When implementing neo-hypervm, calling assemblies will need to be able to attach an event handler(s) to receive instrumentation data created within the neo-hypervm. I think we may want two separate events raised that we could attach separate event handlers to depending on the needs of the calling / referencing assembly.

The two events I could see we need initially:

• Logging events with various classes of event (info, warning, error, etc)
• Performance instrumentation (timing metrics, etc)

We need to integrate libressl for the CHECKSIG and CHECKMULTISIG op commands

https://github.com/libressl-portable

SHA1, SHA256 and RIPEMD160 could be migrated too

Shargon, I think linux static library must be renamed for easier use: "linux-libcrypto-x64.a" => "liblinux-libcrypto-x64.a". This "lib" prefix is very important in linux, it helps path resolution in compiling with "-L".
Congratulations, project is getting great! I'll take a deeper look in the next days/weeks/months :)

We need to be able to compile and execute in ARM processors

Problem

A VM implementation needs to push and pop a lot of data. Using a stretchy buffer (which the std::list is) will need to grow a lot in a single program execution, an intensive operation.

Solution

A double linked list will significantly improve performance. We could use a single linked list, but some DROP instructions will remove list items that are not on the top of the stack. I suggest using a double linked list implementation.

As doing lots of c++ projects in the past I suggest renaming some folders to match a more "standard" way of structuring opensource c++ projects. This is what I see the most in large opensource c++ project.

|
- tests
- bench
- vendor
- interop
- src