Ares2 wrote:Should we put them into a subsection? I think it's best if you move those rules as you added most of them so you know what section they belong to.
I've added another test I'd personally like to perform: https://hg.adblockplus.org/easylist/rev/66dfdd4a5821
||banner.$third-party
||banners.$third-party
(I asked this to Ares2 last night, when I switched subject in a PM again, something I do way to often, but well, to make sure I won't switch subjects here: he was fine with it).
I also spotted a lot of ||affiliate(s).somethi.ng in easylist_thirdparty. Can I add both ||affiliate.$third-party and ||affiliates.$third-party to EasyTest too?
Famlam wrote:I also spotted a lot of ||affiliate(s).somethi.ng in easylist_thirdparty. Can I add both ||affiliate.$third-party and ||affiliates.$third-party to EasyTest too?
A recent discussion among the EasyList authors resulted in general agreement to switch to a less aggressive approach when it comes to filtering words used in common speech (due to false positives, but also due to the ambiguity of those words and possible non ad-related usage in articles, commentary or even simply in negations like no-ads, ads-free, etc).
^^I remember in some filter-list (actually for Opera but still) I had a problem with a URL that read in part "white-house-advisor" because it matched for "house-ad"; I see where you're coming from there.
Michael wrote:Is the update interval actually valid? I didn't detect any headers to force the setting, but is one hour really appropriate? I would recommend daily as a minimum, even for such a small user base.
Wladimir's "expire" header interferes anyway, so I have reset the interval to the one that is approximately how long Adblock Plus will take in reality to auto-update it for now.
Michael wrote:Additionally, where should whitelists be located? In section separated by a comment?
I would place them just beneath the rule they are required for to see at one glance how many whitelists are necessary for it. Generally, if this list ever gets so hard to read that it requires additional sections, it's a strong indication that this list has failed it's purpose (which is: being a very time-limited test environment with the sole purpose of determining the best ways to improve the main lists in the near future).
About 1.5 years after it started, the Google-Analytics experiment has ended.
||google-analytics.com^$third-party and ||google-analytics.com/ga.js were decided to still be too dangerous. Two other scripts they serve were considered safe enough to be blocked. https://hg.adblockplus.org/easylist/rev/fd7a9aca7b9f