Gregory Popovitch

I'm no cmake expert. Would you have a suggestion on how to do this?

Cool, thanks for investigating this @fegennari . Do you have a suggestion to address this issue? If you submit a PR I'll be happy to merge it in.

Oh just read the rest of your comment. Thanks for looking into it further!

Hi @fegennari , please try the new version I just pushed in. You are right that the issue was updating capacity_ before the memory allocation, leaving an inconsistent state if...

Yes, I think you understand the problem and explained it well. I see the issue, but I'm not quite sure how to fix it cleanly. Interestingly the google's [Abseil-cpp](https://github.com/abseil/abseil-cpp/blob/master/absl/container/internal/raw_hash_set.h) hash...

Sorry, I reverted my last change which did not seem to work, even though it did when I tested locally. I'll think it over.

Yes, I did think of doing that, and it is definitely a valid solution, but phmap doesn't throw any exception right now, and I think some users rely on this...

Your use case is perfectly valid, and I do need to fix this!

This is a restriction I inherited from the original Abseil source code. Can you tell me what kind of pointer type you would like to use?

Thanks! Maybe it is doable to support this. I'm wondering how I could test it. Would you have a small example with a boost unordered_map/set using a custom pointer type...