What about forking the Boost.Test library into a new Boost.XTest with the contents of the Boost.Test release and your documentation and taking care of the issues associated to the release branch?
I do not believe this is a good solution. The intersection (name collision) between two libraries is going to be way too big and we'll never know which one code refers to (user complains about). This will lead to no end of confusion.
Richard is free to develop a new library from scratch and eliminate all the things he believes are not important (and for that matter add anything he believe is lacking), pass it through the review and maintain it.
gennadiy, from my understanding of the boost license, richard is free to fork boost.test and you have to right to forbid this ... if his design is better or worse or may or may not clash with boost.test is not your decision to do. besides: competition often leads to better products, imo clang is the best thing that happened to gcc ... i had my troubles with boost.test myself, so also from the perspective of the user it would be very helpful to have two libraries with a similar interface, which could be exchanged by using a different namespace or macro prefix. that said, i have no idea, why you strongly dislike richard's offer (and i don't care), but if you look for help ... what about taking advantage of the recent migration to git: anyone, including richard, could fix issues and submit pull requests. you just have to review the code and press a button. still you may want to close trac tickets if the bugs are fixed. tim