From b6a7eefc9498e116b812c59cbda34ba398859020 Mon Sep 17 00:00:00 2001
From: cfreksen <casper@freksen.dk>
Date: Sun, 29 Oct 2017 02:56:26 +0200
Subject: [PATCH] Add WIP parser.

---
 parser.py | 103 +++++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 87 insertions(+), 16 deletions(-)

diff --git a/parser.py b/parser.py
index 1e6b5f1..e0936a4 100644
--- a/parser.py
+++ b/parser.py
@@ -1,8 +1,19 @@
 """Parser for LLVM--."""
 
-import ply.lex as lex
+from collections import namedtuple
+
+import ply.lex as lex
+import ply.yacc as yacc
+
+# Namedtuples for storing AST
+Program = namedtuple('Program', ['tdecls', 'gdecls', 'fdecls'])
+FunctionDec = namedtuple('FunctionDec', ['return_type', 'name', 'parameters', 'body'])
+FunctionBody = namedtuple('FunctionBody', ['first_block', 'named_blocks'])
+Block = namedtuple('Block', ['insns', 'terminator'])
+
+class LLVMParser(object):
+    # Lexer
 
-class LLVMLexer(object):
     current_string = ""
     line_begin = 0
 
@@ -77,10 +88,6 @@ class LLVMLexer(object):
     t_COMMA = r','
 
 
-    def __init__(self):
-        self.tokens += self.reserved.values()
-
-
     def t_COMMENT(self, t):
         r'(;|declare|target).*'
         pass
@@ -169,20 +176,84 @@ class LLVMLexer(object):
         t.lexer.skip(1)
 
 
+    # Parser
+    def handle_top_decs(self, smth):
+        # TODO: Implement
+        return [], [], smth
+
+    def p_program(self, p):
+        'program : top_decs'
+        tdecls, gdecls, fdecls = self.handle_top_decs(p[1])
+        p[0] = Program(tdecls, gdecls, fdecls)
+
+    def p_topdecs_fdec(self, p):
+        'top_decs : fdec top_decs'
+        p[0] = [p[1]] + p[2]
+
+    def p_topdecs_empty(self, p):
+        'top_decs : '
+        p[0] = []
+
+    def p_fdec(self, p):
+        'fdec : DEFINE ty AtID LPAREN ty_id_list RPAREN LBRACE fbody RBRACE'
+        p[0] = FunctionDec(p[2], p[3], p[5], p[8])
+
+    def p_ty_simple(self, p):
+        '''ty : VOID
+            | I1
+            | I8
+            | I32
+            | I64'''
+        p[0] = p[1]
+
+    def p_ty_id_list_single(self, p):
+        'ty_id_list : ty PercentID'
+        p[0] = [(p[1], p[2])]
+
+    def p_ty_id_list_multiple(self, p):
+        'ty_id_list : ty PercentID COMMA ty_id_list'
+        p[0] = [(p[1], p[2])] + p[4]
+
+    def p_ty_id_list_empty(self, p):
+        'ty_id_list : '
+        p[0] = []
+
+    def p_fbody_one_block(self, p):
+        'fbody : block'
+        p[0] = FunctionBody(p[1], [])
+
+    def p_block_terminator(self, p):
+        'block : terminator'
+        p[0] = Block([], p[1])
+
+    def p_terminator_ret_oper(self, p):
+        'terminator : RET ty operand'
+        p[0] = (p[2], p[3])
+
+    def p_operand(self, p):
+        '''operand : NULL
+            | INT
+            | AtID
+            | PercentID'''
+        p[0] = p[1]
+
+    def __init__(self):
+        self.tokens += self.reserved.values()
+
     def build(self, **kwargs):
         self.lexer = lex.lex(module=self, **kwargs)
-
+        self.parser = yacc.yacc(module=self, **kwargs)
 
     def test(self, data):
-        self.lexer.input(data)
-        for tok in self.lexer:
-            print(tok)
-
+        result = self.parser.parse(data, lexer=self.lexer)
+        print(result)
 
 if __name__ == '__main__':
-    m = LLVMLexer()
-    m.build()
-    data = r'''123 456 c"abc _\2c_ \" \\ and so on" def add sdiv ; some comment
-    qqq
+    p = LLVMParser()
+    p.build()
+    data = r'''
+define void @tigermain (i64 %U_mainSL_8, i64 %U_mainDummy_9) {
+ ret i64 8
+}
     '''
-    m.test(data)
+    p.test(data)