C++ 11 Adoption

[LeStarch]

Community: is there any concern if we were to update to C++11 standards? This would require compiler support for C++11. We'd like to know if anyone out there would be unable to work with this change.

Thanks.

[astroesteban]

Why not skip C++11 and jump to C++14? C++14 fixes several holes in C++11. It's basically a hotfix for 11.

Language Features

• C++11’s version of constexpr implies const for member functions, this is changed in C++14
• C++11 is missing auto return type deduction for regular functions (lambdas have it)
• C++11 does not have auto or variadic lambda parameters
• C++14 adds [[deprecated]] attribute
• C++14 adds ' digit separator, example: 1'000'000
• constexpr functions can be more than just a single return statement in C++14

Library Features

• std::make_unique was added in C++14, which enables the “no raw new” standard
• C++11 doesn’t have std::exchange
• C++14 adds some constexpr support for std::array cbegin, cend, crbegin and crend free functions added for consistency with begin and end free functions and member functions added to standard containers in C++11.”

[Joshua-Anderson]

Since F' has to support a wide variety of embedded systems toolchains, we try to be pretty conservative on what compiler features we require for the core framework. My fear would be excluding compilers with C++11 but not C++14 support. Going from C++03 to C++11 transitions wins us a lot of nice features and allows us to modernize the codebase. The benefits of C++14 are nice but don't seem valuable enough to drop support for older compilers.

[astroesteban]

Good points. Just going to C++11 itself gives the project huge benefits.

[SterlingPeet]

Last I checked C++ 14 didn’t compile correctly with GCC for my platform (ATmega AVRs), but C++11 does. I’m not actively developing for that right now, but I might have a better feel for it in about 2 weeks. Sterling Sent from my mobile. On Jun 26, 2021, at 11:02 AM, Esteban Duran ***@***.***> wrote:  Why not skip C++11 and jump to C++14? C++14 fixes several holes in C++11. It's basically a hotfix for 11. Language Features * C++11’s version of constexpr implies const for member functions, this is changed in C++14 * C++11 is missing auto return type deduction for regular functions (lambdas have it) * C++11 does not have auto or variadic lambda parameters * C++14 adds [[deprecated]] attribute * C++14 adds ' digit separator, example: 1'000'000 * constexpr functions can be more than just a single return statement in C++14 Library Features * std::make_unique was added in C++14, which enables the “no raw new” standard * C++11 doesn’t have std::exchange * C++14 adds some constexpr support for std::array cbegin, cend, crbegin and crend free functions added for consistency with begin and end free functions and member functions added to standard containers in C++11.” — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#777 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAO7KVQ5OZ6SJKZ4EHL3M53TUXTZRANCNFSM47KHXUDQ>.

[astroesteban]

Interesting. I wonder why…

[Joshua-Anderson]

Thanks Sterling, do you happen to know which GCC version you're using for the ATMega?

[SterlingPeet]

Not sure, but I will definitely check on this the next time I’m working on a relevant project. Should be within a week or two. Sterling Sent from my mobile. On Jun 26, 2021, at 2:28 PM, Esteban Duran ***@***.***> wrote:  Interesting. I wonder why… — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#777 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAO7KVUFV55KFNZH4IEFQILTUYL5ZANCNFSM47KHXUDQ>.

[LeStarch]

It sounds like there are no active objections. I am going to mark this as answered.