#!/usr/bin/env python3
"""v4tof18.py -- turn the nucleus's .v4 sources into the system's .f18 blobs.

    v4/tools/v4tof18.py

Run once (MACHINE.md step 2; plan Task 8): reads v4/capsule/*.v4 and writes
capsules/v4/sys/*.f18.  Everything is kept token for token and line for
line except `header NAME flags :`, which text.c lays down itself and the
ROM's language says with its directives:

    set HDR-NFA here  ", NAME  #, F  #, HDR-NFA  #, HDR-LINK  set HDR-LINK here  :

-- the name as a counted string, the flags (immediate 1, data 4, inline 8,
compile-only 16), the name's address and the link to the header before;
HDR-LINK is then the new entry's code address, as text.c's latest is.  The
first file begins `set HDR-LINK 0`.  The tokens are found as text.c's
next_token finds them: `\\` and `(` as whole tokens begin comments; a
header's name is the next run of characters, whatever they are.
"""
import os
import sys

ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), "..", ".."))
SRC = os.path.join(ROOT, "v4", "capsule")
OUT = os.path.join(ROOT, "capsules", "v4", "sys")
FILES = ["core", "input", "dict", "codegen", "compile", "quit", "forth", "numout",
         "words", "system", "qmath", "blocks", "log", "acl"]
FLAGS = {"immediate": 1, "data": 4, "inline": 8, "compile-only": 16}
BLANK = " \t\r\n"


def tokens(text):
    """(token, start, end) for each token text.c's next_token would return,
    and for each header its raw name."""
    i, n = 0, len(text)
    raw_next = False
    while True:
        while i < n and text[i] in BLANK:
            i += 1
        if i >= n:
            return
        j = i
        while j < n and text[j] not in BLANK:
            j += 1
        tok = text[i:j]
        if raw_next:
            raw_next = False
            yield tok, i, j
            i = j
            continue
        if tok == "\\":
            while j < n and text[j] != "\n":
                j += 1
            i = j
            continue
        if tok == "(":
            k = text.find(")", j)
            if k < 0:
                sys.exit("comment not closed")
            i = k + 1
            continue
        yield tok, i, j
        if tok == "header":
            raw_next = True
        i = j


def convert(text, first):
    out, last = [], 0
    toks = list(tokens(text))
    k = 0
    while k < len(toks):
        tok, s, e = toks[k]
        if tok != "header":
            k += 1
            continue
        name = toks[k + 1][0]
        flags, m = 0, k + 2
        while toks[m][0] != ":":
            if toks[m][0] not in FLAGS:
                sys.exit("header %s: %s is not a flag" % (name, toks[m][0]))
            flags |= FLAGS[toks[m][0]]
            m += 1
        colon = toks[m][1]
        span = text[s:colon]
        directive = ('set HDR-NFA here  ", %s  #, %d  #, HDR-NFA  #, HDR-LINK  set HDR-LINK here  '
                     % (name, flags))
        out.append(text[last:s])
        out.append(directive + "\n" * span.count("\n"))
        last = colon
        k = m
    out.append(text[last:])
    body = "".join(out)
    if first:
        body = "set HDR-LINK 0  \\ the newest dictionary entry: none yet\n" + body
    return body


def main():
    os.makedirs(OUT, exist_ok=True)
    for i, f in enumerate(FILES):
        with open(os.path.join(SRC, f + ".v4")) as h:
            text = h.read()
        with open(os.path.join(OUT, f + ".f18"), "w") as h:
            h.write(convert(text, i == 0))
    print("v4tof18: %d files -> %s" % (len(FILES), OUT))


if __name__ == "__main__":
    main()