What the–
It’s quite interesting that, after adding all percentages, the total goes up to 144%. Clearly some fancy math is in play.
I regularly use C++17, C++20 and C++23 with different projects. I am counted in three categories.
If you’re required to use different environments to produce software for multiple platforms, those different environments may come with different compilers that support different C++ standards. And then you’re regularly using more than one C++ standard.
@Redkey @rmam I assume you’re aware of the fact that there exists *only one* C++ standard at a time (C++20 today, C++23 later this year).
You probably mean
* using reduced feature sets
* compiling with non-conforming compilersBut this might sound less nice than “using an obsolete standard xyz”. I’m totally aware why large swaths of the industry are stuck in the past for well-motivated reasons.
there exists only one C++ standard at a time
What? No. That’s not how it works, at all. When a new version of the international standard is published, that does not mean previous versions cease to exist. It just means there’s a new version.
@rmam Older versions of the C++ standard are *withdrawn*.
@rmam Older versions of the C++ standard are withdrawn.
This just goes to show the degree of confusion you’re dealing with. You’re confusing ISO’s systematic review process with the real world. ISO’s withdrawal process is used as a janitorial process regarding documents than require updates and/or maintenance. Just because no one will update C+11 that does not mean no one writes code in C++11 or compiler writers pulled it’s support. ISO’s review process matters nothing.
Are C++ modules still a meme, or do they actually exist in a portable way and solve more problems than they create?
Only properly supported by MSVC currently. I’d expect them to have more widespread support somewhere around C++26 release time.
As soon as CMake support is released it should start spreading more and more. But so far, I think we’ll still need about 3 more years
