"""Find List<Integer>/Set<Integer>.remove(int) misdecompilations.

FernFlower sometimes resolves List.remove(Object) to remove(int index) when the
element type is Integer. The bytecode always boxes with Integer.valueOf first,
so a source-level remove() on an Integer collection without that box is a bug:
it deletes by position, and throws IndexOutOfBounds on an empty list.

    python find_remove_index_bugs.py <src-root>
"""

import io
import os
import re
import sys


FIELD = re.compile(r'\b(?:private|public|protected)?\s*(?:final\s+)?'
                   r'((?:java\.util\.)?(?:List|Set|Collection|ArrayList|LinkedList|Queue|Deque))\s*'
                   r'<\s*(?:java\.lang\.)?Integer\s*>\s+(\w+)\s*[;=]')
REMOVE = re.compile(r'\b(\w+)\.remove\(([^()]*)\)')


def main():
    root = sys.argv[1]
    hits = []
    for base, _dirs, files in os.walk(root):
        for name in files:
            if not name.endswith('.java'):
                continue
            path = os.path.join(base, name)
            with io.open(path, encoding='utf-8') as handle:
                text = handle.read()
            fields = {}
            for match in FIELD.finditer(text):
                fields[match.group(2)] = match.group(1)
            if not fields:
                continue
            for number, line in enumerate(text.split('\n'), 1):
                for match in REMOVE.finditer(line):
                    receiver, argument = match.group(1), match.group(2).strip()
                    if receiver not in fields:
                        continue
                    if 'valueOf' in argument or 'Integer' in argument:
                        continue
                    hits.append((os.path.relpath(path, root).replace(os.sep, '/'),
                                 number, receiver, argument, line.strip()))

    if not hits:
        print('no List<Integer>.remove(int) candidates found')
        return 0
    print('%d candidate(s):' % len(hits))
    for path, number, receiver, argument, line in hits:
        print('  %s:%d  %s.remove(%s)' % (path, number, receiver, argument))
        print('      %s' % line)
    return 0


if __name__ == '__main__':
    raise SystemExit(main())