Adblock decoder ignore some portion when decoding

As reported by @dnmTX at https://github.com/Ultimate-Hosts-Blacklist/dev-center/issues/9:

everything with ##[href^=…

are ignored.

[funilrys]

Hi @jawz101, I’m writing an improvement but we have to admit, it’s impossible to avoid false positive. That’s why whitelisting is more important than blocking.

Now we have to decide between 2 way:

• We extract and test all possible domains. or
• We extract all domains which are (or may be) relevant.

For now, I’m implementing the second way but we may think about the other way in the future or as an extra option.

[keczuppp]

OK, so I’ve just tested the newest PyFunceble dev right now and I’ve noticed that the reported issues mentioned in : https://github.com/funilrys/PyFunceble/issues/13#issuecomment-749607193 and https://github.com/funilrys/PyFunceble/issues/13#issuecomment-803537932 have been fixed.

---

The summarision:

keczuppp: As for the last 3 failures, many of such failures can be found in https://easylist-downloads.adblockplus.org/easylistpolish.txt The list contains about 2961 domains, but only 2459 are found by Adblock Decoder (with --aggressive option), which gives 83% efficiency.

• the current EasyList Polish contains 3190 domains and the newest Pyfunceble has extracted 3106 what gives 97% efficiency compared to 83% previously

keczuppp: currently the decoder extracts about a half of domains, to prove it I copy-pasted all domain= lines from several popular / big adblock filter lists + two polish lists: EasyList, EasyPrivacy, AdGuard Base, AdGuard Tracking Protection, Official Polish Filters for AdBlock, uBlock Origin & AdGuard, EasyList Polish and put into a single file / list (see domain=.zip), which contain about 11425 domains, but Adblock Decoder extracts only about a half (5426 domains)

• the newest PyFunceble has extracted 8298 what gives 73% compared to previously 48%

Good improvement.