Hi, I'm Tomasz

There’s one repeating pattern in all C++ code bases I’ve worked with. shared_ptr is abused - one way or another. There are many reasons, sometimes people just start with shared_ptr instead of unique_ptr out of laziness, sometimes it’s the sole, default smart pointer they rely on. Often, it’s a result of many passes of refactoring and eventual quality degradation with time. This leads to all sorts of problems but there’s are definitely some repeating patterns.

Welcome to the third and concluding part of the series about dependency and package management in C++ projects in which I’m gonna mostly focus on solutions built into build systems themselves. If you haven’t already, I encourage you to have a quick read of the first and the second part.

Overview

In this part I’ll have a closer look on meson wraps, cmake’s FetchContent, bazel’s central registry and, a bit unconventionally, conda.

Recently, I’ve learned about constexpr virtual functions. The support for which was added along with C++20. C++23 additionally allows for constexpr new (with some restrictions). There’s a great article about constexpr virtual on C++ stories blog, followed by another great post about constexpr new. I highly recommend reading Bartek’s posts.

Wanting to add this new feature to my tool belt I was wondering what would be the best application for constexpr virtual and how would I use them in my code.

Introduction of concepts in c++20 brought along a new set of problems requiring solving related to wrapper types (like e.g. std::optional, std::variant).

Specifically, special functions (like constructors, copy constructors, destructors etc) need to have the same type of traits as the type they are wrapping i.e. these have to be copyable if the underlying type is copyable, trivially copyable if the underlying types are trivially copyable, and trivially destructible if the underlying types are trivially destructible.

Sure, = default is great and convenient especially for destructors and other special member functions but when declaring a class that’s meant to be a part of a shared library’s interface I try to avoid using it and here’s my reasons why.

TLDR

Problem definition

Let’s start with a simple abstract class declared in a header file (calc.hpp):

Welcome to the second part of the series about dependency and package management in C++ projects. This time I’m gonna focus on somewhat less popular solutions than the main three.

Overview

Let’s evaluate if any of these are worth the hassle.

Hunter

Hunter caters towards CMake. The way it works is really similar to meson’s wraps (with the only difference that it’s not built into CMake). Usage is fairly simple. You have to export HUNTER_ROOT to a location of choice. Hunter will use it to store its build directories and obtained source code. There’s nothing that has to be installed. It’s kind of “self bootstrapping”. You have to download HunterGate within your CMake project:

This post is a quick reminder to self regarding specifics of RPATH, RUNPATH, LD_LIBRARY_PATH, LD_RUN_PATH and the lookup order.

Refresher (a very short one)

RPATH, RUNPATH - are entries in the ELF header, baked into the binary allowing the dynamic loader to lookup its shared dependencies.

RPATH/RUNPATH can be specified directly, using a linker option (-rpath):

g++ -Wl,-rpath=path/for/rpath

If missing in the command line, RPATH can be set using LD_RUN_PATH variable.