Hacker News Reader

Forked software is usually reasonable, but for me, the sheer complexity of the graphical, interactive web sadly begets an exception for modern browsers. Even if much of the work is offloaded to the corporations backing modern free/libre browsers (e.g. keeping up with new web standards and technologies), it certainly goes without saying that rebasing with upstream or even just applying the latest security fixes is no trivial task I would entrust any browser fork with minimal developer backing with.

This problem is especially worsened with the kind of user base that firefox forks tend to attract, from what I have seen. These users tend to ask very high-level questions e.g. "is waterfox more secure than pale moon" and will usually blindly switch from one fork to the next based on poorly-backed, unsubstantial crowd opinion. No userbase means waning support + maintenance. If you ask me, I think in the very special case of browser technologies, it would be more beneficial if the developers and users of firefox forks directed their energies towards making generally desirable changes in upstream.

Not just the weird habits of the users, either. I have way more confidence that Mozilla will fix issues in a timely manner than the one developer behind Waterfox, or the several behind Palemoon. Especially when the forks are depending on upstream to fix things for them. A few years ago there was a build of Waterfox that was substantially late because the one developer had exams.

It seems like the solution to this would be to have the build system reproducible from a base image with everything other than the code repo required to successfully kick off a build other than the code repo itself. Uploaded (or made available to upload yourself) to any of the cloud providers (or run locally in a VM), it would allow you to sync the repo and kick off a build for any supported architecture.

Dev not available to integrate a pull request and start a build? Download the appropriate build arch image, fire up VirtualBox, sync the repo (and apply a pull request if the dev hasn't had a chance to do that yet) and start the build script.

This doesn't entirely solve the problem, if nobody has submitted a fix yet, and you don't know enough to pull in the upstream fix and merge it yourself, you're at the mercy of some other user having that knowledge and making a pull request. It does close the gap somewhat though.

Are there existing projects to help get to this level of build reproducibility that can serve as a base to use? It would be awesome to know people are already working on making this easy to adopt.

THis is not the issue, I think. See: https://git.archlinux.org/svntogit/packages.git/tree/trunk/P...

The problem is that it is a giant codebase, you need to compile C++, Rust for hours. Even pulling the latest code is an ordeal.

It's not feasible on lowly workstations and garden variety cheap cloud VM builders, the performance is just crap.

I was speaking more to the general case of projects with small bus factors for deploying. And slow is moistly irrelevant to the point, which is making something possible which largely wasn't before.

When I ran Gentoo on my laptop, emerging Firefox took far longer than compiling even the Kernel. Around 90 minutes, I think.

I am using Fennec F-droid and some of the things that the developers did are unlikely to be welcomed in the main branch of the FF tree.

What's the difference between those branches? Any explanatory links, or care to elaborate?

"focused on removing any proprietary bits found in official Mozilla's builds"

Yeah Mozilla isn't going to like that.

Some of the changes are alternative defaults however. That isn't something one can just merge upstream since there are busi ESS and product reasons for Mozilla to reject certain changes

>and these forks claiming that Mozilla is this monstrous organization looking to invade your privacy

Do they claim that? I felt like most of the forks were made in good feeling, appreciating the work that Mozilla has done, but customising it in their own way/aligning vanilla FireFox with their views.

I used to do web dev in chrome but installed ungoogled-chromium after that scandalous update where they forced you to login; in fact, after installing said update, I suddenly found myself "logged in" with a gmail address that I never even created in the first place (something like mycompanyname[at]gmail[dot]com). My trust in Google was already crumbling, this creepy move was the final blow. So far, my experience with ungoogled-chromium has been great, but I must admit I'm not really following up on the project (devs, userbase, security updates, etc).

I've been using Waterfox since Mozilla changed their extension format and it's been incredibly stable. I used it for the Tree Style Tab extension. There's probably less reason to use it just for that anymore but vertical tabs in Firefox still isn't the best experience. I use Firefox with Tree Tabs at work with the extra user UI modification required to hide the top tabs and the sidebar headings. Only allowing one sidebar means when I accidentally open the bookmarks I need to mess around to get my tabs back.

No nee to switch to a fork for that. Tree Style Tab 2.x supports Firefox even after the switch to webextensions.

(Disclaimer: I work for Mozilla but not directly on the browser)

>How will further differentiating yourself from the masses (especially since Firefox is already a minority browser) help with privacy, tracking, and fingerprinting?

I don't think that's what fingerprinting is about. Whether you're 1 in 50k or 1 in 500k - it's not usable from a marketing/data perspective. Stuff like cookie retention & IP addr seem much more dangerous on this front than browser choice - both of those are more like 1 in 1 or 1 in say 20 depending on NAT.