C++20 Feature Talk Heats Up At Latest C++ ISO Meeting

[reavertm]

So C++ committee still taking low hanging syntax fruits and no modules support yet.

[pal666]

Will it be possible to add separate checking in future revisions of the concepts feature in the C++ standard?

it is not clear whether it will be possible to do it at all(concepts are in development for more than 10 years), but there is still hope

LOL

if you find it easy, feel free to participate

[plonoma]

some people wanted to leave concepts out of standard until they get separate checking support. majority decided it is better to have { concepts without separate checking now } than { no concepts at all for indefinite amount of time }

Will it be possible to add separate checking in future revisions of the concepts feature in the C++ standard?

because it is hard

LOL

[pal666]

Then what's with all the talk about separate checking in the blog posts?

some people wanted to leave concepts out of standard until they get separate checking support. majority decided it is better to have { concepts without separate checking now } than { no concepts at all for indefinite amount of time }

Why didn't they add separate checking if they have time enough?

because it is hard

[plonoma]

you are getting wrong impression. they are merging already implemented and tested TSes. non-ts features are trivial additions

Then what's with all the talk about separate checking in the blog posts?
Why didn't they add separate checking if they have time enough?

[carewolf]

if you read it carefully, you will not find mentioning of "plan of merging cairo into stdlib", you will find plan to base 2d graphics ts on cairo, just as i said

I hope they don't do that. The code they have is not generic enough, it is based on how all 2d rendering has looked the last 10 years, but it is lacking modern features like 16-bit per color, gamma-correction, HDR and color-spaces.

First they will need to make it a lot more generic which will make it hard to base on existing code, because all those features have not even been solved in Skia yet, let alone cairo.

[pal666]

I'm getting the impression the standards committee is trying to cram too much new specifications and material into an insufficiently large time-frame.

you are getting wrong impression. they are merging already implemented and tested TSes. non-ts features are trivial additions

It would be better to put the new standard version in 2022-2023.

it would be another c++0x-kind disaster. committee is smarter than you, they develop features in parallel in tses and then merge ready ones into trunk

(A lot of the new stuff touches the standard libraries in huge, deep and drastic ways.)

most of new stuff is libraries and somehow both java and c# managed to get orders of magnitude larger libraries in shorter timeframe

[pal666]

Maybe they can add that feature later?

maybe. it is not easy endeavor. at least in the meantime we have other concepts features, just nine years after c++11. btw, c++11 became c++11 instead of c++0x due to concepts

[pal666]

There was, at least back in 2013-2014.

https://lists.cairographics.org/arch...er/024858.html

if you read it carefully, you will not find mentioning of "plan of merging cairo into stdlib", you will find plan to base 2d graphics ts on cairo, just as i said

[pal666]

Not sure what you mean by "in good state thanks to stl".

i mean that stl is doing good job maintaining ms c++ standard library. stl is a person ( https://nuwen.net/stl.html )

MSVC is always years behind competitors.

Frankly, I'd imagine that's why they've started adopting clang into VC++ IDE.

clang(and msvc for that matter) is a compiler while i was talking about library