Alter base64 encoding before decoding it

[fmarier]

Here is an example of an email link tracker that we could debounce:

https://link.emlmind.com/en/mail_link_tracker?hash=66c5ibqsukgp7ikdzwu4ujmw7y7boyubuqqmaq86u5sjqdohiizkiackas9r171wq1go8tfs5eongbftyy5rh47j45mgawwfzgn66xho&url=aHR0cHM6Ly93d3cubWVsb3R0b2dyb3VwLmNvbS9zY2hlZHVsZQ~~&uid=Njg1OTgwOA~~&ucs=88ff10da1ab90c2435eec8aa3fb9fb04

since the url parameter includes the destination page using the Base64 encoding:

url=aHR0cHM6Ly93d3cubWVsb3R0b2dyb3VwLmNvbS9zY2hlZHVsZQ~~

However, this platform is using ~~ instead of == for padding and so it gets rejected by the Chromium decoding function. We would need to do some kind of string replacement prior to the base64 decoding in order to support this.

[fmarier]

https://github.com/brave/adblock-lists/pull/1858 was a failed attempt at debouncing this using the existing base64 decoder.