Manuel Pöter

I suspect that the main issue with the original spin lock was that it did not implement a proper backoff mechanism. In most cases it is not worth it to...

You might want to take a look at my Master's Thesis [Effective Memory Reclamation for Lock-Free Data Structures in C++](http://www.ub.tuwien.ac.at/dipl/VL/51367.pdf). It describes a large number of different reclamation schemes, and...

Your `PushNotification` seems to contain a pretty large string encoded json value. Do you really want to index that field? This does not seem sensible to me. What do you...

Got it, so this is more of a worst-case scenario test instead of an actual use case. From the code it looks like we create commit WriteBatches with 1024 documents...

Well, as a developer, if I send a request to update a document in the DB, I expect that update to be applied. If I want to update a *specific...

@jsteemann Thanks for the explanation! I wasn't aware that a single update requires multiple writes (but it makes sense now that I think of it, since you have to update...

No, because many "conflicting" operations can be combined without destroying the effects of the other operations. @dothebart already gave an example - two patch operations that change different attributes of...

Obviously I (as the original author) have a very strong opinion regarding this issue. I have opened this ticket when I was just a normal user, but in the meantime...

Which language standard do you use? Just below the line you indicated there is a comment: https://github.com/mpoeter/xenium/blob/e1f4c667f13dfe259e117a382b7451f2a9895fe4/xenium/reclamation/impl/stamp_it.hpp#L513-L520 Apparently gcc performs this check during compile time instead of runtime. Either way,...

I would like to take a look at this, but so far I was not able to reproduce it. I am running `cache_bench -use_clock_cache` in a loop, but so far...