Debian Codegraph @

CHANGELOG

raw
1 0.20 9 November 2019
2 Adoption.
3
4 0.19 20 September 2010
5 Checking tool chain with perl.5.13.5.
6
7 0.18 3 October 2007
8 Hopefully fixed test-suite for none thread enabled Perls. Updated
9 to use new VERSION feature of Devel::Required.
10
11 0.17 7 January 2004
12 Fixed some documentation inconsistencies. Added documentation on what
13 is different from Thread::Queue::Any.
14
15 Made default optimization dependent on the version of Perl used.
16 Added test for it.
17
18 Removed mention of Storable: replaced it with Thread::Serialize.
19 Added Thread::Serialize as a dependency, just for documentation sake
20 really.
21
22 0.16 28 December 2003
23 Added automatic required modules update using Devel::Required.
24
25 Added dependency to load.pm.
26
27 0.15 9 November 2003
28 A problem in (threaded) 5.8.2 made Thread::Conveyor segfault in the
29 test-suite. The problem was fortunately easy to circumvent: just don't
30 use subroutines consisting of {} as an object method (in the case of
31 Thread::Conveyor, the shutdown() method).
32
33 0.14 11 August 2003
34 Cleaned up Makefile.PL and updated copyright info. Verified everything
35 works with 5.8.1. Updated dependency of Thread::Tie.
36
37 0.13 30 September 2002
38 Use "load.pm" instead of AutoLoader. Changed dependency on Thread::Tie
39 to version 0.08 (which also uses load.pm).
40
41 Added check for availability of methods again: load.pm does support
42 that, whereas AutoLoader does not.
43
44 0.12 27 September 2002
45 Ask found another problem in T::C::Tied::clean_dontwait. Fixed.
46
47 Also left warnings enabled in the test-suite last time, which caused
48 warnings in the make test. Changed code so that the warnings are not
49 emitted anymore.
50
51 0.11 26 September 2002
52 Fixed some warnings in T::C::Throttled, spotted by Ask Bjoern Hansen.
53
54 Removed "our" from $VERSION and @ISA, should shave off some bytes in
55 memory usage, as found from testing with Benchmark::Thread::Size.
56
57 0.10 1 September 2002
58 Replaced dependency on Thread::Serialize by dependency on Thread::Tie
59 as the default, memory optimized implementation, now uses Thread::Tie.
60
61 Changed Thread::Conveyor so that it uses T::C::Tied for unthrottled
62 belts when optimizing for memory.
63
64 Changed T::C::Throttled so that it makes its own decision on which
65 belt implementation to be used, either T::C::Array (optimized for
66 CPU) or T::C::Tied (optimized for memory).
67
68 Added "semaphore" object method to T::C::Array for obtaining the
69 lock() semaphore, needed by T::C::Throttled.
70
71 Removed T::C::Thread from the distribution. This shaves about 1200
72 bytes off of the distribution package.
73
74 Replaced Thread::Conveyor::Thread by Thread::Conveyor::Tied, which
75 uses Thread::Tie as its shared array implementation. This has the
76 side-effect to being almost twice as slow but saving 18% of
77 memory (at least in the test-suite). So there is now a true CPU
78 versus memory trade-off. Main difference in CPU is caused by the
79 fact that whole array fetches aren't possible using the tie()
80 interface (this is a loop with FETCH()), wherease T::C::Thread was
81 optimized to deal with that case). Main difference in memory is
82 caused by the fact that only one thread (the Thread::Tie default
83 thread) is used for all belts, whereas before each belt had its own
84 thread. So the savings in memory could in real world situations, be
85 quite a lot more than the 18% seen in the test-suite.
86
87 0.09 30 August 2002
88 Changed all modules to use AutoLoader to defer loading of necessary
89 subroutines to when they are actually needed. This should save
90 memory and CPU for larger programs, or with large numbers of threads.
91 The test-suite only marginally takes more memory at the same CPU
92 usage: overhead of compiling is levelled out with the overhead of
93 cloning pre-compiled routines.
94
95 0.08 20 August 2002
96 Dave Mitchell showed me the error of my ways with regards to client
97 server thread programming. Fixed the handler optimized for memory so
98 that no yield()s are necessary anymore. The whole thing now gets
99 through the test suite more than 25% faster for that part.
100 Scalability to many more threads should now be a lot better too.
101
102 0.07 14 August 2002
103 Removed (possibly premature) optimization from T::C::Thread
104 _handler: it now doesn't zap the local threadspace anymore.
105
106 0.06 13 August 2002
107 Handed over freezing and thawing operations to Thread::Serialize.
108 Added dependency to Thread::Serialize in Makefile.PL.
109
110 0.05 6 August 2002
111 Fixed "onbelt" in T::C::Thread so that it won't hang if the belt is
112 shut down. Needed for Thread::Pool and in general useful.
113
114 5 August 2002
115 Added index functionality to "peek" and "peek_dontwait" to all
116 sub-classes and added documentation for it. Needed for Thread::Pool.
117
118 Finally got the T::C::Thread implementation getting through the
119 test-suite reliably. Problem was caused by my misconception that
120 cond_wait would always regain the lock() after being signalled. It
121 does not. This should have been in big red letters in the
122 threads::shared pod.
123
124 Added methods "shutdown", "thread" and "tid" plus documentation.
125
126 3 August 2002
127 Adapted test-suite to use many more combinations of possible
128 settings.
129
130 2 August 2002
131 Added new "optimize" field parameter to allow you to select an
132 implementation optimized for speed (T::C::Array and T::C::Throttled)
133 or for memory (T::C::Thread).
134
135 Added new implementation of Thread::Conveyor based on using a
136 seperate thread for keeping the belt: Thread::Conveyor::Thread.
137
138 Move basic parts of Thread::Conveyor to Thread::Conveyor::Array.
139 Which should allow for other implementations of the conveyor
140 metaphor.
141
142 0.04 1 August 2002
143 Added internal method "_belt" to get at the actual belt, for both
144 the normal and the throttled belt.
145
146 0.03 31 July 2002
147 Added internal methods "_freeze" and "_thaw" so that we can move
148 to a different serialization scheme or other optimizations without
149 having to change all the sub-classes as well. First user is
150 Thread::Conveyor::Monitored.
151
152 Added "maxjobs" and "minjobs" method in Thread::Conveyor that will
153 die when they're called (throttling only takes place in
154 Thread::Conveyor::Throttled).
155
156 Fixed problem that would cause throttling to not be switched off
157 when the "maxjobs" method was specified with an undef value.
158
159 Added methods "clean" and "clean_dontwait" for cleaning the belt
160 to both Thread::Conveyor and Thread::Conveyor::Throttled. Also
161 added documentation for it.
162
163 Added internal method "_clean" for obtaining frozen contents of
164 the belt to both Thread::Conveyor and Thread::Conveyor::Throttled.
165 This was needed for Thread::Conveyor::Monitored.
166
167 0.02 30 July 2002
168 Adapted Thread::Conveyor to default to a throttled conveyor belt
169 automatically if no specific parameters specified. Now allows for
170 specification of throttling settings in a parameter hash reference.
171 If unthrottling is specifically specified, returns to the old
172 unthrottled behaviour (where the object is just an array reference),
173 thus keeping the old execution speed.
174
175 Added helper class Thread::Conveyor::Throttled to allow for throttled
176 conveyor belts that block if there are too many boxes on the belt.
177
178 Fixed some documentation nits.
179
180 0.01 25 July 2002
181 First version of Thread::Conveyor, which started life as
182 Thread::Queue::Any.