-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.txt
266 lines (205 loc) · 8.24 KB
/
README.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
About sPickle
=============
sPickle is an advanced (and experimental) Pickler for Python. The code
was developed as part of a commercial project and released as free
software under the Apache 2.0 License by science + computing ag. The
software is copyrighted by science + computing ag.
Why did we decide to make sPickle free software? We utilise Python and
other open source products a lot. Therefore we think it is just fair
to release enhancements back to the public.
Requirements
------------
* Python 2.7
* RPyC is used in the unit tests; the test is skipped, if
import rpyc fails.
(Sorry, Python 3 support is still in its beginnings.
See https://github.com/akruis/sPickle/tree/py3port. I'll happily accept
patches.)
Installation
------------
Get easy_install and do "easy_install sPickle".
Or get the latest source code from github.
git clone git://github.com/akruis/sPickle.git
Using sPickle
-------------
The sPickle module provides a class sPickle.Pickler. This class is very
similar to the conventional pickle.Pickler class and can be used as a
drop in replacement. Instead of:
from pickle import Pickler
file = open(filename, "wb")
pickler = Pickler(file, -1)
pickler.dump(object_to_be_serialised)
file.close()
you write:
from sPickle import Pickler
file = open(filename, "wb")
pickler = Pickler(file, -1)
pickler.dump(object_to_be_serialised)
file.close()
You see, just 2 characters difference. Of course, there are more
differences under the hood.
* sPickle supports serialisation of more types than the conventional Pickler.
In fact, most types can be pickled now, if serialisation makes any sense
for the particular type at all. But be warned: unpickling of some types
requires the same Python version.
* A sPickle Pickler has a list-typed attribute "serializeableModules". You can
use this attribute to determine which modules are to be pickled. For
details, read the comments in the source code and look at the examples.
* A sPickle Pickler has some (experimental) features to manipulate the content
of the created pickle and to handle/replace objects, which can not be pickled.
Support
=======
There is currently no support available, but you can drop me a mail.
a [dot] kruis [at] atos [dot] net
or use the facilities provided by https://github.com/akruis/sPickle
Plan
====
No further plans currently
### Changes ###
Version 0.1.11
--------------
2016-07-27
- Improved the pickling of global objects from the Python standard library
with missing or incorrect or misleading values of "__module__" and/or
"__name__". This change adds a new class ObjectDispatchBuilder and new
attributes "object_dispatch" and "object_dispatch_builder" of the pickler.
Version 0.1.10
--------------
2016-02-26
- Removed 'StacklessTaskletReturnValueException' from sPickle.__all__
- Documented constants and exceptions and added them to sPickle.__all__
- Previously a method of the super class was pickled by value, if the
subclass overloaded the method. Now the pickler performs a correct
mro-search to find the implementing class of a method.
Version 0.1.9
-------------
2016-02-19
- Fixed pickling of private methods.
- PEP 8 fixes.
Version 0.1.8
-------------
2015-03-26
- Added support for pickling empty cell objects.
- PEP 8 fixes.
Version 0.1.7
-------------
2014-07-12
- Added the optional argument "logger" to class sPickle.Pickler. Can be used to
customize or suppress logging.
Version 0.1.6
-------------
2014-05-09
- Added limited support for pickling objects returned by io.open().
Version 0.1.5
-------------
2014-04-16
- Cleanups, reformatted source (PEP-8).
- More robust test suite.
- Python 3 compatible syntax.
- I switched the SSH implementation used by the examples back to paramiko.
- No functional changes in package sPickle.
Version 0.1.4
-------------
2014-02-07
New class FailSavePickler. It is a subclass of Pickler, that can replace
otherwise unpickleable objects by surrogates.
Added support for pickling bound and unbound instance methods.
sPickle no longer requires "ssh" and "rpyc". Both modules are
are required to run the examples.
Version 0.1.3
-------------
2013-11-05
Improved pickling of abstract base classes. Every abstract base class
contains two WeakSet attributes _abs_cache and _abc_negative_cache. These
sets cache the result of subclass tests. It is advisable not to pickle the
content of those caches, because it could contain unpickleable objects.
This commit changes sPickle to replace _abs_cache and _abc_negative_cache
by new empty WeakSet instances.
Version 0.1.2
-------------
2013-05-22
Fixed two problems concerning module renaming.
Adapted to Python 2.7.4 and 2.7.5.
Note: Python 2.7.5 broke unpickling of named tuples pickled by Python 2.7.3 or 2.7.4.
This is Python bug http://bugs.python.org/issue18015. You can place the file
http://bugs.python.org/file30338/fix_python_275_issue18015.pth into your
site-packages directory to monkey patch this bug.
Version 0.1.1
-------------
2012-11-06
Support pickling of classes with a custom meta class.
Improved sPickle.Pickler. It no longer modifies sys.modules or modules.
Version 0.1.0
-------------
2012-10-28
Support for plain Python
sPickle no longer requires Stackless Python for most of its work. Of course,
if you want to pickle tasklets you need Stackless, but for functions, classes
and most other objects, the normal CPython 2.7 is good enough.
Version 0.0.9
-------------
2012-10-28
Support pickling of cStringIO objects.
Fixed pickling of recursive collections.OrderedDict objects
Support pickling of <type 'cell'> type objects.
Support pickling of weakref.ReferenceType objects without callback function.
Fixed pickling of modules with a modified __name__ attribute.
Support pickling of super objects.
Fixed classes with unusual __dict__ attribute.
Replaced REDUCE __import__ ... with GLOBAL. Now GLOBAL is used for all imports.
This change makes it possible to get all imported modules
using SPickleTools.getImportList().
Enhancements:
- The new utility method SPickleTools.reducer() simplifies
the implementation of complex __reduce__ methods.
- I Added an option to rename the module name of pickled objects.
- The methods dumps, dumps_with_external_ids and
remotemethod of class SPickleTools got new parameters.
- The class Pickler got a new method analysePicklerStack. It can
be used to extract information about the Pickler state from
the stack or from a traceback.
Version 0.0.8
-------------
2012-07-17
Important bug fix: previous versions failed to pickle some classes.
Added support for member descriptors, getset descriptors and some
dictproxy objects.
Added a debugging aid for "id(obj) not in self.memo" assertions.
Set the environment variable SPICKLE_TRACE_MEMO_ENTRIES to a list of
space separated memo keys (integer numbers). sPickle will log information
about the object hierarchy, when it adds a traced memo entry.
Version 0.0.7
-------------
Minor fixes. Changed the logging to delay log messages while holding the
import lock. Otherwise you might get a dead-lock.
Upgrade the Python extensions used by the examples and switched from
Paramiko to ssh.
Version 0.0.6
-------------
Improved the documentation and improved the method SPickleTools.remotemethod.
Version 0.0.5
-------------
Finally I found a few hours to add a nice documentation using Sphinx.
Besides the documentation, there are a few cleanups.
Version 0.0.4
-------------
Minor improvements, but unfortunately the documentation is still
incomplete. I release this version, because I'm going to present this
version in my talk at EuroPython 2011.
Version 0.0.3
-------------
Renamed example1 to example2 and added a new example1. Example1 now demonstrates
checkpointing of programs, example2 demonstrates RPyC.
Fixed a few problems.
Use sPickle in flowGuide instead of the original flowGuide pickle code.
Version 0.0.2
-------------
Reorganised the directory layout, added this README.txt and fixed a few
problems. Currently we still use a private source code repository as
science + computing, but I plan to switch to github soon.
Version 0.0.1
-------------
Initial version, released by science + computing ag. This version is more or
less a copy of two modules from the flowGuide2 source code. The code works
for the specific requirements of flowGuide2, but has not been tested outside
of the flowGuide2 environment.