Samuel Schetterer

In general, I'm biased towards trie structures for this case since they provide very good performance overall and tend to fare better in many-writer scenarios, whereas hashmaps can have very...

The issue is caused since multiqueue currently stalls the producer when a consumer is behind. I'm looking into an api that allows consumers to remain unmonitored by the producer and...

Can you post the entire code? This is somewhat expected on an oversubscribed system due to the notification method differences that must happen between broadcast and mpmc. The broadcast is...

> One big change was initially I had the queue size at 4096. I read in an LMAX faq that smaller size allows the entire buffer to stay in cache,...

Yeah, your assumption about clone vs add-stream is correct. It's pretty simple to make a best-effort attempt to create a new reader at the current write position, but this will...

I'm not even sure if the solution is correct, it only exists in my head up till this point. If you want to send without crashing, for now you can...

I'm going to make an api change that would support such behavior in a best-effort sense, as I've realized the other way leads to some edge-case race conditions

thanks, pulling this out of hibernation so will look at it. Fixing some bugs I never got around to and that may reveal the cause of this

There is at least one bug with the current implementation, I'm working out a whether this directly competes with my current job before I do any future work though