Evernote webclipper won't clip

[dmodmo]

Was working great until I did a fresh install. I've got the latest of both add-ons:

evernote 4.0.0.87683 noscript 2.0.2.5 FFox 3.69

I've added ^http://www\.evernote\.com/clip\.action$ to the advanced xss section, but Evernote's webclipper simply won't pop up. Suggestions would be great. thanks

[dmodmo]

Bump?

[Giorgio Maone]

Do you get any message in Tools|Error Console when this happens?

[vnavone]

Having the same problem. I click on the icon for the Firefox add-on and nothing happens. Problem goes away when I disable NoScript.
Evernote Web Clipper 6.12.1.1
NoScript 5.0.8.1
Firefox 54.0.1 (64-bit)

Hope you can help!

[vnavone]

P.S. can't find the Error Console in Firefox. There's a bunch of stuff about Evernote in the Debugger, but when I try to paste it here the post gets bounced by your anti-spam filter.

[barbaz]

P.S. can't find the Error Console in Firefox.

Error Console has been gone for a long time. It's replaced by the Browser Console (Cmd-Shift-J).

[GµårÐïåñ]

As stated, not exactly gone, you can open the "Browser Console" and it will show you everything you need and then some.

It has much of the same filters that existed before and some new ones, you can set them properly to filter some of the unnecessary noise and garbage that shows up there.

[vnavone]

Okay, looking in the Browser Console I can't see any errors or other messages about the Evernote Web Clipper. Would it be useful to post the stuff from the Debugger?

[GµårÐïåñ]

Okay, looking in the Browser Console I can't see any errors or other messages about the Evernote Web Clipper. Would it be useful to post the stuff from the Debugger?

Not particularly but can you make sure that you have at the very least, the following: Net > first three, JS > select all three filters, Security > select both, Logging > all of them, and finally Server > errors and then clear the log and try to do what you need with clipper and then see what populates there. The rest of them are pretty much noise at this point. Clipper works in some ways like say Pushbullet, so they do a lot of XHR and behind the scene actions that normally are not exposed in the console, so make sure you have the options above and then PERFORM the action in question and watch and see what goes in there, not all will be relevant but it would be the generally most useful set of filters to see what is happening. Good luck.

[vnavone]

Thanks for the filtering tips. I've applied those and tried to use the Web Clipper on a few pages, including these forums. Poking the Evernote icon doesn't immediately generate any errors, but I am intermittently seeing a these in the console:

Unchecked lastError value: Error: Could not establish connection. Receiving end does not exist.

and

XML Parsing Error: not well-formed
Location: jar:file:///Users/victor/Library/Application%20Support/Firefox/Profiles/ngxib457.default/extensions/%7BE0B8C461-F8FB-49b4-8373-FE32E9252800%7D.xpi!/content/Clipper.js
Line Number 1, Column 1:

Give you any clues?

[GµårÐïåñ]

Unchecked lastError value: Error: Could not establish connection. Receiving end does not exist.

This means the "script" inside the extension is reaching for a server or in some manner that is not completing, this alone tells me that it can't contact the service and it is an Evernote issue.

XML Parsing Error: not well-formed
Location: jar:file:///Users/victor/Library/Application%20Support/Firefox/Profiles/ngxib457.default/extensions/%7BE0B8C461-F8FB-49b4-8373-FE32E9252800%7D.xpi!/content/Clipper.js
Line Number 1, Column 1:

This confirms my belief that this is an Evernote issue for sure. However, I will do an extra step to try and make sure. Can you send me the file at the path referenced above.

- You can bring up run, type in %appdata% and press enter.
- You will have a folder open up within which I need you to locate \Roaming\Mozilla\Firefox\Profiles\ngxib457.default\extensions
- Then open up the XPI file above {E0B8C461-F8FB-49b4-8373-FE32E9252800}.xpi
- using any archive software, WinZip, WinRAR, 7zip or whatever and then within the content folder, save the Clipper.js file

Send me that file, chances are the content will be large, so posting here might be prohibitive but you can try to paste the full content of the file by using the CODE tag and putting in the block when replying and I can see what it is trying to do and why it is throwing an error on the first line of the code. Alternatively if that fails, pastebin it or something similar and post the link to it here. Whichever is fine.

[vnavone]

So I'm on a Mac, not windows, but I was able to locate the .xpi file. I can't open it so I've zipped it and put a copy here:
https://www.dropbox.com/s/s8mr6vy1ot3b4 ... i.zip?dl=0

Incidentally, I've noticed that my 1Password add-on has stopped working as well. When I disable NoScript it works again...

[GµårÐïåñ]

So I'm on a Mac, not windows, but I was able to locate the .xpi file. I can't open it so I've zipped it and put a copy here:
https://www.dropbox.com/s/s8mr6vy1ot3b4 ... i.zip?dl=0

Ok, I missed that but the process is the same despite the navigation differences that you can figure out I am sure. You can just upload the zip that's fine, I can work with it myself, that's fine, I will grab the file in a sec and get back to you on it.

Incidentally, I've noticed that my 1Password add-on has stopped working as well. When I disable NoScript it works again...

The problem is most likely how they are "injecting" their scripts into the page rather than using the extension infrastructure to properly handle things, it is a lazy way that many "developers", and I use the term loosely, use to make it easier for them without having to support the actual devices they are running on and making them quasi-web scripts in nature, reminiscent of script kiddies.

You may need to do their job for them, identify the script and manually whitelist it in NS which is doing its job by ensnaring them, to make them function. But it is bad programming and they should be ashamed of it.

[GµårÐïåñ]

Just a quick heads up, the reason the error shows on line 1 and column 1 is that the script has been minified which means it all appears on one line, so no matter where in it the error occurs, it will show as line one. So that immediately makes the error report useless as it stands. Although I did use error tracking to try and locate the code block that would throw that particular error, but seems they are using the error generically in a few places, so again that makes it less than ideal.

Properly refactored code looks like this:

Code: Select all

/*! Copyright 2009-2017 Evernote Corporation. All rights reserved. */
define(["GlobalUtils", "PageInfo"], function(a, b) {
    function c(a, c, d) {
        function e(b) {
            k += b, j < i.length ? m(i[j++], null, a, e, null, d) : c(k)
        }

        function f(b) {
            g = b, l(g), m(g, null, a, e, null, d, h)
        }
        var g, h = {
                isOutlookMail: b.isOutlookMail()
            },
            i = b.getNextPages(),
            j = 1,
            k = "";
        try {
            if (g = contentPreview.getArticleElement()) return l(g), void m(g, null, a, e, null, d, h)
        } catch (a) {
            log.warn("Couldn't get preview from contentPreview. Trying pageInfo.")
        }
        try {
            return void b.getDefaultArticle(f)
        } catch (a) {
            log.warn("Couldn't get article from pageInfo. Trying default.")
        }
        g = document.body, l(g), m(g, null, a, e, null, d, h)
    }

    function d(a, b, c) {
        l(document.body), m(document.body, null, a, b, null, c)
    }

    function e(c) {
        c('<embed src="' + a.escapeXML(b.getPdfUrl()) + '" type="application/pdf"></embed>')
    }

    function f(a, b) {
        var c = contentPreview.getEmailElement();
        l(c), m(c, null, !0, a, null, b, {
            isEmailClip: !0
        })
    }

    function g(a, b, c, d) {
        if (b && document.querySelector("embed[type='application/pdf']")) return void c(b);
        var e, f;
        try {
            if ((e = contentPreview.ensureSelectionIsShown()) && (f = e.getRangeAt(0))) {
                if (f.commonAncestorContainer.nodeType == Node.TEXT_NODE) {
                    var g = f.commonAncestorContainer.textContent.substring(f.startOffset, f.endOffset);
                    c(n(g))
                } else l(f.commonAncestorContainer), m(f.commonAncestorContainer, f, a, c, null, d);
                return
            }
        } catch (a) {
            c("")
        }
    }

    function h(b, c) {
        var d = document.querySelector("img[src='" + b + "'], img[src$='" + b.replace(document.body.baseURI, "") + "']"),
            e = t.convertImgSrcToBase64IfPossible(d, b, !0);
        e === b && (e = a.escapeXML(e)), c('<img src="' + e + '"></img>')
    }

    function i(a) {
        var b = contentPreview.getUrlElement(function(b) {
            a(b)
        });
        b && a(b)
    }

    function j(a, b) {
        l(document.body);
        for (var c = contentPreview.getClearlyReformat(), d = c.$iframePages[0].getElementsByTagName("span"), e = 0; e < d.length; e++) d.item(e).parentNode.removeChild(d.item(e));
        l(c.$iframePages[0]), m(c.$iframePages[0], null, !1, a, c.iframe.contentWindow, b)
    }

    function k(a, b) {
        var c = contentPreview.getCustomElementContent();
        m(c, null, !1, a, c.ownerDocument.defaultView, b)
    }

    function l(a) {
        for (var b = a.querySelectorAll("a.clearly_highlight_delete_element"), c = 0; c < b.length; c++) b.item(c).parentNode.removeChild(b.item(c));
        for (var d = a.querySelectorAll("em.clearly_highlight_element"), c = 0; c < d.length; c++) d.item(c).outerHTML = "<highlight>" + d.item(c).innerHTML + "</highlight>"
    }

    function m(a, c, d, e, f, g, h) {
        function i(a, b) {
            b ? g(b) : r(a)
        }

        function j(a) {
            t.serialize(o, p, q, i, a, null, h)
        }
        if (o = a, p = c, q = d, r = e, f) t.serialize(o, p, q, i, null, f);
        else try {
            h || (h = {}), h.isOutlookMail = b.isOutlookMail(), serializeFrames(j, h)
        } catch (a) {
            g(a)
        }
    }

    function n(a) {
        return a = a.replace(/&/g, "&"), a = a.replace(/</g, "<"), a = a.replace(/>/g, ">")
    }
    var o, p, q, r, s = {},
        t = new HtmlSerializer;
    return s.clipArticle = c, s.clipEmail = f, s.clipImage = h, s.clipFullPage = d, s.clipPdf = e, s.clipSelection = g, s.clipUrl = i, s.clipClearly = j, s.clipCustom = k, s
});

Now if you are capable of doing it yourself, the best way to properly get the error reporting to work for us ON YOUR SYSTEM, is for you to backup the existing XPI, open the clipper.js file inside and replace it with the refactored code and then repackage it and then run the browser and now we get actual lines the error comes on. If not, then I can go ahead and repackage the XPI for you and have you replace yours with it long enough for us to figure out what's wrong and then you can delete it and restore the old file or just wait until their next update which will overwrite it.

Your call but let me know either way. I really don't want to install, load up and configure an entire Mac VM to do this, and plus it wouldn't replicate your environment either so it is not helpful. SO doing it on your own system will be the best route, but again depends on how much you can do to facilitate or proficiency to do it, so I can do some of the stuff for you and then give it to you and have you load it but ultimately it will not be something I can directly observe, so that will be on you to provide.

One thing to note however is that the XML malformation results in the archaic way they build the XML structure using basically string building, which can introduce a lot of artifacts a long the way and THEN they take that "string" they have built and try to run it through the serializer which is where that XML error is being generated. It would make sense that if at any step along the way, if they get something NaN or undefined, that would simply break the structure they are building and trying to feed into the serializer which would make it effectively malformed and fail.

Also, a question of reference, are you using Outlook mail on your machine? or have you ever installed it? or has it ever been pre-loaded in any form? It has a lot of code that keeps checking for that, which could make that if there is any left over artifacts (if you are not using it but somehow it has existed in some form, let's say you have office installed, it would be) and when it gets an inconsistent reply, the script is failing to properly error check it, leading to a malformed structure, and if you do use it, then something involving permissions could be malforming the response they are getting and it is not handling it gracefully, again leading to a malformed structure that is failing. As you can see, unless on your environment, we have no way to know. I can also write a few "log" lines into the code before repackaging so you can "walk" the script, meaning it shows an alert at each step that you can screenshot or just keep in mind to see what it is doing and where it fails, and that way you can tell me, oh on this line when the alert popped it was wrong or garbled, get it? So, let me now how you want to proceed and we'll try and give it a go somehow.

Or you can just talk to their development team and say, hey it is not working, why?

[vnavone]

Hi, thanks for the lengthy description. I'm happy to try things on my end, but unfortunately I'm going to need a bit more hand-holding. I haven't found any way to open the .xpi file on my Mac to extract the clipper.js file. A quick Google didn't reveal any clues, either. Any advice would be helpful.

Per your question, I do not, nor have I ever used Outlook on my Mac.