On 22 Oct 2013 at 0:21, Edward Diener wrote:
Sure, they're trying to get people off 4.6.x, which is indeed old but trusty. I keep a tarball of it locally.
There is a 4.6.2-1 in the link above, but no 4.6.4. That's why I was wondering why you are specifying 4.6.4.
Probably because I am going senile (I am also in Seattle, and therefore unable to check the exact version, so I relied on my obviously faulty memory).
I had no reason to look. I only cared about AFIO, and therefore only cared about its dependencies ASIO and Thread.
Did you find any tests not working in ASIO and Thread when using gcc 4.7+ on up with MingW ? Or was it just in your own work with AFIO ?
AFIO's tests failed. I didn't run anyone else's tests. The problems came from not AFIO code though, but rather how AFIO instantiated other people's code.
Also, me personally I have zero time for Mingw given MSVC is free and Mingw binaries are not interoperable with MSVC ones. However Boost peer review wants demonstrated portability, and for that alone I delved into Mingw compatibility. Ordinarily I'd never touch it, except to debug unhelpful MSVC error messages (and even there I prefer clang-win).
It sounds like you able to test clang under Windows using the MingW-w64 RTL and the Boost Build clang toolset.
No, I usually use scons for build. I only used Boost.Build for AFIO because I had to, and in truth Paul did almost all of the Jamfile scripting. Niall -- Currently unemployed and looking for work. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/