Proceeding wildcards in rules

[ryanbr]

In Easylist (and Adguard) we'll use, we consider this "optimised" when using wildcard for general block rules with "/ /". Its been like this for the last 15yrs+ for us.

https://github.com/easylist/easylist/blob/master/easylist/easylist_general_block.txt
https://github.com/AdguardTeam/AdguardFilters/blob/master/BaseFilter/sections/general_url.txt

/adsmanager.nsf/*
/cgi-bin/ad/*
/cgi-bin/ads/*

But I've noticed proceeding wildcards in rules Isn't the wildcard redudant here? Seems less optimised.

Should be re-writen:

/ads/autopromo/*$media,xhr,redirect=noopmp3-0.1s,domain=deezer.com

Whats the faster/better method to use in uBO?

[Yuki2718]

IIRC there was a commit to remove such redundant wildcards on compilation.

[gorhill]

The purpose of the wildcard is to prevent treating the filter as a regex -- and trailing wildcard is the preferred approach in such case. I don't know why that commit changed this. It makes no difference performance-wise for uBO whether it's at the start of the end, though the syntax highlighter highlights the wildcard differently when it's at the end as it is considered non-superfluous in case of /.../ pattern.

[stephenhawk8054]

@garry-ut99 Yuki was talking about "tokenized" strings, which is different to this discussion (wildcard to prevent regex)

[ryanbr]

I'll close this since there is a general clarity regarding this.