Daan

Ah, and related to this. It seems that the full statistics are nice for debugging, but here the request is more for monitoring the current memory usage; What would be...

I just tried approach 1 but it does affect the bench marks (just a tiny bit, under 1%). But I realized we can already use this approach with the current...

Secondly, I guess we can expose the now internal `mi_process_info`: ```C void mi_process_info(mi_msecs_t* utime, mi_msecs_t* stime, size_t* peak_rss, size_t* page_faults, size_t* page_reclaim, size_t* peak_commit) { ``` this return OS traced...

I hope the `mi_register_deferred_free` works for you. I have given it more thought but I think it as actually a very nice solution and maybe it should be built in...

I just pushed commit 46ee895 which exposes: ```C void mi_process_info(size_t* user_msecs, size_t* system_msecs, size_t* current_rss, size_t* peak_rss, size_t* current_commit, size_t* peak_commit, size_t* page_faults); ``` This may help getting global statistics...

I would say v3 is production ready as it is already running at large scale on some big services -- I will soon remove the beta tag. However, there are...

Hi @Noxybot : the fix in #1057 is not quite right -- the root cause is that `mi_process_init` is not called (which calls `_mi_page_map_init` which ensures the pagemap is not...

I think I fixed the issue -- let me know if the latest `dev3` fixes the issue.

> No, rev. [02311cb](https://github.com/microsoft/mimalloc/commit/02311cb126c729666295ad66d5fcec5d920c7d15) still crashes. Ah, that is not good. I just tested on freeBSD 14.2, clang 18, [02311cb], debug and release, and it works for me? (both `ctest...

Yikes -- now I'm unable to repro this :-( Maybe you can share a stack trace again? maybe I'll be able to spot what could be the cause.. edit: I...