Adblock decoder ignore some portion when decoding

[funilrys]

As reported by @dnmTX at Ultimate-Hosts-Blacklist/dev-center#9:

everything with ##[href^=....

are ignored.

[dnmTX]

Caught another bug:
This section in the original lists are rules that removing elements(only) from legit sites:

After PyFunceble filter the lists the end result in domain.list is:

Notice how legit sites are being blocked?

[funilrys]

Okay you have to explain me AdBlock then @dnmTX 😸 I'm not a big fan of it as its syntax is confusing.

So how do I differ legit from bad site in adblock ? I though that adblock was only about blocking not whitelisting 🤔

[dnmTX]

OK......
i'll do the basics only to be more clear:
If you want to block domain you need to add || in the front and ^ at the end(it will catch the subdomains as well)
If you want to block just element in that website,you need to find it(chrome dev-tools helps a lot with that) and add ## after the domain name.Example:
Open yahoo.com(i removed soooo many elemnts from that page you wouldn't recognize it).
Now...look at your yahoo page and compare to mine:

Much cleaner,no videos,no annoyances.
Rules examples:
yahoo.com###applet_p_50000278
yahoo.com###applet_p_32209491
yahoo.com###applet_p_50000277
yahoo.com###applet_p_63802
yahoo.com###applet_p_63796
yahoo.com###sticky-lrec2-footer

[funilrys]

Okay so what about this format ? Which of the following mark the domain as a bad or good boy ?

||google.com$script,image
||api.google.com/papi/action$popup
facebook.com###player-above-2
~github.com,hello.world##
@@||cnn.com/*ad.xml
!||world.hello/*ad.xml
!@@||funceble.world/js
yahoo.com,msn.com,api.hello.world#@#awesomeWorld
!funilrys.com##body
hubgit.com|oohay.com|ipa.elloh.dlorw#@#awesomeWorld

I know you will not find them in real world but they are part of the tests for the decoder.

[dnmTX]

The ##[href^=.... it's different.it's embedded in the iframe and this how you blocking those domains

[funilrys]

Okay I'm working on that implementation.

So in this

hubgit.com|oohay.com|ipa.elloh.dlorw#@#awesomeWorld

they are all legit right ?

[dnmTX]

||google.com$script,image -this rule will not allow any scripts or images to be shown or executed on that domain
||api.google.com/papi/action$popup -this rule will stop the popup coming from that link
facebook.com###player-above-2 -this one will hide element(looks like a video player) on that page
~github.com,hello.world## -hmmmm haven't seen this one
@@||cnn.com/*ad.xml -this rule will whitelist that link on the webpage (@@ in front is whitelisting)
!||world.hello/*ad.xml -this will block it(! in the front is comment)
!@@||funceble.world/js -this will whitelist that js script (! in the front is comment)
yahoo.com,msn.com,api.hello.world#@#awesomeWorld -don't know
!funilrys.com##body -this will block element
hubgit.com|oohay.com|ipa.elloh.dlorw#@#awesomeWorld -don't know

[dnmTX]

Stay put,let me do some research on #@# rule cause i'm using AdGuard and haven't seen such a rule there

[dnmTX]

Ok... in the above example the rule #@# allows(whitelists) that particular element on the listed domains,so yes,all those domains are legit

[funilrys]

Okay let me implement this issue first with the current format will then review with you for all tests as those need some hotfix. Never thought about whitelisting 😹

[dnmTX]

I know,it's Java,more complex.Took me a while to get around it but i'm getting there

[dnmTX]

@funilrys make it simple.Everything that has || in front and ^ at the end stays.
Everything that has href in it stays(filtered of course to leave the domain only).The rest should be removed as it's rules that don't really concern any of us who will use the lists in hosts format.

[funilrys]

Yeah but if I do that, I'll invalidate AdBlock/filter list like https://github.com/MajkiIT/polish-ads-filter 😸