Skip navigation

Netty 4.1.85.Final released

by chrisvest
on

We are happy to announce the release of netty 4.1.85.Final. This release is a bug-fix release.

The most important changes are:

  • A bug in FlowControlHandler that broke auto-read has been fixed (#12467)
  • The HTTP/2 HPACK encoder is now faster at encoding headers that have many values (#12665)
  • A potential memory leak bug has been fixed in the pooled allocator (#12897)
  • Fix an issue with the Blockhound integration, which could cause the MacOSDnsServerAddressStreamProvider to be flagged as making blocking calls (#12913)
  • Inconsitencies in how epoll, kqueue, and NIO handle RDHUP have been fixed (#12930)
  • ByteToMessageDecoder now handle situations where the same ByteBuf instance is read multiple times (#12922)
  • The check that ensures the HTTP/1 Content-Length header is unique, now no longer causes headers to be rearranged (change their order) (#12921)
  • Fix a NullPointerException bug with class initialisation order between InternalLogger and InternalThreadLocalMap (#12942)
  • When the netty-resolver-dns-native-macos classes can't load their native bindings, they now only print a short error message instead of the huge stack trace it printed previously. The stack trace is still included if DEBUG logging is enabled (#12950)
  • The Graal native-image meta-data is now placed in the recommended location, and no longer causes warnings to be printed (#12951)
  • The HTTP/1 and HTTP/2 codecs now properly support RFC 8297 Early Hints (#12918)
  • Subclasses of FastThreadLocalThread can now tell the Netty Blockhound integration that they should be allowed to make blocking calls (#12978)
  • Validation of HTTP/2 connection headers have been moved from Http2Headers to HpackDecoder, so that outgoing headers are not validated (#12975)

Thank You

Every idea and bug-report counts, and so we thought it is worth mentioning those who helped in this area.

Please report an unintended omission.