Use `unaligned_load` for `regex` virtual machine

This should hopefully avoid any issues with loading operands greater
than 8 bits on alignment-sensitive platforms.

core/text/regex/compiler/compiler.odin

@@ -1,5 +1,6 @@
 package regex_compiler
 
+import "base:intrinsics"
 import "core:text/regex/common"
 import "core:text/regex/parser"
 import "core:text/regex/tokenizer"
@@ -408,7 +409,7 @@ compile :: proc(tree: Node, flags: common.Flags) -> (code: Program, class_data:
 				break add_global
 
 			case .Rune:
-				operand := (cast(^rune)&code[pc+1])^
+				operand := intrinsics.unaligned_load(cast(^rune)&code[pc+1])
 				inject_at(&code, pc_open, Opcode.Wait_For_Rune)
 				pc_open += size_of(Opcode)
 				inject_raw(&code, pc_open, operand)
@@ -490,20 +491,20 @@ compile :: proc(tree: Node, flags: common.Flags) -> (code: Program, class_data:
 				case .Jump:
 					jmp   := cast(^i16)&code[pc+size_of(Opcode)]
 					if code[cast(i16)pc+jmp^] == .Jump {
-						next_jmp := (cast(^i16)&code[cast(i16)pc+jmp^+size_of(Opcode)])^
+						next_jmp := intrinsics.unaligned_load(cast(^i16)&code[cast(i16)pc+jmp^+size_of(Opcode)])
 						jmp^ = jmp^ + next_jmp
 						do_another_pass = true
 					}
 				case .Split:
 					jmp_x := cast(^i16)&code[pc+size_of(Opcode)]
 					if code[cast(i16)pc+jmp_x^] == .Jump {
-						next_jmp := (cast(^i16)&code[cast(i16)pc+jmp_x^+size_of(Opcode)])^
+						next_jmp := intrinsics.unaligned_load(cast(^i16)&code[cast(i16)pc+jmp_x^+size_of(Opcode)])
 						jmp_x^ = jmp_x^ + next_jmp
 						do_another_pass = true
 					}
 					jmp_y := cast(^i16)&code[pc+size_of(Opcode)+size_of(i16)]
 					if code[cast(i16)pc+jmp_y^] == .Jump {
-						next_jmp := (cast(^i16)&code[cast(i16)pc+jmp_y^+size_of(Opcode)])^
+						next_jmp := intrinsics.unaligned_load(cast(^i16)&code[cast(i16)pc+jmp_y^+size_of(Opcode)])
 						jmp_y^ = jmp_y^ + next_jmp
 						do_another_pass = true
 					}

core/text/regex/compiler/debugging.odin

@@ -1,5 +1,6 @@
 package regex_compiler
 
+import "base:intrinsics"
 import "core:io"
 import "core:text/regex/common"
 import "core:text/regex/virtual_machine"
@@ -9,11 +10,11 @@ get_jump_targets :: proc(code: []Opcode) -> (jump_targets: map[int]int) {
 	for opcode, pc in virtual_machine.iterate_opcodes(&iter) {
 		#partial switch opcode {
 		case .Jump:
-			jmp   := cast(int)(cast(^u16)&code[pc+1])^
+			jmp   := cast(int)intrinsics.unaligned_load(cast(^u16)&code[pc+1])
 			jump_targets[jmp] = pc
 		case .Split:
-			jmp_x := cast(int)(cast(^u16)&code[pc+1])^
-			jmp_y := cast(int)(cast(^u16)&code[pc+3])^
+			jmp_x := cast(int)intrinsics.unaligned_load(cast(^u16)&code[pc+1])
+			jmp_y := cast(int)intrinsics.unaligned_load(cast(^u16)&code[pc+3])
 			jump_targets[jmp_x] = pc
 			jump_targets[jmp_y] = pc
 		}
@@ -46,18 +47,18 @@ trace :: proc(w: io.Writer, code: []Opcode) {
 			operand := cast(rune)code[pc+1]
 			io.write_encoded_rune(w, operand)
 		case .Rune:
-			operand := (cast(^rune)&code[pc+1])^
+			operand := intrinsics.unaligned_load(cast(^rune)&code[pc+1])
 			io.write_encoded_rune(w, operand)
 		case .Rune_Class, .Rune_Class_Negated:
 			operand := cast(u8)code[pc+1]
 			common.write_padded_hex(w, operand, 2)
 		case .Jump:
-			jmp   := (cast(^u16)&code[pc+1])^
+			jmp   := intrinsics.unaligned_load(cast(^u16)&code[pc+1])
 			io.write_string(w, "-> $")
 			common.write_padded_hex(w, jmp, 4)
 		case .Split:
-			jmp_x := (cast(^u16)&code[pc+1])^
-			jmp_y := (cast(^u16)&code[pc+3])^
+			jmp_x := intrinsics.unaligned_load(cast(^u16)&code[pc+1])
+			jmp_y := intrinsics.unaligned_load(cast(^u16)&code[pc+3])
 			io.write_string(w, "=> $")
 			common.write_padded_hex(w, jmp_x, 4)
 			io.write_string(w, ", $")

core/text/regex/virtual_machine/virtual_machine.odin

@@ -1,5 +1,6 @@
 package regex_vm
 
+import "base:intrinsics"
 @require import "core:io"
 import "core:slice"
 import "core:text/regex/common"
@@ -121,12 +122,12 @@ add_thread :: proc(vm: ^Machine, saved: ^[2 * common.MAX_CAPTURE_GROUPS]int, pc:
 
 		#partial switch vm.code[pc] {
 		case .Jump:
-			pc = cast(int)(cast(^u16)&vm.code[pc + size_of(Opcode)])^
+			pc = cast(int)intrinsics.unaligned_load(cast(^u16)&vm.code[pc + size_of(Opcode)])
 			continue
 
 		case .Split:
-			jmp_x := cast(int)(cast(^u16)&vm.code[pc + size_of(Opcode)])^
-			jmp_y := cast(int)(cast(^u16)&vm.code[pc + size_of(Opcode) + size_of(u16)])^
+			jmp_x := cast(int)intrinsics.unaligned_load(cast(^u16)&vm.code[pc + size_of(Opcode)])
+			jmp_y := cast(int)intrinsics.unaligned_load(cast(^u16)&vm.code[pc + size_of(Opcode) + size_of(u16)])
 
 			add_thread(vm, saved, jmp_x)
 			pc = jmp_y
@@ -236,7 +237,7 @@ add_thread :: proc(vm: ^Machine, saved: ^[2 * common.MAX_CAPTURE_GROUPS]int, pc:
 			vm.top_thread += 1
 
 		case .Wait_For_Rune:
-			operand := (cast(^rune)&vm.code[pc + size_of(Opcode)])^
+			operand := intrinsics.unaligned_load(cast(^rune)&vm.code[pc + size_of(Opcode)])
 			if vm.next_rune == operand {
 				add_thread(vm, saved, pc + size_of(Opcode) + size_of(rune))
 			}
@@ -409,7 +410,7 @@ run :: proc(vm: ^Machine, $UNICODE_MODE: bool) -> (saved: ^[2 * common.MAX_CAPTU
 				}
 
 			case .Rune:
-				operand := (cast(^rune)&vm.code[t.pc + size_of(Opcode)])^
+				operand := intrinsics.unaligned_load(cast(^rune)&vm.code[t.pc + size_of(Opcode)])
 				if current_rune == operand {
 					add_thread(vm, t.saved, t.pc + size_of(Opcode) + size_of(rune))
 				}
@@ -482,7 +483,7 @@ run :: proc(vm: ^Machine, $UNICODE_MODE: bool) -> (saved: ^[2 * common.MAX_CAPTU
 				vm.top_thread += 1
 
 			case .Wait_For_Rune:
-				operand := (cast(^rune)&vm.code[t.pc + size_of(Opcode)])^
+				operand := intrinsics.unaligned_load(cast(^rune)&vm.code[t.pc + size_of(Opcode)])
 				if vm.next_rune == operand {
 					add_thread(vm, t.saved, t.pc + size_of(Opcode) + size_of(rune))
 				}
@@ -558,7 +559,7 @@ run :: proc(vm: ^Machine, $UNICODE_MODE: bool) -> (saved: ^[2 * common.MAX_CAPTU
 						break escape_loop
 
 					case .Jump:
-						t.pc = cast(int)(cast(^u16)&vm.code[t.pc + size_of(Opcode)])^
+						t.pc = cast(int)intrinsics.unaligned_load(cast(^u16)&vm.code[t.pc + size_of(Opcode)])
 
 					case .Save:
 						index := vm.code[t.pc + size_of(Opcode)]