Merge pull request #174 from baloo/baloo/features/decode-gtid-from-packet

Decode gtid from network packet

Author: julien-duponchelle

pymysqlreplication/gtid.py

@@ -3,6 +3,7 @@
 import re
 import struct
 import binascii
+from io import BytesIO
 
 
 class Gtid(object):
@@ -16,17 +17,18 @@ def parse_interval(interval):
         else:
             return (int(m.group(1)), int(m.group(2)))
 
-
     @staticmethod
     def parse(gtid):
-        m = re.search('^([0-9a-fA-F]{8}(?:-[0-9a-fA-F]{4}){3}-[0-9a-fA-F]{12})((?::[0-9-]+)+)$', gtid)
+        m = re.search('^([0-9a-fA-F]{8}(?:-[0-9a-fA-F]{4}){3}-[0-9a-fA-F]{12})'
+                      '((?::[0-9-]+)+)$', gtid)
         if not m:
             raise ValueError('GTID format is incorrect: %r' % (gtid, ))
 
         sid = m.group(1)
         intervals = m.group(2)
 
-        intervals_parsed = [Gtid.parse_interval(x) for x in intervals.split(':')[1:]]
+        intervals_parsed = [Gtid.parse_interval(x)
+                            for x in intervals.split(':')[1:]]
 
         return (sid, intervals_parsed)
 
@@ -39,18 +41,18 @@ def __init__(self, gtid):
     def __str__(self):
         return '%s:%s' % (self.sid,
                           ':'.join(('%d-%s' % x) if isinstance(x, tuple)
-                                                 else str(x)
-                                                 for x in self.intervals))
+                                   else str(x)
+                                   for x in self.intervals))
 
     def __repr__(self):
         return '<Gtid "%s">' % self
 
     @property
     def encoded_length(self):
-        return (16 + # sid
-                8 + # n_intervals
-                2 * # stop/start
-                8 * # stop/start mark encoded as int64
+        return (16 +  # sid
+                8 +  # n_intervals
+                2 *  # stop/start
+                8 *  # stop/start mark encoded as int64
                 len(self.intervals))
 
     def encode(self):
@@ -79,25 +81,72 @@ def encode(self):
 
         return buffer
 
+    @classmethod
+    def decode(cls, payload):
+        assert isinstance(payload, BytesIO), \
+            'payload is expected to be a BytesIO'
+        sid = b''
+        sid = sid + binascii.hexlify(payload.read(4))
+        sid = sid + b'-'
+        sid = sid + binascii.hexlify(payload.read(2))
+        sid = sid + b'-'
+        sid = sid + binascii.hexlify(payload.read(2))
+        sid = sid + b'-'
+        sid = sid + binascii.hexlify(payload.read(2))
+        sid = sid + b'-'
+        sid = sid + binascii.hexlify(payload.read(6))
+
+        (n_intervals,) = struct.unpack('<Q', payload.read(8))
+        intervals = []
+        for i in range(0, n_intervals):
+            start, end = struct.unpack('<QQ', payload.read(16))
+            if end == start + 1:
+                intervals.append(start)
+            else:
+                intervals.append((start, end))
+
+        return cls('%s:%s' % (sid.decode('ascii'), ':'.join([
+            '%d-%d' % x
+            if isinstance(x, tuple)
+            else '%d' % x
+            for x in intervals])))
+
 
 class GtidSet(object):
     def __init__(self, gtid_set):
+        def _to_gtid(element):
+            if isinstance(element, Gtid):
+                return element
+            return Gtid(element.strip(' \n'))
+
         if not gtid_set:
             self.gtids = []
+        elif isinstance(gtid_set, (list, set)):
+            self.gtids = [_to_gtid(x) for x in gtid_set]
         else:
             self.gtids = [Gtid(x.strip(' \n')) for x in gtid_set.split(',')]
 
     def __str__(self):
         return ','.join(str(x) for x in self.gtids)
 
     def __repr__(self):
-        return '<GtidSet "%s"' % ','.join(repr(x) for x in self.gtids)
+        return '<GtidSet %r>' % self.gtids
 
     @property
     def encoded_length(self):
-        return (8 + # n_sids
+        return (8 +  # n_sids
                 sum(x.encoded_length for x in self.gtids))
 
     def encoded(self):
         return b'' + (struct.pack('<Q', len(self.gtids)) +
-                b''.join(x.encode() for x in self.gtids))
+                      b''.join(x.encode() for x in self.gtids))
+
+    encode = encoded
+
+    @classmethod
+    def decode(cls, payload):
+        assert isinstance(payload, BytesIO), \
+            'payload is expected to be a BytesIO'
+        (n_sid,) = struct.unpack('<Q', payload.read(8))
+
+        return cls([Gtid.decode(payload) for _ in range(0, n_sid)])

pymysqlreplication/tests/test_basic.py

@@ -1,6 +1,7 @@
 # -*- coding: utf-8 -*-
 import time
 import sys
+import io
 if sys.version_info < (2, 7):
     import unittest2 as unittest
 else:
@@ -715,6 +716,24 @@ def test_gtidset_representation_newline(self):
         myset = GtidSet(mysql_repr)
         self.assertEqual(str(myset), set_repr)
 
+    def test_gtidset_representation(self):
+        set_repr = '57b70f4e-20d3-11e5-a393-4a63946f7eac:1-56,' \
+                   '4350f323-7565-4e59-8763-4b1b83a0ce0e:1-20'
+
+        myset = GtidSet(set_repr)
+        payload = myset.encode()
+        parsedset = myset.decode(io.BytesIO(payload))
+
+        self.assertEqual(str(myset), str(parsedset))
+
+        set_repr = '57b70f4e-20d3-11e5-a393-4a63946f7eac:1,' \
+                   '4350f323-7565-4e59-8763-4b1b83a0ce0e:1-20'
+
+        myset = GtidSet(set_repr)
+        payload = myset.encode()
+        parsedset = myset.decode(io.BytesIO(payload))
+
+        self.assertEqual(str(myset), str(parsedset))
 
 if __name__ == "__main__":
     import unittest