I'm blocked, help me!

[Walter Bright]

On 1/9/2019 12:45 AM, Brad Roberts wrote:
> Much like for the os, it's not reasonable to expect the minimum supported c++ 
> compiler to be a c++17.  It _might_ be reasonable to expect c++11/14, but 
> someone would have to survey what's actually there to be sure.  Regardless, 
> conditionalizing the support is a more usable longer term solution.  Exactly how 
> to do that within the existing test suites, I can't answer that since I haven't 
> touched it in a few years and really don't know the state of the art for them.

If the c++17 compiler has switches that revert it to older behaviors, that might 
be reasonable.