Weird, because since the latest versions screen updates seem to get suspended while Opera is not on the foreground. I can't really comment on power usage, because I tend to close apps whenever I'm done with them, but to me it seems that it should behave better with current versions, not worse.

Could be a mix of all other apps i have i suppose. I have incepted, and run fastern9 with it, Dunno if thats a problem or not, but there is no problem with the app in it self. Its just when idling my phone with it running background/forground it makes my n9 very warm and it empties battery within a couple of hours.

Is there a app for messuring power usage? Or a command in terminal that gives an hint. For all i know it could be something else reacting to opera and not opera reacting with something else. I have something called energy profiler , from meecatalog but that only give me how much current its use with everything.

Thanks for all the hard work on keepeng opera mobile labs updated for N9. Only thing I am missing is a option to lock screen rotation to landscape, to enable comfortable surfing in bed (what im doing right now).

Out of curiosity I tried to run the SunSpider JavaScript Benchmark from http://www.webkit.org/perf/sunspider/sunspider.html with the latest N9's Opera Mobile 12.10.1 .

Unfortunately Opera got stuck downloading the 1.3 MB Javascript benchmark blob code wrapped into an HTML page.

Does OM has problems loading huge JS files over the network?

Loading the benchmark HTML page locally as file:/// did work and gave a result of ca. 5000 ms, somehow a little bit worse than the N9's builtin webkit browser (4250 ms). The best performance had Firefox Mobile 14 with less then 4000 ms.
I don't find the difference significant that I would use only one of the browsers and JavaScript performance of course is just one part of the browser's user experience.

Are there some JavaScript performance tweaks in the latest Opera update as on a much slower non-overclocked N900 with a slightly newer OM (12.10.16) I got SunSpider around 6000 ms? Or is this just a matter of the compiler's version and the compilation options?

Best, M.

mick3_de,

I just tried the same test I got around 4,300ms...It ran without having to download the file and run it locally...

Screenshot:

I took another run and got now a slightly faster 4,750 ms. For Firefox Mobile 15 I got now 3,750 ms from the locally saved benchmark page.

With Sunspider not loaded maybe there is a problem with my wireless network connection. But from all the browsers Opera has the most difficulties successfully loading and running the benchmark.

I posted this in another thread, but it's probably worth putting here too - you can lock the screen orientation by temporarily disabling the accelerometer. For some reason initctl doesn't work when called from a shell script, but you can add an alias to shorten the command.

Add the following to ~/.profile:

alias om='devel-su -c "/sbin/initctl stop sensord && devel-su -c \"/opt/operamobile-labs/bin/operamobile-labs > /dev/null 2>&1\" - user && /sbin/initctl start sensord"'

Save, restart the terminal, and type "om". Enter the root password, and opera will launch locked in that orientation. When you exit, it'll unlock again. It's a hacky solution (note the nested devel-su ), but it'll do for now!

edit: a small change so it'll still work even if sensord isn't running to begin with:

alias om='devel-su -c "/sbin/initctl stop sensord > /dev/null 2>&1 || true && devel-su -c \"/opt/operamobile-labs/bin/operamobile-labs > /dev/null 2>&1\" - user && /sbin/initctl start sensord" && exit'

The path to opera might need changing depending on the version - I've stuck with 12.00.8 as it seems to be less buggy, which has the path /opt/operamobile-labs/operamobile.