#include <threaded_job_scheduler.hpp>

Inheritance diagram for moost::thread::threaded_job_batch:

Collaboration diagram for moost::thread::threaded_job_batch:

Public Member Functions
	threaded_job_batch ()
virtual void	add (job_t job)
void	run (worker_group &wg)
size_t	count () const
const std::vector< std::string > &	errors () const
Private Types
typedef std::queue< job_t >	jobs_t
Private Member Functions
void	do_one (worker_group &wg)
bool	done () const
void	run_one (job_t &job)
Private Attributes
jobs_t	m_jobs
boost::condition_variable	m_cond
boost::mutex	m_mx
volatile size_t	m_todo
volatile size_t	m_count
std::vector< std::string >	m_errors
boost::mutex	m_mx_errors

Detailed Description

A job batch that runs jobs in a set of worker threads

This implementation of a job_batch runs jobs in a set of worker threads provided by the threaded_job_scheduler class.

Definition at line 53 of file threaded_job_scheduler.hpp.

Member Typedef Documentation

typedef std::queue<job_t> moost::thread::threaded_job_batch::jobs_t [private]

Definition at line 119 of file threaded_job_scheduler.hpp.

Constructor & Destructor Documentation

moost::thread::threaded_job_batch::threaded_job_batch ( ) [inline]

Definition at line 58 of file threaded_job_scheduler.hpp.

Member Function Documentation

virtual void moost::thread::threaded_job_batch::add ( job_t job ) [inline, virtual]

Add a new job to the batch

This method adds a job to the concrete implementation. When this job gets run is determined by the implementation. However, new jobs may be added even while jobs from the current batch are being run as long as the batch doesn't run out of jobs. In particular, this means that it's always safe to add new jobs from within jobs of the current batch.

Parameters:

job	The job to be added.

Implements moost::thread::job_batch.

Definition at line 64 of file threaded_job_scheduler.hpp.

size_t moost::thread::threaded_job_batch::count ( ) const [inline]

Get the number of jobs executed in this batch

Returns:: The total number of jobs.

Definition at line 88 of file threaded_job_scheduler.hpp.

void moost::thread::threaded_job_batch::do_one ( worker_group & wg ) [inline, private]

Enqueue one job to worker threads

This method is supposed to be called by threaded_job_scheduler only. It pops the next job off the job queue and enqueues it to run in the next available worker thread.

Definition at line 128 of file threaded_job_scheduler.hpp.

Here is the call graph for this function:

Here is the caller graph for this function:

bool moost::thread::threaded_job_batch::done ( ) const [inline, private]

Check if all jobs are done

This method is supposed to be called by threaded_job_scheduler only. Does more or less what it says on the tin, though it'll block if the job queue is empty. This is to ensure that new jobs can still be added while other jobs are being run.

Definition at line 150 of file threaded_job_scheduler.hpp.

Here is the caller graph for this function: