Differently optimized code between `let else` and `if`

Hi,

Those two pieces of code do the same thing, but end up having different assembly 👀

#[inline(never)]
pub fn foo(x: i32) -> bool {
    let (1 | 3 | 99) = x else { return false; };
    
    true
}

#[inline(never)]
pub fn bar(x: i32) -> bool {
    x == 1 || x == 3 || x == 99
}

playground::foo: # @playground::foo
# %bb.0:
	cmpl	$1, %edi
	je	.LBB0_4
# %bb.1:
	cmpl	$99, %edi
	je	.LBB0_4
# %bb.2:
	cmpl	$3, %edi
	jne	.LBB0_3

.LBB0_4:
	movb	$1, %al
                                        # kill: def $al killed $al killed $eax
	retq

.LBB0_3:
	xorl	%eax, %eax
                                        # kill: def $al killed $al killed $eax
	retq
                                        # -- End function

playground::bar: # @playground::bar
# %bb.0:
	movb	$1, %al
	cmpl	$1, %edi
	je	.LBB1_4
# %bb.1:
	cmpl	$3, %edi
	je	.LBB1_4
# %bb.2:
	cmpl	$99, %edi
	je	.LBB1_4
# %bb.3:
	xorl	%eax, %eax

.LBB1_4:
                                        # kill: def $al killed $al killed $eax
	retq
                                        # -- End function

(playground link)

Curiously, they do get optimized the same sometimes (e.g. for (1 | 3 | 5) or even (1 | 2 | 3 | 5..=10)).

Start with the playground reproduction and compare the generated assembly for foo and bar in release mode. Investigate why the let-else pattern and equivalent boolean expression receive different optimization, then verify whether the cases produce equivalent assembly after the fix.