On Tue, 11 Feb 2025 10:51:52 GMT, Jaikiran Pai <j...@openjdk.org> wrote:
>> Fixes stream cancellation in `Http2Connection::putStream`. See comments in >> the ticket for details on the issue. `tier1,2` results are attached to the >> ticket too. > > Hello Volkan, so from what I understand, before this change, we used to raise > a generic `IOException` if a new request couldn't be initiated because the > `Http2Connection` on which it is being raised already had some failure. With > this new change, we propagate any original exception that was registered on > the `Http2Connection` thus allowing any special handling of that original > exception type before it gets propagated to the application code. Did I > understand this correctly? > > The linked JBS issue has a `noreg-self`. I think it should instead be > `noreg-hard`, since `noreg-self` is used when we change a regression test > itself, which isn't the case here. @jaikiran, thanks so much for the review, much appreciated! 🙇 As discussed with you and @dfuch, - Changed the ticket labels to `httpclient` and `intermittent` - Added the JBS issue ID to the `@bug` list of `AsyncExecutorShutdown` - Wrapped the `cause` in an `IOException` in `putStream()` ------------- PR Comment: https://git.openjdk.org/jdk/pull/23540#issuecomment-2650571807
