RFC: Python 3 and speedup

setup.py

@@ -1,5 +1,5 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 from distutils.core import setup
 

xortool/__init__.py

@@ -1,5 +1,5 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 __all__ = ["args", "colors", "libcolors", "routine"]
 __version__ = "0.98"

xortool/args.py

@@ -1,31 +1,54 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 from docopt import docopt
 
-from xortool.routine import parse_char
 from xortool.charset import get_charset
 
 
 class ArgError(Exception):
     pass
 
 
+def parse_char(ch):
+    """
+    'A' or '\x41' or '0x41' or '41'
+    '\x00' or '0x00' or '00'
+    """
+    if ch is None:
+        return None
+    if len(ch) == 1:
+        return bytes([ord(ch)])
+    if ch[0:2] in ("0x", "\\x"):
+        ch = ch[2:]
+    if not ch:
+        raise ValueError("Empty char")
+    if len(ch) > 2:
+        raise ValueError("Char can be only a char letter or hex")
+    return bytes([int(ch, 16)])
+
+
+def parse_int(i):
+    if i is None:
+        return None
+    return int(i)
+
+
 def parse_parameters(doc, version):
     p = docopt(doc, version=version)
     p = {k.lstrip("-"): v for k, v in p.items()}
     try:
         return {
-            "input_is_hex": bool(p["hex"]),
-            "max_key_length": int(p["max-keylen"]),
-            "known_key_length": int(p["key-length"]) if p["key-length"] else None,
-            "most_frequent_char": parse_char(p["char"]) if p["char"] else None,
             "brute_chars": bool(p["brute-chars"]),
             "brute_printable": bool(p["brute-printable"]),
-            "text_charset": get_charset(p["text-charset"]),
-            "frequency_spread": 0,  # to be removed
             "filename": p["FILE"] if p["FILE"] else "-",  # stdin by default
             "filter_output": bool(p["filter-output"]),
+            "frequency_spread": 0,  # to be removed
+            "input_is_hex": bool(p["hex"]),
+            "known_key_length": parse_int(p["key-length"]),
+            "max_key_length": parse_int(p["max-keylen"]),
+            "most_frequent_char": parse_char(p["char"]),
+            "text_charset": get_charset(p["text-charset"]),
         }
     except ValueError as err:
         raise ArgError(str(err))

xortool/charset.py

@@ -1,8 +1,10 @@
-#!/usr/bin/env python
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 import string
 
+import numpy as np
+
 
 class CharsetError(Exception):
     pass
@@ -23,7 +25,7 @@ class CharsetError(Exception):
 }
 
 
-def get_charset(charset):
+def _get_charset_string(charset):
     charset = charset or "printable"
     if charset in PREDEFINED_CHARSETS.keys():
         return PREDEFINED_CHARSETS[charset]
@@ -32,5 +34,10 @@ def get_charset(charset):
         for c in set(charset):
             _ += CHARSETS[c]
         return _
-    except KeyError as err:
-        raise CharsetError("Bad character set")
+    except KeyError:
+        raise CharsetError("Bad character set: ", charset)
+
+
+def get_charset(charset):
+    charset_string = _get_charset_string(charset)
+    return np.array(list(bytes(charset_string, 'utf8')), dtype=np.uint8)

xortool/colors.py

@@ -1,5 +1,5 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 from xortool.libcolors import color
 
@@ -17,3 +17,17 @@
 C_KEY = color("red", attrs="bold")
 C_BOLD = color(attrs="bold")
 C_COUNT = color("yellow", attrs="bold")
+
+COLORS = {
+    'C_RESET': C_RESET,
+    'C_FATAL': C_FATAL,
+    'C_WARN': C_WARN,
+    'C_KEYLEN': C_KEYLEN,
+    'C_PROB': C_PROB,
+    'C_BEST_KEYLEN': C_BEST_KEYLEN,
+    'C_BEST_PROB': C_BEST_PROB,
+    'C_DIV': C_DIV,
+    'C_KEY': C_KEY,
+    'C_BOLD': C_BOLD,
+    'C_COUNT': C_COUNT,
+}

xortool/libcolors.py

@@ -1,5 +1,5 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 import os
 
@@ -18,7 +18,7 @@
 
 def _main():
     header = color("white", "black", "dark")
-    print();
+    print()
 
     print(header + "       " + "Colors and backgrounds:      " + color())
     for c in _keys_sorted_by_values(BASH_COLORS):
@@ -37,7 +37,6 @@ def _main():
                c1 + "red text" + color() + "     " +
                c2 + "white text" + color()))
     print()
-    return
 
 
 def color(color=None, bgcolor=None, attrs=None):
@@ -53,16 +52,14 @@ def color(color=None, bgcolor=None, attrs=None):
             ret += ";" + BASH_ATTRIBUTES[attr]
 
     if color:
-        if color in BASH_COLORS:
-            ret += ";" + BASH_COLORS[color]
-        else:
+        if color not in BASH_COLORS:
             raise ValueError("Unknown color: " + color)
+        ret += ";" + BASH_COLORS[color]
 
     if bgcolor:
-        if bgcolor in BASH_BGCOLORS:
-            ret += ";" + BASH_BGCOLORS[bgcolor]
-        else:
+        if bgcolor not in BASH_BGCOLORS:
             raise ValueError("Unknown background color: " + bgcolor)
+        ret += ";" + BASH_BGCOLORS[bgcolor]
 
     return ret + "m"
 

xortool/routine.py

@@ -1,98 +1,47 @@
-#!/usr/bin/env python2
-#-*- coding:utf-8 -*-
+#!/usr/bin/env python3
+# -*- coding:utf-8 -*-
 
 import os
 import sys
-import string
 
 
 class MkdirError(Exception):
     pass
 
 
-def load_file(filename):
-    if filename == "-":
-        return sys.stdin.read()
-    fd = open(filename, "rb")
-    contents = fd.read()
-    fd.close()
-    return contents
-
-
-def save_file(filename, data):
-    fd = open(filename, "wb")
-    fd.write(data)
-    fd.close()
-    return
-
-
 def mkdir(dirname):
     if os.path.exists(dirname):
         return
     try:
         os.mkdir(dirname)
     except BaseException as err:
         raise MkdirError(str(err))
-    return
 
 
 def rmdir(dirname):
     if dirname[-1] == os.sep:
         dirname = dirname[:-1]
     if os.path.islink(dirname):
         return  # do not clear link - we can get out of dir
-    files = os.listdir(dirname)
-    for f in files:
-        if f == '.' or f == '..':
+    for f in os.listdir(dirname):
+        if f in ('.', '..'):
             continue
         path = dirname + os.sep + f
         if os.path.isdir(path):
             rmdir(path)
         else:
             os.unlink(path)
     os.rmdir(dirname)
-    return
-
-
-def decode_from_hex(text):
-    only_hex_digits = "".join([c for c in text if c in string.hexdigits])
-    return only_hex_digits.decode("hex")
-
-
-def parse_char(ch):
-    """
-    'A' or '\x41' or '41'
-    """
-    if len(ch) == 1:
-        return ord(ch)
-    if ch[0:2] == "\\x":
-        ch = ch[2:]
-    if not ch:
-        raise ValueError("Empty char")
-    return ord(chr(int(ch, 16)))
 
 
 def dexor(text, key):
     ret = list(text)
     mod = len(key)
     for index, char in enumerate(ret):
-        ret[index] = chr(ord(char) ^ ord(key[index % mod]))
+        ret[index] = chr(char ^ ord(key[index % mod]))
     return "".join(ret)
 
 
 def die(exitMessage, exitCode=1):
     print(exitMessage)
     sys.exit(exitCode)
-
-
-def is_linux():
-    return sys.platform.startswith("linux")
-
-
-def alphanum(s):
-    lst = list(s)
-    for index, char in enumerate(lst):
-        if char in (string.letters + string.digits):
-            continue
-        lst[index] = char.encode("hex")
-    return "".join(lst)