#!/usr/bin/env python3
"""gapdiff.py -- score a candidate re-export against the canonical one.

    python tools/gapdiff.py [oldDecompDir] [newDecompDir]
        defaults: reference/decomp   reference/decomp/reexport

Both dirs must already have been censused (tools/gapcensus.py <dir>).
Reports: index/coverage deltas, dark regions closed, and -- the part that
matters for reconstruction -- which previously-dark addresses the KB already
cites (past raw-disasm digs) now have real pseudocode, plus how much of the
never-touched game-side dark came into the light.
"""
import os
import re
import sys
import glob

ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
OLD = os.path.abspath(sys.argv[1]) if len(sys.argv) > 1 else os.path.join(ROOT, 'reference', 'decomp')
NEW = os.path.abspath(sys.argv[2]) if len(sys.argv) > 2 else os.path.join(ROOT, 'reference', 'decomp', 'reexport')

# addresses the KB/port cite (a past dig reached them) -- the practical audience
CITE = set()
_pat = re.compile(r'(?:@|FUN_|0x)0{0,2}(4[0-9a-f]{5})\b', re.I)
for base in ('game', 'context', 'docs'):
    for dp, _d, fs in os.walk(os.path.join(ROOT, base)):
        for fn in fs:
            if fn.lower().endswith(('.cpp', '.hpp', '.h', '.md')):
                try:
                    for m in _pat.finditer(open(os.path.join(dp, fn), encoding='utf-8',
                                                errors='replace').read()):
                        CITE.add(int(m.group(1), 16))
                except OSError:
                    pass


def load_index(d):
    idx = {}
    p = os.path.join(d, 'functions_index.tsv')
    for line in open(p, encoding='utf-8', errors='replace'):
        f = line.rstrip('\n').split('\t')
        if len(f) < 4:
            continue
        try:
            idx[int(f[0], 16)] = (int(f[1]), f[2], f[3])
        except ValueError:
            pass
    return idx


def load_regions(d):
    p = os.path.join(d, 'gap_census.tsv')
    out = []
    for line in open(p, encoding='utf-8', errors='replace').read().splitlines()[1:]:
        f = line.split('\t')
        if len(f) < 8:
            f += [''] * (8 - len(f))
        out.append({'lo': int(f[0], 16), 'hi': int(f[1], 16), 'size': int(f[2]),
                    'code': int(f[3]), 'visited': f[6], 'tu': f[7]})
    return out


def exported_addrs(d):
    got = set()
    hdr = re.compile(r'/\* @(00[0-9a-f]{6}) ')
    for p in sorted(glob.glob(os.path.join(d, 'all', 'part_*.c'))):
        for m in hdr.finditer(open(p, encoding='utf-8', errors='replace').read()):
            got.add(int(m.group(1), 16))
    return got


def main():
    oi, ni = load_index(OLD), load_index(NEW)
    orr, nr = load_regions(OLD), load_regions(NEW)
    oe, ne = exported_addrs(OLD), exported_addrs(NEW)

    o_dark = sum(r['code'] for r in orr)
    n_dark = sum(r['code'] for r in nr)
    o_game = sum(r['code'] for r in orr if 'bt/' in r['tu'] or 'bt_l4/' in r['tu'])
    n_game = sum(r['code'] for r in nr if 'bt/' in r['tu'] or 'bt_l4/' in r['tu'])

    new_fns = sorted(set(ni) - set(oi))
    # which new functions land in what USED to be dark
    def in_old_dark(a):
        for r in orr:
            if r['lo'] <= a < r['hi']:
                return True
        return False
    new_in_dark = [a for a in new_fns if in_old_dark(a)]
    cited_new = [a for a in new_in_dark if a in CITE]

    print('=== GAP DIFF: %s -> %s' % (os.path.basename(OLD), os.path.basename(NEW)))
    print('indexed functions : %d -> %d  (+%d)' % (len(oi), len(ni), len(ni) - len(oi)))
    print('exported pseudocode: %d -> %d  (+%d)' % (len(oe), len(ne), len(ne) - len(oe)))
    print('dark regions      : %d -> %d  (%+d)' % (len(orr), len(nr), len(nr) - len(orr)))
    print('dark REAL code    : %.1f KB -> %.1f KB  (%+.1f KB, %.1f%% recovered)'
          % (o_dark / 1024, n_dark / 1024, (n_dark - o_dark) / 1024,
             100.0 * (o_dark - n_dark) / o_dark if o_dark else 0))
    print('  game-side dark  : %.1f KB -> %.1f KB  (%+.1f KB)'
          % (o_game / 1024, n_game / 1024, (n_game - o_game) / 1024))
    print('new functions in previously-DARK bytes: %d' % len(new_in_dark))
    print('  ...of which the KB/port already cites (past raw-disasm digs): %d' % len(cited_new))
    if cited_new:
        print('  ' + ' '.join('%x' % a for a in cited_new[:24])
              + (' ...' if len(cited_new) > 24 else ''))

    # celebrity check -- the addresses that historically burned us
    fam = [(0x4c05c4, 'VehicleDead handler'), (0x4b838c, 'Searchlight ToggleLamp'),
           (0x4a07b8, 'TakeDamage death tail'), (0x4a9b5c, 'master perf'),
           (0x4b8be3, 'myomer integrator'), (0x4aa011, 'duck consumer'),
           (0x486467, 'btmssn dispatch cluster'), (0x4c090a, 'btplayer hole')]
    print('--- celebrities (function created? exported?)')
    for a, what in fam:
        # nearest containing function in the new index
        owner = max((x for x in ni if x <= a and x + ni[x][0] > a), default=None)
        print('  %-24s %s' % (what, ('IN %06x (%s)%s' % (owner, ni[owner][2],
                                                         ' +pseudocode' if owner in ne else ''))
                              if owner else 'still dark'))


if __name__ == '__main__':
    main()