Fixed https://github.com/robmcmullen/omnivore/issues/209

* force unicode on uuid string conversion in py2

atrcopy/segments.py

@@ -8,12 +8,11 @@ from builtins import range
 from builtins import object
 import bisect
 import io
-import uuid
 
 import numpy as np
 
 from .errors import *
-from .utils import to_numpy, to_numpy_list
+from .utils import to_numpy, to_numpy_list, uuid
 from functools import reduce
 
 user_bit_mask = 0x07
@@ -364,7 +363,7 @@ class DefaultSegment(object):
         self.verbose_name = verbose_name
         self.page_size = -1
         self.map_width = 40
-        self.uuid = str(uuid.uuid4())
+        self.uuid = uuid()
         if memory_map is None:
             memory_map = {}
         self.memory_map = memory_map
@@ -453,7 +452,7 @@ class DefaultSegment(object):
         added to a segment, a default value should be applied here.
         """
         self.memory_map = dict(state.pop('memory_map', []))
-        self.uuid = state.pop('uuid', uuid.uuid4())
+        self.uuid = state.pop('uuid', uuid())
         self.can_resize = state.pop('can_resize', self.__class__.can_resize_default)
         self.restore_missing_serializable_defaults()
         self.__dict__.update(state)

atrcopy/utils.py

@@ -4,6 +4,7 @@ from builtins import str
 from builtins import range
 from builtins import object
 import types
+import uuid as stdlib_uuid
 
 import numpy as np
 
@@ -17,6 +18,18 @@ except NameError:
     _xd = False
 
 
+def uuid():
+    u = stdlib_uuid.uuid4()
+
+    # Force it to use unicode(py2) or str(py3) so it isn't serialized as
+    # future.types.newstr.newstr on py2
+    try:
+        u = unicode(u)
+    except:
+        u = str(u)
+    return u
+
+
 def to_numpy(value):
     if type(value) is np.ndarray:
         return value

test/test_serialize.py

@@ -1,6 +1,7 @@
 from __future__ import print_function
 from builtins import range
 from builtins import object
+from builtins import str
 import os
 
 import numpy as np
@@ -15,7 +16,14 @@ class TestSegment(object):
         r = SegmentData(data)
         self.segment = DefaultSegment(r, 0)
 
-    def test_s1(self):
+    def test_getstate(self):
+        state = self.segment.__getstate__()
+        for k, v in state.items():
+            print("k=%s v=%s type=%s" % (k, v, type(v)))
+        byte_type = type(str(u'  ').encode('utf-8'))  # py2 and py3
+        assert type(state['uuid']) == byte_type
+
+    def test_extra(self):
         s = self.segment
         s.set_comment([[4,5]], "test1")
         s.set_comment([[40,50]], "test2")
@@ -51,4 +59,4 @@ class TestSegment(object):
 if __name__ == "__main__":
     t = TestSegment()
     t.setup()
-    t.test_s1()
+    t.test_getstate()