We don't imply outlives-bounds from the own explicit bounds of a GAT projection

Not sure if worth tracking/discussing. I'm basically revisiting this because of IATs and LTAs. I know that T-types would like to avoid implying outlives-bounds in more places and under more circumstances at least until various soundness issues related to them have been fixed (LTAs prolly being the exception) (via: #t-types > implied bounds for the more recent features? @ 💬 (2023)).

Presently, when inferring outlives-bounds we only consider the explicit predicates of the corresp. trait when looking at projections as noted by this comment (which was written by me):

https://github.com/rust-lang/rust/blob/77101febcc9662a076ff43887497cd5b30674d93/compiler/rustc_hir_analysis/src/outlives/implicit_infer.rs#L207-L209

This means:

trait Trait<'a, T: 'a> { type Project; }
impl<'a, T: 'a> Trait<'a, T> for () { type Project = (); }

struct Env<'a, T>(<() as Trait<'a, T>>::Project);
//                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ from this projection, we imply `T: 'a`

//fn env<'a, T>() { let _: Env<'a, T>; } //~ ERROR the parameter type `T` may not live long enough

trait Trait { type Project<'a, T: 'a>; }
impl Trait for () { type Project<'a, T: 'a> = (); }

struct Env<'a, T>(<() as Trait>::Project<'a, T>); //~ ERROR the parameter type `T` may not live long enough
//                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ we **don't** imply anything from this projection!

Start in compiler/rustc_hir_analysis/src/outlives/implicit_infer.rs around lines 207–209, then compare the two GAT projection examples in the issue. Trace how explicit predicates from the corresponding trait are considered for projections and verify the current diagnostics with focused compiler tests. Done means the intended outlives-bound behavior is decided and covered by tests, accounting for the linked IAT, LTA, and soundness context.