Re: Boost test matrix

From: Peter Dimov (pdimov_at_[hidden])
Date: 2023-11-03 15:56:06

René Ferdinand Rivera Morell wrote:
> Yes, the testing scripts should be more reliably resilient. As for b2..
> 
> (1) It can't be fixed in b2. As it's not possible for b2 to read
> user's minds. Hence the existence of the multiple ways to tell b2 what
> toolset to build with.
> 
> (2) When we have the Boost modular arrangement it will be possible to
> not use "b2 headers" at all. And hence that problem will be "fixed".

The fact that the automatic header symlink generation works for normal
targets and fails for targets passed to configure.builds leads me to think
that there is a problem in b2 somewhere, but I have no idea what or where.