Oh yeah, it appears r144 was broken in that way. I just got notified of r144.2 and it fixed this issue. Thanks.

Link to .capx file (required!):

megamanrevolution.com/other/bugged.capx

Steps to reproduce:

1. Open the CAPX.

2. Build a Node-webkit export.

3. Try running the EXE export.

Observed result:

The game in question does not open at all aside from running in the background and it just stays there forever. This CAPX example worked before beta r144 and now that it's opened in r144 and exported, it refuses to function.

I don't know if this is coincidence but when I tried building this example CAPX to EXE in an old copy our actual project's node-webkit export, the old game played instead of this, almost as if Construct 2 didn't even include the game's files whatsoever.

In our actual project, I did manage to get the EXE to work, but only after deleting every single event sheet, which is obviously not a viable option.

Expected result:

The game was supposed to run as an EXE.

Browsers affected:

Chrome: N/A

Firefox: N/A

Internet Explorer: N/A

Operating system & service pack:

This has been tested on Windows 8 64-bit, Windows XP SP2 32-bit and Windows 7 64-bit.

Construct 2 version:

r144 64-bit with that patch Ashley released to fix OR blocks since they didn't work without it.

Thank you, Ashley for your fine work. C2 continues to amaze me at the raw power and flexibility it has and this allows me to use the beta you released without having to worry about my OR events being totally broken.

I was searching through the forums out of curiosity and I came across hardly any results on the matter. We have implemented an array obstacle detection system for static objects. Why has no one else thought of this? It handles hundreds of moving enemies checking hundreds of static obstacles at 45-55 FPS and I'm sure it can still be improved because my array checks are quite plentiful at the moment and could stand to be cut down a bit. With using collisions in this same setting, we were over one million collision checks per second for enemies and obstacles alone, even with disabling collisions for things off screen and such, which of course slowed the game down to 14fps or worse even on my computer that isn't a weak computer.

I just wonder why this alternative to collision checks hasn't been brought up before much. It works quite well after we set up the image points to handle "collisions" on the array.

Well I added for it to only check collisions when on step and that improved things back to 60FPS. Now it seems the game is slowing down immensely from the set custom movement speed action no matter what I do once that mass of enemies come on screen because of all the collision checks I take it.

What's an effective way to reduce collision checks?

Oh sorry, I forgot to remove the plugins so it can be opened. I tried doing that, if it still gives you the same error, which plugins is it saying are being used?

The debugger says the player controls group and some other groups are using up a lot of the CPU, but it doesn't specify which part of them are using the CPU.

EDIT: It appears the engine is using up 72~% of the CPU.

Here is a link to our CAPX to show we actually have tried with this issue. The enemies used to use tile by tile movement the player has still, which was faster than this, but we're transitioning to pixel by pixel.

dropbox.com/s/q996hnxmcp4ccan/game.capx

We had an experimental CAPX where we were able to have 800 enemies on screen without slow down, complete with dynamic shadows for each enemy... but for some reason, even the enemies being on screen with no code is a major slow down. Go directly downward to witness the 250 enemies slowing the game down a huge amount. This happens to me where it drops to around 7 FPS and I have a quad core computer with 8GB of RAM and a decent video card, so my computer is not the issue, especially since we had that experiment with 800 enemies. I left all the event sheets open in the CAPX that I believe are the problem. Gnashers are the enemy that is slowing it down badly.

Also the Gnashers are extremely glitchy in their movement, and once they are on screen, the game never recovers from the FPS drop even when they're all off screen. How do we fix that as well?

So is this still being developed? I was interested in putting it in my game with a scream attack, along the lines of microphone from Kiby, where when you yell into the microphone (if you have one, otherwise it just automatically does damage) it does damage to all enemies on the screen based on how loud you were.

I have an enemy type with 73 lines of code and they each have one detector to determine how to move properly in a tile by tile movement engine which forces the actors to walk back in line with the invisible tile grid if they aren't lined up with it properly because of wind blowing them pixel by pixel or what have you. I can't seem to find a way to connect the detector to the enemy without using for each loops, which drags the FPS of my game down to below 10 when there are more than 100 on screen at once and I for each through all the enemies. Is it possible to pull this system off without a for each loop?

I did some digging around on the forums and found that the issue with gamepad support in C2 can be eliminated by using drivers from the x360ce bundle. My question is, since it says I can freely distribute the DLLs with my game, if I included these DLLs in the folder my EXE is in, will the gamepad support work automatically for all gamepads when that EXE is executed?

Ah thanks a bunch! I thought I remembered a project setting somewhere that you had to set for this. I appreciate the help.

Hi all, I've been using C2 for about a week, just bought the standard license today. I have to say, coming from Multimedia Fusion 2 that I am highly impressed with this program! C2 can do so much more than MMF2 can even with less features. I have done things already that I wouldn't have even dreamed of doing in MMF2. I regret that I hadn't known about C2 before I got MMF2 in fact.

As for my question:

I've tried setting C2's default preview browser in the preferences to custom and even node but it continues to insist on running in Firefox. What am I doing wrong here? I would prefer testing with node because it runs much much MUCH faster than in Firefox.