chore: encode: use FlushError instead of Flush (#6168)

Co-authored-by: Francis Lavoie <lavofr@gmail.com>
This commit is contained in:
WeidiDeng 2024-03-11 11:04:35 +08:00 committed by Matthew Holt
parent 898ef9f212
commit c5d6b22f37
1 changed files with 5 additions and 5 deletions

View File

@ -202,15 +202,15 @@ func (enc *Encode) Match(rw *responseWriter) bool {
return enc.Matcher.Match(rw.statusCode, rw.Header()) return enc.Matcher.Match(rw.statusCode, rw.Header())
} }
// Flush implements http.Flusher. It delays the actual Flush of the underlying ResponseWriterWrapper // FlushError is an alternative Flush returning an error. It delays the actual Flush of the underlying
// until headers were written. // ResponseWriterWrapper until headers were written.
func (rw *responseWriter) Flush() { func (rw *responseWriter) FlushError() error {
if !rw.wroteHeader { if !rw.wroteHeader {
// flushing the underlying ResponseWriter will write header and status code, // flushing the underlying ResponseWriter will write header and status code,
// but we need to delay that until we can determine if we must encode and // but we need to delay that until we can determine if we must encode and
// therefore add the Content-Encoding header; this happens in the first call // therefore add the Content-Encoding header; this happens in the first call
// to rw.Write (see bug in #4314) // to rw.Write (see bug in #4314)
return return nil
} }
// also flushes the encoder, if any // also flushes the encoder, if any
// see: https://github.com/jjiang-stripe/caddy-slow-gzip // see: https://github.com/jjiang-stripe/caddy-slow-gzip
@ -221,7 +221,7 @@ func (rw *responseWriter) Flush() {
} }
} }
//nolint:bodyclose //nolint:bodyclose
http.NewResponseController(rw.ResponseWriter).Flush() return http.NewResponseController(rw.ResponseWriter).Flush()
} }
// Hijack implements http.Hijacker. It will flush status code if set. We don't track actual hijacked // Hijack implements http.Hijacker. It will flush status code if set. We don't track actual hijacked