Yes, I would trust an actively developed fork of a TCP stack that is 2 years ahead of its forked project more than the original. Especially for something as critical as TCP, and equally so for BGP routing. Why use a dead project that hasn't gotten bug fixes for years?