On Tue, 10 Jun 2025 09:54:47 GMT, Daniel Fuchs <dfu...@openjdk.org> wrote:
>> The java/net/httpclient/DigestEchoClient.java fails intemittently with >> IOException: HTTP/1.1 header parser received no bytes. >> >> Analysis shows that this is caused by the CleanupTrigger which receives data >> after the reused connection has been taken out of the HTTP/1.1 clear pool >> (Caused by: java.io.IOException: Data received while in pool). This should >> not happen. >> >> The fix for [JDK-8338569](https://bugs.openjdk.org/browse/JDK-8338569) has >> improved the situation but apparently didn't fix the issue completely. >> The issue is that the write publisher manages to come active before the read >> subscriber has been switched, which opens a window in which the old read >> subscriber receives the data that should have been passed to the new one. >> >> To fix that, this change proposes to pause reading at the selector level >> before connecting the flows with the new publisher /subscriber pair. This >> should ensure that data doesn't reach the "old" subscriber, if the new write >> publisher manage to send data before the new read subscriber is subscribed. > > Daniel Fuchs has updated the pull request incrementally with one additional > commit since the last revision: > > Update src/java.net.http/share/classes/jdk/internal/net/http/SocketTube.java > > Co-authored-by: Daniel Jelinski <djelins...@gmail.com> Marked as reviewed by djelinski (Reviewer). ------------- PR Review: https://git.openjdk.org/jdk/pull/25416#pullrequestreview-2912955239
