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
|
Just before a thread ends it may have produced some results. These results may
have to to be communicated to other threads. In multi threaded programs
several classes and functions can be used that produce
hi(shared state)em(shared states), making it easy to communicate results
to other threads. Results could be values, objects or exceptions.
Objects that contain such shared states are called
hi(asynchronous return object)em(asynchronous return objects). However,
due to the nature of multi threading, a thread may request the results of an
asynchronous return object before these result are actually available. In
those cases the requesting thread blocks, waiting for the results to become
available. Asynchronous return objects offer tt(wait) and tt(get) members
which, respectively, em(wait) until the results have become available, and
em(produce) the asynchronous results once they are available. The phrase that
is used to indicate that the results are available is `the shared state has
been made ready'.
Shared states are made ready by
hi(asynchronous provider)em(asynchronous providers). Asynchronous
providers are simply objects or functions providing results to shared
states. Making a shared state ready means that an asynchronous provider
itemization(
it() marks its shared state as being ready, and
it() unblocks any waiting threads (e.g., by allowing blocking members,
like tt(wait), to return).
)
Once a shared state has been made ready it contains a value, object, or
exception which can be retrieved by objects having access to the shared
state. While code is waiting for a shared state to become ready the value or
exception that is going to be stored in the shared state may be computed. When
multiple threads try to access the same shared state they must use
synchronizing mechanisms (like mutexes, cf. section ref(MUTEX)) to prevent
access-conflicts.
Shared states use reference counting to keep track of the number of
asynchronous return objects or asynchronous providers that hold references to
them. These return objects and providers may release their references to these
shared states hi(shared state: releasing) (which is called `releasing the
shared state). This happens when a return object or provider holds the last
reference to the shared state, and
hi(shared state: destruction) the shared state is destroyed.
On the other hand, an asynchronous provider may also
hi(shared state: abandon)em(abandon) its shared state. In that case the
provider, in sequence,
itemization(
it() stores an exception object of type
hi(future_error)tt(std::future_error), holding the error condition
tt(std::broken_promise) in its shared state;
it() makes its shared data ready; and
it() releases its shared data.
)
Objects of the class tt(std::future) (see the next section) are asynchronous
return objects. They can be produced by the tt(std::async) (section
ref(ASYNC)) family of functions, and by objects of the classes
tt(std::packaged_task) (section ref(PACKAGE)), and tt(std::promise) (section
ref(PROMISE)).
|
