Is there a way to set the window width and height with the NW.js object independent of window border width and height? I need pixel-perfect sizes for window modes and it's hard to get it right when I need to take the window borders into account. Do they even stay the same pixel size for different OS'es?

This would be a useful addition to the platform behaviour I think. Enables us to do things like multiple wall jumps or upwards swimming more easily.

I think that post is out of date and nVidia allowed support for high-performance GPU in browsers eventually.

Yep, still it's good to have a solution for people with old-ish dual gpu's.

Still, the next version of C2 has a new "rendererDetail" expression which in Chrome and IE can return the renderer hardware info e.g. "NVIDIA GeForce GTX 660", which should make testing this easier.

That'll come in handy, thanks!

I found this:

http://superuser.com/questions/645918/h ... rd-optimus

[quote:3b9l7d3x]According to NVIDIA, chrome and firefox both have been disabled from using the Nvidia GPU by design.

[quote:3b9l7d3x]new nvidia drivers allow running with nvidia gpus, but you still have to run chrome with

--disable-gpu-driver-bug-workarounds

parameters. otherwise backend always uses intel gpu. you can check this on

chrome://gpu

page and see for yourself.

Changing the flag worked great, hopefully I'm not disabling some important bug-workarounds in the process <img src="{SMILIES_PATH}/icon_razz.gif" alt=":P" title="Razz">

Thanks! I found out I can just open nw.exe from the install folder and type in chrome://gpu there. And I see it sticks to using intel integrated no matter what I try to override it. Balls.

There's no doubt quite a few people with this dual-gpu thing in their laptops out there, and if those computers refuse to switch to the dedicated GPU for NW.js apps then that's a performance problem. I guess there's not much we or Scirra can do about that?

I have a dual gpu in my laptop that's supposed to switch to a dedicated Geforce chip for gpu-hungry apps, but I suspect NW.js is always using the integrated chip. There's an easy way to tell in Chrome with "chrome://gpu" so is there anything like that for NW.js too?

Wow! I made a post like five seconds ago and it's already being pushed down the list by spam.

Like — said.

There's no true fullscreen to be had with NW.js, which is a shame cos it impacts performance somewhat even with the low quality/layout scale trick.

In case you haven't tried, in Windows 7 at least you can just grab the window while it's loading and drag it to the secondary monitor before it goes fullscreen.

NW.js?

Node-Webkit.

In my research all I could find was a blog post by a Microsoft engineer saying "10,000 should be enough for anyone"

Wow. "64 kilobytes of system memory should be enough for anyone" - Bill Gates in the 80's. MS does not learn

We did not know about this before making C2, so it is very difficult to do anything about it. The only workaround is to move spritesheeting in to the editor instead of only upon exporting, but this is an architectural change which I think we will only be able to take in to account for C3.

That would make me very happy.

>

> PS: Here's another thing I don't quite understand with HTML5 games: 60FPS for a game and depending on the jank spike, anything that dips below that becomes unplayable? Seriously? I know that it's 30FPS that's supposedly playable, minimum at a reasonable 24. So how did it come to this?

>

I'd like to know what people are talking about when they're using that word- Do they mean things go wrong with the gameplay that make the games literally impossible to play properly, or do they just mean they really don't like the way the janking looks and they don't want to release full games with that sort of issue?

Like Aphrodite said, it's both. The lag spikes look bad *and* they affect gameplay. Collision checks fail, platform behaviours behave inconsistently etc.

Starred btw.