uBlock Lite is already in testing.

https://chrome.google.com/webstore/detail/ublock-origin-lite...

EDIT: Discussion from 6 months ago

https://news.ycombinator.com/item?id=32911640

I wished it had been named "uBlock Neutered" or something that more strongly highlighted the relationship to uBlock Origin.

Or uBlock Chrome Edition. When it doesn't work as good as uBO on FF, let it be known whose fault it is.

Chrome will still be faster even if ublock doesn't block as much as FF does. In fact the new declarative filters will most likely be faster than the current ones.

In the same way that a car is faster if you make it lighter by removing parts that aren't technically needed to drive down the road :) Airbags, ABS, Crossmembers? Meh.

How much faster does it need to be? I have some complaints about FF, but not on the account of speed. I don't care about microbenchmarks, for me it works fast enough to not notice any lag.

Besides, I really don't care how fast it it when displaying ads and tracking me. I'd prefer not to see them, thank you very much.

uBlock Reduced Functionality exclusive to Chrome

uBlock Gutted imo.

"uBlock Crippled Edition" sounds good.

>CNAME uncloaking is essential

>Chrome doesn't have an API for that.

It rolled out a year ago.

https://bugs.chromium.org/p/chromium/issues/detail?id=115104...

>Chrome because Chrome doesn't allow extensions to use wasm.

Chrome doesn't allow loading remote web assembly in mv3 extentions. You can't run remote javascript either. It's fine if you bundle the wasm with your extension.

Loading remote wasm after installing should be allowed. Your extension would still need bindings in javascript for whatever the wasm is doing.

Just knowing the bindings isn't enough to be able to classify a wasm file as being malicious.

>remote javascript

When Google says "remote" they really mean "remote to Google". You cant execute any of your own code from local drive, everything must be controlled by Google.

How can a malware scan be effective if all the code can't be scanned? Imagine if antimalware only scanned exe files and ignored dll files.

There is a solution for Chromium based browsers:

https://github.com/brave/adblock-rust

but you will only have it built-in in Brave currently.

"Instead you need a gigantic list of filter rules, one for each domain, making loading web pages slower."

By not using recursive DNS, including open resolvers or other third party DNS providers, I avoid this problem. I do not need to maintain a list of domains to block,^1 I only need to maintain a list of domains that I actually want to access. I do not even make remote DNS lookups because generally I already have the DNS data stored (gathered in bulk beforehand), so the web becomes faster, not slower. One of the requirements for the web to suck the way it does, and for so-called "tech" companies to proliferate by collecting data and selling ad services, is that the web browser and webpages must be give free reign to lookup any domain name. IME over nearly 20 years of controlling DNS, restricting that ability makes the web instantly readable, even so-called "dumpster fire" webpages.

1. AFAICT no one actually does this anyway. They delegate the task to third parties, "blocklist" maintainers.

while I do some dnsmasq tweaks to raise minimum ttl on my terms, doing bulk gathering ahead of time and sinkhole anything not cached (I assume) is interesting. What's your refresh interval? source of domains you query in bulk (browser history / bookmarks?). Any place you described your process to read into?

I rarely have problems with IP addresses that stop working for a given domain name. And when these changes occur I want to know. Refresh interval is at least once daily. The format for the file that stores the data is simple

  domainname IP-address # timestamp

The file is sorted by timestamp. The proxy will use the first occurrence, and therefore the most recent data.

The source of domains is usually HTML, XML sitemaps or proxy log but it can be anything. I have small utilties I wrote to extract URLs and domains from any input. For example, I retrieve data for all the domains that appear on HN, excluding the ones in comments. I see there are close to 68K domains from the past 6-9 months.

do you mind sharing your setup for this? sounds like umatrix but on the dns level

While I have always used local authoritative DNS to avoid remote DNS queries, long before adblockers existed, other parts of the steup have changed over time. Currently, I use

(a) HTTP/1.1 pipelining and DoH servers for bulk DNS data

(b) localhost-bound authoritative DNS server for redirecting all domain names to forward proxy IP address

(c) localhost-bound forward proxy that stores domain-to-IP address mappings in memory

NB. Without modification, this setup is probably not going to work well for people who like to use graphical web browsers for everything. For recreational web use I am not using a graphical browser that auto-fetches resources. I am using a text-only browser. Thus, I only need DNS data for the domain names in the URLs I visit, not for ones found in web pages pointing to fonts, CSS, images, Javascript, and so forth. This makes the DNS data retrieval process very straightforward: one domain name for each URL. In the times I have tested recreational browsing using a graphical browser, it can be surprising how many web pages are quite readable without using resources from other domains, but for folks that want to view myriad web pages that are comprised of resources from a variety of hosts requiring numerous domain name lookups, bulk DNS data retrieval will require more work. In any event, there are many, many possible variations of a setup like this and the one I use is suited for me; I am not fond of the large graphical browsers and I rely on the proxy for a number of tasks. Other variations might suit other users' personal preferences.

Hope that helps.

But DNS has its own weakness, that's dealing with first party ads and anti-adblockers, which websites use a tons nowadays.