forked from repoze/repoze.retry
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES.txt
139 lines (83 loc) · 3.14 KB
/
CHANGES.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
``repoze.retry`` Changelog
==========================
After 1.0
---------
- Fix handling of case where environ['CONTENT_LENGTH'] is an empty string.
- Added possibility of tracebacks being written to wsgi.errors only after
the specified number of tries.
1.0 (2010-08-09)
----------------
- Moved documentation to Sphinx.
- Micro-optimization in lookup of 'wsgi.errors' from WSGI environ.
0.9.4 (2010-03-01)
------------------
- Fixed bug where wsgi.input read errors were not being caught for payloads
large enough to cause a chunked read. Also expanded the error handling for
this part to catch IOError in addition to socket.error, since mod_wsgi has
been observed to raise IOError in some cases.
0.9.3 (2009-09-30)
------------------
- Don't write a temporary file unless the request content length is
greater than 2MB (use a StringIO instead).
- Make ZPublisher.Publish:Retry exceptions retryable (via a soft dependency).
This makes repoze.retry work the same as the Zope 2 publisher when that
module is installed.
- 100% test coverage.
- Change documentation to show proper retryable exception syntax in
paste config.
0.9.2 (2008-07-30)
------------------
- Close the app_iter at appropriate points to silence lint errors.
- Return a Bad Request error if we get a socket error while reading
input.
- Fix traceback output to wsgi.errors (it was going to console).
- Assert that downstream app must call start_response before successfully
returning.
0.9.1 (2008-06-18)
------------------
- Seek wsgi.input back to zero before retrying a request due to a
conflict error.
0.9 (2008-06-15)
----------------
- Fixed concurrency bug whereby a response from one request might be
returned as result of a different request.
- Initial PyPI release.
0.8
---
- Added WSGI conformance testing for the middleware.
0.7
---
- Made the retryable exception(s) configurable, removing the hardwired
dependency on ZODB3.
0.6
---
- Relaxed requirement for ZODB 3.7.2, since we might need to use
the package with other verions.
0.5
---
- Depend on PyPI release of ZODB 3.7.2. Upgrade to this by doing
bin/easy_install -U 'ZODB3 >= 3.7.2, < 3.8.0a' if necessary.
0.4
---
- Write retry attempts to 'wsgi.errors' stream if availabile.
- Depend on rerolled ZODB 3.7.1 instead of zopelib.
- Add license and copyright, change trove classifiers.
0.3
---
- We now buffer the result of a downstream application's
'start_response' call so we can retry requests which have already
called start_response without breaking the WSGI spec (the server's
start_response may only be called once unless there is an exception,
and then it needs to be called with an exc_info three-tuple,
although we're uninterested in that case here).
0.2
---
- The entry point name was wrong (it referred to "tm"). Change it so
that egg:repoze.retry#retry should work in paste configs.
- Depend on zopelib rather than ZODB 3.8.0b3 distribution, because the
ZODB distribution pulls in various packages (zope.interface and ZEO
most notably) that are incompatible with stock Zope 2.10.4 apps and
older sandboxes. We'll need to revisit this.
0.1
---
- Initial release.