xtofl

I suggest defining it as ```C++ T& reinterpret_as(gsl::span s) noexcept; ``` The proposed `T* reinterpret_as(gsl::span s, OffsetType offset) noexcept` does two things instead of one: find a portion of the...

@MikeGitb are you referring to `subspan`? In general, it's wise to reverse the membership question: why would you make it a member? It's hard enough to make a class correct;...

I can't wait for https://en.wikipedia.org/wiki/Uniform_Function_Call_Syntax.

It seems that pycycle does not find package content: ```shell (venv) /tmp/tmp.pSbwX4HflM$ find p -type f | xargs cat # p/a.py import b b.foo() # p/b.py import a a.bar() #...

Also, I would expect an 'incremental' behavior: if folder `a/b/c` contains a cycle, so does `a/b`. This is currently not the case: ```shell $ git describe v0.0.7-2-gb93822c # GOOD: $...

"then you get the whole of /include added to your include paths"... There are two world views with regard to include paths and install prefixes: 1. I don't care because...

You're right; I first thought 'compiler vendors' should solve, but why would they if the compilation model leaves this issue in the open? So I aimed a little higher and...

Definitely a good idea. The `std::expected` type may be used here, once it arrives. http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n4109.pdf

By the way, the fact that pointer values are _known_ to be possibly `null_ptr` is an idiom that should probably be stated explicitly, too.

Thanks! That explains it. I read over it - silly. Shall I add it to the dependencies in the rockspec?