right now, i just have a broken mess and a theoretical concept that lacks the knowledge of how to implement it in an algorithm.

The best approach i had till now was to make a function that gets four parameters (X & Y of the tile to check, direction and an loopindex to controll the distance of view).

The function then checks if the tile is ground or wall, if ground then this tile is seeable and if the loopindex isnt too high it will call the function again to check the next tile in that direction. That works for four directions, if a blocking element is found it will stop to look further in that direction. But there arent just four surrounding tiles, there are 8 and all my approaches to also check the other tiles AND stop looking further when a blocking element where found didnt work.

Hi,

i need to check if NPCs on my tilemap can see each other.

my problem is that the built in line of sight behaviour is way to cpu consuming for my project.

here is how i did it, make a second tilemap with all the sight blocking obstacles, make this tilemap solid, then use LOS behaviour and bam this game isnt build for running on mobiles :/

i think the los bevaiour is way to powerfull for my needs and im pretty sure there are many many algorithms that do that more efficiently (bresenham, raycasting) and when i read about it, it seemed so easy but im to dumb to understand any of them to do them in construct2. My floodfill approach did either too much or to less, i dont get how to check if the tile isnt just empty but also seeable from the position, so yeah, i did understand the raycasting approach but not how to build it.

My dream would be that i overlooked an plugin where i just specify the range of sight in tiles, blocking obstacles and then use it

can anybody help me? im not a native speaker so its possible i overlooked some good tutorials, the main problem in finding information was to me that there where so many and most of them where with complicted code that i didnt understand.

hello all!

I represent Adtoapp, In short - our company is a smart ad mediation platform that accumulates largest ad networks like Admob, Vungle, Chatsboost e.t.c and can increase revenue from your app in 2 times and make you 100% fill rate with highest eCPM with no manual work on managing networks.

Lets discuss?

Do you have an construct2 plugin?

because the process of just trying it out would take less time that way

In case you didn't see my blog post, you can get WebGL running on a Raspberry Pi 2, although it takes a bunch of hacks. It seems to work fine for Construct 2 games!

could you publish an image of your system configured to run webgl?

the load image from url command replaces the current frame of the object so every object gets a new image, you can avoid this by giving the sprite an additional frame then just put an event like "sprite set frame to frame x" and then put the load image from url event, so the other sprites can keep the original image.

you could just add a condition to the button event "button/layer/etc. is visible", then the event only fires when you want it to

You need to make sure you have the most up-to-date version of VS2013. In the installer, make sure you don't uncheck the support for Windows Store apps, or anything related to Javascript/HTML/web support.

did you manage to install and run your exported win10 apps on an win10 mobile device?

..android 5.0 doesnt' support webGL...

thats not true

it doesnt make any sense how it is now, i think anyone understood that "feature" as copy any combinationation of tiles and repaste them instead of how you descibe it

very nice game!