Struct SseDecoder 

pub struct SseDecoder { /* private fields */ }

Incremental SSE frame decoder: feed raw body chunks, drain complete frames. Tolerates CRLF, multi-line data:, and frames split across chunk boundaries — including chunks split inside a multi-byte UTF-8 character. Comments, id: and retry: lines are framing noise (dropped).

Implementations

impl SseDecoder

pub fn new() -> Self

Create a decoder with the default untrusted-input limits.

pub fn with_limits(limits: SseLimits) -> Self

Create a decoder with explicit frame and buffer limits.

pub fn push(&mut self, chunk: &[u8]) -> Result<Vec<SseFrame>, TransformError>

Append a chunk and return all complete (blank-line-terminated) frames. SSE is text by definition; genuinely invalid UTF-8 is replaced lossily, while an incomplete trailing sequence waits for the next chunk.

Frames are parsed by BORROWING slices of the buffer behind an advancing cursor; consumed bytes are drained once at the end. (The former per-frame drain(..).collect::<String>() walked the frame char by char AND memmoved the buffer tail per frame — the top CPU hotspot of every streaming profile.)

pub fn finish(&mut self) -> Result<Option<SseFrame>, TransformError>

Drain a trailing, unterminated frame at end of stream (some upstreams omit the final blank line).

Trait Implementations

impl Debug for SseDecoder

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SseDecoder

fn default() -> SseDecoder

Returns the “default value” for a type.