Firefox 13 became Aurora this week. Firefox 13 is the first release that contains a significant amount of Snappy improvements. Some work got backported to 11 & 12, but 13 is where the bulk of work landed. I am very proud of how much we accomplished in 13.
Distribution of Snappy fixes
I suspect the unnamed bar mostly consists of bugs that did not get version-tagged, but those should also be FF13.
Smooth Scrolling
About 10 years ago Avi Halachmi embarked on perfecting scrolling via his smoothwheel extention. It may be that Firefox was the first browser to feature kinetic scrolling (with his extension). This Tuesday Avi landed some of those smarts in Firefox; see bug 206438. I’m sad that it took us 10 years to notice an awesome UX improvement like this. However, it is awesome to work with community developers who can focus on in on a particular feature and perfect it to their liking. To me this is what open source is about: enabling amazing people to perfect their software experience (and have everybody else benefit). Thanks, Avi!
My secret plan is to clean up a few primary user experiences and freeze the rest of the browser while the user is interacting via bug 712478. Scrolling is one such activity, watching video is another, typing, etc.
Peptest
There is now a graph for peptest results. Mark is figuring out how to deal with outlier, and considering integrating telemetry histograms.
Cycle Collector
Olli reduced CC lag observed after closing tabs like gmail, bug 734057.
WIP
We have fixed a lot of the low-hanging fruit and are working on larger projects at the moment. Going forward I’ll stick to mentioning projects when they get started, finish, or post builds for people to test.
Plans for Firefox 14
I’m very proud of Snappy fixes in Firefox 13, as a result I have high expectations for Firefox 14. The following are within reach:
- Get rid of remaining main thread SQL (including Local Storage).
- Make background maintenance tasks interruptable
- Fix themes to not ruin our startup
- Teach Firefox to suspend background processing during user interaction
- Ensure that graphics acceleration does not regress browser responsiveness
- Exit quickly
- De-jank animations
- Make our network cache more responsive
I can tell you that I am very happy to see this kind of progress for such a critical project for the product. Great work guys. But I found this **VERY** annoying as hell bug in firefox. All you need to do to reproduce it is load a page that is rather very populated with content and watch as the spinner slows down, freezez and or just spins slowly while the page loads all the content. Its annoying and does not exist in google chrome. If you can’t fix the spinner just replace it with a small bar above the tab, a progress bar has the right to slow down, but a spinner really does not. Here is what I mean by progress **BAR** look at the third tab:
http://www.stephenhorlander.com/images/blog-posts/incontent-ui/mac-extension-isolated.jpg
Its not at all a bad proposal, and it does give the user more info weather the page is almost done loading or not. And on top of that you don’t need to fix the slowing down part since its a progress bar.
That feature in bugzilla that generates bargraphs for bugs and their status is really good. It really gives a good visual representation of whats going on with the bugs.
Yeah that’s something we are aiming to fix. This falls under de-janking animation work.
Good to hear work is coming along. Any progress on the freezez that occure when loading flash games in facebook. Firefox can hang for up to 25 seconds during this time. Turning hardware accelarion off does help a little, but that defeats the point. Have switched to IE for gaming in facebook. Must say that IE 10 is awesome alround. Finding myself using that more often.
When will work resume on D2D performance?
IE10 is not just kicking your behinds, not just nuking your behinds of this face of the earth, but it is cutting your behinds off – throwing them into an alternative dimension and destroying that dimension.
I don’t have any data on IE doing anything to our behinds, however we are actively improving our acceleration architecture.