C# and Dotnet 2.1, the CTO and another senior dev have years of pro experience with it. I am learning the repository method and such, pretty neat stuff. It compiles to native (which I want to play around with in a lambda setting). The stack is dotnetcore 2.1, c#, k8s, etc., on AWS. Pretty boring stack I guess.
I like C#. I spent a year-ish writing a lot of it and it felt like we produced easy to understand and maintain code. I think the MS/Linux divide is probably the biggest reason more people don’t use it.
I’m not the person you’re asking, but C++ has been the language easiest for me to manage larger projects in. Additionally, the focus on duck typing, generics, and containers makes it a natural language to port to from Python.
I like writing C++ just fine, but managing dependencies drove me crazy. Am I missing something there? When I want to do something simple like import an HTTP lib and make some requests it always feels like a hassle and half the time I get some weird linker error.
I effectively use github as a package manager. I place all dependencies as submodules and handle building the ones which need compilation in my Makefile. Most of the time it works, but boost libraries which aren’t header-only are an absolute nightmare, so I avoid those.
Okay I’m glad to hear that I’m not the only one who found complicated boost libraries to be a total mess. I spent a while banging on managing those. Header-only libraries tended to just work for me.
I actually used Boost for a project, but I only needed header files for the subset I was using. I made each of these a submodule as in [0] and was able to benefit from Boost without all the pains of linking.
So before deciding you can’t use it, see if there’s a way to make it header-only.
