Big GIFs can cause application crash

[lorenzos]

I don't know if all big files, or only specific ones, but this:

var mi = require('node-metainspector');
new mi("http://media.giphy.com/media/euBj15T6nrp6g/giphy.gif", {}).fetch();

Crashes the application with:

node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:47
  _.each(dom, function(elem) {
                      ^
RangeError: Maximum call stack size exceeded
    at /home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:47:23
    at Array.forEach (native)
    at Function._.each._.forEach (/home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/node_modules/underscore/underscore.js:78:11)
    at exports.connect (/home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:47:5)
    at /home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:64:7
    at Array.forEach (native)
    at Function._.each._.forEach (/home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/node_modules/underscore/underscore.js:78:11)
    at exports.connect (/home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:47:5)
    at /home/pi/hangouts-bots/node_modules/node-metainspector/node_modules/cheerio/lib/parse.js:64:7
    at Array.forEach (native)

If this is hard to solve, I will be glad to know if there is some workaround in order to, at least, keep my application running.

[alexymik]

I had the same issue, seems like it's to do with large requests. I set a request size limit which seems to have fixed the issue for now:

new MetaInspector(param, { limit: 3000000 } );

[19h]

Why would you crawl inspect images? This is for meta information about websites.

[alexymik]

From the project description:

"You give it an URL, and it lets you easily get its title, links, images, description, keywords, meta tags...."

It should not crash when given a URL, even if it's given a URL to an image. The Ruby MetaInspector gem also returns image resolution, so a future feature would be to port that functionality over as well.

[19h]

Here's the source code so PR maybe?

[sshen81]

I've run into similar issues with non-HTML URIs causing errors. Specifically, passing non-HTML body data into the cheerio module can cause call stack exceptions. I've submitted PR #25 to help address this.