xiiregexbuilder

 1import sys
 2import re
 3
 4# Enhanced Assembler for Regex CPU
 5# Layout: [31:24] char | [23:16] next1 | [15:8] next2 | [7:4] mid | [3] term | [0] any
 6
 7
 8def pack(char_val, next1, next2, mid, term, any_bit):
 9    return (
10        (char_val << 24)
11        | (next1 << 16)
12        | (next2 << 8)
13        | (mid << 4)
14        | (term << 3)
15        | any_bit
16    )
17
18
19def main():
20    input_file = sys.argv[1] if len(sys.argv) > 1 else "regexes.rasm"
21    output_file = sys.argv[2] if len(sys.argv) > 2 else "imem.hex"
22
23    mem = [0] * 256
24    with open(input_file, "r") as f:
25        for line in f:
26            line = line.strip()
27            if not line or line.startswith("#"):
28                continue
29
30            m = re.match(r"(\d+):\s+(\w+)(?:\((.*)\))?", line)
31            if not m:
32                continue
33
34            pc, instr, args_str = int(m.group(1)), m.group(2), m.group(3)
35            args = []
36            if args_str:
37                for arg in re.split(r",\s*", args_str):
38                    arg = arg.strip()
39                    if arg.startswith("'") and arg.endswith("'"):
40                        args.append(ord(arg[1:-1]))
41                    else:
42                        args.append(int(arg))
43
44            if instr == "CHAR":
45                # Support: CHAR('a') OR CHAR('a', n1, n2, mid, term, any)
46                c = args[0]
47                n1 = args[1] if len(args) > 1 else pc + 1
48                n2 = args[2] if len(args) > 2 else 0
49                mid = args[3] if len(args) > 3 else 0
50                term = args[4] if len(args) > 4 else 0
51                any_b = args[5] if len(args) > 5 else 0
52                mem[pc] = pack(c, n1, n2, mid, term, any_b)
53            elif instr == "SPLIT":
54                mem[pc] = pack(0, args[0], args[1], 0, 0, 0)
55            elif instr == "JMP":
56                mem[pc] = pack(0, args[0], 0, 0, 0, 0)
57            elif instr == "MATCH":
58                mem[pc] = pack(0, 0, 0, args[0], 1, 0)
59            elif instr == "ANY":
60                n1 = args[0] if len(args) > 0 else pc + 1
61                mem[pc] = pack(0, n1, 0, 0, 0, 1)
62
63    with open(output_file, "w") as f:
64        for val in mem:
65            f.write(f"{val:08x}\n")
66    print(f"Assembled {input_file} to {output_file}")
67
68
69if __name__ == "__main__":
70    main()