Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Test with Chromium backend #8
Hello @a-m-s, more work to do ;-)
How to get Chromium backend:
What to focus on:
The second recipe running the
There's no trouble with the flash player or HTML5 opt-in (which seems to have stopped working in the webkit version).
Also, live TV and radio now works, where only catch-up programmes worked before. It's also much less crashy.
I can't seem to get the ADK recipe to work though? I installed the master branch, ran the ADK shell, set the environment variable, and ran
I'm not sure how to best compare the memory usage when chrome has so many processes, and many of them probably share some of the apparent memory usage. Both use a lot; gigabytes of virtual memory.
The performance difference between WebKit and Chromium is vast. The WebKit version barely works. It often hangs up, spinners spin forever, and live streams don't work at all. The Chromium version just works, and works fast.
The only usage issue I've noticed so far is the full-screen thing. The integration appears just fine.
Launch Chrome with only a single tab with BBC iPlayer and then sum up all Chrome processes in GNOME system monitor. Then sum up all Nuvola's processes (e.g. nuvola-app-whatever, ValacefSubprocess, eu.tiliado.Nuvola, nuvola, etc.).
Virtual memory does not really matter. If you use
BBC World Service didn't work for me. I'll try again.
I've still not found a satisfactory way to read the true memory usage for a multi-process application, however, the sum of the "memory" sdfs from the system monitor are as follows:
eu.tiliado.NuvolaAdk + descendants: 294MB
In both cases the browser was on the same page, with the same video paused at around the same spot.
However, it's probably an unfair comparison because chrome is running some extensions. Besides which, I'm pretty sure that the processes within the application will be sharing some of that memory, so some of it will have been double accounted.
I've not given numbers for the WebKitGTK memory usage because the website did not work well in that browser, and how efficiently it didn't run it is not that interesting.