Why not looking on support of rust ?

[svalat]

This might be a simple patch just wrapper the rust jemalloc in addition to the standard allocator. Could be an issue if they wrap it with hash in function name, but if not the patch should be trivial.

[lu-zero]

I can tell you that the current rust works fine with the normal malt (I already found and fixed issues in rav1e using it).

We can do maybe better by leveraging the #[global_allocator] here the details and probably I can help :)

[svalat]

Hi, thanks very much for the feedback, I still didn't had the time to check, good to know.

For the #[global_allocator] that's exactly the kind if improvement I was thinking about. That should be a really simple patch do be done directly in the GUI. I will give a look to give you pointers if you want, I might not have time really soon to do this myself.

In principle doing it here should be enougth if you have the exact name you want to skip (or at least arround here): https://github.com/memtt/malt/blob/master/src/webview/server-files/MaltProject.js#L915

[lu-zero]

I'll send a patch to skip the allocator signatures hopefully tomorrow.

I was thinking about instrumenting rust software by directly overriding the global allocator at compile time.

[svalat]

No need, I have the same issue with new/delete in C++ and when apps implement their own operators and that's far easier to handle at this location than sooner. Also we keep the info if we need.