Closed jgluck closed 9 years ago
Also, hope you're having a good mini-vacation ;)
I'm back :)
So I crated a new version which consumes less CPU. Actually it consumes no CPU after the comments are removed and no new comments loaded.
I also tweaked "matches"
part of manifest file where I removed youtube.com, githubusercontent.com and apis.google.com thinking I don't need them anymore. Published a new version, seemed to work, came home, updated the extension and it stopped working. Re-added githubusercontent.com, apis.google.com to manifest again and it started working again, yay. Does this explain well enough?
EDIT: The extension needs an access to raw.githubusercontent.com to read the blacklist and an access to apis.google.com and plus.googleapis.com to remove the comments (the comments reside inside an iframe).
Ah, brilliant! Might want to copy paste that up to a changelog.md or something.
I was thinking you could probably do something similar to this by building up a language model to identify if a comment is similar enough to a common fedora user.
Everybody keeps suggesting I make it smarter with some sort of recognition. I wanna say challenge accepted but right now I would prefer a pull request :)
I'm considering it. Do you have a corpus of Le' Reddit posts?
I don't.
Care to explain the new perms?
Not that I'm distrustful, was just wondering why you want access to github...