I've noticed. Thats why I came to the same conclusion and put many objects into one spritesheet so it exists only on 1 texture. This is why both Unity and Gamemaker have ways to intelligently merge sprites onto one texture. This is why I've requested an optimization feature for C3 (that got denied).

Maybe only the more advanced users come to this conclusion and that's why you don't hear it often.

...

- Everyone in 2011

This is looking back with rose tinted glasses. Flash absolutely blew away HTML5 then. So at the time, we couldn't export to the best web format or PC. Other engines could export to Flash and PC and then they added HTML5 later when it became viable. Isn't that a better solution than risking everything on HTML5? So sure you made a good prediction on HTML5 becoming a good web format but it took a few years. And still we've only recently gotten a PC export that works properly with Steam, an export to XBox that doesn't work, no PS4 and no Switch. So we are still dealing with the repercussions of having an HTML5 export.

I'd be willing to bet most people here use Construct in spite of it being html5 than because it is html5. I'm here for the event system. A lot of posts have been made vindicating the choice to use html5, but i feel regardless Scirra is still a bit blind to its down sides and the on going issues it presents. Just my opinion.

Exactly this.

I have global variables representing all the actions in my game, for example, key_jump. When the user wants to change controls, I wait for input using keyboard.LastKeyCode or gamepad.LastButton and assign that to key_jump. I then call those variables using keyboard.keycode or gamepad(0).buttonindex. That all works fine. However the mouse doesn't seem to have these features. I can't seem to assign an index of a mouse button to a variable. I then can't say on mouse index pressed - do something.

For keyboard and gamepad objects, I can get the last key or button pressed and assign that input to a variable to customize controls. I can't seem to find any similar actions for the mouse object so I can't let a user change those inputs. Am I missing something or are there any workarounds for this?

Thanks

-Lou

That seems like a crippling feature of C3 to me. I guess I just can't imagine not managing and optimizing graphics assets. Why would I want to load frames from a giant boss creature when I want to fire a bullet that may be auto sprite sheeted into the same texture. For another example, in my game I include a bullet graphic plus fx for the bullet all into the same sprite object so they will be in the same sprite sheet at around a 64x64 texture. In C2 I can load that on the fly no problem. No jank. In C3, if the system decides to split that off into different sprite sheets, then it will load in a sprite sheet that could be 1024x1024 for the bullet, then it might need to load in even more large sheets to get the other bullet fx in. Wouldn't that also be more jank?

TBH unless your player animations are absolutely huge with loads of frames, splitting things out in to separate objects is probably a micro-optimisation that doesn't make much difference. It also can jank the game as it loads textures mid-frame if the object isn't already preloaded by being placed on the layout. And currently the engine doesn't release textures until the end of the layout, so the peak memory usage will be the same.

I am probably in the minority, but I am working on a larger game with lots of animations in larger layouts. I depend on being able to smartly load in textures. Peak memory usage would not be the same, because as in my previous example, instead of loading in 10 death animations, I only load in the one that I need when it is needed. I never need all 10 death animations. Also there is no jank, even if there was, it wouldn't matter that there's jank when you die.

I also don't understand all of this jank concern. Ori and the Blind Forest had jank when you got to new areas because it would stream them in instead of loading. Who really cares? Feeling like you were in a huge connected world was amazing. As long as you smartly load in textures in areas where it doesn't matter, the majority of the gamers won't mind. I guess thats personal preference though

I've already had to make changes to my game design because I can't have large layouts (due to not being able to unload graphics.) .

The auto sprite sheet feature sounds pretty neat but I don't understand how it actually works for managing memory. In the blog post I see a sprite sheet with 2 enemies and bullets in it, which is great if they are all needed on that layout. But what happens if you need them separated because those 2 enemies don't ever exist together?

For example, in my current game I have a lot of death animations for the player. Each animation is a different sprite objects so they exist on separate sprite sheets. When the player dies, I spawn one of the death animation sprites.

This way I don't load 10 different death animations into memory that aren't needed. Will the system in C3 automatically lump all the sprites together into one sheet? Is there a system to tag objects that can be spritesheeted together and separate others? Or am I completely confused about whats happening?

rexrainbow Is there any way to get this to work with the Pin Behavior? I tried both effect and behavior modes and they don't seem to work.

Edit: Nevermind...just figured it out. The behavior has to be added after the pin behavior. I had it before.

Surprised no one has replied yet. This looks pretty awesome and I didnt notice any frame drops on my Surface Pro 4. Are you going to release this so we can experiement with it? Nice job man!

Thanks for answering my questions. One more came to mind... Can multiple babylon files be loaded at runtime? For example, can I have the player mesh in one babylon file and the level in another and then spawn the player into the level? Or would the player mesh\animations and level all have to exist in the same babylon file?

This plugin looks pretty awesome and I'm interested in buying, but I had a few questions first.

1) Are shadows supported on lights and if so, is there a limit to the number of lights\shadows?

2) Is Character Studio for 3D studio Max supported?

3) Can you change the camera's FOV?

4) How exactly are meshes loaded? Can I create a huge 3D asset warehouse, bring that into C2, then only load the meshes that I need at the time or will the whole scene file be loaded in?

5) Can sprites be rendered over the top of the 3D canvas? Seems like it can based on the platformer text over the canvas, but I wanted to make sure.

6) Is there a list of all the actions and events for each plugin?

Thanks!

Thanks for the reply — I understand the issue a little bit better now. So even if we could snap the collision polygon to exact numbers it still might not merge into one single collision box?

I will try some other solutions to work around the problem for now.

I'm having a problem creating straight collision polygons. My problem is in the way my tile maps are designed, the collision polygon needs to cut through the center of the tile. This creates inaccurate collisions because the polygon line can never be perfectly straight.

Here is a r239 capX showing the problem: http://www.louisferina.com/games/testCollision.capx There is a sprite with the platform behavior and a tilemap with the solid behavior. No events.

• Jump into the right wall by holding right arrow and pressing up repeatedly. Do not let go of the right arrow while jumping. (Works as expected)
• Jump into the left wall by holding left arrow and pressing up repeatedly. Do not let go of the left arrow while jumping. (Does not work as expected)

Jumping against the left wall causes the sprite to hit the center tile collision polygon. It seems impossible to create an exact line where this would not happen. Jumping against the right wall works as expected because the border of the tile can be lined up perfectly.

It seems like there needs to be either some leniency where the sprite does not hit a collision polygon that is smaller than a 0.5 pixel difference or a way to snap collision polygons to exact pixels.

Is this a bug, an editor request or am I just missing something?

Thanks for any help!