NetOne - yeah that's what I worry about. Would this affect the GPU fillrate? Because if it does, it's already hampering slower machines and needs to be taken care of.

NetOne - right. That's where I read that. I guess my question is "will doing all the legwork to fix it be worth it?"

After (of course) developing 50% of my pretty large game, I've finally read something about tilemaps that worried me. I was using them to get tons of variation on almost every tile (ie platform floors with 3-4 variants, BG walls with 6-7 variants, windows that are 4-6 tiles wide/long, etc.), but I'm realizing for each next different tile it has to issue a whole new draw call. So a window in the background, even though its really only 4 tiles, it has to render 4 draw calls instead of having a single sprite of a window and only doing 1 draw call.

Now, my levels aren't HUGE, per se, but some of them have a lot of variations going on and nothing is really being "tiled". Is this something I should go back in and look at, or because it's pixel art it's not a huge deal? Or am I (once again) misunderstanding how the tilemaps are supposed to be used?

UPDATE: Problem found - it was script mini-fication.

Problem Description

When in full NWjs export (works fine in preview), if you open a dialog box (save or open) and then do a "on dialog cancel or OK", that trigger will never be called in the export.

Attach a Capx

https://dl.dropboxusercontent.com/u/28104568/NW%20Bug.capx

(this requires you to export a nwjs build to work; win32 tested)

Description of Capx

Single button that opens save dialog. Clicking cancel should cause the browser to close. You must export this to NW (it will work fine in preview)

Steps to Reproduce Bug

Add button to click to open NW save dialog Add trigger of NWjs on "save dialog CANCEL" and add browser close action

Observed Result

The trigger isn't being called - no matter what you put in the "on save CANCEL". This also happens on "save OK"

Expected Result

The event should work outside of preview.

Affected Browsers

NW: No (nw specific)

Operating System and Service Pack

Windows 10 Pro

Construct 2 Version ID

C2 Personal stable (244)

Most recent C2 NWjs

I'm curious about everyone's thoughts on good practices for battling with GPU fillrate issues. This is mainly geared towards high powered laptops that have insane resolutions (or mobile devices, if you will).

R0J0hound - sorry, one more question. For some reason I got it to work using the int(), but I've just tried to recreate something I have in a different layout (which is working) and its literally the same code, but I can't figure out why it's not working in this new instance.

I'm trying to just get the rgb of a canvas object based on the mouse position (the canvas object is being filled by gradient). I'm just doing "set text box to "rgb("&userGradient.redAt(Mouse.X, Mouse.Y)&","&userGradient.greenAt(Mouse.X, Mouse.Y)&","&userGradient.blueAt(Mouse.X, Mouse.Y)&")", but it's always returning rgb(0,0,0).

I've tried int() on both individual mouse.x and mouse.y, as well as the entirety of the redAt, greenAt, etc.

It just keeps returning 0 for all the paramenters.

EDIT AGAIN (sorry)

AhHA! Got it - I didn't know this breaks when you resize a canvas. For those who were wondering you fix by setting the mouse.x - canvas.x, and mouse.y - canvas.y for each.

R0J0hound - unfortunately the default variables were already integers and not floats. It just keeps returning black (but when I set the "rgb(x,x,x)" to a text object, it looks the exact same as if I just set the color to "rgb(x,x,x)" using numbers (without variables)

EDIT - I'm wrong. That worked strangely enough. Thank you!

I do find the article Huemerson posted a bit ... interesting. Basically what I'm getting from the performance tips there is the best way to optimize your game is not to make the game at all.

I'm not sure my fps slowdown is from global layers, but I am using a lot of them, too, and have over 400 invisible objects on all of them combined. Has anyone else had issues with this?

R0J0hound - ah I'll give that a shot - thanks!!

mercuryus - not sure I follow on the best way to do this. Would you use timers or variables?

I've found if you have multiple events that use the shake command on a camera sprite, if you set something like "shake 10px for 8 seconds", and then have another command of "shake 3px for 0.5 seconds", the 8 second shake is stopped.

Is there a way to avoid that from happening?