jjp-memcache-client 1.8.7
Sign up to get free protection for your applications and to get access to all the features.
- data/FAQ.rdoc +32 -0
- data/History.rdoc +288 -0
- data/LICENSE.txt +28 -0
- data/README.rdoc +70 -0
- data/Rakefile +42 -0
- data/bin/memcached_top +56 -0
- data/lib/continuum_native.rb +41 -0
- data/lib/memcache.rb +1210 -0
- data/lib/memcache/event_machine.rb +173 -0
- data/lib/memcache/version.rb +6 -0
- data/lib/memcache_util.rb +105 -0
- data/performance.txt +143 -0
- data/test/test_benchmark.rb +142 -0
- data/test/test_event_machine.rb +110 -0
- data/test/test_mem_cache.rb +1323 -0
- metadata +72 -0
data/FAQ.rdoc
ADDED
@@ -0,0 +1,32 @@
|
|
1
|
+
= Memcache-client FAQ
|
2
|
+
|
3
|
+
== Does memcache-client work with Ruby 1.9?
|
4
|
+
|
5
|
+
Yes, Ruby 1.9 is supported. The test suite should pass completely on 1.8.6 and 1.9.1.
|
6
|
+
|
7
|
+
|
8
|
+
== I'm seeing "execution expired" or "time's up!" errors, what's that all about?
|
9
|
+
|
10
|
+
memcache-client 1.6.x+ now has socket operations timed out by default. This is to prevent
|
11
|
+
the Ruby process from hanging if memcached or starling get into a bad state, which has been
|
12
|
+
seen in production by both 37signals and FiveRuns. The default timeout is 0.5 seconds, which
|
13
|
+
should be more than enough time under normal circumstances. It's possible to hit a storm of
|
14
|
+
concurrent events which cause this timer to expire: a large Ruby VM can cause the GC to take
|
15
|
+
a while, while also storing a large (500k-1MB value), for example.
|
16
|
+
|
17
|
+
You can increase the timeout or disable them completely with the following configuration:
|
18
|
+
|
19
|
+
Rails:
|
20
|
+
config.cache_store = :mem_cache_store, 'server1', 'server2', { :timeout => nil } # no timeout
|
21
|
+
|
22
|
+
native:
|
23
|
+
MemCache.new ['server1', 'server2'], { :timeout => 1.0 } # 1 second timeout
|
24
|
+
|
25
|
+
|
26
|
+
== Isn't Evan Weaver's memcached gem faster?
|
27
|
+
|
28
|
+
The latest version of memcached-client is anywhere from 33% to 100% slower than memcached in various
|
29
|
+
benchmarks. Keep in mind this means that 10,000 get requests take 1.8 sec instead of 1.2 seconds.
|
30
|
+
In practice, memcache-client is unlikely to be a bottleneck in your system but there is always going
|
31
|
+
to be an overhead to pure Ruby. memcache-client does have the advantage of built-in integration into
|
32
|
+
Rails and should work on non-MRI platforms: JRuby, MacRuby, etc.
|
data/History.rdoc
ADDED
@@ -0,0 +1,288 @@
|
|
1
|
+
= 1.8.7 (2011-10-13)
|
2
|
+
|
3
|
+
* Added to MemCacheError "Value too large..." information about the cache_key under which such too large value was about to be stored. This way chasing code which is responsible for that error is possible whenever that error gets logged, no matter the logger level. Useful on production servers.
|
4
|
+
|
5
|
+
= 1.8.6 (HEAD)
|
6
|
+
|
7
|
+
* Use String#bytesize so non-ascii characters write correctly to the socket.
|
8
|
+
|
9
|
+
= 1.8.5 (2010-07-05)
|
10
|
+
|
11
|
+
* Fix bad release
|
12
|
+
|
13
|
+
= 1.8.4 (2010-07-02)
|
14
|
+
|
15
|
+
* Fix unfibered usage of memcache-client in EM
|
16
|
+
* Remove nag message (tenderlove)
|
17
|
+
|
18
|
+
= 1.8.3 (2010-04-26)
|
19
|
+
|
20
|
+
* Don't allow blank keys. (Bill Horsman)
|
21
|
+
|
22
|
+
= 1.8.2 (2010-04-03)
|
23
|
+
|
24
|
+
* Fix concurrency issues with eventmachine support.
|
25
|
+
|
26
|
+
= 1.8.1 (2010-03-20)
|
27
|
+
|
28
|
+
* Only require SystemTimer if the Ruby VM looks like MRI.
|
29
|
+
* Remove VERSION.yml usage as we should avoid using files outside of lib at runtime. (josh)
|
30
|
+
|
31
|
+
= 1.8.0 (2010-03-05)
|
32
|
+
|
33
|
+
* Add support for EventMachine-based connections.
|
34
|
+
* Add support for raw values in get_multi
|
35
|
+
* Add memcached_top binary for gathering server statistics
|
36
|
+
|
37
|
+
= 1.7.8 (2010-02-03)
|
38
|
+
|
39
|
+
* Fix issue where autofix_keys logic did not account for namespace length. (menno)
|
40
|
+
* Fix issue when using memcache-client without rubygems. (anvar)
|
41
|
+
* Fix issue when using the cas method with raw=true (Takahiro Kikumoto)
|
42
|
+
|
43
|
+
= 1.7.7 (2009-11-24)
|
44
|
+
|
45
|
+
* Fix invalid delete request in memcached 1.4.x. The expiry parameter to MemCache#delete is
|
46
|
+
now ignored as memcached 1.4.x has dropped support for this feature.
|
47
|
+
|
48
|
+
= 1.7.6 (2009-11-03)
|
49
|
+
|
50
|
+
* Reworked socket timeout code due to several customer complaints about timeouts not
|
51
|
+
working 100% of the time since 1.7.3.
|
52
|
+
* Add option to configure the namespace separator string, for interop with Perl
|
53
|
+
which does not use a separator character:
|
54
|
+
MemCache.new(servers, :namespace_separator => '')
|
55
|
+
* Move to jeweler and gemcutter for RubyGem support.
|
56
|
+
|
57
|
+
= 1.7.5 (2009-09-09)
|
58
|
+
|
59
|
+
* Fix ruby warnings (josh)
|
60
|
+
* Make megabyte value size limit optional since Tokyo Tyrant can accept values larger than 1MB.
|
61
|
+
Use :check_size => false to disable the size check. (jsl)
|
62
|
+
* Ruby 1.9 support for recent I/O changes.
|
63
|
+
* Fix duplicate value marshalling on server error. (rajiv)
|
64
|
+
* Added option :autofix_keys (disabled by default) to replace long keys with md5 hashes (sd)
|
65
|
+
|
66
|
+
= 1.7.4 (2009-06-09)
|
67
|
+
|
68
|
+
* Fix issue with raising timeout errors.
|
69
|
+
|
70
|
+
= 1.7.3 (2009-06-06)
|
71
|
+
|
72
|
+
* Remove SystemTimer support, refactor I/O to use nonblocking operations. Speeds up
|
73
|
+
performance approx 100%. Timeouts basically have no overhead now! (tenderlove)
|
74
|
+
* Update load logic to support SystemTimer running in Ruby Enterprise Edition. Thanks
|
75
|
+
to splattael on github for the comment.
|
76
|
+
|
77
|
+
= 1.7.2 (2009-04-12)
|
78
|
+
|
79
|
+
* Rollback socket timeout optimization. It does not work on all operating systems
|
80
|
+
and was a support headache.
|
81
|
+
|
82
|
+
= 1.7.1 (2009-03-28)
|
83
|
+
|
84
|
+
* Performance optimizations:
|
85
|
+
* Rely on higher performance operating system socket timeouts for low-level socket
|
86
|
+
read/writes where possible, instead of the (slower) SystemTimer or (slowest,
|
87
|
+
unreliable) Timeout libraries.
|
88
|
+
* the native binary search is back! The recent performance tuning made the binary search
|
89
|
+
a bottleneck again so it had to return. It uses RubyInline to compile the native extension and
|
90
|
+
silently falls back to pure Ruby if anything fails. Make sure you run:
|
91
|
+
`gem install RubyInline` if you want ultimate performance.
|
92
|
+
* the changes make memcache-client 100% faster than 1.7.0 in my performance test on Ruby 1.8.6:
|
93
|
+
15 sec -> 8 sec.
|
94
|
+
* Fix several logging issues.
|
95
|
+
|
96
|
+
= 1.7.0 (2009-03-08)
|
97
|
+
|
98
|
+
* Go through the memcached protocol document and implement any commands not already implemented:
|
99
|
+
- cas
|
100
|
+
- append
|
101
|
+
- prepend
|
102
|
+
- replace
|
103
|
+
|
104
|
+
Append and prepend only work with raw data since it makes no sense to concatenate two Marshalled
|
105
|
+
values together. The cas functionality should be considered a prototype. Since I don't have an
|
106
|
+
application which uses +cas+, I'm not sure what semantic sugar the API should provide. Should it
|
107
|
+
retry if the value was changed? Should it massage the returned string into true/false? Feedback
|
108
|
+
would be appreciated.
|
109
|
+
|
110
|
+
* Add fetch method which provides a method very similar to ActiveSupport::Cache::Store#fetch,
|
111
|
+
basically a wrapper around get and add. (djanowski)
|
112
|
+
|
113
|
+
* Implement the flush_all delay parameter, to allow a large memcached farm to be flushed gradually.
|
114
|
+
|
115
|
+
* Implement the noreply flag, which tells memcached not to reply in operations which don't
|
116
|
+
need a reply, i.e. set/add/delete/flush_all.
|
117
|
+
|
118
|
+
* The only known functionality not implemented anymore is the <flags> parameter to the storage
|
119
|
+
commands. This would require modification of the API method signatures. If someone can come
|
120
|
+
up with a clean way to implement it, I would be happy to consider including it.
|
121
|
+
|
122
|
+
= 1.6.5 (2009-02-27)
|
123
|
+
|
124
|
+
* Change memcache-client to multithreaded by default. The mutex does not add significant
|
125
|
+
overhead and it is far too easy, now that Sinatra, Rails and Merb are all thread-safe, to
|
126
|
+
use memcache-client in a thread-unsafe manner. Remove some unnecessary mutexing and add
|
127
|
+
a test to verify heavily multithreaded usage does not act unexpectedly.
|
128
|
+
|
129
|
+
* Add optional support for the SystemTimer gem when running on Ruby 1.8.x. This gem is
|
130
|
+
highly recommended - it ensures timeouts actually work and halves the overhead of using
|
131
|
+
timeouts. Using this gem, Ruby 1.8.x is actually faster in my performance tests
|
132
|
+
than Ruby 1.9.x. Just "gem install SystemTimer" and it should be picked up automatically.
|
133
|
+
|
134
|
+
= 1.6.4 (2009-02-19)
|
135
|
+
|
136
|
+
* Remove native code altogether. The speedup was only 10% on Ruby 1.8.6 and did not work
|
137
|
+
on Ruby 1.9.1.
|
138
|
+
|
139
|
+
* Removed memcache_util.rb from the distribution. If you are using it, please copy the code
|
140
|
+
into your own project. The file will live in the github repository for a few more months
|
141
|
+
for this purposes. http://github.com/mperham/memcache-client/raw/7a276089aa3c914e47e3960f9740ac7377204970/lib/memcache_util.rb
|
142
|
+
|
143
|
+
* Roll continuum.rb into memcache.rb. The project is again a single Ruby file, with no dependencies.
|
144
|
+
|
145
|
+
= 1.6.3 (2009-02-14)
|
146
|
+
|
147
|
+
* Remove gem native extension in preference to RubyInline. This allows the gem to install
|
148
|
+
and work on JRuby and Ruby 1.8.5 when the native code fails to compile.
|
149
|
+
|
150
|
+
= 1.6.2 (2009-02-04)
|
151
|
+
|
152
|
+
* Validate that values are less than one megabyte in size.
|
153
|
+
|
154
|
+
* Refactor error handling in get_multi to handle server failures and return what values
|
155
|
+
we could successfully retrieve.
|
156
|
+
|
157
|
+
* Add optional logging parameter for debugging and tracing.
|
158
|
+
|
159
|
+
* First official release since 1.5.0. Thanks to Eric Hodel for turning over the project to me!
|
160
|
+
New project home page: http://github.com/mperham/memcache-client
|
161
|
+
|
162
|
+
= 1.6.1 (2009-01-28)
|
163
|
+
|
164
|
+
* Add option to disable socket timeout support. Socket timeout has a significant performance
|
165
|
+
penalty (approx 3x slower than without in Ruby 1.8.6). You can turn off the timeouts if you
|
166
|
+
need absolute performance, but by default timeouts are enabled. The performance
|
167
|
+
penalty is much lower in Ruby 1.8.7, 1.9 and JRuby. (mperham)
|
168
|
+
|
169
|
+
* Add option to disable server failover. Failover can lead to "split-brain" caches that
|
170
|
+
return stale data. (mperham)
|
171
|
+
|
172
|
+
* Implement continuum binary search in native code for performance reasons. Pure ruby
|
173
|
+
is available for platforms like JRuby or Rubinius which can't use C extensions. (mperham)
|
174
|
+
|
175
|
+
* Fix #add with raw=true (iamaleksey)
|
176
|
+
|
177
|
+
= 1.6.0
|
178
|
+
|
179
|
+
* Implement a consistent hashing algorithm, as described in libketama.
|
180
|
+
This dramatically reduces the cost of adding or removing servers dynamically
|
181
|
+
as keys are much more likely to map to the same server.
|
182
|
+
|
183
|
+
Take a scenario where we add a fourth server. With a naive modulo algorithm, about
|
184
|
+
25% of the keys will map to the same server. In other words, 75% of your memcached
|
185
|
+
content suddenly becomes invalid. With a consistent algorithm, 75% of the keys
|
186
|
+
will map to the same server as before - only 25% will be invalidated. (mperham)
|
187
|
+
|
188
|
+
* Implement socket timeouts, should fix rare cases of very bad things happening
|
189
|
+
in production at 37signals and FiveRuns. (jseirles)
|
190
|
+
|
191
|
+
= 1.5.0.5
|
192
|
+
|
193
|
+
* Remove native C CRC32_ITU_T extension in favor of Zlib's crc32 method.
|
194
|
+
memcache-client is now pure Ruby again and will work with JRuby and Rubinius.
|
195
|
+
|
196
|
+
= 1.5.0.4
|
197
|
+
|
198
|
+
* Get test suite working again (packagethief)
|
199
|
+
* Ruby 1.9 compatiblity fixes (packagethief, mperham)
|
200
|
+
* Consistently return server responses and check for errors (packagethief)
|
201
|
+
* Properly calculate CRC in Ruby 1.9 strings (mperham)
|
202
|
+
* Drop rspec in favor of test/unit, for 1.9 compat (mperham)
|
203
|
+
|
204
|
+
= 1.5.0.3 (FiveRuns fork)
|
205
|
+
|
206
|
+
* Integrated ITU-T CRC32 operation in native C extension for speed. Thanks to Justin Balthrop!
|
207
|
+
|
208
|
+
= 1.5.0.2 (FiveRuns fork)
|
209
|
+
|
210
|
+
* Add support for seamless failover between servers. If one server connection dies,
|
211
|
+
the client will retry the operation on another server before giving up.
|
212
|
+
|
213
|
+
* Merge Will Bryant's socket retry patch.
|
214
|
+
http://willbryant.net/software/2007/12/21/ruby-memcache-client-reconnect-and-retry
|
215
|
+
|
216
|
+
= 1.5.0.1 (FiveRuns fork)
|
217
|
+
|
218
|
+
* Fix set not handling client disconnects.
|
219
|
+
http://dev.twitter.com/2008/02/solving-case-of-missing-updates.html
|
220
|
+
|
221
|
+
= 1.5.0
|
222
|
+
|
223
|
+
* Add MemCache#flush_all command. Patch #13019 and bug #10503. Patches
|
224
|
+
submitted by Sebastian Delmont and Rick Olson.
|
225
|
+
* Type-cast data returned by MemCache#stats. Patch #10505 submitted by
|
226
|
+
Sebastian Delmont.
|
227
|
+
|
228
|
+
= 1.4.0
|
229
|
+
|
230
|
+
* Fix bug #10371, #set does not check response for server errors.
|
231
|
+
Submitted by Ben VandenBos.
|
232
|
+
* Fix bug #12450, set TCP_NODELAY socket option. Patch by Chris
|
233
|
+
McGrath.
|
234
|
+
* Fix bug #10704, missing #add method. Patch by Jamie Macey.
|
235
|
+
* Fix bug #10371, handle socket EOF in cache_get. Submitted by Ben
|
236
|
+
VandenBos.
|
237
|
+
|
238
|
+
= 1.3.0
|
239
|
+
|
240
|
+
* Apply patch #6507, add stats command. Submitted by Tyler Kovacs.
|
241
|
+
* Apply patch #6509, parallel implementation of #get_multi. Submitted
|
242
|
+
by Tyler Kovacs.
|
243
|
+
* Validate keys. Disallow spaces in keys or keys that are too long.
|
244
|
+
* Perform more validation of server responses. MemCache now reports
|
245
|
+
errors if the socket was not in an expected state. (Please file
|
246
|
+
bugs if you find some.)
|
247
|
+
* Add #incr and #decr.
|
248
|
+
* Add raw argument to #set and #get to retrieve #incr and #decr
|
249
|
+
values.
|
250
|
+
* Also put on MemCacheError when using Cache::get with block.
|
251
|
+
* memcache.rb no longer sets $TESTING to a true value if it was
|
252
|
+
previously defined. Bug #8213 by Matijs van Zuijlen.
|
253
|
+
|
254
|
+
= 1.2.1
|
255
|
+
|
256
|
+
* Fix bug #7048, MemCache#servers= referenced changed local variable.
|
257
|
+
Submitted by Justin Dossey.
|
258
|
+
* Fix bug #7049, MemCache#initialize resets @buckets. Submitted by
|
259
|
+
Justin Dossey.
|
260
|
+
* Fix bug #6232, Make Cache::Get work with a block only when nil is
|
261
|
+
returned. Submitted by Jon Evans.
|
262
|
+
* Moved to the seattlerb project.
|
263
|
+
|
264
|
+
= 1.2.0
|
265
|
+
|
266
|
+
NOTE: This version will store keys in different places than previous
|
267
|
+
versions! Be prepared for some thrashing while memcached sorts itself
|
268
|
+
out!
|
269
|
+
|
270
|
+
* Fixed multithreaded operations, bug 5994 and 5989.
|
271
|
+
Thanks to Blaine Cook, Erik Hetzner, Elliot Smith, Dave Myron (and
|
272
|
+
possibly others I have forgotten).
|
273
|
+
* Made memcached interoperable with other memcached libraries, bug
|
274
|
+
4509. Thanks to anonymous.
|
275
|
+
* Added get_multi to match Perl/etc APIs
|
276
|
+
|
277
|
+
= 1.1.0
|
278
|
+
|
279
|
+
* Added some tests
|
280
|
+
* Sped up non-multithreaded and multithreaded operation
|
281
|
+
* More Ruby-memcache compatibility
|
282
|
+
* More RDoc
|
283
|
+
* Switched to Hoe
|
284
|
+
|
285
|
+
= 1.0.0
|
286
|
+
|
287
|
+
Birthday!
|
288
|
+
|
data/LICENSE.txt
ADDED
@@ -0,0 +1,28 @@
|
|
1
|
+
Copyright 2005-2009 Bob Cottrell, Eric Hodel, Mike Perham.
|
2
|
+
All rights reserved.
|
3
|
+
|
4
|
+
Redistribution and use in source and binary forms, with or without
|
5
|
+
modification, are permitted provided that the following conditions
|
6
|
+
are met:
|
7
|
+
|
8
|
+
1. Redistributions of source code must retain the above copyright
|
9
|
+
notice, this list of conditions and the following disclaimer.
|
10
|
+
2. Redistributions in binary form must reproduce the above copyright
|
11
|
+
notice, this list of conditions and the following disclaimer in the
|
12
|
+
documentation and/or other materials provided with the distribution.
|
13
|
+
3. Neither the names of the authors nor the names of their contributors
|
14
|
+
may be used to endorse or promote products derived from this software
|
15
|
+
without specific prior written permission.
|
16
|
+
|
17
|
+
THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS
|
18
|
+
OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
|
19
|
+
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
20
|
+
ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
|
21
|
+
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY,
|
22
|
+
OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
|
23
|
+
OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
|
24
|
+
BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
|
25
|
+
WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
|
26
|
+
OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
|
27
|
+
EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
28
|
+
|
data/README.rdoc
ADDED
@@ -0,0 +1,70 @@
|
|
1
|
+
= memcache-client
|
2
|
+
|
3
|
+
A ruby library for accessing memcached.
|
4
|
+
|
5
|
+
== memcache-client is deprecated as of August 2010. It will be supported through 2010 but new code should use Dalli[http://github.com/mperham/dalli] instead.
|
6
|
+
|
7
|
+
|
8
|
+
Source:
|
9
|
+
|
10
|
+
http://github.com/mperham/memcache-client
|
11
|
+
|
12
|
+
== Installing memcache-client
|
13
|
+
|
14
|
+
Just install the gem:
|
15
|
+
|
16
|
+
$ sudo gem install memcache-client
|
17
|
+
|
18
|
+
== Using memcache-client
|
19
|
+
|
20
|
+
With one server:
|
21
|
+
|
22
|
+
CACHE = MemCache.new 'localhost:11211'
|
23
|
+
|
24
|
+
Or with multiple servers:
|
25
|
+
|
26
|
+
CACHE = MemCache.new %w[one.example.com:11211 two.example.com:11211]
|
27
|
+
|
28
|
+
|
29
|
+
== Tuning memcache-client
|
30
|
+
|
31
|
+
The MemCache.new method takes a number of options which can be useful at times. Please
|
32
|
+
read the source comments there for an overview. If you are using Ruby 1.8.x and using
|
33
|
+
multiple memcached servers, you should install the RubyInline gem for ultimate performance.
|
34
|
+
|
35
|
+
|
36
|
+
== Using memcache-client with Rails
|
37
|
+
|
38
|
+
Rails 2.1+ includes memcache-client 1.5.0 out of the box. See ActiveSupport::Cache::MemCacheStore
|
39
|
+
and the Rails.cache method for more details. Rails 2.3+ will use the latest memcache-client
|
40
|
+
gem installed.
|
41
|
+
|
42
|
+
== Using memcache-client with EventMachine
|
43
|
+
|
44
|
+
memcache-client 1.8.0 added support for native EventMachine connections using
|
45
|
+
Ruby 1.9. If you are using an EventMachine-based application (e.g. thin), you can
|
46
|
+
activate the EventMachine support like so:
|
47
|
+
|
48
|
+
require 'memcache'
|
49
|
+
require 'memcache/event_machine'
|
50
|
+
|
51
|
+
EM.run do
|
52
|
+
Fiber.new do
|
53
|
+
m = MemCache.new('localhost:11211')
|
54
|
+
m.set 'abc', 'xyz'
|
55
|
+
m.get 'abc'
|
56
|
+
end.resume
|
57
|
+
end
|
58
|
+
|
59
|
+
== Questions?
|
60
|
+
|
61
|
+
memcache-client is maintained by Mike Perham and was originally written by Bob Cottrell,
|
62
|
+
Eric Hodel and the seattle.rb crew.
|
63
|
+
|
64
|
+
Email:: mailto:mperham@gmail.com
|
65
|
+
Twitter:: mperham[http://twitter.com/mperham]
|
66
|
+
WWW:: http://mikeperham.com
|
67
|
+
|
68
|
+
If my work on memcache-client is something you support, please take a moment to
|
69
|
+
recommend me at WWR[http://workingwithrails.com/person/10797-mike-perham]. I'm not
|
70
|
+
asking for money, just a electronic "thumbs up".
|
data/Rakefile
ADDED
@@ -0,0 +1,42 @@
|
|
1
|
+
# vim: syntax=Ruby
|
2
|
+
require 'rubygems'
|
3
|
+
require 'rake/rdoctask'
|
4
|
+
require 'rake/testtask'
|
5
|
+
|
6
|
+
require File.dirname(__FILE__) + "/lib/memcache/version.rb"
|
7
|
+
begin
|
8
|
+
require 'jeweler'
|
9
|
+
Jeweler::Tasks.new do |s|
|
10
|
+
s.name = "memcache-client"
|
11
|
+
s.version = MemCache::VERSION
|
12
|
+
s.summary = s.description = "A Ruby library for accessing memcached."
|
13
|
+
s.email = "mperham@gmail.com"
|
14
|
+
s.homepage = "http://github.com/mperham/memcache-client"
|
15
|
+
s.authors = ['Eric Hodel', 'Robert Cottrell', 'Mike Perham']
|
16
|
+
s.has_rdoc = true
|
17
|
+
s.files = FileList["[A-Z]*", "{lib,test}/**/*", 'performance.txt']
|
18
|
+
s.test_files = FileList["test/test_*.rb"]
|
19
|
+
s.executables = ['memcached_top']
|
20
|
+
end
|
21
|
+
Jeweler::GemcutterTasks.new
|
22
|
+
rescue LoadError
|
23
|
+
puts "Jeweler not available. Install it for jeweler-related tasks with: sudo gem install jeweler"
|
24
|
+
end
|
25
|
+
|
26
|
+
|
27
|
+
Rake::RDocTask.new do |rd|
|
28
|
+
rd.main = "README.rdoc"
|
29
|
+
rd.rdoc_files.include("README.rdoc", "FAQ.rdoc", "History.rdoc", "lib/memcache.rb")
|
30
|
+
rd.rdoc_dir = 'doc'
|
31
|
+
end
|
32
|
+
|
33
|
+
Rake::TestTask.new do |t|
|
34
|
+
t.warning = true
|
35
|
+
t.libs = ['lib', 'test']
|
36
|
+
end
|
37
|
+
|
38
|
+
task :default => :test
|
39
|
+
|
40
|
+
task :rcov do
|
41
|
+
`rcov -Ilib test/*.rb`
|
42
|
+
end
|