choksi2212 commented on PR #4497: URL: https://github.com/apache/iggy/pull/4497#issuecomment-6173993359
Rebased onto master @ e9e5e10 and force-updated — thanks for the heads-up, the multi-leader rework touched exactly the helper territory this PR had refactored. How the conflicts resolved: - **Every transport test stays on its (new) multi-leader helpers.** The earlier sweep that pointed the tcp test files at `fakevsr` is gone — `tcp_testing_test.go` and friends are byte-identical to master again, so the multi-leader assertions (new client per failover, reset request counters, roster reads) are untouched. - **The shared-certificate ask from your review survives**: `fakevsr.SelfSignedCert` is now the single implementation, and the transport suite reaches it through the local `selfSignedCert` wrapper in `tcp_connect_test.go`. - **`fakevsr` shrank to what the connection-string suite needs** (fake server + reply builders + the cert helper), and its roster reply gained the `ReplicaID` the multi-leader client expects — everything else was pruned rather than left half-dead. Verification on the new head: - `go build ./...` clean, `go vet ./client/... ./internal/fakevsr/...` clean, gofmt clean on all five changed files. - `go test ./client/... -race -count=1`: **both `client` and `client/tcp` pass** — all 14 `TestParseConnectionString_*` cases against the multi-leader client (connect-and-login against the fake server still observes register → cluster-metadata, so the login sequence the issue cares about is unchanged), plus the full transport suite. - The `internal/vsr` protocol-parity tests fail in my local checkout, but identically on pristine master (they parse the Rust sources and my Windows checkout's line endings break the regexes) — untouched by this diff either way. The five review points are all still in: `nodelay` defaults to on, sub-microsecond intervals rejected with boundary tests, spaces/`µs` accepted like the node SDK, the carry-block simplification, and the shared cert helper. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
