Here are two or three very annoying little things since the very last update of gdevelop on ipadOS.
1 - you have to fight with the ok button at the bottom right when you want to validate an action either in the conditions/actions window or in the action window itself. That is to say that pressing ok makes the iPad virtual keyboard reappear the only way to validate is to click several times frantically on the OK button
2 - About the keyboard precisely when you want to copy paste an action or a condition on an empty line then after “copying” the keyboard opens then you have to hide it to finally copy with the GDevelop menu which appears in popup.
3 - So it’s not really a bug but I would like you to take my request into account, after unsuccessful requests for help on the forum and the discord, it seems impossible to use a classic collision between an object player and an object on a background with a parallax effect because for some reason no one knows the parallax object’s collision box is very far away from the object. So my request is: “could we have an action or condition that would take into account objects that touch or overlap in collision even if their cameras are not the same”. I don’t have a solution to progress in my game. I’m giving you an example to understand, the little yellow star must be recovered by the ship but the star is on the background and must remain there and scroll at the same speed as background objects.
I hope you will read me and that my wish will come true
Thanks for reporting the issue, I’ve changed some of the tags on the thread to help it get looked at by the right people.
As far as the request:
From my understanding, that’s not how collision works because thats not how layers work. All layers have the same x/y coordinates. When you are moving the camera for parallax, you aren’t moving the objects, you are moving the viewport of that layer (what part of the layer that the player sees).
An object at 1040, 800 on layer “bg”, will still be at 1040,800 even if the camera is scrolled way over to 4000,4000. So if the player character is on the base layer at 1040,800 they will collide with that object even if it isn’t visible to the player. Likewise, just because they are visibly at the same place on the viewport, their actual positions could be wildly different and they will not be logically in collision. This is how it works in all engines I am aware of (Defold, unity, GMS2)
Your options would either be to have all the objects on the same layer (but different z orders) and manually move their positions based off the main layer camera, or add duplicate objects on the main layer (at their respective positions) and hide them at the start of the scene. Both will involve a decent amount of math (that you have to decide on). I’d recommend adding the objects to a single object group so you only need a few events without repeating them.
Because each layer has a unique camera, there isn’t really a feasible way for there to link the viewport to the positioning logic, a toggle would not cover a majority of use cases.
This doesn’t mean this won’t ever be done, this is just me giving you context and detail on why it works this way, and why the request may not ever be added.
Ok I understand better, but then is it feasible that my star is on the base layer with the rest of the interactive objects but that it follows the movement of the objects on the “BG” layer? But hey, it may be the same problem because the coordinates of the point will not be in the right place because of the paralax.
