Nick Tindall

Results 22 issues of Nick Tindall

(when using the UberHandler at least) It's called here https://github.com/OpenHFT/Chronicle-Network/blob/5d53d322e4b1713c6e8ce24326038f4e382a0b53/src/main/java/net/openhft/chronicle/network/cluster/handlers/UberHandler.java#L157 and here https://github.com/OpenHFT/Chronicle-Network/blob/5d53d322e4b1713c6e8ce24326038f4e382a0b53/src/main/java/net/openhft/chronicle/network/cluster/handlers/HeartbeatHandler.java#L227-L228

The javadoc for EventHandler#loopFinished states that > Event loop implementations call this once only, from the event loop's execution thread. But there are lots of places we call it from...

This could be a genuine bug. The error is: ``` org.junit.ComparisonFailure: expected: but was: Expected :[2, 1, 3] Actual :[2,1,3] ``` When you attach a debugger to it, the LHS...

flaky

When a pre-toucher is running on a queue and `earlyAcquireNextCycle` is enabled, any Tailer reading from it can skip over messages when it hits the end of a roll cycle...

bug

`EventGroup#start` blocks until the core event loop has started. There have been multiple reports of `performStart` timing out and throwing an exception. This is unexpected and hasn't been reproduced reliably....

bug

This implements a prototype approach for detecting and correcting partial updates that can occur as a result of an abrupt termination Summary If we acquire a write lock and `writeInProgress`...

#1323 was reported. It's possible to write to arbitrary cycles in the past if you set the system clock to point to those cycles. This is an experiment that forces...

We are an append only queue so we shouldn't allow this.

review