A change I needed for my project

[michaelboccara]

Settings.velocityThresold needs to be modifiable, as it is in other languages.

I needed that because my world has no gravity, and objects without friction and with full restitution. Having velocityThreshold caused my bodies to stick to walls.

[dmurph]

I just made all settings non-final :)
You should see performance improvements as well on master.

[michaelboccara]

Hi Daniel,

Great! Can't wait to see, after work ;-).
My game can add infinite number of bodies (upon user request), and they
bounce without loss of energy for ever, so I'm always challenging
performance. Therefore any performance improvement will increase my number
of supported bodies.

BTW, I saw that you tried the JNI option. Did you really see no improvement
comparing to pure Java ?

On Wed, Feb 12, 2014 at 3:09 AM, Daniel Murphy notifications@github.comwrote:

I just made all settings non-final :)
You should see performance improvements as well on master.

Reply to this email directly or view it on GitHubhttps://github.com/dmurph/pull/1#issuecomment-34828395
.

[dmurph]

It was a somewhat large disappointment :( I should probably double check my numbers though. Most of the issue came from needing the callbacks to java (proceed in the query or not), and crossing the java-c barrier is rather slow. I had some ideas for removing that for the most common cases, but I haven't gotten around to it yet.

If you want to see some cool stuff, check out the liquidfun branch :)

[michaelboccara]

I wonder whether GWT+V8 would be faster than JNI ;-)

On 12 February 2014 20:53, Daniel Murphy notifications@github.com wrote:

It was a somewhat large disappointment :( I should probably double check
my numbers though. Most of the issue came from needing the callbacks to
java (proceed in the query or not), and crossing the java-c barrier is
rather slow. I had some ideas for removing that for the most common cases,
but I haven't gotten around to it yet.

If you want to see some cool stuff, check out the liquidfun branch :)

—
Reply to this email directly or view it on GitHubhttps://github.com/dmurph/pull/1#issuecomment-34902223
.

[dmurph]

I've seen a couple of people do comparisons, apparently javascript is still slower.