WIP

Primary usage of this repo is as a submodule to my MACHINE_WIDE repo.

DBJ CPP TOP

In IT systems in general, the programming language is not of primary importance. Architecture is.

IT system (not your desktop experiments) is made of components. Each component has a well-defined interface and set of operational requirements. E.g. size, speed, security, and such. If your component does meet operational requirements I really do not care if you have written it in Quick Basic, Julia, Rust, Elm, Awk, Bash Script or C++.

Just after that, it will be rigorously tested for meeting the functional requirements.

Also please do not (ever) forget the economy of IT systems: how much money will it take to develop it and more importantly how much money will it take to maintain what you did, and for the next 10+ years. -- DBJ 2020 Dec.

This is the collection of my top level C++ headers. Meaning they are or should be used in all of my C++ code. But why do I develop C++ code as I do?

The Reality and consequences

Operational requirements are economy and market driven.

Run Time

• Windows
  • no point in avoiding Win32
  • Win32 in (undefined, C89 like) C API
  • SEH is a reality and intrinsic to Windows
    • patented M$FT technology
• Legacy is a big problem
  • UTF-16

Compile Time

What compiler? clang-cl.exe And why:

• CL
  • heavy burden of legacy
• clang-cl
  • much better untroubled and legacy unburdened compiler
• MS STL
  • works rather well with c++ exceptions switched off
  • That is a game changer
  • Makes me free to avoid the technical debt ammased in the std lib
    • iostreams
    • exceptions
    • RTTI
    • new/delete
• C11/17
  • officaly are in cl.exe but one can not mix C/C++ in cpp files using cl.exe
  • clang-cl has no issues, it "just works"
    • clang uses some C++ machinery available in compiler while the code is C.
      • same as in GCC that is done throught __attribute__((att name here))
    • I can and will rely on modern C costructs and features
    • like for example VLA typedefs

    typedef char (*arr_of_3_char_ptr)[3] ;
    
    arr_of_3_char_ptr a3p = 
        (arr_of_3_char_ptr)
           malloc( sizeof arr_of_3_char_ptr);

    • or arrays as arguments

    void use_buff ( int size, char buff[size]);

Do not develop it if it exist

• generaly yes but
  • that requires time for search and testing
  • sometimes so much time it is more feasible to develop it in house

Note on Orthodox C++

While I agree in principle I do not agree if I smell the dogmatic approach.