On 12/6/24 17:01, Peter Dimov via Boost wrote:
Claudio DeSouza wrote:
As I mentioned, we briefly considered requiring C++14 so that the latter isn't an issue, but the feedback on Slack was negative. People love C++11 for some reason or other. Either way, this still wouldn't have solved the GCC issue.
But is Slack representative of a broader C++ audience. It depends which channel you asked at, because if you asked at #boost, you will have a biased response in favour of C++11.
Well, we can ask here.
Do people care about C++11 support in Hash2? Would it be acceptable to impose a minimum requirement of C++14?
There has to be a cost/benefit analysis. That is, if C++14 is the baseline, will that significantly affect the library design and implementation? And what is the estimate of C++11-only user base that will be cut off? On the first question, library authors are probably the ones who can comment. The second one is more difficult to answer, but I think someone posted here survey results a while ago. Stock compiler versions in the major OSes is another indication. Personally, I'm fine with anything up to and including C++17.