mwashburn2 commented on PR #2697: URL: https://github.com/apache/activemq/pull/2697#issuecomment-6267536371
> LGTM, thanks @mwashburn2! > > Making `socketBufferSize=0` an explicit "let the OS auto-tune" option is useful, and the change keeps existing behavior for all other values. > > A few minor points (non blocking): > > * `testPositiveSetsBufferWithoutLogging`: asserting `> 0` doesn't really check anything, since an untouched socket already reports positive sizes. Could you assert `>= 128 * 1024` instead? Linux doubles the value, so that bound is still safe there. > * Log level for negative values: I would keep it at WARN rather than ERROR. The setting is ignored and the broker keeps working, and this is logged on every connection, so ERROR could trigger log alerts. > * Missing newline (nit): please add a trailing newline to the end of the test file. > > While reviewing I also found some existing issues in `initialiseSocket`. They are not related to this PR, so we can open separate issues for them: > > * the `socket.*` options are applied to the `socket` field instead of the `sock` parameter. > * the 64KB `socketBufferSize` default overwrites any `socket.receiveBufferSize` / `socket.sendBufferSize` set in the URI. > * buffers are set after `connect()`, which is too late for sizes above 64KB to enlarge the TCP window. Addressed the minor points. Since I'm here I I'll open a PR for the other issues. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] For further information, visit: https://activemq.apache.org/contact
