AlexFrancois where did you heat that? I'm 99.99% sure that's not true.

EDIT: I misread your post a little bit, I guess you're talking about mac OS's? I can't really comment on that, I haven't tried on a lot of different OSX's.

AlexFrancois what makes you think webGL doesn't work on OSX? In all my testing it works fine, as it should.

ErekT Oh, sorry, I did. I'll attach it or reupload it.

EDIT: NVM, I thought it might have been a dropbox link, but in fact it had been attached. Looks like the Scirra website has had trouble with attachments lately, I hope they aren't gone forever.

Animmaniac thanks Paulo. Interesting that the loops are unrolled, but I guess it makes sense.

I guess I'll just go with this hack for now and if I hear reports of bad performance I'll resort to the multiple shaders method.

Pros:

-The pokemon are cute

Cons:

-literally everything else

OK, maybe that's a bit too far, but I really don't see many redeeming qualities in pokemon or most other similar RPG's. You're basically doing the same thing over and over again 1000 times until you win the game. The type rock/paper/scissors thing is cool until you get past fire/water/grass and realize it's all incredibly arbitrary and completely undocumented. Also, the random encounters system is the worst thing ever. And yeah, even the story is shallow and boring.

szymek I think that specific bug has been mostly fixed but Chrome still suffers from shocking vsync issues below 60fps. As well as NW issues that are preventing deployment to mac and linux. I think Ashley is slightly understating the third party issues.

[I will insert here that I have no experience with app development so anything I say may not apply in the mobile marketplace, where commercial-scale games can be a lot smaller and simpler]

As I've said before, If Scirra want to continue with a small number of staff then one exporter is really the way to go from their perspective. Otherwise it's a maintenance nightmare. Maintaining parallel codebases is absolutely horrible, but there's a reason GM and Unity do it.

Basically, from my perspective as a commercial developer, I can't afford to rely on so many third parties. Ashley made a valid point that you'll always be relying on third parties to some extent. However I'd rather rely on the 550 Unity employees. The difference isn't the reliance on third parties, it's who can afford to take responsibility. For example, if Unity runs badly on, say, PS4, or NVIDIA GPU's, Unity can take responsibility for that and fix it themselves. Ashley on the other hand cannot afford to take responsibility for this sort of stuff, and hence the engine is in HTML5, which takes the burden off of him and puts it on Google or whoever. The problem comes when Google or Apple or a mobile wrapper developer doesn't want to prioritize getting large-scale games to work correctly, due to the tiny portion of their userbase developing games like that.

I don't mean this as a slight to C2. It's without a doubt one of the best 2D engines out there for hobbyist/learning developers. But in my opinion it's not ready for commercial scale development due to reliance on third parties. [AGAIN, I'M TALKING PC/CONSOLES HERE] Not that it can't be done, but it can turn into a nightmare very fast when you want to release your game but for some unknown reason it janks on supercomputers, or runs horribly bad on all intel processors, or doesn't run at all on current-gen consoles.

tl;dr: HTML5 is the best choice for C2 but it means commercial PC developers will have a bad time.

OK, I got it mostly working, however I think I'm in a spot of trouble. R0J0hound if you have any ideas that would be fantastic.

My shader has a for loop that loops from a specific value to a specific value. e.g.

for (int i = START_VALUE; i <  END_VALUE; i++)[/code:36gm9ilc] 

In shadertoy, these were just set with consts and life was good. However, it's problematic in C2. Basically, I want to be able to see these loop values at edit-time, presumably via an effect parameter. However GLSL simply will not allow you to use non const loop values at all, even if they are in reality constant.

The reason these need to be set at edit time is that I will need a few instances of this shader, all with different loop values.

Ugh, I have a feeling this won't be possible without an addition to C2's effect system 

EDIT: I suppose I could use something like [url=http://gmc.yoyogames.com/index.php?showtopic=616107]this hack[/url], but I have a bad feeling that on some hardware I'd pay dearly

Yeah, it's really weird that Scirra has written literally 0 documentation on effects. I guess it's not really high on the priority list.

R0J0hound thanks!

I think I'm missing something though - the compiler doesn't seem to agree that you can omit the precision.

EDIT: Never mind, turns out I need to add the line

precision mediump float;[/code:19tugnqz] to the beginning of the program.

— Thanks! I'm worried though for the sake of others that this post will be lost in the aether. Maybe you could put it in a tutorial or something?

OK, finally got my mac up and running. Same issue. I've submitted a bug report to nw.js. Hopefully they can do something with that. If you guys could maybe add a comment saying it's effecting you too that would be awesome!

Hi all,

I'm trying to write a shader for my game. I got it up and running in shadertoy, and now I'm trying to port it across to C2.

However, looking at C2 shaders, they seem to be formatted quite differently to the GLSL code on Shadertoy. Specifically, the variable declarations seem to need a lot more specifiers like 'varying', 'uniform', and 'mediump'. Do I really need to specify varying and precision for every single variable? And how do I specify things like 'const'?

Cheers!