Great topic, talk could use some work IMO. I would have liked a bit more structure to the presentation. Thanks for the work on cppfront! Looking forward to seeing what you achieve!
It's called dynamic for a reason. I think the implementation consists of reading vtables so unless those can be constructed at compile-time, it's probably not possible.
Guys how about you stop? Why don’t you fix what there is as of now in modern c++ before you move on? The modules in c++ do NOT work! When I use them and use string class in a node which is being exported I get ICE. I reverted back to using headers. This is such a nonsense, and this is just one example of broken feature. I am already considering using other language for completely new independent project. I’m disappointed by this crap.
I don’t think the priority of most contributors to the C++ ecosystem is to get everything working before exploring further advancements. The ecosystems of some other languages do prioritize such things. It’s not obvious to me why C++ should match those priorities, especially if it’s at the expense of advancing things like CppFront. But I’m confident that’s a false dichotomy anyway. Is CppFront development really coming at the expense of fixing compiler bugs or modules development or so many other things worth doing? I doubt it.
?? What are you talking about?? Do you think this guy is the CEO of C++ and this talk is him getting distracted by a shiny new idea instead of working on someone else's compiler? Modules don't work right now because they completely upend the foundation of the C++ build system with something entirely different, and compiler devs need time to implement that. What do you want this guy to do about it?
C++ is not some homogeneous entity with some single group developing all compilers. If modules don't work for you look at which compiler you're using and look at their official progress. MSVC supposedly is supposed to fully work by now but still has a lot if bugs that get reported and fixed. Clang and GCC both have people who are actively working on the parts of modules they haven't completely implemented yet, but since it's a huge feature that takes a lot of effort to implement and both compilers are open source projects it takes a while to get it fully functioning. Cppfront in itself doesn't have anything to do with any compiler and is completely separate. It is mostly just a group of people (originally mainly just herb sutter) working on a transpiler from a new language to cpp that implements some of the features they would love to see in future versions of C++. Any developments on CPPfront do not take any time away from developments being done on GCC, LLVM and MSVC so you complaining about "guys" "moving on" does not really make any sense here
Great Talk!
I'm wondering if there's a place or forum to discuss cppfront. I had trouble using it with ranges and specifically piping range views.
Github project. It has issues tab. Also stackoverflow.
there's a github discussions page on the cppfront repo
Great topic, talk could use some work IMO. I would have liked a bit more structure to the presentation. Thanks for the work on cppfront!
Looking forward to seeing what you achieve!
constexpr version of dynamic_cast when?
Now, actually! It is part of the C++20 standard!
It's called dynamic for a reason. I think the implementation consists of reading vtables so unless those can be constructed at compile-time, it's probably not possible.
it's static_cast ?
What C programmers think of when they hear the word "template":
Guys how about you stop? Why don’t you fix what there is as of now in modern c++ before you move on? The modules in c++ do NOT work! When I use them and use string class in a node which is being exported I get ICE. I reverted back to using headers. This is such a nonsense, and this is just one example of broken feature. I am already considering using other language for completely new independent project. I’m disappointed by this crap.
I don’t think the priority of most contributors to the C++ ecosystem is to get everything working before exploring further advancements.
The ecosystems of some other languages do prioritize such things. It’s not obvious to me why C++ should match those priorities, especially if it’s at the expense of advancing things like CppFront. But I’m confident that’s a false dichotomy anyway. Is CppFront development really coming at the expense of fixing compiler bugs or modules development or so many other things worth doing? I doubt it.
?? What are you talking about?? Do you think this guy is the CEO of C++ and this talk is him getting distracted by a shiny new idea instead of working on someone else's compiler? Modules don't work right now because they completely upend the foundation of the C++ build system with something entirely different, and compiler devs need time to implement that. What do you want this guy to do about it?
everybody knows they don't work, go complain to you compiler vendor
C++ is not some homogeneous entity with some single group developing all compilers.
If modules don't work for you look at which compiler you're using and look at their official progress. MSVC supposedly is supposed to fully work by now but still has a lot if bugs that get reported and fixed.
Clang and GCC both have people who are actively working on the parts of modules they haven't completely implemented yet, but since it's a huge feature that takes a lot of effort to implement and both compilers are open source projects it takes a while to get it fully functioning.
Cppfront in itself doesn't have anything to do with any compiler and is completely separate. It is mostly just a group of people (originally mainly just herb sutter) working on a transpiler from a new language to cpp that implements some of the features they would love to see in future versions of C++.
Any developments on CPPfront do not take any time away from developments being done on GCC, LLVM and MSVC so you complaining about "guys" "moving on" does not really make any sense here
What led you to the conclusion that this guy is in any way responsible for getting modules to work?