Paul A. Bristow
This isn't just a two-sides issue. There are also a lot of Boost users and developers who have been unhappy with Boost.Test documentation, evolution, and maintenance for over about a decade.
Really? I can appreciate that you seem to be frastrated for some reason, but I frankly not sure what exactly is a cause. The only thing that comes to mind is that you always prefered franction tolerance over percent. You might be interested to know that new floating point comparison interfaces improve on that.
It is clear to me that Gennadiy is not willing to cooperate with anyone.
Can you please, please, show me a single post, here I claim anything along these lines. I am not going to cooperate on making the library worse (in any aspect). You can cite me on that.
2 I'd like to see a much simpler faster lighter-weight header-only version of Boost.Test with a different name and a different maintenance *team*.
It appears you just want to use something else. That's fine and many people do. That said: * I did not see yet a single library which is simpler * No one ever expressed a problem with Boost.Test being slow * Boost.Test is header only (has an ability) Gennadiy