5 ms·

Great, now we just have to wait for about 7-8 years or so for all of these features to be widely adopted, and then we're good to go!

dgellow 6y ago
C++20 compiler support is already quite good among the main compilers (GCC, Clang, MSVC): https://en.cppreference.com/w/cpp/compiler_support https://en.cppreference.com/w/cpp/compiler_support
- einpoklum 6y ago
  I'm talking about: 1. Adoption by users, including the larger, sluggish, corporate ones (and I'm being lenient with 8 years). 2. Transition from headers to modules.
  - dgellow 6y ago
    > 2. Transition from headers to modules. Yeah, this one is likely to take a while...
je42 6y ago
nope. I am using C++ 20 features in production already.
superkuh 6y ago
Even worse, all the devs jumping on using the C++20 features right now (like the other commentors in this thread) will make software that's near impossible to compile on a linux distro more than 4 years old. And even if someone compiles on a modern machine with a modern build toolchain there's no certainty that the binaries will run on the aforementioned slightly old distro. C++xx new features and improvements are good. But they come very fast and that futureshock is causing the fever that is containerization "solutions".
- dgellow 6y ago
  I honestly cannot say if that's a GPT-3 comment or not. Are you implying that the 3 years cycle of the C++ committee is the reason why people are using containers? Without being disrespectful, that doesn't seem to make sense.
  - superkuh 6y ago
    Not exclusively c++xx but more like the rapid pace of new features in all of the big libraries (glibc, etc). And it's not the only reason for containers but it's probably the largest. I'm going to bet you haven't tried to use an older distro in the last decade and have kept up to date. Trying to run software from outside your repos that's written with the latest features is a serious problem on even mildly old distros. It wasn't such a serious problem with the slower rate of development 2000-2010. GCC 4 spanned an entire decade. I get that on the big corporate server side there are different reasons for containers; for scaling, deployment, and all that jazz. But on the desktop, or in academia, it's the primary reason. It's futureshock and containers are the fever.
    - einpoklum 6y ago
      Actually, a compiler is one of the less problematic things to keep up-to-date, since compilers (at least C/C++ compilers) typically have few dependencies, and can live just fine "on the side" of the system's main compiler. But some libraries or apps need half the world and its mother to run, including packages with subtle clashes, and then you're in trouble.
notacoward 6y ago
The bigger problem is that you'll have to wait infinity years for people to use these features wisely. Neophiles will sprinkle them everywhere as though they're magic fairy dust. Luddites will avoid them forever. Very few will explore the space in between.

dgellow 6y ago

C++20 compiler support is already quite good among the main compilers (GCC, Clang, MSVC): https://en.cppreference.com/w/cpp/compiler_support https://en.cppreference.com/w/cpp/compiler_support

einpoklum 6y ago
I'm talking about: 1. Adoption by users, including the larger, sluggish, corporate ones (and I'm being lenient with 8 years). 2. Transition from headers to modules.
- dgellow 6y ago
  > 2. Transition from headers to modules. Yeah, this one is likely to take a while...

je42 6y ago

nope. I am using C++ 20 features in production already.

superkuh 6y ago

Even worse, all the devs jumping on using the C++20 features right now (like the other commentors in this thread) will make software that's near impossible to compile on a linux distro more than 4 years old. And even if someone compiles on a modern machine with a modern build toolchain there's no certainty that the binaries will run on the aforementioned slightly old distro. C++xx new features and improvements are good. But they come very fast and that futureshock is causing the fever that is containerization "solutions".

dgellow 6y ago
I honestly cannot say if that's a GPT-3 comment or not. Are you implying that the 3 years cycle of the C++ committee is the reason why people are using containers? Without being disrespectful, that doesn't seem to make sense.
- superkuh 6y ago
  Not exclusively c++xx but more like the rapid pace of new features in all of the big libraries (glibc, etc). And it's not the only reason for containers but it's probably the largest. I'm going to bet you haven't tried to use an older distro in the last decade and have kept up to date. Trying to run software from outside your repos that's written with the latest features is a serious problem on even mildly old distros. It wasn't such a serious problem with the slower rate of development 2000-2010. GCC 4 spanned an entire decade. I get that on the big corporate server side there are different reasons for containers; for scaling, deployment, and all that jazz. But on the desktop, or in academia, it's the primary reason. It's futureshock and containers are the fever.
  - einpoklum 6y ago
    Actually, a compiler is one of the less problematic things to keep up-to-date, since compilers (at least C/C++ compilers) typically have few dependencies, and can live just fine "on the side" of the system's main compiler. But some libraries or apps need half the world and its mother to run, including packages with subtle clashes, and then you're in trouble.

notacoward 6y ago

The bigger problem is that you'll have to wait infinity years for people to use these features wisely. Neophiles will sprinkle them everywhere as though they're magic fairy dust. Luddites will avoid them forever. Very few will explore the space in between.