Ok, thanks, I figured as much. I'm a bit of a computer ignoramus, but will try to report as an edge bug.

In debug mode, chrome is awful. At least the end product should not be this bad. It runs better in the profile mode of the debugger. Nonetheless, edge seems faster when not using the debugger. There are a number of potentially recommended things to look at in chrome to improve performance (none worked for me). Ultimately, if the end user suffers issues playing your game they will not blame chrome, or try and make chrome work, they'll blame the game

I load sounds in both formats, just the one is not good enough for edge. M4a is supported.

It is also my experience of consistently better fps performance by edge.

I haven't read the license here, but the majority of commercial game assets are not to be distributed in a way in which they can be extracted by others. I would think that selling a capx would breach these conditions. It actually presents a problem in uploading capx in a public forum, eg. In a bug report, as I think you should remove licensed assets first (depending on license specifics).

Problem Description

When an immovable Physics behaviour tilemap is in a layout 'key down' events fail in Edge

Attach a Capx

https://1drv.ms/u/s!AnzkE7yFwDAQgehlNrznFaMQqZNP9A

Description of Capx

Play and this capx displays a text message about whether the W key is held down or not. There are also some tiles of a physics behaviour tilemap in the same layout. In Edge, the keydown events are often missed.

Steps to Reproduce Bug

• Step 1: Hold W down for a few seconds and observe text
• Step 2: release W for a few seconds and observe text
• Step 3: repeat above (6-10 times should be sufficient) IMPORTANT: do not press other keys, touch the mouse or touch screen during this test as it corrects the defect.

Observed Result

In Chrome, FF, and IE, the capx behaves as expected. Edge, however, tends to miss key down/up events.

Expected Result

games in Edge to know when keys are up or down with a Physics tilemap present

Affected Browsers

• Edge: (YES)
• Chrome: (NO)
• FireFox: (NO)
• Internet Explorer: (NO)

Operating System and Service Pack

Windows 10 - note it recently installed the anniversary update

Construct 2 Version ID

The latest non-beta, I think, r233. 64bit

Having said what I said, I have just found a browser specific fail with key presses when physics tilemaps are present - will make a capx and file.

Ok, so I reread, you are changing tiles in gameplay!

With static tilemaps I haven't had a problem.

newt

Tilemaps aren't made to work with physics since they don't move.

You could try using the Chipmunk Physics behavior, but I would guess the outcome would be the same.

What can be done is using the Platform behavior on a sprite, and the Solid behavior on your tilemap.

I am confused by this since physics behaviour on tilemaps seems to work fine for me (set to immovable of course) and provides a barrier for physics objects?

Disable hardware acceleration in Chrome settings and restart the browser, See if it fixes the lag.

Tip: enter this url => chrome://flags

Disabling hardware acceleration via the settings reduces the FPS further and iirc doesn't allow webGL either.

Helladen

Ashley

So I do see the reported sound issue on browser first opening only in Chrome and FF, but not in Edge or IE, so it is browser specific. The same defect is in NW.js as expected. A very noticeable frame rate drop is seen on my device (Surface pro 3, i7 + Win 10) with the first trigger event that doesn't play sound and this is associated with the inspector estimating very heavy CPU usage. It does appear to be an issue?

RasmusTheRaccoon

No Answer I am afraid. I assume you mean when running the Inspector, which I simply do not do in Chrome or NW.js anymore unless I have a very good reason. I think Edge and IE are still faster and less burdened than Chrome without the Inspector, but the inspectors (Inspect mode) in Chrome is seemingly a waste of time.

I am not a coder, but on point 2, I think else is often misunderstood or misused and perceived not to work (maybe I now prove myself guilty). In c2 i'd use else when appropriate and assume that is more efficient. So in your example, if more than one object has the variable as an instant and you use else, the else line will only run if the condition fails for all objects. Not using else, but an invert or different number means all objects respond appropriately. Caveat: I did say I wasn't a coder.