[Bug 1945950] Re: Enable WebExtensions Experiments in the packages

[KOLANICH]

>why do you think the Ubuntu package should enable it?

GNU/Linux is a free software ecosystem, and reserving certain features
for exclusive use of the monopolyst is not freedom.

I have developed 3 addons which would likely highly disapproved by
Mozilla (in fact, they have disabled `about:config` in their Fenix ans I
remember several that several times Mozilla employees were posting
comments on their Bugzilla in favour of phasing out about:config).

1. An extension exposing the API to manipulate about:config prefs to other WebExtensions. Like in old good XPCOM exts times. This API was trivial to implement, but Mozilla hasn't implemented it on purpose. https://github.com/KOLANICH-WebExts/experiment-config.xpi/tree/master/experiment
2. A WebExtension to semi-automatically apply prefs from Arkenfox - a set of Firefox settings intended to improve security and privacy. I guess it was the real rationale about non-exposing the API, because extensions like this one would make disabling some antifeatures (i.e. telemetry, i.e. disallowing privacy preserving exts (ΝoScript, μMatrix, μBΟ, Canvas Blocker) operation on Mozilla own websites, ads built into the browser...) introduced and enabled by default by Mozilla easier. https://github.com/KOLANICH-WebExts/experiment-config.xpi/tree/master/arkenfox
3. An extension exposing the API to parse JavaScript source into AST. Of course, one can use some existing pure JS parser (the AST format of SpiderMonkey has become a de-facto standard for JS AST and pure JS libs (and their python ports) parse into the same objects hierarchy), but it is a big maintainment issue (one has to keep these stuff upgraded as JS gets new features. The parser within Firefox upgrades with Firefox.) It is also a performance and bloat issue. The excuse they have given about not exposing it is that they want to be free to change it as they wish. IMHO - it is not an excuse, they can just expose that and document that they don't guarantee that it is stable, and let the devs to decide which is more suitable for them. https://github.com/KOLANICH-WebExts/experiment-parse.xpi

>Do you know of the rational for upstream to not enable it?

Mozilla's rationale is that Firefox is their brand and they don't want
users to be able to disrupt the browser operation and then blame
Mozilla. They have named the same rationale for not allowing unsigned
addons. For Windows they provide some unbranded installers where signing
can be made non-mandatory, also in the branded Developers` Edition these
prefs work (on Windows machines I use D. E. exclusively). But on Linux
platform they have ceased to pursue signing in release builds pretty
long ago (though IDK if they have continued and if it was the cause of
https://bugs.launchpad.net/bugs/1945947, or if it was just some change
in the code of fx that just broke the patches. I have googled the news
and found nothing, searching in commit logs in Debian repo also has
resulted in nothing.) I **hypothesize** it is because they have very low
market share, and if Linux distros would report IceWeasel instead of
Firefox, it would cause their browser share further be dropped, and so
the revenues search engines pay them. So I **guess** they are not
currently in position to dispute allowing unsigned addons and Mozilla-
monopolized features in the builds built by distros, because if they
insist, the distros instead of satisfying Mozilla ultimatum can just say
"OK, we just rebrand back to Iceweasel", and this would harm Mozilla
more than the distros.

-- 
You received this bug notification because you are a member of Mozilla
Bugs, which is subscribed to firefox in Ubuntu.
https://bugs.launchpad.net/bugs/1945950

Title:
  Enable WebExtensions Experiments in the packages

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1945950/+subscriptions