StormByte::Buffer::Pipeline Class Reference

A class designed to execute a sequence of functions asynchronously in a pipeline. More...

#include <pipeline.hxx>

Public Member Functions
	Pipeline () noexcept=default
	Default constructor Initializes an empty pipeline buffer.
	Pipeline (const Pipeline &other)=default
	Copy constructor Creates a new Pipeline that shares the same underlying buffer as the original.
	Pipeline (Pipeline &&other) noexcept=default
	Move constructor Moves the contents of another Pipeline into this instance.
	~Pipeline () noexcept=default
	Destructor.
Pipeline &	operator= (const Pipeline &other)=default
	Copy assignment operator.
Pipeline &	operator= (Pipeline &&other) noexcept=default
	Move assignment operator.
void	AddPipe (const PipeFunction &pipe)
	Adds a new pipe function to the pipeline.
void	AddPipe (PipeFunction &&pipe)
	Moves a new pipe function to the pipeline.
Consumer	Process (Consumer buffer) const noexcept
	Processes the pipeline buffer.

Detailed Description

A class designed to execute a sequence of functions asynchronously in a pipeline.

The Pipeline class allows users to define a chain of processing functions, where each function:

• Takes two parameters:
  1. A Consumer buffer to retrieve input data.
  2. A Producer buffer to store processed output data.
• Returns a bool indicating the success or failure of the operation.

Functions are executed in the order they are added to the pipeline. Each function runs asynchronously in a detached thread, allowing all functions in the chain to execute concurrently. The output of one function is fed as the input to the next function in the chain. If any function in the chain returns false, the output buffer is marked with a status of Status::Error. It is the responsibility of each subsequent function in the chain to check the input buffer's status and return false if the status is Status::Error. This ensures that the error condition is propagated consistently until the end of the pipeline.

Buffer Status Handling:

• There is no need to manually update the buffer status inside the function. The pipeline will handle this automatically:
  • A return value of true will set the output buffer's status to Status::EoF, signaling to the next function that the input has finished.
  • A return value of false will set the output buffer's status to Status::Error, ensuring that subsequent functions return false accordingly.

Thread Synchronization:

• The pipeline relies on the Async buffer's shared ownership model to manage thread synchronization and lifetime.
• Each detached thread processes its assigned function and updates the buffer's status accordingly.
• The returned buffer from the Process function represents the final output of the pipeline. Its status indicates the state of the processing:
  • Status::Ready: Data is still being processed in at least one thread.
  • Status::EoF: All threads have completed successfully.
  • Status::Error: An error occurred in one of the threads, and processing was terminated.

Important Notes:

• Functions must not throw exceptions.
• If any function does not return a bool or does not finish execution, the behavior of the pipeline is undefined.
• The lifetime of intermediate buffers is managed automatically through std::shared_ptr. Buffer are destroyed when no longer needed.

Constructor & Destructor Documentation

Pipeline() [1/2]

StormByte::Buffer::Pipeline::Pipeline ( const Pipeline &  other )

default

Copy constructor Creates a new Pipeline that shares the same underlying buffer as the original.

Parameters

other

Pipeline to copy

Pipeline() [2/2]

StormByte::Buffer::Pipeline::Pipeline ( Pipeline &&  other )

defaultnoexcept

Move constructor Moves the contents of another Pipeline into this instance.

Parameters

other

Pipeline to move

Member Function Documentation

AddPipe() [1/2]

void StormByte::Buffer::Pipeline::AddPipe

(

const PipeFunction &

pipe

)

Adds a new pipe function to the pipeline.

Parameters

pipe	Pipe function to add

AddPipe() [2/2]

void StormByte::Buffer::Pipeline::AddPipe

(

PipeFunction &&

pipe

)

Moves a new pipe function to the pipeline.

Parameters

pipe	Pipe function to move

operator=() [1/2]

Pipeline & StormByte::Buffer::Pipeline::operator= ( const Pipeline &  other )

default

Copy assignment operator.

Parameters

other

Pipeline instance to copy from

Returns

Reference to the updated Pipeline instance

operator=() [2/2]

Pipeline & StormByte::Buffer::Pipeline::operator= ( Pipeline &&  other )

defaultnoexcept

Move assignment operator.

Parameters

other

Pipeline instance to move from

Returns

Reference to the updated Pipeline instance

Process()

Consumer StormByte::Buffer::Pipeline::Process ( Consumer  buffer ) const

noexcept

Processes the pipeline buffer.

The provided buffer is passed through the chain of functions in the pipeline. Each function processes the buffer and passes its output to the next function in the chain. All functions are executed asynchronously in detached threads, allowing concurrent processing of the pipeline stages.

The returned buffer represents the final output of the pipeline. Its status indicates the state of the processing:

• Status::Ready: Data is still being processed in at least one thread.
• Status::EoF: All threads have completed successfully.
• Status::Error: An error occurred in one of the threads, and processing was terminated.

Execution Details:

• Each function in the pipeline runs in a detached thread, and the Async buffer ensures proper synchronization and lifetime management.
• The pipeline automatically propagates errors by marking the buffer with Status::Error if a function fails.
• The returned buffer is the last buffer in the chain, ensuring that its status reflects the state of the entire pipeline.

Parameters

buffer

Buffer to process

Returns

Processed buffer

---

The documentation for this class was generated from the following file:

• /home/runner/work/StormByte/StormByte/lib/public/StormByte/buffer/pipeline.hxx