penge/block-site

includes function is too broad for URL matching

Closed this issue · 7 comments

ItsNickBarry commented

Some URLs may include the addresses of blocked sites, despite being separate sites. For example, I was accessing a YouTube video via an RSS reader, and the tab was closed. Arguably this is intended behaviour, but I think it's too strict.

Instead, maybe match window.location.host to the restricted sites, instead of using includes?

penge commented

Will be implemented in next version! (Definitely a great suggestion, Thanks!)


I have already checked on window.location.host inside background.js and during development it returns a really strange value, like gipebbenhibhdifokjmpdnpeigjfgopa (probably running through a local proxy).

I understand your idea and will make it work! Not based on window.location.host but will check the host as needed.

After then. Opening addresses like:

https://my-website.com/something/facebook.com

Will work!

penge commented

Actually. There is already a way how to solve it right now! :)))

In Options, you may define more specific domains, like:

https://web.facebook.com
https://instagram.com
https://www.instagram.com
https://www.reddit.com

And so on!

I will try to make it more intelligent though. To assume http(s), www, no-www, and to not look beyond the first slash.

Challenge accepted!

penge commented

Update:

And if there is NO http(s), www, no-www, present in the URL, I'll make sure it works both ways :))

In that case, a combination of options like this:

https://facebook.com
instagram.com

Will also work! :)

penge commented

I think I'll make it easier and just look at the beggining of the string, right around the //

Simply as: * // * your blocked address

That's it! :D

penge commented

Actually found a very best solution: (no need for regex)

Here it is:

// 1. You get the url from the tab

var url = tab.url;


// 2. This url is tricky! It has some flow of updates.
// It can be URL from Chrome Extensions.
// Some fancy other url with no www, http or whats-ever (basically not a valid URL)!
// And even, it can get to be "chrome://newtab" !!!


// 3. But no worries, you make a simple filter
if (url.startsWith("http") === false) {
  return;
}


// 4. Then, you ARE actually able to get a hostname (URL is valid at this point)
function getDomain(myUrl) {
  var url = new URL(myUrl);
  var hostname = url.hostname;
  return hostname;
}


// 5. Lastly, you just compare the domains
getDomain(url).includes(blockedItem);

:)

penge commented

Hi @ItsNickBarry,

It is now resolved. Releasing version 3.0.