0.1.5
0.1.5
- Added memflow::prelude::v1 and memflow_win32::prelude::v1 modules
- Added new fields to FFI
- Improved consistency of these function names in C FFI:
phys_read_raw
->phys_read_raw_into
,page_size
->arch_page_size
. - Added C++ bindings for the FFI
- Fixed core errors not displaying the full error message when wrapped in a win32 error
- Changed windows inventory search path from [user]/.local/lib/memflow to [user]/Documents/memflow
- Added {PWD} to inventory search path
Transitioning from C FFI to C++ FFI:
memflow.h
, andmemflow_win32.h
becomememflow_cpp.h
, andmemflow_win32_cpp.h
.- The headers still depend on
memflow.h
, andmemflow_win32.h
. They are just wrappers for safety, and ergonomics.
- The headers still depend on
- Types transition from
Type *
toCType
. EveryCType
include automatic object destruction, so there is no need for thetype_free
methods. CType
contains aType *
inside. The pointer can still benull
. Checking whether object is valid is still the same:if (CType != NULL)
- Methods are implemented as class members. Most methods loose their prefix. The change looks like this:
process_module_info(Win32Process *process, const char *name)
becomesCWin32Process::module_info(this, const char *name)
.- Calling methods changes into calling a function on the object, instead of with the object. Example:
process_module_info(proc, "ntdll.dll")
becomesproc.module_info("ntdll.dll")
. - Exception to this are
virt
, andphys
read/write functions. They do not loose their prefix, because they do have the prefix in the Rust library. So,virt_read_u64(mem, addr)
becomesmem.virt_read_u64(addr)
.
- Calling methods changes into calling a function on the object, instead of with the object. Example:
- There are extra convenience functions that utilize STL's
string
, andvector
containers. Getting process/module names, and lists becomes much simpler.