Trait ImageAdapter 

pub trait ImageAdapter {
    type CodecId: Copy + Eq + Debug;
    type PixelFormat: Clone + Eq + Debug;
    type PacketExtra;
    type FrameExtra;
}

Backend vocabulary for a compressed/decoded still image — cover art, an embedded thumbnail, a poster frame.

A separate vocabulary from VideoAdapter rather than a reuse of it, because the two disagree about the one thing an adapter exists to name: what rides the frame. A still’s extras are EXIF, an ICC profile, an orientation — not a picture type, not a field order, not a best-effort timestamp. Bending VideoAdapter::FrameExtra over both would give every motion frame seats that only a still fills and every still seats that only motion fills.

The compressed side is an AttachmentPacket: a still image inside a container is an attachment, one whole file, no timeline. PacketExtra is therefore the attachment’s extras, and a backend that also implements DemuxAdapter will normally bind the same type in both seats — the packet a demuxer hands out is the packet an image decoder is fed.

Required Associated Types

type CodecId: Copy + Eq + Debug

Codec identifier — the still’s own codec (MJPEG, PNG, BMP, …), in the same namespace the rest of the backend uses.

type PixelFormat: Clone + Eq + Debug

Pixel format identifier.

Clone, not Copy, for the same reason VideoAdapter::PixelFormat is.

type PacketExtra

Backend-specific extras carried on the AttachmentPacket an image decoder is fed.

type FrameExtra

Backend-specific extras carried on every ImageFrame — EXIF, ICC profile, side data.

Dyn Compatibility

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors