Much discussed last few days<p><a href="https://news.ycombinator.com/item?id=38369820">https://news.ycombinator.com/item?id=38369820</a><p><a href="https://news.ycombinator.com/item?id=38361758">https://news.ycombinator.com/item?id=38361758</a><p><a href="https://news.ycombinator.com/item?id=38361758">https://news.ycombinator.com/item?id=38361758</a><p><a href="https://news.ycombinator.com/item?id=38301801">https://news.ycombinator.com/item?id=38301801</a><p><a href="https://news.ycombinator.com/item?id=38298502">https://news.ycombinator.com/item?id=38298502</a>
Two questions I haven't seen addressed by any coverage of this change:<p>1. Will the ultimate removal of Manifest V2 support affect other Chromium-based browsers, or only Chrome itself?<p>If the support for Manifest V2 <i>isn't</i> removed upstream in Chromium, but only disabled in Chrome, then I would expect that we will end up in a world where other browsers (e.g. Edge, Brave, Opera) continue to allow the installation of Manifest V2 extensions, esp. from their own first-party verified-extension hosting platforms. So even if the Chrome Web Store also ceases to host Manifest V2 extensions, users of these other Chromium-based browsers could still get uBlock Origin from "Edge Add-ons" or "Opera Addons" etc.<p>2. Would it be possible for some random developer to put in a PR to <i>the upstream Chromium project</i>, to introduce one or more <i>Manifest V3</i> capabilities (new strings for the manifest.json "permissions" key) that, when added, would allow the extension to do all the stuff that Manifest V2 let extensions do by default, that uBO and others depend on: increased request-filter list size, async periodic network data-file updates, etc? Would such a PR have any chance of being accepted?<p>My own guess is that such a PR <i>wouldn't</i> be accepted, because I get the impression that the <i>nominal</i> goal of Manifest V3 is to allow V3 extensions to run under a streamlined extension "runtime" that has fewer hook-points into the browser runtime, and so fewer places where the browser runtime must call back to the extension runtime; where adding such capabilities would require adding all these additional hook-points and callbacks back in, which would defeat the purpose. Correct me if I'm wrong!<p>I would also guess that even if such a PR <i>were</i> accepted, Chrome would still disable the use of those capabilities downstream, and also reject any extension that used them from the Chrome Web Store. So at best, such a change would just mean that uBO and friends wouldn't be stuck as "legacy" Manifest V2 extensions, but could instead just be "modern" Manifest V3 extensions with a few capabilities that Chrome and only Chrome forcibly rejects.
Ok. Goodbye chrome. I will be switching to whatever privacy-focused browser allows me to keep not seeing ads.<p>I do wonder how long it will be before we see browser browsers, software that takes a browser instance and sanitizes it. Maybe chrome will continue as a daemon allowed to run inside a sandbox within a browser's browser that actually displays content to a human.
I fear this scenario in a few years:<p>1- non hacker users too starting to realize corporate friendly browsers like Chrome and many derivatives can't be used anymore for painless surfing, then flocking to Firefox.<p>2- corporations and advertising companies pushing for a new closed HTTP standard that requires their browser, or an old browser using a closed extension that doesn't allow adblockers when using a given service or page.<p>Open browsers work because they still connect to open web servers, and the industry already ruined the mobile environment by forcing users to run apps instead of navigating web pages (that is, installing a hundred application for a hundred services instead of just one that speaks a standard protocol); I have no doubt they'll attempt the same in the desktop world too. We have to fight to keep protocols open.
> Nevertheless, Firefox said it will adopt Manifest V3 in the interest of cross-browser compatibility.<p>The article makes it sound like Firefox will have the same ad blocker limitations as Chrome. The article fails to mention that Mozilla is implementing MV3 APIs in Firefox, but not removing the MV2 APIs like Chrome is:<p><a href="https://blog.mozilla.org/addons/2022/05/18/manifest-v3-in-firefox-recap-next-steps/" rel="nofollow noreferrer">https://blog.mozilla.org/addons/2022/05/18/manifest-v3-in-fi...</a>
The error is to use Chrome in the first place.<p>Use Firefox. uBlock Origin works best in Firefox:<p><a href="https://github.com/gorhill/uBlock/wiki/uBlock-Origin-works-best-on-Firefox">https://github.com/gorhill/uBlock/wiki/uBlock-Origin-works-b...</a>
> But Google has decided that block and allow are not that easily abused so it will allow up to 30,000 rules to be added dynamically.<p>Can someone give an example of what a good number would be? How many dynamic rules are currently used?<p>> Also, extension developers are limited in what regular expressions they can use, along with other technical limitations.<p>Does this meaningfully impact rules? Just curious.<p>> According to Firefox’s Add-on Operations Manager, most malicious extension that manage to get through the security review process, are usually interested in simply observing the conversation between your browser and whatever websites you visit. The malicious activity happens elsewhere, after the data has already been read. So in their mind, what would really help security is a more thorough review process, but that’s not something Google says it has plans for.<p>I don't see how one follows from the other. Attackers are using malicious extensions to eavesdrop on networks... therefore we need better reviews and not restricted APIs? I get why you might want to advocate for the latter over the former, but certainly it seems like restricting APIs also has positive impact.
Can anyone explain why there isn't a robust and thriving adblocking solution available at the router or OS level? Why are we all forced to grasp at the straws of the browser?
Google's master plan to ditch V2 might be more than just a power move. It's like they're trying to reshape the Chromium landscape, making it a tricky playground for any V2 loyalists. We're looking at a future where maintaining V2 support is like trying to keep a vintage car running in a world of electric vehicles. Nostalgic but increasingly impractical.
What's the long term impact on this?<p><pre><code> * More (Advertising) revenue for Youtube streamers?
* More people switching to Edge browser to use Ad Blockers?
* People using Pi-hole style Ad blocking?</code></pre>
I'd like clarification on something. I've spent an hour or two trying to figure this out to no avail, so I suspect many other people might be wondering the same thing I am.<p>Examining the Manifest V3 changes more closely (<a href="https://developer.chrome.com/docs/extensions/mv3/intro/mv3-overview/" rel="nofollow noreferrer">https://developer.chrome.com/docs/extensions/mv3/intro/mv3-o...</a>), and comparing/contrasting to what uBlock Origin themselves say about it (<a href="https://support.ublock.org/hc/en-us/articles/11749958544275-Google-s-Manifest-V3-What-it-is-and-what-it-means-for-uBlock-Users-" rel="nofollow noreferrer">https://support.ublock.org/hc/en-us/articles/11749958544275-...</a>), I can understand the cause of <i>one</i> of uBO's problems with V3... but not the other.<p>The cause of "Allow List Limits" is clear: uBO Lite will be forced to use declarativeNetRequest; and declarativeNetRequest imposes limits on the size of the ruleset you can "declare".<p>But I'm confused about uBO's point on "Ad Blocking Quality". It seems that Manifest V3 only restricts 1. the use of eval(), and 2. the loading of remote-origin scripts into the DOM and/or as service-worker modules. It doesn't restrict the use of remote-origin-loaded data files generally; which I would presume means that uBO would still be able to use its service-worker to periodically fetch and update its filter lists.<p>Is there some part of the way uBO uses these filter lists, that requires arbitrary remote code execution (and for which the only true substitute is burning in the lists locally?) If so: why, exactly? (Not a rhetorical question; I'm not doubting that they <i>do</i> need it. I just can't figure out where the need comes from, and I'd like to know!)<p>It might <i>seem</i> at first blush that the literal answer is this feature: <a href="https://github.com/gorhill/uBlock/wiki/Static-filter-syntax#scriptlet-injection">https://github.com/gorhill/uBlock/wiki/Static-filter-syntax#...</a> ... but it actually isn't, as you don't write actual JS to be eval()ed in these rules, but rather just name a function that's already burned into the extension locally as part of its "scriptlet resource library".<p>Is it instead, just the way that these rules get "baked down" into in-page logic? Does uBO compile the lists into a bunch of Javascript source-code, and then have the page evalScript() that code?<p>And if that <i>is</i> the blocking issue — and I'm still not clear that it is — then wouldn't there be other workarounds for this?<p>For example, sticking the generated JS code into a data: URL and then dropping it into the page as a <script> tag. Or even, at worst, swapping out feeding the page "JS source code", for feeding the page a (static!) <i>interpreter</i>, and then having that <i>interpreter</i> receive instructions as regular ol' data from the uBO service-worker? (Maybe that'd violate uBO's performance goals, I suppose? But it wouldn't have to do it on every page; only on pages that it knows from the ruleset can't be blocked entirely declaratively.)
So, manifest v3 is out there, and does allow some form of adblocking. are there any adblockers actually implemented with it, so i can see for myself what the adblocking performance is like?
Watch the already limited traffic I send to Google servers go even lower. I am not entangled in their mess. None of the faangs can touch me because I wasn't stupid enough to put my entire digital life in the hands of a company.<p>I'm looking ten or more years in the future, though. By that time, the Firefox/Chrome duopoly will be broken by alternatives that don't compromise between the user and business models.
I thought Brave made very little sense when it was first launched. Who would want an ersatz Chrome browser? Now it's starting to look very smart. Sure, it's open source so anyone can fork Chromium. But it does take a lot of sophistication to do it properly and add back the usability bits in a nice way. And if you could basically have Chrome but without losing the ad blocking, that starts to sound pretty compelling.
I've used Chrome for a decade and there's no way I pick it up if they actually commit to this.<p>That said, they already announced plans to do this once and then backed down for a year due to pushback around MV3.<p>Guess we'll see what happens.
Can we all finally make a concerted effort to switch back to Firefox? I get it, it was slow and bloated when Chrome initially came out and everyone switched.<p>Well, it's not slow or bloated anymore and Chrome is now officially evil. It's time. Don't just switch your own browser. Switch the browsers of all the non-technical folks that come to you for questions.
I like using Chrome and even went back to it after trying FF and Brave 2-3 years ago. But this would really make me look for alternatives, even if those are inferior.<p>It's not even about Youtube. I pay for Youtube Premium. But not being able to block third party cookies would be a deal breaker.
I like Firefox a lot (with Privacy Badger and uBlock) but sometimes the browser slows to a crawl. New tabs are slow to open, takes way too long to select and activate a text input field, and so on. A restart fixes this, but it's still annoying.<p>I think I might go back to Safari. I like the way it looks and it feels snappier all the way around.
I think this is a good opportunity to bypass chrome and use a standalone adblocker, if you're forced to use Google or Microsoft (ad|spy)ware as a browser.
The war against general-purpose computing is but one front in the class war.<p>Daily reminder that economics is a political theory, not a science, that capitalism is the most incidious form of oligarchy, and that the US is no longer a democracy (if it ever was).
In a few years:<p>LLM: Hi!<p>Me: Please develop a browser that is full HTML, etc, etc compliant<p>LLM: No problem... download the source here.<p>Me: Thank you, but could you please optimize it for speed?<p>LLM: No problem, done.<p>Me: I have only 5 more minutes, could you please write a version in Rust, and two more in Go and C++? Ah, and support Linux, MacOS, Windows, Android and iOS. Don't forget to use the native WebView in iOS.<p>LLM: done.<p>Me: Could you please do me a favor? Remove all ads.<p>LLM: done.
in wake of the EU ruling that youtube isn't allowed to block adblockers, i can understand this move.<p>imo the eu needs to stay out of this. its a competition between people trying to block ads and trying to force you to see them, which i see nothing wrong with.