|
0
|
1 2007-07-17: version 1.24
|
|
|
2
|
|
|
3 * update the stats method, including tests for it
|
|
|
4 (Ronald J Kimball <rkimball+memcached@pangeamedia.com>)
|
|
|
5
|
|
|
6 * arguments to 'new' constructor can be %hash or $hashref now
|
|
|
7 (previously was only $hashref)
|
|
|
8
|
|
|
9 * work around a Perl segfault (Matthieu PATOU <mp@oxado.com>)
|
|
|
10 see http://lists.danga.com/pipermail/memcached/2007-June/004511.html
|
|
|
11
|
|
|
12 2007-06-19: version 1.23
|
|
|
13
|
|
|
14 * add 'remove' as an alias for 'delete' (Dave Cardwell <dave@davecardwell.co.uk>)
|
|
|
15
|
|
|
16 2007-06-18: version 1.22
|
|
|
17
|
|
|
18 * lost connection handling broken due to wrong %sock_map indexing
|
|
|
19 http://rt.cpan.org/Public/Bug/Display.html?id=27181
|
|
|
20 fix from RHESA
|
|
|
21
|
|
|
22 * let parser_class be configured as a constructor option,
|
|
|
23 defaulting to XS if available, else regular. (unless
|
|
|
24 $ENV{NO_XS} is set, in which case the default is regular)
|
|
|
25
|
|
|
26 2007-05-02: version 1.21
|
|
|
27
|
|
|
28 * new faster optional interface for GetParser subclasses. doing
|
|
|
29 this release so upcoming Cache::Memcached::GetParserXS can
|
|
|
30 depend on this. otherwise this release isn't interesting.
|
|
|
31
|
|
|
32 2007-04-16: version 1.20
|
|
|
33
|
|
|
34 * fix "Warning produced when flush_all called" from CDENT
|
|
|
35 http://rt.cpan.org/Public/Bug/Display.html?id=22181
|
|
|
36
|
|
|
37 * support access via unix domain sockets. (Nathan Neulinger <nneul@umr.edu>)
|
|
|
38
|
|
|
39 * abstract out response parsing into own class, and add XS-module
|
|
|
40 detection, so if you have the XS (C) version, things'll be faster.
|
|
|
41 that part's not done yet.
|
|
|
42
|
|
|
43 2006-07-03
|
|
|
44 * don't use dual scalar/glob sockets. makes it all profilable
|
|
|
45 again under SmallProf, DProf, and Devel::Profiler, all three
|
|
|
46 of which used to barf on those weird sockets previously
|
|
|
47 * only init_buckets once, when servers are changed
|
|
|
48 * don't call sock_to_host and get_sock as much: cache closer
|
|
|
49 in get_multi
|
|
|
50 * more internal caching (buck2sock, etc)
|
|
|
51 * fast paths for namespaces/single sock/etc in a few more places
|
|
|
52 * general micro-speedups all over
|
|
|
53
|
|
|
54 2006-06-27
|
|
|
55 * patch from Maxim Dounin <mdounin@rambler-co.ru> to fix a typo
|
|
|
56 which caused no_rehash flag to not work.
|
|
|
57 * release 1.18
|
|
|
58
|
|
|
59 2006-04-29
|
|
|
60 * flush_all command from Patrick Michael Kane <modus-memcache@pr.es.to>
|
|
|
61 * document namespaces
|
|
|
62 * release 1.17
|
|
|
63
|
|
|
64 2006-04-29
|
|
|
65 * fix stats method (people have only been asking since 2004 :-/)
|
|
|
66 * add tests
|
|
|
67 * move Memcached to lib/Cache directory to be more CPAN-friendly
|
|
|
68 * release 1.16
|
|
|
69
|
|
|
70 2005-09-20
|
|
|
71 * configurable connect delays and callback on connect failure (brad)
|
|
|
72 * release 1.15
|
|
|
73
|
|
|
74 2005-08-09
|
|
|
75 * _connect_sock never works in blocking mode because of a bug in setting
|
|
|
76 the default timeout. (Gaal)
|
|
|
77
|
|
|
78 2004-07-27
|
|
|
79 * release 1.14
|
|
|
80
|
|
|
81 2004-07-27
|
|
|
82 * kill buggy, slow ord() _hashfunc, replace with crc32.
|
|
|
83 this adds String::CRC32 as a dependency. thanks to
|
|
|
84 everybody's feedback on the mailing list.
|
|
|
85
|
|
|
86 2004-07-19
|
|
|
87 * don't use pos() because it doesn't seem to work in
|
|
|
88 taint mode. use $+[0] instead. (Dave Evans <..@djce.org.uk>)
|
|
|
89
|
|
|
90 2004-06-19
|
|
|
91 * support multiple paths to memcache nodes (Brad)
|
|
|
92 see 'set_pref_ip'
|
|
|
93
|
|
|
94 2004-05-30
|
|
|
95 * release version 1.13
|
|
|
96
|
|
|
97 2004-05-26 (Whitaker <whitaker@danga.com>)
|
|
|
98 * quiet warning
|
|
|
99
|
|
|
100 2004-05-25 (Whitaker <whitaker@danga.com>)
|
|
|
101 * get_multi shouldn't modify caller's @_
|
|
|
102
|
|
|
103 2004-05-18 (Michael <ged@danga.com>)
|
|
|
104 * namespace support
|
|
|
105 * use fields
|
|
|
106
|
|
|
107 2004-05-16 (Alexei Kozlov <kozlov@w-m.ru>)
|
|
|
108 * remove warnings with vec
|
|
|
109
|
|
|
110 2004-04-09 (brad)
|
|
|
111 * in perl 5.6, trap errors dethawing 5.8 storable objects
|
|
|
112 and instead treat it like a cache miss
|
|
|
113
|
|
|
114 2004-04-01
|
|
|
115 * use $! and not %! for perl 5.6 compat (Dave Evans <..@djce.org.uk>)
|
|
|
116 * don't mark whole IP dead anymore when a node is down (Jason Titus <jtitus@postini.com>)
|
|
|
117 * start version numbering (Jamie McCarthy <jamie@mccarthy.vg>)
|
|
|
118
|
|
|
119 2004-03-09 (Brad/Avva)
|
|
|
120 * _oneline can return more than one line (but always on a line break),
|
|
|
121 so caller must decide when it's really time to quit. had to modify
|
|
|
122 run_command to know that. (which is used by stats)
|
|
|
123
|
|
|
124 2004-03-05 (Dave Evans <..@djce.org.uk>)
|
|
|
125 * Here's a really trivial patch for the Perl binding,
|
|
|
126 Cache::Memcached. The bug is that the module assumes that the
|
|
|
127 currently select()ed filehandle is STDOUT, but this might not be
|
|
|
128 the case. So this patch ensures that the select()ed filehandle is
|
|
|
129 preserved, not forced to STDOUT.
|
|
|
130
|
|
|
131 2004-02-29 (Brad)
|
|
|
132 * add readonly option
|
|
|
133
|
|
|
134 2004-02-27 (Avva)
|
|
|
135 * Cleaner handling of the case when _oneline is called without a
|
|
|
136 line parameter (i.e. not to send anything, just read a line from
|
|
|
137 the socket). Make it depend on $line being defined only,
|
|
|
138 regardless of its content (thanks Brad!).
|
|
|
139
|
|
|
140 2004-02-25 (Avva)
|
|
|
141 * Asyncify all I/O, finally get rid of alarm() yuckiness, unify all
|
|
|
142 one-liner command/responses into a single internal API.
|
|
|
143
|
|
|
144 2004-02-17
|
|
|
145 * document in POD the delete method
|
|
|
146
|
|
|
147 2004-02-03
|
|
|
148 * fix bug with 2k read boundaries falling in the middle
|
|
|
149 of "VALUE ..." or "END" lines, thus halting future
|
|
|
150 parsing and responses. (eek!)
|
|
|
151 * version 1.0.12
|
|
|
152
|
|
|
153 2003-12-01
|
|
|
154 * merge stats/stats_reset patch from Jamie McCarthy
|
|
|
155 * trailing whitespace cleanup
|
|
|
156
|
|
|
157 2003-11-08
|
|
|
158 * work on Solaris/BSD where there's no MSG_NOSIGNAL.
|
|
|
159 the expense is extra syscalls to change the local
|
|
|
160 SIGPIPE handler all the time. in the future, it'd
|
|
|
161 be nice to have an option so Solaris/BSD callers
|
|
|
162 can say, "Hey, I've turned off SIGPIPE globally,
|
|
|
163 don't worry about it."
|
|
|
164
|
|
|
165 2003-10-26
|
|
|
166 * add a test file, so automated CPAN test hosts are happy
|
|
|
167 * check MSG_NOSIGNAL immediately on module load, not on use,
|
|
|
168 so Solaris dies early. (still on TODO to fix, but better
|
|
|
169 to fail loudly)
|
|
|
170 * version 1.0.11
|
|
|
171
|
|
|
172 2003-10-25
|
|
|
173 * version 1.0.10, rename to Cache::Memcached, upload to CPAN
|
|
|
174
|
|
|
175 2003-10-18
|
|
|
176 * implement read/write timeouts everywhere. Now the client shouldn't
|
|
|
177 hang if the server machine goes down unexpectedly. (avva)
|
|
|
178
|
|
|
179 2003-10-16
|
|
|
180 * use Storable::nfreeze instead of freeze, so hosts from different
|
|
|
181 architectures can all use the same data. (all must use Perl, though.
|
|
|
182 the different memcache APIs all store/pickle/serialize data differently)
|
|
|
183 Suggestion by Jason Titus <jtitus@postini.com>
|
|
|
184
|
|
|
185 2003-10-06
|
|
|
186 * fix _incrdecr to return complete number, not just first
|
|
|
187 digit (thanks to Ryan T. Dean)
|
|
|
188 * release version 1.0.9
|
|
|
189
|
|
|
190 2003-10-04
|
|
|
191 * document expiration times in POD (thanks to Tim Bunce
|
|
|
192 for noting the omission)
|
|
|
193 * release version 1.0.8
|
|
|
194
|
|
|
195 2003-10-03
|
|
|
196 * add connect timeout of 0.25s, for dead host detection.
|
|
|
197 We had 1 second a couple revs ago, but lost it when
|
|
|
198 ditching IO::Socket module. (avva)
|
|
|
199
|
|
|
200 2003-10-02
|
|
|
201 * fix _incrdecr with explicit-hashvalue keys (whitaker)
|
|
|
202
|
|
|
203 2003-10-01
|
|
|
204 * add run_command API call. TODO: document, and document
|
|
|
205 the $exptime on the setters
|
|
|
206
|
|
|
207 2003-09-30
|
|
|
208 * use send instead of print, so we can set MSG_NOSIGNAL
|
|
|
209 and not get SIGPIPES, which avoids 3 syscalls of localizing
|
|
|
210 $SIG{PIPE} and sends everything at once, instead of 4k
|
|
|
211 stdio chunks. in review: stdio buffered in, send unbuffered
|
|
|
212 out. TODO: setvbuf so reads are buffered at more than 4k.
|
|
|
213
|
|
|
214 2003-09-29
|
|
|
215 * yet faster parsing
|
|
|
216 * switch to stdio/perlio instead of raw io: more correct,
|
|
|
217 simpler parsing code.
|
|
|
218
|
|
|
219 2003-09-28
|
|
|
220 * prevent some warnings
|
|
|
221 * faster get() call that doesn't use get_multi()
|
|
|
222 * optimizations for single-server case
|
|
|
223 * use socket APIs directly, instead of uber-slow IO::* modules
|
|
|
224 * new faster _load_items parsing
|
|
|
225
|
|
|
226 2003-09-04
|
|
|
227 * emit debug when set/add/replace fails, in addition to succeed
|
|
|
228
|
|
|
229 Version 1.0.7
|
|
|
230 -- compression support (Brad Whitaker)
|
|
|
231
|
|
|
232 Version 1.0.6
|
|
|
233 -- incr/decr client support
|
|
|
234 -- make delete optionally take second argument (server now supports
|
|
|
235 a delay time on delete)
|
|
|
236 -- doc updates from Jamie McCarthy
|
|
|
237 -- better hashing after dead host detection: new requests go to different
|
|
|
238 remaining hosts, instead of all to the same one.
|
|
|
239
|
|
|
240 Version 1.0.2
|
|
|
241 -- initial release, about.
|
