##// END OF EJS Templates
Proper HISTORY file
Marcin Kasperski -
r155:335c0c13 0.7.0 default
parent child Browse files
Show More
@@ -0,0 +1,198 b''
1 0.7.0
2 ~~~~~~~~~~~~~~~~~~~
3
4 Delaying keyring module import until passwords are really needed. It
5 can noticeably improve Mercurial (non pull/push) performance in some
6 cases (no longer slow hg status because D-Bus is busy an keyring tries
7 to activate KDE Wallet through it…).
8
9 0.6.7
10 ~~~~~~~~~~~~~~~~~
11
12 #46 Fixed syntax of smtp.tls configuration setting (current Mercurials
13 doesn't handle "true" anymore, TortoiseHG crashed with mercurial
14 keyring enabled while currently recommended starttls/smtps/none values
15 were in use).
16
17 0.6.6
18 ~~~~~~~~~~~~~~~~~
19
20 #44 Handling some more mercurial versions in demandimport-detection
21 logic.
22
23 0.6.5
24 ~~~~~~~~~~~~~~~~~
25
26 #36 Shutting up warning about no logging handlers.
27
28 0.6.4
29 ~~~~~~~~~~~~~~~~~
30
31 #44 Pre-2.9.1 Mercurials compatibility (probing for active
32 demandimport differently).
33
34 0.6.3
35 ~~~~~~~~~~~~~~~~~
36
37 #41 Fix for incorrect demandimport activity check logic, which could
38 cause various problems with imports after mercurial_keyring is
39 imported.
40
41 0.6.2
42 ~~~~~~~~~~~~~~~~~
43
44 #33 Fix for UnicodeDecodeErrors happening on some backends (especially
45 Vault) when passwords with non-ascii characters are in use and native
46 locale is not utf-8. Passwords are no longer saved to keyring backends
47 as-entered, they are now decoded from local encoding (whichever is
48 detected by Mercurial), then encoded to unicode.
49
50 0.6.1
51 ~~~~~~~~~~~~~~~~~
52
53 #30 Yet another demandimport conflict fixed.
54
55 0.6.0
56 ~~~~~~~~~~~~~~~~~
57
58 #28 Disable demandimport completely during keyring import. Mayhaps it
59 will resolve (most) demandimport conflict errors.
60
61 0.5.7
62 ~~~~~~~~~~~~~~~~~
63
64 #27 Some more demandimport ignores.
65
66 0.5.6
67 ~~~~~~~~~~~~~~~~~
68
69 #24, #25 Demandimport fixes (import failures in specific cases).
70
71 Better way of demandimport-ignoring modules. In particular, we append
72 more of them if gobject happens to be on the list.
73
74 0.5.5
75 ~~~~~~~~~~~~~~~~~
76
77 Fix for gnome keyring import problems.
78
79 0.5.4
80 ~~~~~~~~~~~~~~~~~
81
82 #22 Some more demandimport ignores (fix import failures).
83
84 SMTP password was not cleared properly (after detecting that it is
85 invalid).
86
87 Clarified license to be modified BSD style license.
88
89 0.5.3
90 ~~~~~~~~~~~~~~~~~
91
92 Remove useless import which caused problems on Mercurial 2.3 when
93 demandimport was not enabled
94
95 0.5.1
96 ~~~~~~~~~~~~~~~~~
97
98 Add help text to output for hg help.
99
100 0.5.0
101 ~~~~~~~~~~~~~~~~~
102
103 Improved bad password detection. Internally: extension is now able to
104 properly differentiate between an authentication failure and a new
105 request to the same url.
106
107 Fixes in debug message
108
109 Further debug messages patching
110
111 Improving debug messages handling.
112
113 Mercurial Keyring debug messages are now prefixed with
114 [HgKeyring] to make distinguishing them easier
115
116 0.4.6
117 ~~~~~~~~~~~~~~~~~
118
119 More compatibility (changed signature of httpconnection.readauthforuri
120 , introduced post Mercurial 1.9 - since hg.0593e8f81c71)
121
122 Fix compatibility code which did not work due to demandimport issues
123 (attempts to catch ImportErrors on "from mercurial.url import
124 readauthforuri" were not working properly).
125
126 0.4.5
127 ~~~~~~~~~~~~~~~~~
128
129 Mercurial 1.9 compatibility (readauthforuri has been moved into new
130 httpconnection module).
131
132 0.4.4
133 ~~~~~~~~~~~~~~~~~
134
135 Mercurial 1.8 compatibility (passwordmgr.readauthtoken() has been
136 moved into mercurial.url.readauthforuri).
137
138 0.4.3
139 ~~~~~~~~~~~~~~~~~
140
141 Keyring fork no longer is needed as keyring releases are available
142 again.
143
144 Workaround for gnomekeyring mercurial.demandimport incompatibility:
145 mercurial.demandimport, which is enabled while in a mercurial
146 extensions, prevents the correct import of gobject._gobject and
147 consequently doesn't allow the loading of the gnomekeyring module,
148 which can be used by keyring. This just adds the proper module to
149 demandimport ignore list.
150
151 0.4.2
152 ~~~~~~~~~~~~~~~~~
153
154 No longer raising an error when username is specified both in ~/.hgrc
155 and <repo>/.hg/hgrc if it is the same in both places.
156
157 Docs recommend sborho keyring fork.
158
159 0.4.1
160 ~~~~~~~~~~~~~~~~~
161
162 Some tweaks and docs related to prefix handling.
163
164 Explicit information that keyring is not used due to lack of username.
165
166 0.4.0
167 ~~~~~~~~~~~~~~~~~
168
169 Store and lookup prefix from [auth] so that password is shared amongst
170 shared auth entries
171
172 0.3.3
173 ~~~~~~~~~~~~~~~~~
174
175 Better error message
176
177 0.3.2
178 ~~~~~~~~~~~~~~~~~
179
180 Doc tweaks
181
182 0.3.1
183 ~~~~~~~~~~~~~~~~~
184
185 Introduced and documented PyPi package, added setup.py
186
187 0.2.0
188 ~~~~~~~~~~~~~~~~~
189
190 Added handling of SMTP passwords (tested on patchbomb extension but
191 should work on anything what utilizes mercurial.mail)
192
193 Docstrings mention Debian keyring packages.
194
195 0.1.1
196 ~~~~~~~~~~~~~~~~~
197
198 Initial public release
@@ -185,6 +185,11 b' The extension is monkey-patching the mer'
185 185 to replace the find_user_password method. Detailed order of operations
186 186 is described in the comments inside `the code`_.
187 187
188 History
189 ==========
190
191 See `HISTORY.txt`_.
192
188 193 Development
189 194 ===========
190 195
@@ -197,3 +202,5 b' Additional notes'
197 202
198 203 Information about this extension is also available
199 204 on Mercurial Wiki: http://mercurial.selenic.com/wiki/KeyringExtension
205
206 .. _HISTORY.txt: http://bitbucket.org/Mekk/mercurial_keyring/src/HISTORY.txt
General Comments 0
You need to be logged in to leave comments. Login now