https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106808
--- Comment #8 from andysem at mail dot ru --- (In reply to Jonathan Wakely from comment #7) > (In reply to andysem from comment #6) > > So do you think this is a problem in Boost.Filesystem? > > I don't know yet, I can't reproduce it with the Boost in Fedora 36, and I > haven't looked further. I have added a workaround to the current Boost.Filesystem develop and master, and Boost 1.80 (the latest official release) didn't support std::string_view, so is not affected either. In order to reproduce you would need the revision I posted in the original problem description. > > I would say this is a regression in string_view as the code is valid in > > pre-C++23, and I would expect it to stay valid in C++23 onwards. > > That range constructor is explicit in the C++23 draft now, but that change > hasn't been backordered to GCC 12 yet, and apparently doesn't change > anything if you still see this in trunk. > > > Shouldn't > > the range constructor be simply disabled when fs::path::iterator is not > > defined? > > Do you want ODR violations? Because that's how you get ODR violations. I understand this, but my point is that this is a breaking change, apparently even with the constructor being marked explicit, and it breaks in a rather surprising way. Also, apparently MSVC is able to not break somehow as it doesn't have this issue. So maybe there is a solution that does not introduce ODR violations and yet still works. I think it would be preferable to fix std::string_view rather than all its uses like in std/Boost.Filesystem.
