Thibaut Goetghebuer-Planchon

Effectively I went a bit too fast with the proposed solution, it will not work in multiple cases. I'll check this evening to see how we can do that.

Hi, I checked a bit and currently I see the following possible solutions: * Keeping the same behaviour as `std::unordered_map`. * Don't wrap around, we just add a buffer at...

Adding a counter would also be possible but the problem is that you also need to initialize it in other methods than `begin()`. For example on insertion you need to...

Hi, I have plan to update my benchmarks as the hash table landscape has evolved quite a bit in these last few years with Google releasing a [new hash map](https://abseil.io/docs/cpp/guides/container#hash-tables)...

Hi, Effectively, due to the nature of the data structure most time will be spend on allocations on insertion. Currently there is no support for a custom allocator due to...

Hi, Thank you for the suggestion. It's effectively an undefined behaviour. As specified in the documentation, any operation modifying the hash map may invalidate the iterators and references to the...

Thank you, I checked a bit further the issue and it seems to be effectively possible without too much troubles. I'll be a bit busy the coming week but I'll...

Hello, Working a bit on the issue I realized it's not really possible for the following code to work with an open-adressing implementation: ```c++ int main () { tsl::robin_map v;...

The main reason to use `std::deque` as default is that it allows faster rehashes as we don't have to move all the values when growing the container. Is there any...

Thank you for the report. From a first glance I have troubles to understand how `m_load_threshold` ends up with such a large value. The only place where the variable is...