regex-fied domain

[Yuki2718]

(transferred from internal discussion https://github.com/orgs/uBlockOrigin/teams/ublock-filters-volunteers/discussions/423 so some details are outdated)

krystian3w notified me that AG is implementing this: AdguardTeam/CoreLibs#1550

I think this will be game changer, for example we'll be able to address those pixsera crap like

@@*$ghide,domain=/img[a-z]{3,4}\.buzz/
/img[a-z]{3,4}\.buzz/##+js(aeld, , _0x)

The format is just my example idea, AG seems to be going with the same format as $path but we don't have to follow. Although I expect pixsera will then fully randomize domain, the regex will also be useful to other cases such as xhamster proxy sites or some pirate sites.

[gorhill]

@@*$ghide,domain=/img[a-z]{3,4}\.buzz/

gorhill/uBlock@b1de8d3

[gorhill]

/img[a-z]{3,5}\.buzz/##+js(nowoif)

gorhill/uBlock@8149847

[Yuki2718]

Ty for all your time and continuous effort!

[v1gnesh]

Sincere thanks for everyone's work on this... seriously!

I dream for a day when I can automatically convert my long list of dynamic filtering rules to static filtering rules.
And on that day, when all rules are static filters, baking uBO-core into some network-wide filtering web proxy will be very close to achievable.

[v1gnesh]

Why confused, @krystian3w ?

[uBlock-user]

[krystian3w]

works both:

! "negation" only com
/example\.(?!com)/##body > *

! "negation" normal 'regex'
~/example\.com/##body > * 

[uBlock-user]

No, I meant ~/www.google.com/,/com/##+js(nowoif) and yes it works as expected.

[MasterKia]

Added to wiki:

https://github.com/uBlockOrigin/uBlock-issues/wiki/Static-filter-syntax#domain
https://github.com/uBlockOrigin/uBlock-issues/wiki/Static-filter-syntax#to

https://github.com/uBlockOrigin/uBlock-issues/wiki/Resources-Library#general-purpose-scriptlets (Last point)

[uBlock-user]

This new syntax does not apply to static extended filters.

Remove this line, support has been added - gorhill/uBlock@8149847

[MasterKia]

Done.

[krystian3w]

See these maybe regex allow change from declined declined to fixed issue has been addressed : #957 / fyi @gwarser

[gwarser]

in uBlock's My Rules panel.

He asked for rule syntax.

[krystian3w]

But is "inactive" after 11 january 2023.

[krystian3w]

I guess "negated" regex don't allow ,. #2234 (reply in thread)

example.*,~/example\.([a-z]{1,2}|[a-z]{4,16})/##body > *

Workaround can be:

example.*##body > *
/example\.([a-z]{1,2}|[a-z]{4,16})/#@#body > *

by virtue of the fact that the parser does not have a noop state for cosmetic filters like other silinics.

---

and maybe |:

$all,~doc,domain=example.*|~/example\.([a-z]{1,2}|[a-z]{4,16})/

Lack of workaround to avoid mark allowed to load and block latter with ,important modifier, I doubt that it can be used with ,bafilter, denyallow or to=.

[stephenhawk8054]

Currently when I use to=~/youtube/, the filter automatically translates to to=~~/youtube/ and cannot exclude domain with youtube inside.

Steps to reproduce:

1. Add

*$frame,from=plainlight.com,to=~/youtube/

2. Access https://www.plainlight.com/how-to-embed-youtube-videos.html
3. The filter becomes

*$frame,from=plainlight.com,to=~~/youtube/

and cannot exclude youtube-nocookie.com domain: https://user-images.githubusercontent.com/66517106/232954843-e6dc927b-0e0c-47be-8cf6-b6f20d2a6ce7.png

---

If I use

*$frame,from=plainlight.com,to=/youtube/

then it blocks the youtube domain successfully as expected.: https://user-images.githubusercontent.com/66517106/232955214-c399a4e2-2478-4cb6-ab67-6185a3a6c38c.png

[gorhill]

Fixed with gorhill/uBlock@3dddf4c. Unfortunate that I published stable release yesterday before this was found.

[stephenhawk8054]

Confirm it's fixed now: