They are for privileged code, and searching for EXTRA_JS_MODULES.backgroundtasks finds the updater, which we disable with that flag, and telemetry... So things we don't exactly want, but that we disable also in other ways .
My main concern here is that background tasks could prevent firefox.exe from being deleted ASAP because it's running background tasks...
But I am not sure this is really a possibility.
One of the main reasons we don't want the update service is that it's privileged code.
However, background tasks should be run as a user.
Apart from this:
My main concern here is that background tasks could prevent firefox.exe from being deleted ASAP because it's running background tasks... But I am not sure this is really a possibility.
do they have other disadvantages?
Possibly disk leaks?
But from what I understand, the user won't have a Firefox process that keeps running forever. Rather, from time to time it does some maintenance, then stop.
Sounds like we should do an audit of the existing background tasks and verify we don't need them for anything. Hopefully we don't need them (though I suspect this will be a losing battle long term if Mozilla decides they are a super neat place to put functionality).
If we don't need them, I'd like the system's backend ifdef'd out if possible. Ideally we would trigger a build-time error if it gets used in the future but we'll see.
Otherwise, if you want to remove the updater, too, we could entirely disable the feature (--disable-backgroundtasks).
Please notice that we currently disable the updater's background task on Windows because of a Rust/mingw bug, that we could even solve, since we rebuild the compiler, for our builds.
So, the updater as background task would work only on macOS.
I think it is okay to have it, since it doesn't run anything with privileges?
I think it's disabled by defualt when mingw windows builds are detected (though I may be mis-remembering). Assuming this doesn't interfere with the updater let's go ahead and explicitly disable it and then revisit this for esr115 next year.
I was able to reproduce this issue on the latest Tor Browser alpha build (12.0a3). my test environment was a HP Laptop 15 – bs0xx running Windows 10 Home, Version 21H2 19044 x 64.
Steps To Reproduce
Start tor browser.
Go to "open application menu" at the upper right of the browser.
Click on settings and scroll down to Tor Browser Update.
Will the issue be closed after shipping the patch in 12.0a4
Usually they are, this won't be, because we want to follow up on it next year when switching to 13.0 based on 115.
We know that this feature is there, and if it's a feature or an antifeature for us.
So, we'd like you to go deeper with it, not just to say "it's there", but if it's really a disk leak, and what are the negative effects of it in general (if any), or if disabling it was overcautious.
what are the blockers been experienced ?
I think I don't understand the question. Anyway, the blockers were discussed in the latest release meeting.