e3cead085a | ||
---|---|---|
_locales | ||
assets | ||
css | ||
dist | ||
doc/img | ||
img | ||
js | ||
lib | ||
1p-filters.html | ||
3p-filters.html | ||
LICENSE.txt | ||
README.md | ||
about.html | ||
background.html | ||
dashboard.html | ||
manifest.json | ||
popup.html | ||
settings.html |
README.md
µBlock for Chromium
See Change log for latest changes.
An efficient blocker for Chromium-based browsers. Fast and lean.
The above screenshot was taken after running my reference benchmark plus a bit of random browsing. All blockers were active at the same time, thus they had to deal with exactly the same workload. Before the screenshot was taken, I left the browser idle for many minutes so as to let the browser's garbage collector kicks in.
New myths flying around
- "µBlock does not support element hiding".
- Yes it does. If you don't believe it, try entering
twitter.com##body
in the "Your filters" text area and see what happens when you visit twitter.com...
- Yes it does. If you don't believe it, try entering
- "The memory usage isn't actually ABP's fault, EasyList is like 40,000+ lines of rules that all have to be parsed by ABP".
- Hum, µBlock also parse EasyList. And EasyPrivacy. And Malware domains lists. And Peter Lowes's Ad server list. Out of the box. Yet uses less than half the memory of ABP.
And for a fun memory test, you can try yourself the infamous vim test, once with only ABP as the active extension, and once as only µBlock as the active extension.
Regarding reviews in various web store
Some articles about the origin of the source code behind µBlock
Forums
Installation
From the Chrome store, or manually.
To benefit from the higher efficiency, it is of course not advised to use an inefficient blocker at the same time. µBlock will do as well or better than the popular blockers out there.
Documentation
I think it is pretty obvious, except for this I suppose:
The big fat power button is to disable/enable µBlock for the specific hostname which can be extracted from the URL address of the current page. (It applies to the current web site only, it is not a global power button.) The state of the power switch for a specific site will be remembered.
Benchmarks
I ran my reference benchmark to compare against other popular blockers.
Results -- figures are 3rd party / all:
µBlock 0.1.0.4
- Domains: 66 / 67
- Hosts: 117 / 171
- Scripts: 239 / 321
- Outbound cookies: 8 / 42
- Net requests: 1,035 / 1,877
Adblock Plus 1.8.3
- Domains: 72 / 73
- Hosts: 124 / 177
- Scripts: 243 / 328
- Outbound cookies: 8 / 44
- Net requests: 1,041 / 1,913
Ghostery 5.3.0
- Domains: 83 / 84
- Hosts: 140 / 211
- Scripts: 239 / 343
- Outbound cookies: 17 / 57
- Net requests: 1,046 / 1,930
Adguard 1.0.2.9
- Domains: 89 / 90
- Hosts: 145 / 217
- Scripts: 262 / 349
- Outbound cookies: 18 / 68
- Net requests: 1,064 / 1,904
Disconnect 5.18.14
- Domains: 95 / 96
- Hosts: 163 / 239
- Scripts: 283 / 381
- Outbound cookies: 18 / 74
- Net requests: 1,087 / 1,989
No blocker
- Domains: 380 / 381
- Hosts: 566 / 644
- Scripts: 490 / 592
- Outbound cookies: 245 / 315
- Net requests: 1,950 / 2,871
The figures show the number of requests allowed, thus lower numbers are better. The point is to show how many 3rd-party servers are hit on average after running the reference benchmark (three repeats in the current instance).
The less hits on 3rd-party servers, the better. All blockers were configured in such a way as to compare apples-vs-apples:
- µBlock: out-of-the-box settings -- no change.
- Adblock Plus: "EasyList", "EasyPrivacy", "Malware Domains" checked. "Acceptable ads" unchecked.
- Ghostery: "Advertising", "Analytics", "Beacons", "Privacy" checked. "Widgets" not checked. "GhostRank" unchecked.
- Adguard: "English", "Spyware and tracking", "Phishing and malware protection" checked. "Social media" not checked. "Acceptable ads" unchecked.
- Disconnect: out-of-the-box settings -- no change.
About
µBlock is born out of HTTP Switchboard. All the niceties of HTTPSB have been removed, and what is left is a straightforward blocker which support EasyList and the likes, and also support host files. Cosmetic filters ("element hiding") are supported.
There is nothing more to it. But it does what popular blockers out there do, at a fraction of CPU and memory usage for the same blocking power.
Free. Open source. No donations sought. For users by users.