Google's 7-year slog to improve Chrome extensions still hasn't satisfied developers Google's overhaul of Chrome's extension architecture continues to pose problems for developers of ad blockers, content filters, and privacy tools. This story starts in 2019 when Google detailed its plans to improve extensions’ security and privacy features with a project it called Manifest V3 (MV3) that changes the way …
Google’s relentless pivot towards serving big corporations - manifested in the crippling restrictions of Manifest V3 (khm) - is nothing short of a betrayal to the very developers and users who built Chrome into a 90% market leader. This isn’t just about limiting extensions; it’s about Google prioritizing its own profit margins and ad revenue over innovation and genuine user interests. By deliberately stifling independent development and enforcing policies that favor corporate giants, Google is effectively sacrificing its original ethos for cold, hard cash.
In contrast, Firefox continues to champion user privacy, open-source innovation, and true developer empowerment, making the shift to Firefox not only a smart alternative but a necessary stand against Google’s corporate overreach.
And where has the vast preponderance of Firefox's money come from?
If Google is forced to hive off Chrome - or indeed forbidden to buy the default search engine slot - as some regulators are contemplating, it might prove difficult to fund two mainstream unbundled browsers. Of course, the regulators may not survive the current political maelstrom either, but Firefox is actually a beneficiary of the status quo.
Although the major part of Mozilla funding comes from Google by setting Google as the default search provider, something that general users actually appear to want, it's also a fact that most of this money doesn't go to the development of firefox.
So if Mozilla drops the ball with firefox development, I consider it highly likeable that it could self-fund via donations and what not.
See: https://assets.mozilla.net/annualreport/2024/mozilla-fdn-2023-fs-final-short-1209.pdf
We’ll exactly
relating to stripping out googly redirection links, I read “We can't do it the correct way because when Google engineers design the [chrome.declarativeNetRequest API], they fail to think of this scenario”
I’m sure the Googly engineers absolutely didn’t fail to think about it at all. I’m further sure that they deliberately broke it for obvious reasons
Chromium is the FOSS part derived from webkit. Crome is Chromium with added propiertary propiertary bits and blobs (+trademarks) that belong to google.
webkit is the FOSS part derived from konkeror. Safari is webkit with added propiertary bits and blobs (+trademarks) that belong to apple.
Thank-you. But where does this put Manifesto v3 ? Does it apply to Chromium, or Chrome ? Because by the look of it, it's difficult to imagine that Chrome and Chromium would use different extension systems. In which case that would mean that Google is also in charge of Chromium, which gets us back to square one : how did that happen ? How come that a (US) megacorp was able to hijack a (European) open-source project ?
Google is in charge on Chromium in as much as most of the developers working on Chromium are google employees. Recently, the linux foundation set up a comitee to steer chromuim ( https://www.theregister.com/2025/01/10/google_linux_foundation_chromium/ )
From the article
«According to Sreekanta, Google last year made more than 100,000 commits to the Chromium code base, representing about 94 percent of the contributions. Google's hope is that other organizations building their browsers on Chromium will step up their contributions.»
Of course, Manifest V3 is a Google Initiative, If anyone wants "chrome with manifest V2 too", they are free to fork Chromium, but so far, no group of developers has taken on the chalenge, mostly because maintaining a browser is hard AND costly, and no group hast the £€¢¥$ and ammount of developers necesary to go at it. And we are talking about chromium especifically, not chrome. Brave browser, for example, is chromium based, with a few extra twists, but is not a "fork of chromium", ditto Microsoft's "Edgium"
Same thing with Android™, AOSP is the Open source part of Android™, Android™ is AOSP + a bunch of propiertary blobs, and services and trademarks that belong to google. Most developers working on AOSP are Google employees, so Google can steer the project any way they see fit. If some group wants to steer AOSP in a different direction, they can fork it, so far no group of developers have done so in a significant way. Many a phone maker (especially in china, where Google does not operate) releases AOSP phones, sand google's propiertary bits, and with propiertary bits and services of their own, but no one hass "forked AOSP".
Same thing with WebKit, most of the developers working on WebKit, are apple employees, so Apple can steer WebKit any way they see fit...
And all of this is allowed by the respective licenses, in particular the GPL, as long as you release the GPL source, you are golden. In 99% of the developers of the FOSS project work for you, guess who is the "community" in the first place, and who will steer the project...
In my country we have a saying: "The Golden rule is: S/he who has the Gold makes the rules"
Very informative, even if disturbing. In my country we'd say that the one who pays the band decides the music.
I still have some reservations : where is the source code repository ? https://source.chromium.org/chromium and a whois lookup says that it belongs to Google. And when I follow the link to get the source code, it says that don't do a git-clone but follow some instructions ... which don't lead nowhere. Which means I cannot get the source code with the prefered method through the link that they advertise. How is this not a violation of the GPL if I cannot get the source-code from the main online repository ? (you might argue that my Firefow browser is tied down quite strictly, but that won't save their arse from the GPL).
I then go to a get-the-code page I found somewhere (NOT through the repo) which leads me trough hops and loops to tell me I have to use some Nix shell without which it won't work. After some digging I find the nix.shell ... and now tell me with a straight face that this is not obfuscation !
Google is clearly a monopoly, and with this Manifest V3 clearly an abusive one. So why doesn't anybody go after them, like the EU for example, and force them to leave Chromium development ? Would be quite easy to do I think by forbidding them to have commit rights to the source code, and they must go through an independent developer organisation. The EU went after Microsoft to force them to unbundle InternetExplorer from Windows, but we leave Google violating the GPL by obfuscating the source-code they got from KDE ?
maintaining a browser is hard AND costly
why ? If they stopped messing with it it could be quite stable. And Konqeror still works today.
NOTE to TheRegister journalists (Liam may-be ?) : would you make an investigation about what it takes to actually compile Chromium from source, only source, and what Richard Stalman thinks about it ? Is it even possible ?
NOTE to self : when reading the GitHub instructions about how to build, it says :
Run the hooksOnce you've run install-build-deps at least once, you can now run the Chromium-specific hooks, which will download additional binaries and other things you might need:
$ gclient runhooks
and you have to use a tool called GN which says:
For Chromium and Chromium-based projects, there is a script in depot_tools, which is presumably in your PATH, with this name. The script will find the binary in the source tree containing the current directory and run it.
How is this even remotely compatible with the LGPL of khtml ?
Blink (the beating heart of chrome) DOES NOT have to be compatible with the LGPL of khtml.
If anything, it has to be compatile with the (L)GPL of Webkit.
We do not know (and I am not going out of my way to find out, I'd rather play Robocop: Rogue City) if apple linked to khtml (and therefore the lgpl applied) or if they lifted the soruce code wholesale (and therefore, the GPLv2 applied).
Remeber blink is a fork of webkit, not a fork of khtml. As long as bllink is compatible license wise with webkit, all is honki-dori, it is webkit the ones that have to worry about being license-compatible with khtml.
Finally, I suspect that, if google were in serious violation of their licenses obligation by obfuscating their source code:
1.) the linux foundation would not have got an agreement with them, even with all the money they put on the table.
2.) Small companies such as microsoft, Brave and Opera would be protesting, as their browsers are based on chromium.
3.) Other small companies sucha as valve and Epic games would protest also (as their launches are electron apps, and therefore, based on chromium).
4.) Distros like GalliumOS, Deepin Linux, ZorinOS, eLive, Lubuntu, Knoppix would complain about it, because they ship chromium by default.
The fact that none of these four canaries in the mine has signaled anything, canaries that dela with chromium day in and out, much more than we do, and have entire legal departments looking into licenscing stuff, makes me heavily suspect that google complies with the licenses obligations of chromium well enough.
Thank-you for your extended answer. What I take from it is that the entire reasoning rests on : "if Google were breaching the (L)GPL or the WebKit license (a mix of LGPL and BSD) then someone would have noticed it " ... and would have taken on to fight the megalodon. I hope you don't mind if I don't find that very convincing.
PS: of course I have Chrome since I have an Android phone
à propos the WebKit license :
Licensing WebKitWebKit is open source software with portions licensed under the LGPL and BSD licenses available here
what portion is licensed under what license ? This is again obfuscation.
So this is my understanding of it : Apple's business model is to be the underdog. It wasn't always like this, during the Apple ][ and Macintosh period they were the best. But during the 1995-2000 there was much competition – Win95, WinNT, OS/2, Be OS, AIX, SunOS, HPUX ... – and Apple nearly died. They were saved by Microsoft who needed a valid competitor to not be dismembered like Southern Oil, and prepped-up Apple with an MS Office version. Since then Apple's business model is "Think Different" , which means they are the underdog of their master. This master was Microsoft, now it's Google. Apple's business model – and revenue – depends on a single monopolist provider, against which they can pretend to be different. Apple cannot afford to have many competitors, like BlackBerry or Nokia.
So Apple nicked the (L)GPL licensed khtml source, put some obfuscated license on it, and let Google derive an even more obfuscated version to rule the world, where Apple could pretend to be clean. A classical good-cop/bad-cop arrangement. Nothing-more, nothing less. Simple US imperialistic stuff.
The GPLv2 (i.e. the relevant license here) clearly states that if a company distributes programs in executable form, that company must give the source code to those people to whom they gave the executable, NOT to everybody, even people who are not customers... (i.e. world + dog)
¿You do not have the chrome executable? Then Google has no obligation to give you the source code, as you are not their customer.
¿You do have Chrome installed in your machine? The use that to retrieve the source code, no hoops.
Do not blame me, I did not write the license. Blame Stallman. Read the license and you will see.
Regarding browser maintenance costs and hardness, there are various types of changes that a browser needs:
1.) Self inflicted (for example, changing the UI for the sake of change)
2.) Evolving standards, the W3 foundation changes/expands/deprecate standards and functions from time to time,and some other standards bodies do to (like MIME, MPEG, et al).... someone has to implement said changes.
3.) Correcting bugs and security vulns.
4.) Soemone figures out a better/faster/more correct way to do things.
5.) The world around the browser changes (win11 2xH2 is released, new MacOS is released, new Ubuntu LTS is released....)
Just to name a few.
That is expensive. Even if you eliminate 1 and four (and I would severely argue against eliminating 4), 2, 3 and 5 still there, and are still hard and expensive...
But also entirely useable in Lynx... Boy, those were the days, when you could knock up an entire website using just a basic text editor in an afternoon, and have it all fit comfortably on a single floppy disc because it wasn't so dependent on all the MBs of non-HTML content that now seems to be utterly essential for even the simplest of pages to load.
I use Firefox ESR. And will keep using it until:
1.) It dies
2.) Another, better browser comes along.
But browser land is very dull, as there are only 2 rendering engine families:
The khtml family (khtml->WebKit->Blink)
The Gecko Family (Gecko -> goana && gecko to -> servo)
The other rendering engines are rounding errors.
The only browser that piques my interest is LadyBird (https://ladybird.org/) . It brings a novel architectural and security concept, as well as a new rendering engine, and license diversity to boot (BSD instead of the GPL of the others).
For now is to early to tell, but I'll be keeping a close watch, it'd behoove y'all to do so too.
as I use Firefox since 2009, Firefox ESR as soon as it was available.
I also moved a few select non-computer literate people that do trust me (with ad-blocks to boot) and the experience was transformative for them.
As a matter of fact, I also have chrome (for chromecast related duties), and the difference between ad-blocking solutions is notable to say the least.
Having said that, is true that Manifest V2 had serious performance and security flaws that needed addressing. IMHO (and YMMV) the changes in Manifest V3 were done mainly to address those shortcommings, the extra difficulties for ad-blockers and privacy tools were a "possitive side-effect" from google's point of view. The "cherry on top" if you will.
Whatever minor complaints I may have about Firefox, they are overcome by the ability to install extensions like uBlock Origin and Privacy Badger (and NoScript for when you absolutely, positively just need site annoyances to just Go Away). Every so often, I need a Chrome-based browser to accomplish something, but it's really quite rare.
Same here. When absolutely necessary (like M$ Teams or Google-docs, some people don't seem to care), I use ungoogled-chromium with all permissive settings – open-bar for Google and Microsoft – but for the session only, and when I quit it destroys all cookies and history. Well, at least that's what I hope it does.
To date, I can count the number of times I needed a Chrome-based browser on the fingers of my fifth hand.
- If it doesn't work in Seamonkey with NoScript, I try Pale Moon
- If it doesn't work in Pale Moon, I try Firefox
So far only a very few sites have been so horrific as to even need loading in actual Firefox. (Not that several haven't committed other atrocities like hijacking middle-click ad reinterpreting it as left-click.)
I don't think so. Chrome (its rendering engine) has gobbled up quite all the browsers market. I don't mean to say that it's impossible for it to become irrelevant, but it won't happen in the near future. Maybe someday Microsoft will switch to a different rendering engine, and this could make a dent.
It strikes me that this is a situation that would be best resolved with a class action lawsuit against Google.
All users of Chrome and an adblocker are victims of Google's malicious behavior, and the remedy is simple, stop Google from abandoning Manifest V2 until Manifest V3 has adequate functionality for adblockers. Or just stop Manifest V3 permanently. There's got to be numerous grounds for the suit, off the top of my head restraint of trade and unlawful abuse of monopoly power come to mind.
I mean, sure, I ban Chrome from everywhere I can, but users keep installing it, usually with a claim of "but this website I need won't work with Firefox" - except 99% of the time when I demand they show me it works fine, and I delete Chrome, and I come back to find they've reinstalled it. I've even got a bit of "user training software" that crashes Chrome at a periodic interval, and I'm thinking of adding functionality like automatically deleting Chrome's bookmarks with every crash. But I'm only one guy, and that cancer is all over the place.
I feel your pain. A long while ago, I installed firefox (with all the ad-blocking goodness) in some very underpowered machines and told the duhsers:
Do not install chrome.
Do not instyall antivirus, windows defender is doing the work.
¿2 months later? Chrome with the google toolbar + 3 antivirus competing for supremacy.
I ask ¿how do you feel your machine? The answer was "sligthly slower than when you first gave them to us"
I wonder why....
This is why I use Brave. It's Chrome-based but its filtering technology, Brave Shields, is compiled into the browser rather than being an extension, so the move to Manifest V3 will have essentially no effect. They can import and keep up to date with EasyList, AdGuard, etc. so the filtering looks just like I'd get on another browser with an adblocker installed.
https://brave.com/blog/brave-shields-manifest-v3/
int main(enter the void)... The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
