I basically agree with you. I sincerely hope Ashley is still working on such a command. But, I see render cells as a stepping stone, not a replacement. I also see it as a companion to a 'reload layout/layer' command.

Yes, as you say, you would have to handle zone loading on a per-layer basis to get any benefit from render cells, but I still think it would be preferable in some circumstances.

For example: I envisage a zelda style game with seamless transitions between all zones, where tilemaps and immobile objects (scenery, houses, that sort of thing) are simply spawned in and left alone. Render cells should take care of any overhead they might pose. Then, all that has to be spawned in are active objects, which can all be on their own layer (or 2 or 3 layers).

On a slightly different note: what about the initial load? Let's say you have a massive layout with 30,000 objects in it. Now: what about mitigating the initial delay as those objects spawn in? And then, afterwards, we have to delete all but the ones right around us? Hmmm...

I don't think render cells + reset persisted objects is a replacement, but it's a start. Just brainstorming, but here goes:

Think about a 'Reload Layout Zone' command. One of the problems will be avoiding a 'halt' every time a chunk is loaded/unloaded due to creating/destroying all those objects. One way to handle this is some sort of lazy loading; a 'timeToSpawn' setting perhaps. OTOH, what if we didn't have to create/destroy so many objects in the first place?

One way to handle this would be to have another command: 'Reload Layer Zone'. This would only reload the objects on a particular layer, within a particular zone.

This way, stationary objects (tilemaps, decorations, that sort of thing) can be put on their own layer w/ render cells enabled, and simply be left alone. The only objects that have to be spawned in/out are 'active' ones.

Thoughts?

Semi-OT: Somebody ought to bump that old 'Room Object' thread; some good stuff on there.

I don't know, but I'd be interested to hear some opinions. I've been trying to test my current project by installing the apk into the android sdk emulator, and while I managed to get it installed, it won't boot my game. Planning to install andy later tonight and give it a try.

newt

Last time I checked, they're good with milk.

Seriously, though, I think there was a C2 plugin that let you set normal cookies, but they aren't secure ie., the user can easily see and modify them (my hi scor iz 9,999,999,999,007!!!).

So, that's a drawback.

Looked into this a little, sounds like an interesting project. Unfortunately, it looks like this would be difficult data to get at from a javascript framework.

There was a W3C recommendation back in 2009 for the Compass API, which included the fieldStrength attribute, a perfect fit for what you are doing.

http://dev.w3.org/2009/dap/system-info/compass.html

Unfortunately, it doesn't look like fieldStrength was exposed in any version of phonegap I can find. It doesn't look like CocoonJS implemented any compass features at all.

Apparently, not all devices that offer compass data use a magnetometer either. However, all iPads, and all iPhones since the 3GS, do have true magnetometers. Many android devices have them, but some don't.

Here are links to the low-level API documentation for magnetometer data on android and iOS:

Android:

http://developer.android.com/guide/topi ... rs-pos-mag

iOS:

https://developer.apple.com/library/ios ... index.html

Guillermo

Dunno. DL r191 and tell us.

Noticed that r191 included the following changelog notice:

[quote:1mflhdmz]WebStorage: worked around crash caused by the fact Chrome Web Store apps have removed local storage support. WebStorage now does nothing in Chrome Web Store apps. We have requested that Google add support back.

This seems like kind of a big deal.<img src="{SMILIES_PATH}/icon_e_confused.gif" alt=":?" title="Confused">

I can't test this myself ATM, but I was wondering if anyone else has noticed this or would be willing to confirm it?

Here's Ashley's chromium bug report; please chime in if you are seeing the same issue:

https://code.google.com/p/chromium/issu ... ?id=439027

Danwood

Opacity...not sure about that one. Angle, yes, that seems like it might cancel it out, because you are 'moving' the object, or at least, the bounding box is changing (unless the object has symmetrical dimensions). Seems like the best thing to do is disable the behaviors as objects move on/offscreen.

EDIT: whoops, guess Ashley stepped in while I was writing and gave a definitive answer.

Ashley

Does 'Is Onscreen' now take advantage of render cells? On a similar note, would there be a way to use render cells to provide a 'Pick by distance from point' command?

Danwood

For more info on render cells, Ashley has a new, in depth blog post:

https://www.scirra.com/blog/ashley/14/h ... cells-work

Tested again in Chrome (Normal/Canary), FF (Normal/Developer), IE11, and Node-Webkit.

The only platform that showed any 'jerking' was Firefox (of course).

At high speeds the movement of the ball begins to look less smooth because of the high contrast between it and the background, and the lack of natural motion blur.

On my monitor there is a prominent, distracting after image, but that is the nature of viewing fast motion on an IPS screen, not a glitch.

Unknow01

Have you tried some of the built in template games (flappy bird, space shooter, auto-runner, etc...)? Do they 'jerk'?

ErekT

Ah, I see, you are talking about batch processing itself. For gimp, there's BIMP:

Graphics look good. Congrats on your first game!