MemoryPatches: Fix instruction patches

[JoshuaMKW]

For a long while, multiple patched instructions at a time could only be restored in reverse order of patching. This PR fixes the oversight, allowing users to restore patched instructions in any order they like.

[dreamsyntax]

Can you please squash these commits to one?

Looks great btw, I've personally been affected by this issue

[JoshuaMKW]

Sure. It's pretty funny too because I intentionaly split up my changes based on context thinking that was more favored :P

[AdmiralCurtiss]

Separate commits are fine, actually, but we don't like merge commits or 'fixup'-style commits. If you can have separate commits that actually make sensible changes per-commit feel free to leave them.

[dreamsyntax]

Separate commits are fine, actually, but we don't like merge commits or 'fixup'-style commits. If you can have separate commits that actually make sensible changes per-commit feel free to leave them.

I was referring to these:
Merge branch 'master' of https://github.com/JoshuaMKW/dolphin

But I was under the impression Dolphin PRs were 1 commit only.
Good to know that is not the case 👍

[JoshuaMKW]

Alright, I figured out squashing and it's been cleaned up!

On another note, what would have me become auto-trusted? It's of course not that important but seeing the "All checks have failed" messes with my head lol.

[Pokechu22]

On another note, what would have me become auto-trusted?

People who have been around for a while can get added once they've submitted a few PRs. Otherwise someone just needs to manually trigger the CI build. I wouldn't worry about it too much for now.

[Pokechu22]

Tested by enabling "boot to pause", then launching super mario sunshine and nopping out the first 2 function calls, and then restoring the first one. Before restoring would do nothing, now it works. Code looks good to me too.