catapult::thread Namespace Reference

Namespaces
	detail

Classes
class	future
	Provides a way to access the result of an asynchronous operation. More...
class	IoThreadPool
	Represents a thread pool that shares a single io context across multiple threads. More...
class	MultiServicePool
class	promise
	Stores the result of an asynchronous operation. More...
class	Scheduler
	A scheduler. More...
class	StrandedTimedCallback
class	StrandOwnerLifetimeExtender
	Wraps a strand and automatically augments handlers to extend the lifetime of an owning object. More...
struct	Task
	A task that can be dispatched to the scheduler. More...

Typedefs
using	TaskCallback = supplier< thread::future< TaskResult > >
	Task callback that is invoked by the scheduler. More...
using	DelayGenerator = supplier< utils::TimeSpan >
	Supplier that generates delays. More...

Enumerations
enum	TaskResult { TaskResult::Continue, TaskResult::Break }
	Result of a task. More...

Functions
template<typename T >
future< T >	make_ready_future (T &&value)
	Produces a future that is ready immediately and holds the given value. More...
template<typename T , typename E >
future< T >	make_exceptional_future (E ex)
	Produces a future that is ready immediately and holds the given exception (ex). More...
template<typename T >
future< std::vector< future< T > > >	when_all (std::vector< future< T >> &&allFutures)
	Returns a future that is signaled when all futures in allFutures complete. More...
template<typename T >
future< std::vector< future< T > > >	when_all (future< T > &&future1, future< T > &&future2)
	Returns a future that is signaled when both future1 and future2 complete. More...
template<typename TSeed , typename TCreateNextFuture , typename TResultFuture = std::invoke_result_t<TCreateNextFuture, future<TSeed>&&>, typename TResultType = decltype(TResultFuture().get())>
auto	compose (future< TSeed > &&startFuture, TCreateNextFuture createNextFuture)
template<typename T >
std::vector< T >	get_all_ignore_exceptional (std::vector< future< T >> &&futures)
template<typename T >
std::vector< T >	get_all (std::vector< future< T >> &&futures)
std::unique_ptr< IoThreadPool >	CreateIoThreadPool (size_t numWorkerThreads, const char *name)
template<typename TItems , typename TWorkCallback >
thread::future< bool >	ParallelForPartition (boost::asio::io_context &ioContext, TItems &items, size_t numPartitions, TWorkCallback callback)
template<typename TItems , typename TWorkCallback >
thread::future< bool >	ParallelFor (boost::asio::io_context &ioContext, TItems &items, size_t numPartitions, TWorkCallback callback)
std::shared_ptr< Scheduler >	CreateScheduler (const std::shared_ptr< IoThreadPool > &pPool)
	Creates a scheduler around the specified thread pool (pPool). More...
DelayGenerator	CreateUniformDelayGenerator (const utils::TimeSpan &delay)
	Creates a uniform delay generator that always returns delay. More...
DelayGenerator	CreateIncreasingDelayGenerator (const utils::TimeSpan &minDelay, uint32_t numPhaseOneRounds, const utils::TimeSpan &maxDelay, uint32_t numTransitionRounds)
Task	CreateNamedTask (const std::string &name, const TaskCallback &callback)
	Creates an unscheduled task with name and callback. More...
size_t	GetMaxThreadNameLength ()
	Gets the maximum supported thread name length (excluding NUL-terminator). More...
void	SetThreadName (const std::string &name)
std::string	GetThreadName ()
	Gets a thread name in a platform-dependent way. More...
template<typename TCallback , typename... TCallbackArgs>
auto	MakeTimedCallback (boost::asio::io_context &ioContext, TCallback callback, TCallbackArgs &&... timeoutArgs)

Typedef Documentation

DelayGenerator

using catapult::thread::DelayGenerator = typedef supplier<utils::TimeSpan>

Supplier that generates delays.

TaskCallback

using catapult::thread::TaskCallback = typedef supplier<thread::future<TaskResult> >

Task callback that is invoked by the scheduler.

Enumeration Type Documentation

TaskResult

enum catapult::thread::TaskResult

strong

Result of a task.

Enumerator
Continue	Task should repeat.
Break	Task should not repeat.

Function Documentation

compose()

template<typename TSeed , typename TCreateNextFuture , typename TResultFuture = std::invoke_result_t<TCreateNextFuture, future<TSeed>&&>, typename TResultType = decltype(TResultFuture().get())>

auto catapult::thread::compose	(	future< TSeed > &&	startFuture,
		TCreateNextFuture	createNextFuture
	)

Composes a future (startFuture) with a second future returned by a function (createNextFuture) that is passed the original future upon its completion.

Note

The future returned by this function will be the same type as the one returned by createNextFuture.

Here is the caller graph for this function:

CreateIncreasingDelayGenerator()

DelayGenerator catapult::thread::CreateIncreasingDelayGenerator	(	const utils::TimeSpan &	minDelay,
		uint32_t	numPhaseOneRounds,
		const utils::TimeSpan &	maxDelay,
		uint32_t	numTransitionRounds
	)

Creates a three-phase increasing delay generator:

• phase 1: minDelay for numPhaseOneRounds rounds
• phase 2: delay is linearly increased from minDelay to maxDelay over numTransitionRounds rounds
• phase 3: maxDelay for all other rounds

Here is the call graph for this function:

CreateIoThreadPool()

std::unique_ptr< IoThreadPool > catapult::thread::CreateIoThreadPool	(	size_t	numWorkerThreads,
		const char *	name = nullptr
	)

Creates an io thread pool with the specified number of threads (numWorkerThreads) and the optional friendly name used in logging.

Here is the caller graph for this function:

CreateNamedTask()

Task catapult::thread::CreateNamedTask	(	const std::string &	name,
		const TaskCallback &	callback
	)

Creates an unscheduled task with name and callback.

Here is the caller graph for this function:

CreateScheduler()

std::shared_ptr< Scheduler > catapult::thread::CreateScheduler

(

const std::shared_ptr< IoThreadPool > &

pPool

)

Creates a scheduler around the specified thread pool (pPool).

CreateUniformDelayGenerator()

DelayGenerator catapult::thread::CreateUniformDelayGenerator

(

const utils::TimeSpan &

delay

)

Creates a uniform delay generator that always returns delay.

get_all()

template<typename T >

std::vector<T> catapult::thread::get_all

(

std::vector< future< T >> &&

futures

)

Evaluates all futures and returns the results of all.

Note

This function will throw if at least one future has an exceptional result.

Here is the call graph for this function:

Here is the caller graph for this function:

get_all_ignore_exceptional()

template<typename T >

std::vector<T> catapult::thread::get_all_ignore_exceptional

(

std::vector< future< T >> &&

futures

)

Evaluates all futures and returns the results of all non-exceptional ones.

Note

Exceptional results are ignored and filtered out.

Here is the call graph for this function:

GetMaxThreadNameLength()

size_t catapult::thread::GetMaxThreadNameLength

(

)

Gets the maximum supported thread name length (excluding NUL-terminator).

Here is the caller graph for this function:

GetThreadName()

std::string catapult::thread::GetThreadName

(

)

Gets a thread name in a platform-dependent way.

Here is the call graph for this function:

make_exceptional_future()

template<typename T , typename E >

future<T> catapult::thread::make_exceptional_future

(

E

ex

)

Produces a future that is ready immediately and holds the given exception (ex).

make_ready_future()

template<typename T >

future<T> catapult::thread::make_ready_future

(

T &&

value

)

Produces a future that is ready immediately and holds the given value.

Here is the caller graph for this function:

MakeTimedCallback()

template<typename TCallback , typename... TCallbackArgs>

auto catapult::thread::MakeTimedCallback	(	boost::asio::io_context &	ioContext,
		TCallback	callback,
		TCallbackArgs &&...	timeoutArgs
	)

Wraps callback with a timed callback using ioContext. On a timeout, the callback is invoked with timeoutArgs.

Here is the caller graph for this function:

ParallelFor()

template<typename TItems , typename TWorkCallback >

thread::future<bool> catapult::thread::ParallelFor	(	boost::asio::io_context &	ioContext,
		TItems &	items,
		size_t	numPartitions,
		TWorkCallback	callback
	)

Uses ioContext to process items in numPartitions batches and calls callback for each item. A future is returned that is resolved when all items have been processed.

Here is the call graph for this function:

ParallelForPartition()

template<typename TItems , typename TWorkCallback >

thread::future<bool> catapult::thread::ParallelForPartition	(	boost::asio::io_context &	ioContext,
		TItems &	items,
		size_t	numPartitions,
		TWorkCallback	callback
	)

Uses ioContext to process items in numPartitions batches and calls callback for each partition. A future is returned that is resolved when all items have been processed.

Here is the caller graph for this function:

SetThreadName()

void catapult::thread::SetThreadName

(

const std::string &

name

)

Sets a thread name in a platform-dependent way.

Note

Depending on platform capabilities, the name might be truncated.

Here is the call graph for this function:

Here is the caller graph for this function:

when_all() [1/2]

template<typename T >

future<std::vector<future<T> > > catapult::thread::when_all	(	future< T > &&	future1,
		future< T > &&	future2
	)

Returns a future that is signaled when both future1 and future2 complete.

Here is the call graph for this function:

when_all() [2/2]

template<typename T >

future<std::vector<future<T> > > catapult::thread::when_all

(

std::vector< future< T >> &&

allFutures

)

Returns a future that is signaled when all futures in allFutures complete.

Here is the call graph for this function:

Here is the caller graph for this function: