On 30 May 2015 at 21:32, Andrey Semashev wrote:
Trying to emulate Boost.Test is not the solution (and we already have lightweight_test.hpp for that). I think, bpm has to distinguish between the three sets of dependencies and let the user decide what he wants to install.
Sorry, I wasn't being clear. I was coming from the perspective of what is needed to persuade the steering committee and the release managers that a bpm based Boost is close enough to invest pushing it over the finish line. I wasn't suggesting replacing Boost.Test. I was suggesting standing in an emulation good enough to allow the unit tests to compile and link (not necessarily run), and therefore gain a method of formally proving that the other two sets of dependencies (use and build) are working as intended. I would imagine an automated script which bpm checked out every combination possible of Boost library (132 x 131?) and successfully compiled and linked all the libraries and unit tests would be a pretty good proof that a bpm based fully modular Boost distro is close. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/