Feedback on clang change (#53194)

[Mojca Miklavec]

On 1 February 2017 at 16:20, Marcus Calhoun-Lopez wrote:
> Yes, we are talking about the libstdc++ that is shipped with gcc6.

OK.

Something that's not entirely clear to me yet is what's the benefit of
having clang compile software against gcc's libstdc++ vs. gcc6
compiling it.

And what's the relation of all this to the magic
-D_GLIBCXX_USE_CXX11_ABI=0 flag.

> I should have made clear that a major downside of choice #2 is that it requires a library dependency on gcc6.

Thanks, I missed that. And I agree that this is a bit of a downside.
What kind of dependency is that exactly? Is it only a dependency
because the port needs to make sure that gcc's libstdc++ is present at
the time of compilation or are there some other more hardwired
constraints?

> That is something that most (the vast majority of?) users will not want.
> So perhaps choice #3 or choice #2  with the default variant only for systems prior to OS X Mavericks?

We could certainly limit this to pre-Mavericks systems if that doesn't
make things super complicated. The other question is whether we need
to have this enabled for all versions of clang or would a single
compiler suffice.

(I would help if I better understood all consequences and a couple of
use cases.)

We just have to be careful to make sure that we don't create an even
bigger and more incompatible mess :)

And we should finally address binary builds for libc++. That would
likely solve 99.9% of our C++11-related problems.

Mojca