The State Of Various Firefox Features

[Luke]

I'm just glad EME support will include a GUI to disable it-and disable Firefox making connections to download the binaries. I have gone out of my way to disable all "phone home" URL's in about:config by setting blank strings for them. At least with the GUI we "freetards" that won't pay for netflix and cable don't have to worry about being logged by Hollywood if we don't have the knowledge to use about:config to block the paid stuff. I'll still kill the URL's in my own browsers simply as an extra precaution.

[hourwatch]

2) The GTK3 bug (https://bugzilla.mozilla.org/show_bug.cgi?id=627699) Hasnt been updated in basically a year.

I think that is not true. This bug has been in heavy development as you can see in the modifications history. It only depends in a few bugs now.

https://bugzilla.mozilla.org/show_activity.cgi?id=627699

[Delgarde]

I noticed with e10s opening new tabs and loading pages is really slow.
But I think it got a bit faster, it was worse before.

My understanding is that it's not *actually* slower with e10s, if you measure it. But it feels slower... something to do with how quickly the tab itself opens vs when the content is ready to display, causes it to be perceived as slow by the user. There was a blog post by one of the Firefox developers last week, talking about the subject, and how they're trying to improve both actual and perceived time.

[Rakl?dder]

Everything except for MSE and WebM VP9 works great.

[Delgarde]

While that's what we want to happen, the current C++ codebase for Firefox and Chrome have had an enormous amount of work put into optimizations and efficiency. I would love to be wrong, but I find it hard to believe that even an extremely well-done clean sheet redesign will be able to improve on performance any significant amount.

Maybe, maybe not. They've done a lot of optimization work, but that work is still restricted by limitations in the design, things they can't easily fix in the existing code base without either huge effort, badly breaking compatibility with extensions, or both. And knowing what those limitations are, a clean redesign may be able to make big gains by avoiding repeating them.

[Delgarde]

2) The GTK3 bug (https://bugzilla.mozilla.org/show_bug.cgi?id=627699) Hasnt been updated in basically a year.

More accurately, no new comments have been added to the bug for almost a year. But the actual bug is just a tracker entry for all the related tasks that need to be completed, and those related tasks are slowly being checked off the list (the tracker was updated just a few days ago for updating dependencies and things).

[Marc Driftmeyer]

The last 3 releases have steadily gotten less responsive and more resource intensive.

[zouboulou]

e10s still has plenty of bugs, but when its off nightly is rock solid. been using it for years, updating daily and it rarely broke anything, certainly hasnt in the past few month.

As far as sandboxing goes there is seccomp support and it seems chroot and namespaces are coming soon according to public bugs/commits.

[Serge]

Thank you for taking the time to write that up, Eric. Did you get a chance to check out progress on Shumway, Mozilla's Flash player?

[CrystalGamma]

37.0.2, YouTube.

That explains it, then. I've been using Nightly for over a year and haven't had any crashes from Youtube in a while. (In fact, barely any crashes whatsoever)

If this is a problem, I can wholeheartedly recommend Nightly, where it works at least on Intel and AMD GPUs (OSS drivers) with gstreamer-libav and all plugin packages ... I've never had any issues with it that weren't also in stable.