Strange onebox failure due to malformed 404

Product Support

AI-generated summary

Strange onebox failure due to malformed 404

merefield encountered a onebox failure when trying to preview a link from NBC San Diego. The error message indicated an unexpected character in the JSON response. Despite the link previewing fine in iframely and Facebook, the onebox failed.

joffreyjaffeux suspected that the issue was due to the server expecting JSON but receiving HTML instead. After investigating, joffreyjaffeux found that the NBC San Diego server was returning a 404 error with HTML content, but with a 200 HTTP status code, which caused the JSON parser to fail.

The issue is likely due to a bug on NBC San Diego’s side, as the URL for the oembed content comes from their page source. joffreyjaffeux and zogstrip agreed that it’s not worth investing time to fix this issue, as it’s not a problem with the onebox code. merefield also tried to report the issue to NBC San Diego through their website feedback form, but was unable to submit the report due to a faulty Captcha system.

1

relates to this site:

https://www.nbcsandiego.com/news/national-international/CVS-to-Sell-CBD-Products-in-800-Stores-in-8-States-507484191.html

I get a onebox failure in logs:

Failed to onebox https://www.nbcsandiego.com/news/national-international/CVS-to-Sell-CBD-Products-in-800-Stores-in-8-States-507484191.html unexpected character () at line 9, column 2 [parse.c:704]

more info:

REQUEST_URI /onebox?url=https%3A%2F%2Fwww.nbcsandiego.com%2Fnews%2Fnational-international%2FCVS-to-Sell-CBD-Products-in-800-Stores-in-8-States-507484191.html&refresh=false

This link previews fine in iframely, facebook etc.

If anyone is interested in pursuing this let me know if you need more info.

2

Feels like something is expecting JSON and getting HTML.

3

Yes, I’ve seen a related error. I did try to byebug it through … let me see if I can find that … will update

4 
2019-03-22 10:42:42 - MultiJson::ParseError - 767: unexpected token at '<head>
<title>404 Not Found</title>
</head><body>
<h1>404 Not Found</h1>
  <p>The provided content is neither a Video Release nor does it contain a Video Release in Related Media.<br>Example Request Format: https://www.nbcnewyork.com/services/oembed/?url=https://www.nbcnewyork.com/sample/path/Sample-Article-Example-12345678.html</p>
</body></html>

      ':
	/Users/joffreyjaffeux/.rubies/ruby-2.6.1/lib/ruby/2.6.0/json/common.rb:156:in `parse'
	/Users/joffreyjaffeux/.rubies/ruby-2.6.1/lib/ruby/2.6.0/json/common.rb:156:in `parse'
	/Users/joffreyjaffeux/.gem/ruby/2.6.1/gems/multi_json-1.13.1/lib/multi_json/adapters/json_common.rb:14:in `load'
	/Users/joffreyjaffeux/.gem/ruby/2.6.1/gems/multi_json-1.13.1/lib/multi_json/adapter.rb:21:in `load'
	/Users/joffreyjaffeux/.gem/ruby/2.6.1/gems/multi_json-1.13.1/lib/multi_json.rb:122:in `load'
	/Users/joffreyjaffeux/Projects/onebox/lib/onebox/oembed.rb:6:in `initialize'

We are trying to get oembed content from CVS to Sell CBD Products in 800 Stores in 8 States – NBC New York which is a 404

The thing is, this URL comes from their page source:

<link rel="alternate" type="application/json+oembed"
    href="https://www.nbcnewyork.com/services/oembed/?url=https://www.nbcsandiego.com/news/national-international/CVS-to-Sell-CBD-Products-in-800-Stores-in-8-States-507484191.html"
    title="CVS to Sell CBD Products in 800 Stores in 8 States - NBC 7 San Diego " />

So I would day the bug is from them, not us.

5

Thank you, that doesn’t surprise me at all!

Forgive me if I sound stupid here: there’s no way we could make this more resilient and tolerate this as basic goal is to retrieve just title, excerpt, thumbnail, etc.?

6

You can always make things more resilient, proof is: it’s working on iFramely.

Now I don’t know if we want to invest time fixing blatant issues like this one. @zogstrip ?

7

Yeah, nooooo.

They’re sending back the HTTP status code 200 with the HTML content of a 404 on a JSON request… :man_facepalming:

8

FYI I even tried to tell them using their website feedback form … and the submission gets stuck at an unresponsive Captcha dialogue :man_facepalming: