perf: elide some jumps in jump

[DaniPopes]

1. legacy bytecode check (nullptr check); replace expect with unwrap_unchecked (still panics in debug mode)
2. saturate instead of early fail in jump_inner, since usize::MAX is always an invalid jump regardless
3. cleanups (use TryFrom impl in ruint since I optimized it, same code gen)

JUMPI diff:

diff --git a/a.s b/d.s
index 379024bb..49d5f819 100644
--- a/a.s
+++ b/d.s
@@ -8,10 +8,9 @@ revm_interpreter::instructions::control::jumpi::<revm_interpreter::interpreter::
 	.cfi_offset rbp, -16
 	mov rbp, rsp
 	.cfi_def_cfa_register rbp
-	sub rsp, 32
 	mov r8, qword ptr [rdi + 192]
 	cmp r8, 2
-	jb .LBB148_8
+	jb .LBB148_6
 	mov r9, qword ptr [rdi + 184]
 	mov rsi, r8
 	add r8, -2
@@ -30,49 +29,38 @@ revm_interpreter::instructions::control::jumpi::<revm_interpreter::interpreter::
 	or r9, qword ptr [rbp - 8]
 	or r8, qword ptr [rbp - 16]
 	or r8, r9
-	je .LBB148_7
+	je .LBB148_5
 	or rcx, rdx
 	or rcx, rsi
-	jne .LBB148_10
-	cmp dword ptr [rdi], 1
-	jne .LBB148_9
-	mov rcx, qword ptr [rdi + 8]
-	cmp rax, qword ptr [rcx + 64]
-	jae .LBB148_10
-	mov rdx, qword ptr [rcx + 56]
-	mov rsi, rax
+	mov rcx, -1
+	cmove rcx, rax
+	mov rax, qword ptr [rdi + 8]
+	cmp rcx, qword ptr [rax + 64]
+	jae .LBB148_7
+	mov rdx, qword ptr [rax + 56]
+	mov rsi, rcx
 	shr rsi, 3
 	movzx edx, byte ptr [rdx + rsi]
-	mov esi, eax
+	mov esi, ecx
 	and esi, 7
 	bt edx, esi
-	jae .LBB148_10
-	add rax, qword ptr [rcx + 24]
-	mov qword ptr [rdi + 96], rax
-.LBB148_7:
-	add rsp, 32
+	jae .LBB148_7
+	add rcx, qword ptr [rax + 24]
+	mov qword ptr [rdi + 96], rcx
+.LBB148_5:
 	pop rbp
 	.cfi_def_cfa rsp, 8
 	vzeroupper
 	ret
-.LBB148_8:
+.LBB148_6:
 	.cfi_def_cfa rbp, 16
-	add rsp, 32
 	pop rbp
 	.cfi_def_cfa rsp, 8
 	jmp qword ptr [rip + <revm_interpreter::interpreter::Interpreter>::halt_underflow@GOTPCREL]
-.LBB148_10:
+.LBB148_7:
 	.cfi_def_cfa rbp, 16
 	mov esi, 42
-	add rsp, 32
 	pop rbp
 	.cfi_def_cfa rsp, 8
 	vzeroupper
 	jmp qword ptr [rip + <revm_interpreter::interpreter::Interpreter>::halt@GOTPCREL]
-.LBB148_9:
-	.cfi_def_cfa rbp, 16
-	lea rdi, [rip + .Lanon.393361ac5c29a9e6e6e4d251883b8a2b.38]
-	lea rdx, [rip + .Lanon.393361ac5c29a9e6e6e4d251883b8a2b.40]
-	mov esi, 19
-	vzeroupper
-	call qword ptr [rip + core::option::expect_failed@GOTPCREL]

[codspeed-hq]

Merging this PR will degrade performance by 3.04%

❌ 1 regressed benchmark
✅ 172 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	MSTORE_50	19.3 µs	19.9 µs	-3.04%

---

_{Comparing DaniPopes:dani/jump-conds (bf9925b) with main (69083f1)}