On Mon, 6 Jan 2025 05:43:31 GMT, Jaikiran Pai <j...@openjdk.org> wrote:
> Can I please get a review of this change which proposes to address the > test-only issue noted in https://bugs.openjdk.org/browse/JDK-8347000? > > As noted in that issue, the test issues HTTP requests with `Content-Length` > set to `0` implying no request body. However, the code unintentionally sends > additional bytes (14 bytes each request) as part of the request. The JDK's > `HttpServer` implementation currently doesn't cause errors for such requests > (although the HTTP/1.1 RFC expects an error to be raised). > > The change in this PR merely addresses the test code to not send these > additional bytes. The test continues to pass after this change. Change looks fine. We probably should have a test of the server that checks it handles pipelined requests correctly (if not optimally). Because a thread is basically assigned to a connection for the duration of a request, it means that pipelined requests will be (or should be in this impl) handled in sequence. Like I said, this isn't optimal, but then again, HTTP/1.1 pipelining has limitations that have been ameliorated by the newer versions of the protocol h2 and h3 (QUIC). ------------- Marked as reviewed by michaelm (Reviewer). PR Review: https://git.openjdk.org/jdk/pull/22921#pullrequestreview-2534037822
