The problem with the current approach is that any invalid rule breaks the content blocker completely. Also, there's no way to learn what exact rule is invalid, we even had to come up with a script that uses some kind of a "binary search" algorithm to find problematic rules. Overall, this experience is really painful for people who are trying to create content blockers.

Here's what I suggest:

1. Ignore problematic rules instead of throwing out the whole content blocker.
2. Extend the setContentBlocker callback parameter and put a list with the problematic rules there.