From a7308280cb5daaff1f42eedbd450e6f3fc3c93b6 Mon Sep 17 00:00:00 2001
From: Peter McGoron <code@mcgoron.com>
Date: Tue, 7 Feb 2023 16:38:04 +0000
Subject: [PATCH] start assembler

---
 README.md |   7 ++++
 asm.py    | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+)
 create mode 100644 asm.py

diff --git a/README.md b/README.md
index d8de334..cf682b3 100644
--- a/README.md
+++ b/README.md
@@ -21,3 +21,10 @@ The first four bytes determine the type:
 All other values are reserved. Overlong values are allowed, and for some
 argument values they are necessary. All lines are terminated by a byte
 of all zeros.
+
+## Design Philsophy
+
+Creole is small but not minimal. It is easy to add instructions, and
+the amount of labels, registers, and stack space can changed at runtime.
+System calls can be added dynamically, but static instructions that take
+at most 3 arguments should be hard-coded.
diff --git a/asm.py b/asm.py
new file mode 100644
index 0000000..cd6a94d
--- /dev/null
+++ b/asm.py
@@ -0,0 +1,109 @@
+#!/usr/bin/python3
+
+from enum import Enum
+
+class ArgType(Enum):
+	TYPE_IMM = 1
+	TYPE_REG = 2
+	TYPE_VAL = 3
+	TYPE_LAB = 4
+
+	def gettype(s):
+		if s.isnumeric():
+			return (TYPE_IMM, int(s))
+		elif s[0] == 'r' and is_num(s[1:]):
+			return (TYPE_REG, int(s[1:]))
+		elif s[0] == 'l' and is_num(s[1:]):
+			return (TYPE_LAB, int(s[1:]))
+		else:
+			return None
+
+	def typecheck(self, s):
+		t = gettype(s)
+		if t is None:
+			return None
+		if self == TYPE_VAL:
+			return t[0] == TYPE_REG or t[0] == TYPE_IMM
+		else:
+			return t[0] == self
+
+class Instruction:
+	def __init__(self,  opcode, argtypes):
+		self.opcode = opcode
+		assert self.opcode < 0x80 and self.opcode >= 0
+		self.argtypes = argtypes
+	def typecheck(self, sargs):
+		rargs = []
+		if len(sargs) != len(self.argtypes):
+			return None
+		for i in range(0, len(sargs)):
+			if not self.argtypes[i].typecheck(sargs[i]):
+				return None	
+			rargs.append(ArgType.gettype(sargs[i]))
+		return rargs
+
+instructions = {
+"NOP" : Instruction(0, []),
+"PUSH" : Instruction(1, [ArgType.TYPE_REG]),
+"POP" : Instruction(2, [ArgType.TYPE_REG]),
+"ADD" : Instruction(3, [ArgType.TYPE_VAL, ArgType.TYPE_VAL, ArgType.TYPE_VAL]),
+"MUL" : Instruction(4, [ArgType.TYPE_VAL, ArgType.TYPE_VAL, ArgType.TYPE_VAL]),
+"DIV" : Instruction(5, [ArgType.TYPE_VAL, ArgType.TYPE_VAL, ArgType.TYPE_VAL]),
+"JL" : Instruction(6, [ArgType.TYPE_LAB, ArgType.TYPE_VAL, ArgType.TYPE_VAL]),
+"CLB" : Instruction(7, [ArgType.TYPE_LAB]),
+"SYS" : Instruction(8, [ArgType.TYPE_VAL])
+}
+
+encoding_types = {
+#           start      mask   A  B
+	2: (0x7F,       0xC0, 7),
+	3: (0xFFF,      0xE0, 12),
+	4: (0x1FFFF,    0xF0, 16),
+	5: (0x3FFFFF,   0xF8, 21),
+	6: (0x7FFFFFF,  0xFC, 26),
+	7: (0xFFFFFFFF, 0xFE, 36),
+#	A : number of bits in start byte
+#	B : Total number of bits excluding high bits
+}
+
+def encode_pseudo_utf8(n, high_bits, to):
+	if to > 8 or to < 0:
+		return None
+	elif to == 1:
+		if n < 0x80:
+			return bytes([n])
+		else:
+			return None
+
+	(maxval, start_byte, n_tot) = encoding_types[to]
+	if n > maxval or high_bits > 15:
+		return None
+	n = n | (high_bits << n_tot)
+	all_bytes = []
+	for i in range(0, to - 1):
+		all_bytes.append(0x80 | (n & 0x3F))
+		n >>= 6
+	all_bytes.append(start_byte | n)
+	return bytes(reversed(all_bytes))
+
+"""
+class Line:
+	def __init__(self, opcode, args):
+		self.opcode = opcode
+		self.args = args
+	def render(self):
+
+class Program:
+	def asm_push_line(self, ins, args):
+		self.asm.append((ins, args))
+
+	def parse_asm_line(self, line):
+		line = line.split(' ')
+		if line[0] not in instructions:
+			raise Exception
+		else:
+			r = instructions[line[0]].typecheck(line[1:])
+			if r is None:
+				raise Exception
+			asm_push_instruction(instructions[line[0]].opcode, r)
+"""