Collision detection and physics in a second pixel platformer

I am transforming physics in a second motion platformer I am growing (no engine, c++ / sdl). Here is the overall motion logic for all characters:

1. I calculate the offset I want to use to the character
2. I test collision of my character towards all colliders with solely X offset utilized. If I detect a collision with any of them, I have a look at how excessive ought to I transfer the character to depart it. If this worth is low sufficient (the restrict is round 1.2 * xOffset to deal with most slopes), I transfer character upwards and hold X offset. In any other case, it is most probably a wall and I deal with it as one
3. I deal with Y axis as if every little thing is a wall (there are not any slope ceilings and also you should not get pushed downwards into the slope)
4. If the character is left airborne, I search for the bottom beneath and if the space is brief sufficient, regulate Y pos to remain on it

It really works positive generally, however there are a number of circumstances which I do not know methods to correctly deal with. Here is the one I’ve questions on:

Let’s assume that the offset is {10, 0} (10 pixels to the proper). If every little thing works correctly, the character ought to transfer 7 pixels to the proper and keep proper beneath the ceiling. If I totally apply the offset, the character will keep on the third collider and can overlap the 4th. For example we course of 4th after third, how ought to we deal with this collision? Logically we must always transfer character down the slope (2), however we aren’t standing on it, relying on the order we probably did not even collide with it.

So, what’s one of the best ways to implement physics to deal with this type of situations (the place motion alongside one axis could cause motion alongside one other axis and collision that impacts the motion alongside first axis)? My finest concept thus far is to maneuver character by its full width at most and if it is required to maneuver it apart, transfer it per pixel (inside that character-size vary), however I do not understand how optimum is that this, it most probably will get approach worse if I am going to determine to extend characters’ dimension, pace and the quantity of them