }
    return replacement;
  },

[gijsk]

Can you provide more detailed STR? If I open the page in Firefox and enter reader mode, it works. What are you doing differently?

[rNeomy]

I am porting this to Chrome. To regenerate you can follow this

1. Open the page in Chrome
2. Open dev tools
3. Paste the latest Readability.js
4. paste this code

var loc = document.location;
var uri = {
  spec: loc.href,
  host: loc.host,
  prePath: loc.protocol + '//' + loc.host,
  scheme: loc.protocol.substr(0, loc.protocol.indexOf(':')),
  pathBase: loc.protocol + '//' + loc.host + loc.pathname.substr(0, loc.pathname.lastIndexOf('/') + 1)
};

var documentClone = document.cloneNode(true);
var article = new Readability(uri, documentClone).parse();

var iframe = document.createElement('iframe');
document.body.appendChild(iframe);
iframe.contentDocument.open();
iframe.contentDocument.write(article.content);
iframe.contentDocument.close();

which causes:

Uncaught DOMException: Failed to execute 'setAttribute' on 'Element': 'adsbygoogle.js"' is not a valid attribute name.
    at Readability._setNodeTag (<anonymous>:477:19)
    at Readability._grabArticle (<anonymous>:774:25)
    at Readability.parse (<anonymous>:2062:31)
    at <anonymous>:11:51

[gijsk]

I can no longer reproduce this with the latest version of Readability and Chrome, so I assume this has been fixed since. :-)

[MrOrz]

@gijsk I can reproduce the Uncaught DOMException: Failed to execute 'setAttribute' on 'Element' error on different pages.

Here are the reproduction steps:

1. Visit https://opinion.udn.com/opinion/story/10124/3561413 on Google Chrome
2. Open Devtool JS console
3. paste in latest readability.js
4. execute new Readability(location.href, document).parse();

Expected: article being parsed
Actual: thown error Uncaught DOMException: Failed to execute 'setAttribute' on 'Element': '0' is not a valid attribute name.

I have a bunch of URLs that can be used to reproduce the above behavior:
https://docs.google.com/spreadsheets/d/1TwDZmocHxlcSPm2zQoGJF167Uj9RNgqHREEx3PEa580/edit?usp=sharing

Apparently it's caused by malformed HTML, but it should not crash the entire readability.

[gijsk]

Well, this is fun... I filed whatwg/html#4275 for the inconsistency between how HTML is parsed and how the DOM APIs validate exactly the same thing. I should have a patch for this using the workaround mentioned there (using element.attributes.setNamedItem() instead of element.setAttribute()).

[gijsk]

Yum, and cloneNode() is not supported for Attr in jsdom... jsdom/jsdom#2478

[gijsk]

OK, so maybe the simpler solution is just to ignore those additional attributes that we can't even set; it's not like we need them for anything...

[gijsk]

OK, so PR is up, as well as issues found in jsdom and HTML/DOM as a result. Thanks @MrOrz for the clear steps to reproduce!