bytearray speedup - SecurityInnovation#87

docs/source/changelog.rst

@@ -5,6 +5,15 @@
 Changelog
 *********
 
+v0.2.2
+======
+
+PGPy v0.2.2 is primarily a bugfix release.
+
+Bugs Fixed
+----------
+ * Switched the main parse loop to use a bytearray instead of slicing a bytes, resulting in a ~160x speedup in parsing large blocks of pasing.
+
 v0.2.1
 ======
 

pgpy/packet/fields/keyfields.py

@@ -70,7 +70,7 @@ def parse(self, packet, pktype, alg, sec=False):
             mend = pos + bytelen
 
             getattr(self, field)['bitlen'] = bitlen
-            getattr(self, field)['bytes'] = packet[pos:mend]
+            getattr(self, field)['bytes'] = bytes(packet[pos:mend])
 
             pos = mend
 
@@ -268,7 +268,7 @@ def parse(self, packet):
                 pos = 13
 
         if self.id != 0:
-            self.iv = packet[pos:(pos + int(self.alg.block_size / 8))]
+            self.iv = bytes(packet[pos:(pos + int(self.alg.block_size / 8))])
 
     def derive_key(self, passphrase):
         # we use the fields stored here along with the RFC 4880 String-to-Key usage description

pgpy/packet/packets.py

@@ -419,7 +419,7 @@ def parse(self, packet):
             mend = -2
             if self.stokey.id == 254:
                 mend = len(packet)
-            self.enc_seckey_material = packet[pos:mend]
+            self.enc_seckey_material = bytes(packet[pos:mend])
 
         if self.stokey.id in [0, 255]:
             self.checksum = packet[pos:]

pgpy/pgp.py

@@ -160,7 +160,7 @@ def parse(self):
             for key, val in [ (h[i], h[i + 1]) for i in range(0, len(h), 2) ]:
                 self.ascii_headers[key] = val
 
-            self.data = base64.b64decode(k[2].replace('\n', '').encode())
+            self.data = bytearray(base64.b64decode(k[2].replace('\n', '').encode()))
             self.crc = bytes_to_int(base64.b64decode(k[3].encode()))
 
             # verify CRC
@@ -169,10 +169,9 @@ def parse(self):
 
         # dump fields in all contained packets per RFC 4880, without using pgpdump
         if self.data != b'':
-            pos = 0
-            while pos < len(self.data):
-                pkt = Packet(self.data[pos:])
-                pos += len(pkt.header.__bytes__()) + pkt.header.length
+            while len(self.data) > 0:
+                pkt = Packet(self.data)
+                del self.data[:(len(pkt.header.__bytes__()) + pkt.header.length)]
                 self.packets.append(pkt)
 
     def crc24(self):
@@ -185,7 +184,7 @@ def crc24(self):
         # The accumulation is done on the data before it is converted to
         # radix-64, rather than on the converted data.
         if self.data == b'':
-            self.data = self.__bytes__()
+            self.data = bytearray(self.__bytes__())
 
         crc = self.crc24_init
         sig = [ ord(i) for i in self.data ] if type(self.data) is str else self.data
@@ -210,7 +209,7 @@ def extract_pgp_ascii_block(self):
         # this is binary data; skip extracting the block and move on
         else:
             self.bytes = b''
-            self.data = data
+            self.data = bytearray(data)
             return
 
         # are there any ASCII armored PGP blocks present? if not, we may be dealing with binary data instead