Warning that should be removed

[tcbrindle]

Warning that should be removed

[tcbrindle]

Thanks for the info. What compiler/warning level combination triggers a warning in this case?

I think I'd prefer std::size_t dynamic_extent = std::size_t(-1), but looking at the standard text it actually uses std::numeric_limits<std::size_t>::max(), so perhaps that's what we should go for.

[brandl-muc]

With MSVC with warning level 4 this leads to the following:
warning C4245: 'initializing': conversion from 'int' to 'size_t', signed/unsigned mismatch
Also see this online example.

We compile with /W4 and warnings as errors, so we can't use span out of the box. (thanks for the work btw)

Although there don't seem to be any implementations that don't use two's complement and there are afaik ideas to make the abstract machine two's compement currently using -1 is non-compliant. Same goes for ~0 of course.
I'd propose using SIZE_MAX or in case you don't want to add the extra header std::size_t{} - 1u should also work because modular arithmetics are guaranteed for unsigned types.