ml::ScreenSource Class Reference

Grabs the desktop as pixels, using the OS's own capture API. More...

#include <Malena/Screen/ScreenSource.h>

Public Member Functions
	ScreenSource ()
	ScreenSource (const ScreenSource &)=delete
	~ScreenSource ()
int	displayIndex () const
std::optional< sf::Image >	grab ()
const std::string &	lastError () const
ScreenSource &	operator= (const ScreenSource &)=delete
void	setDisplayIndex (int index)
void	setMaxDimension (unsigned int pixels)

Static Public Member Functions
static std::optional< sf::Image >	decode (const std::uint8_t *data, std::size_t size)
static std::optional< std::vector< std::uint8_t > >	encodeJpeg (const sf::Image &frame)
static bool	isSupported ()

Detailed Description

Grabs the desktop as pixels, using the OS's own capture API.

This is the capture half of screen sharing with no third-party media runtime: no GStreamer, nothing for the user to install. Each platform uses its native call (CGDisplayCreateImage on macOS, BitBlt on Windows, XGetImage on X11), and the result is handed back as a plain sf::Image so the rest of the pipeline is platform-agnostic.

Deliberately a pull API: the caller decides the frame rate by how often it calls grab(). A shared screen only needs a handful of frames per second, and pulling keeps the cost proportional to what is actually sent.

ml::ScreenSource src;
if (auto frame = src.grab())
    if (auto jpeg = ml::ScreenSource::encodeJpeg(*frame))
        send(*jpeg);

Definition at line 38 of file ScreenSource.h.

Constructor & Destructor Documentation

ScreenSource() [1/2]

ml::ScreenSource::ScreenSource

(

)

~ScreenSource()

ml::ScreenSource::~ScreenSource

(

)

ScreenSource() [2/2]

ml::ScreenSource::ScreenSource ( const ScreenSource & )

delete

Member Function Documentation

decode()

std::optional< sf::Image > ml::ScreenSource::decode ( const std::uint8_t * data, std::size_t size )

static

Decode a frame produced by encodeJpeg (or any format SFML reads).

displayIndex()

int ml::ScreenSource::displayIndex

(

)

const

encodeJpeg()

std::optional< std::vector< std::uint8_t > > ml::ScreenSource::encodeJpeg ( const sf::Image & frame )

static

Encode to JPEG in memory. Uses SFML's own encoder, so it needs no extra dependency; quality is fixed by SFML. Returns nullopt on failure.

grab()

std::optional< sf::Image > ml::ScreenSource::grab

(

)

Grab one frame, or nullopt if capture failed / is unavailable here.

isSupported()

bool ml::ScreenSource::isSupported ( )

static

True when this platform has a capture backend compiled in.

lastError()

const std::string & ml::ScreenSource::lastError

(

)

const

Human-readable reason capture is unavailable, or "" when it works.

operator=()

ScreenSource & ml::ScreenSource::operator= ( const ScreenSource & )

delete

setDisplayIndex()

void ml::ScreenSource::setDisplayIndex

(

int

index

)

Which display to capture (0 = main). Out-of-range falls back to main.

setMaxDimension()

void ml::ScreenSource::setMaxDimension

(

unsigned int

pixels

)

Cap the long edge of captured frames, preserving aspect (0 = no scaling). A Retina desktop is enormous; downscaling before encoding is the single biggest win for both bandwidth and encode time.

---

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

• Screen/ScreenSource.h