Daan
Daan
I might have fixed the issue .. can you try this with the latest `dev` or `dev-slice` (v2.x) branch? I created a dylib from: ```C++ #include extern "C" int myentry()...
... I hope the latest release may work for you this time (v1.8.2 or v2.1.2). Can you try? Otherwise we need to look into this further where I can repro...
Yikes! Thanks for the report -- I just pushed a fix.
Hi @beyonddream, very interesting -- but tricky... recently we added the `mi_manage_os_memory` which can be used to assign a large memory arena; So, in the producer, one can - allocate...
Ouch -- that is not good :-(. Can you repro the problem so I can take a look? In 2.1.2 I made simplifications to the code (and v1.8.2) by removing...
Any progress on this? As I understand, the error happens with 2.1.1, 1.8.1 and 1.7.2 ? If I read this, it seems to me that there is a heap block...
You may well be right -- the use of static TLS on Windows is still a bit experimental (but it really improves codegen for `mi_malloc` so it would be nice...
I pushed an update where we always use the second user reserved slot but test at runtime if we indeed can allocate that slot. If we are initialized first (as...
Hi @res2k: ah I was a bit too enthusiastic -- it looks like a great trick, very clever use of `selectany`! However, the pipeline failed and after more testing on...
> in spitting distance of the best solution (fixed TLS slot), but quite a bit more robust (especially if the desired TLS slot is not available). So, I'll think I'll...