wsflate: handle short compressor EOF blocks

[sb123sb123]

Fixes #221

Problem

wsflate.Writer assumes that a compressor's Close leaves the four-byte stored-block tail (00 00 ff ff) in its buffer. Go 1.27's compress/flate emits a short fixed-Huffman EOF block instead, so closing after the required PMCE flush can move buffered bytes to the destination and return a false wsflate: bad compressor error.

Cause

The PMCE sync-flush marker and the compressor's final EOF block are different protocol concerns, but Close checked and forwarded them through the same four-byte tail buffer.

Fix

Track whether the last successful operation was Flush. When closing a flushed writer, still call the compressor's Close method, but discard only its close-time output so the PMCE message remains free of the final block. A subsequent Write clears the state and requires a new flush.

The regression test models the short fixed-Huffman EOF block and verifies that the compressor is closed, no error is reported, and the flushed payload is unchanged.

Tests

• go test -count=1 -run '^TestWriterCloseAfterFlushWithShortFinalBlock$' ./wsflate (passed with task-local dependency replacements)
• go test ./... (passed with task-local dependency replacements)
• go vet ./wsflate (passed)
• gofmt -d wsflate/writer.go wsflate/writer_test.go (clean)
• git diff --check (clean)

Limitations

The remote Windows host runs Go 1.26.8, so it could not execute the Go 1.27 compressor directly. The Go module proxy and direct module fetches were unavailable; I used temporary local copies of the two small dependencies for testing and removed them before committing. Full-module vet also reports two pre-existing findings outside this change: wsutil/writer_test.go:135 (reflect.SliceHeader) and example/autobahn/autobahn.go:58 (discarded context.WithTimeout cancel function). The race test was not run because the host has no gcc C compiler.

AI assistance disclosure

This PR was prepared with assistance from OpenAI Codex. Codex performed issue triage, implementation, regression-test authoring, and validation; the final patch and reported results were checked against the repository and command output.

[cristaloleg]

w.flushed = w.err == nil ?

[sb123sb123]

Thanks for the review and approval. The Go CI matrix for this head is green. The separate Autobahn workflow fails while building its test image, before the Autobahn suite starts: installing autobahntestsuite==0.8.2 fails in the Docker build because pycparser is missing from the cffi build requirements; the later docker run ws-autobahn then cannot find the image. The PR changes only wsflate code and tests, so this appears isolated to the test image setup. Would you prefer a separate CI fix, or is there a recommended way to handle this check for this PR?

Failing run: https://github.com/gobwas/ws/actions/runs/36542593276