From 2a9ea9548f4627e993ebeedb164a3b288955f712 Mon Sep 17 00:00:00 2001
From: Russ Olsen <russ@cognitect.com>
Date: Tue, 14 Apr 2020 12:44:37 -0400
Subject: [PATCH] Initial checkin.

---
 README.md                |   9 ++
 sallyforth/compiler.py   |  28 ++++++
 sallyforth/kernel.py     | 115 ++++++++++++++++++++++
 sallyforth/lex.py        |  55 +++++++++++
 sallyforth/sallyforth.py |   6 ++
 sallyforth/stack.py      |  29 ++++++
 sallyforth/words.py      | 200 +++++++++++++++++++++++++++++++++++++++
 7 files changed, 442 insertions(+)
 create mode 100644 README.md
 create mode 100644 sallyforth/compiler.py
 create mode 100644 sallyforth/kernel.py
 create mode 100644 sallyforth/lex.py
 create mode 100644 sallyforth/sallyforth.py
 create mode 100644 sallyforth/stack.py
 create mode 100644 sallyforth/words.py

diff --git a/README.md b/README.md
new file mode 100644
index 0000000..cf8b0e9
--- /dev/null
+++ b/README.md
@@ -0,0 +1,9 @@
+SallyForth: A simple Forth-like language implemented in Python.
+##############################################################
+
+SallyForth is a simple hobby implementation of the FORTH programming
+language. Possibly the most interesting thing about SallyForth is the
+name, which 
+[Michael Nygard suggested](https://twitter.com/mtnygard/status/1249781530219642883)
+as a name for a FORTH implementation
+at exactly the same time that I happened to be writing this code.
diff --git a/sallyforth/compiler.py b/sallyforth/compiler.py
new file mode 100644
index 0000000..24f4572
--- /dev/null
+++ b/sallyforth/compiler.py
@@ -0,0 +1,28 @@
+from lex import forth_prompt
+from stack import Stack
+
+class Compiler:
+    def __init__(self, name=None):
+        self.name = name
+        self.instructions = []
+        self.offsets = Stack()
+
+    def add_instruction(self, ins):
+        self.instructions.append(ins)
+
+    def offset(self):
+        return len(self.instructions)
+
+    def push_offset(self):
+        self.offsets.push(self.offset())
+
+    def pop_offset(self):
+        return self.offsets.pop()
+
+    def _str__(self):
+        result = f'Compiler {name} {immediate} '
+        for i in self.instructions:
+            result += str(i)
+            result += ' '
+        return result
+
diff --git a/sallyforth/kernel.py b/sallyforth/kernel.py
new file mode 100644
index 0000000..3f512b3
--- /dev/null
+++ b/sallyforth/kernel.py
@@ -0,0 +1,115 @@
+import sys
+from os import path
+from words import *
+from lex import forth_prompt, read_tokens, is_string
+from stack import Stack
+
+def to_number(token):
+    try:
+        return int(token)
+    except ValueError:
+        try:
+            return float(token)
+        except ValueError:
+            return None
+
+def push_value_f(value):
+    def x(f):
+        f.stack.push(value)
+        return 1
+    return x
+
+class Forth:
+    def __init__(self, startup=None):
+        self.stack = Stack()
+        self.dictionary = {
+                'true': const_f(True),
+                'false': const_f(False),
+                'nil': const_f(None),
+                '0': const_f(0),
+                '1': const_f(1),
+                '2': const_f(2),
+                ';': w_semi,
+                ':': w_colon,
+                '+': w_add,
+                '+': w_add,
+                '-': w_sub,
+                '/': w_div,
+                '>': w_gt,
+                '<': w_lt,
+                '<=': w_le,
+                '>=': w_ge,
+                '=': w_eq,
+                'dup': w_dup,
+                'swap': w_swap,
+                '.': w_dot,
+                'nl': w_nl,
+                'dump': w_dump,
+                'idump': w_idump,
+                'stack': w_stack,
+                'begin': w_begin,
+                'until': w_until,
+                'if': w_if,
+                'then': w_then}
+
+        self.compiler = None
+        if startup:
+            execute_startup(startup)
+
+    def compiling(self):
+        return self.compiler
+
+    def process_line(self, readline_f=forth_prompt):
+        tokens = read_tokens(readline_f)
+        for token in tokens:
+            if not self.compiling():
+                self.interpret_token(token)
+            else:
+                self.compile_token(token)
+
+    def compile_token(self, token):
+        if self.compiler.name == None:
+            self.compiler.name = token
+            return
+
+        if is_string(token):
+            self.compiler.add_instruction(push_value_f(token[1::]))
+            return
+
+        if token in self.dictionary:
+            word = self.dictionary[token]
+            if 'immediate' in word.__dict__:
+                #print("before immediate word:", self, self.dictionary)
+                word(self)
+                #print("after immediate word:", self, self.dictionary)
+            else:
+                self.compiler.add_instruction(self.dictionary[token])
+            return
+
+        n = to_number(token)
+        if n == None:
+            self.compiler = None
+            print(token, "? Compile terminated.")
+        else:
+            self.compiler.add_instruction(push_value_f(n))
+
+    def interpret_token(self, token):
+        if is_string(token):
+            self.stack.push(token)
+            return
+
+        if token in self.dictionary:
+            self.dictionary[token](self)
+            return
+
+        n = to_number(token)
+        if n == None:
+            print(token, "?")
+        else:
+            self.stack.push(n)
+
+    def dump(self):
+        print("Forth:", self)
+        print("Stack:", self.stack)
+        print("Dictionary:", self.dictionary)
+        print("Compiler:", self.compiler)
diff --git a/sallyforth/lex.py b/sallyforth/lex.py
new file mode 100644
index 0000000..b223750
--- /dev/null
+++ b/sallyforth/lex.py
@@ -0,0 +1,55 @@
+import sys
+from os import path
+
+def is_string(token):
+    #print("is string:", token, token[0], token[0] == '"')
+    return token[0] == '"'
+
+def is_space(ch):
+    return ch == " " or ch == "\t" or ch == "\n"
+
+def tokenize(s):
+    state = 'start'
+    token = ''
+    tokens = []
+    for ch in s:
+        #print(f'Loop state {state} token {token} ch {ch}')
+        if state == 'start' and is_space(ch):
+            continue
+        elif state == 'start' and ch == '"':
+            token = ch
+            state = 'string'
+        elif state == 'start':
+            token = ch
+            state = 'word'
+        elif state == 'string' and ch == '"':
+            tokens.append(token)
+            state = 'start'
+            token = ''
+        elif state == 'word' and is_space(ch):
+            tokens.append(token)
+            state = 'start'
+            token = ''
+        elif state == 'word' or state == 'string':
+            token += ch
+        else:
+            print(f'State: [{state}] token: [{token}] ch: [{ch}]???')
+            state = 'start'
+    if len(token) > 0:
+        tokens.append(token)
+    return tokens
+
+def read_ch():
+    return sys.stdin.read(1)
+
+def read_tokens(read_f):
+    line = read_f()
+    return tokenize(line)
+
+def forth_prompt():
+    return input("SallyForth>> ")
+
+def file_read_f(f):
+    def read_it():
+        return f.readline()
+    return read_it
diff --git a/sallyforth/sallyforth.py b/sallyforth/sallyforth.py
new file mode 100644
index 0000000..c9fa480
--- /dev/null
+++ b/sallyforth/sallyforth.py
@@ -0,0 +1,6 @@
+from kernel import Forth
+
+f = Forth()
+while True:
+    f.process_line()
+
diff --git a/sallyforth/stack.py b/sallyforth/stack.py
new file mode 100644
index 0000000..ca27995
--- /dev/null
+++ b/sallyforth/stack.py
@@ -0,0 +1,29 @@
+class Stack:
+    def __init__(self):
+        self.top = -1
+        self.stack = 100 * [None]
+
+    def push(self, x):
+        # print("stack push", x)
+        self.top += 1
+        self.stack[self.top] = x
+        return x
+
+    def pop(self):
+        result = self.stack[self.top]
+        # print("stack pop", result)
+        self.top -= 1
+        if self.top < -1:
+            print("stack overpop")
+            self.top = -1;
+        return result
+
+    def peek(self):
+        return self.stack[self.top]
+
+    def __str__(self):
+        result = ''
+        for i in range(self.top + 1):
+            result += str(self.stack[i])
+            result += ' '
+        return result
diff --git a/sallyforth/words.py b/sallyforth/words.py
new file mode 100644
index 0000000..8426fe7
--- /dev/null
+++ b/sallyforth/words.py
@@ -0,0 +1,200 @@
+from compiler import Compiler
+
+def execute_f(name, instructions):
+    # print("execute_f:", len(instructions))
+    def inner(forth, debug=False):
+        # print("inner f:", name)
+        # print("inner f:", len(instructions))
+        i = 0
+        while i >= 0:
+            # print(i, "=>", instructions[i])
+            delta = instructions[i](forth)
+            i += delta
+        return 1
+    return inner
+
+def ifnot_jump_f(n):
+    def ifnot_jump(forth):
+        # print("If not jump:")
+        x = forth.stack.pop()
+        # print("==>value:", x)
+        if not x:
+            # print("==>", x, " is false")
+            # print("==>returning", n)
+            return n
+        # print("==>returning 1")
+        return 1
+    return ifnot_jump
+
+def const_f(value):
+    def x(f):
+        f.stack.push(value)
+        return 1
+    return x
+
+def w_gt(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b > a)
+    return 1
+
+def w_lt(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b < a)
+    return 1
+
+def w_eq(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(a==b)
+    return 1
+
+def w_le(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b<=a)
+    return 1
+
+def w_ge(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b>=a)
+    return 1
+
+def w_add(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b+a)
+    return 1
+
+def w_mul(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b*a)
+    return 1
+
+def w_sub(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b-a)
+    return 1
+
+def w_div(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(b/a)
+    return 1
+
+def w_dot(f):
+    a = f.stack.pop()
+    print(a, end='')
+    return 1
+
+def w_dup(f):
+    x = f.stack.peek()
+    f.stack.push(x)
+    return 1
+
+def w_swap(f):
+    a = f.stack.pop()
+    b = f.stack.pop()
+    f.stack.push(a)
+    f.stack.push(b)
+    return 1
+
+def w_nl(f):
+    print()
+    return 1
+
+def w_return(f):
+    return -9999999999
+
+def w_colon(f):
+    f.compiler = Compiler()
+
+def w_semi(forth):
+    forth.compiler.add_instruction(w_return)
+    name = forth.compiler.name
+    word_f = execute_f(name, forth.compiler.instructions)
+    forth.dictionary[name] = word_f
+    forth.compiler = None
+    return 1
+
+w_semi.__dict__['immediate'] = True
+
+def w_should_not_happen(forth):
+    print("Should not execute this word!")
+    raise ValueError
+
+def w_if(forth):
+    print("w_if")
+    compiler = forth.compiler
+    compiler.push_offset()
+    compiler.add_instruction(w_should_not_happen)
+    return 1
+
+w_if.__dict__['immediate'] = True
+
+def w_then(forth):
+    compiler = forth.compiler
+    if_offset = compiler.pop_offset()
+    end_offset = compiler.offset()
+    delta = end_offset - if_offset
+    compiler.instructions[if_offset] = ifnot_jump_f(delta)
+    return 1
+
+w_then.__dict__['immediate'] = True
+
+def w_do(forth):
+    print("w_do")
+    compiler = forth.compiler
+    compiler.push_offset()
+    compiler.add_instruction(w_should_not_happen)
+    return 1
+
+w_do.__dict__['immediate'] = True
+
+def w_while(forth):
+    compiler = forth.compiler
+    do_offset = compiler.pop_offset()
+    while_offset = compiler.offset()
+    delta = do_offset - while_offset
+    compiler.instructions[if_offset] = ifnot_jump_f(delta)
+    return 1
+
+w_while.__dict__['immediate'] = True
+
+def w_begin(forth):
+    compiler = forth.compiler
+    compiler.push_offset()
+    return 1
+
+w_begin.__dict__['immediate'] = True
+
+def w_until(forth):
+    compiler = forth.compiler
+    begin_offset = compiler.pop_offset()
+    until_offset = compiler.offset()
+    delta = begin_offset - until_offset
+    print("Delta:", delta)
+    compiler.instructions.append(ifnot_jump_f(delta))
+    return 1
+
+
+w_until.__dict__['immediate'] = True
+
+def w_dump(f):
+    f.dump()
+    return 1
+
+def w_idump(f):
+    f.dump()
+    return 1
+
+w_idump.__dict__['immediate'] = True
+
+def w_stack(f):
+    print(f.stack)
+    return 1
+