Meanwhile Mozilla is attempting to prevent users with administrative access from installing private extensions in the release version of Firefox, extensions for which the source code has not been handed over to Mozilla for signing.
This continues the long trend of browsers not trusting the underlying OS to actually protect users from soft-malware.
I feel like the people we consider ”regular users” to be really different.
Like we’re talking about a subset of the population that want’s to install their own private extensions or sideload someone else’s but doesn’t understand how to uninstall the release version and install the developer edition when prompoted.
It's also possible to make this setting configurable only from the browser UI, where you get a chance to educate and properly warn your users.
They also want to restrict the config option on Linux, where the adware problem you describe is not really present.
It would be more productive to allow others to learn about the issue and give us a chance to have a public discourse about it, where technical details can be explained and their threat model shared, and possibly allow for feedback by the community to shape the final implementation.
This site is full of threads talking about how friends, family, and coworkers’ browsers are riddled with badware that was installed behind their backs.
Not even Google is this heavy-handed, they allow installing local extensions in Chrome after users enable an option, although a warning is shown on browser restarts about the presence of external extensions, which can be dismissed.
Why not have sensible defaults, then educate and warn users about certain actions, instead of treating them like cattle?
And has been pointed out in most of these threads, the reason for not allowing people to circumvent some things is because that also allows malicious software to circumvent it.
Extensions loaded in about:debugging are not persistent, they are removed at the end of the browsing session.
> And has been pointed out in most of these threads, the reason for not allowing people to circumvent some things is because that also allows malicious software to circumvent it.
Yes, and the discussion mostly ends after people point out that malware if free to circumvent extension signing in any number of ways when it has root access on the device.
The threat model is flawed, or at the very least they're placing minimal security benefits above important user liberties.
Forcing extension signing by default is a great initiative, but we must be given ways to override defaults, guarded by administrative access and appropiate warnings, and be able to install local extensions in the browser we love, and that is the release version of Firefox.
But that's a big if. Extensions can be installed without root access, and can wreck a lot of havoc. I'd also dare to state the the organisation with the best insight into the extent of security benefits is the organisation that has data on the number and nature of exploits, i.e. Mozilla.
> we must be given ways to override defaults, guarded by administrative access and appropiate warnings, and be able to install local extensions
Guarded by administrative access sounds reasonable, and I think that's even already possible with Firefox ESR? I'd also consider an unbranded version of Firefox equal to the branded one, with the unbrandedness being an "appropriate warning". It does need to support auto-updating first, though.
Malware extensions is a real problem that affects literally millions more people than people who want to use the standard Firefox release with an unsigned extension.
And it's such a lazy thing to say "oh they can just replace the Firefox binary" when that's clearly a significantly more complex task than just writing an extension.
What's so secret about your extension code that you want to be able to distribute it to users without getting it reviewed by Mozilla?
You posted on that Bugzilla post but gave no details as to what your use case is.
Also why can't a user take the code for your extension, create an account on the add-on marketplace and privately get it signed for themselves. It's a multistep process but works.
Though its worth noting that we do not need reasons to expect our privacy being respected. And this issue is not just about me and my needs, but about user liberties, so please refrain from making it personal.
> Also why can't a user take the code for your extension, create an account on the add-on marketplace and privately get it signed for themselves. It's a multistep process but works.
No, it does not work, as the maintainer of Page Translator can attest.
https://github.com/jeremiahlee/page-translator/issues/26
This change is hostile to users and concentrates power in the hands of the browser vendor. Consider that they are making third-party extension stores impossible.
Ouch. I'm generally in favour of having add-ons signed, but I do have an extension that injects an external script into the browser and web pages - for a completely non-nefarious use case.
I do understand the security risk and why they'd block it, but I'm not looking forward to the day my extension is no longer flying under the radar and gets pulled.
Edit: Of note is that the Chrome web-store does allow my extension, but requires a multi-week review every time I upload a new version. Somehow they're fine with the remote code, even though it could change after their review.
Given that the leadership of the Mozilla Corporation is completely useless, users have the responsibility to raise these issues so we can push them to fix them.