Ashley's Forum Posts

  • The end result is a single HTML file, so it's only one HTTP request.

    No way. I'd ask if they're joking, but it sounds like they mean it. Tell your client that is far slower to download, will use much more memory, and will be much slower to load. I am not kidding, that is a whole new level of deoptimisation beyond turning off spritesheets; inlining loads of binary files as strings is way less efficient.

    As for spritesheet memory optimisation, we have dozens of critical things to be working on right now, and I have seen precisely zero cases where spritesheeting has caused memory problems in a game, so it's not a priority right now.

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads

  • Construct makes 2D games, and WebVR is a 3D technology, so I'm not sure how Construct could usefully integrate it.

  • It's almost certainly to do with your particular third party plugins. That is why nobody else is reporting it, and plugin developers very frequently make this type of mistake. I guess failing to report the errors is a separate issue.

    There's nothing wrong with cf_deprecated or af_deprecated on their own, but ef_deprecated is invalid on its own because C2 still needs to know the return type of the expression since it can still be used by existing projects! So it must be combined with one of the return type flags, but plugin developers frequently replace the flag and make that mistake.

    Similarly a plugin can't say it is a "world" type if it specifies pf_singleglobal, probably just another addon developer mistake.

  • They want to have the art assets hidden and reduce the amount of HTTP requests (the individual sprites I can easily encode as data URIs, but currently having to reverse the spritesheeting makes it a bit more cumbersome).

    This makes no sense. Spritesheeting does reduce the number of HTTP requests. Besides, HTTP2 makes it a moot point since it makes the cost of additional HTTP requests so low. It sounds like someone's confused about what's going on or what's important to optimise.

  • Which is the first build that it stops working in? That is always a critical thing to know in a bug report.

  • None of the related code changed between r245 and r256. They make exactly the same plugin checks. I've no idea why one would report an error but not the other.

  • We're working with a client, who wishes to have all art assets as a BLOB (Binary large object).

    You could provide the art assets to them separately. Why would they insist the exported version be like that? I'm afraid we need a good answer to that to justify spending development time on it when we have dozens of other things to be working on.

  • Maybe file it on our feature suggestions platform.

  • This doesn't happen for me with a fresh install, and this type of error is almost always caused by third party addons. I'd double check what you have installed. I'm not sure why it wouldn't reproduce in r245 though...

  • I'd add that this demo involves moving an immovable object, which is presumably not something the physics engine expects and should not be surprising if it produces odd results. This may be closed as won't fix, I haven't had much time to take a look yet.

  • The Samsung Browser is not exactly the same as Chrome, which runs the native web view. Please check chrome://gpu like I suggested.

  • No, because this will de-optimise your project. Why would you want that?

  • I'm afraid I don't think there's a way to do this. You may have to keep existing apps on Crosswalk and only use the native webview for new projects.

  • We'd like to submit C3 to the Chrome Web Store, but not until C3 is out of beta.

    You can already use the "add to shelf" feature to add C3 to your Chromebook.

  • Simply store its angle at the last tick in an instance variable. If the current angle is clockwise of the last angle, it's rotating clockwise, otherwise anticlockwise (or not rotating).

    Note if you rotate 270 degrees clockwise, that will count as 90 degrees anticlockwise.