Re:Trying out C++20's modules with Clang and Make
by david2ndaccount on Sunday October 15, 2023 @04:09PM
(#37892970)
I do not expect standardization of features that do not exist.
Re:Trying out C++20's modules with Clang and Make
by rewmie on Sunday October 15, 2023 @04:32PM
(#37893150)
> I do not expect standardization of features that do not exist.
But that makes absolutely no sense, doesn't it? I mean, think about it for a moment. Specifying a standard behavior is not about picking a flawed winner from a list of ad-hoc implementation or settle with the least common denominator. Specifying a standard behavior is about providing the best possible solution that meets the design requirements following the input from the whole community.
Re:Trying out C++20's modules with Clang and Make
by kazinator on Monday October 16, 2023 @02:24AM
(#37896445)
> How do you expect to standardize a feature that does not exist yet but the whole community is demanding?
Implement it in a popular compiler or library.
I know we are talking about C++, but the ISO C committee, over the past several decades, has repeatedly ignored working, good features in GCC and invented crap in their place.
Re:Trying out C++20's modules with Clang and Make
by maccard on Monday October 16, 2023 @04:52AM
(#37897232)
This is part of the modules mess in the first place. Modules were implemented in MSVC and clang, and neither agreed with the others implementation.
Also if MSVC came along with an "extension" to c++, I can only imagine the uproar suggesting that this is now step 2 of EEE.
How do you expect to standardize a feature that does not exist yet but the whole community is demanding?
It sounds like the C++ standardization committee designed a feature following a process where all stakeholders had a say. Is this worse than being force-fed a vendor-specific solution?