EDIT: also fooled by changelog timing. Original comment kept: It is a little odd to add it in now. If they had it before but were hesitant to purge it from the

EDIT: also fooled by changelog timing. Original comment kept: It is a little odd to add it in now. If they had it before but were hesitant to purge it from the codebase, that would be less of a head-scratcher. In many cases there is in fact value gained by adding older algorithms to software that lets the two parties negotiate a specific one. Libraries have to strike a balance between offering sane (and secure) defaults but also offering choice for advanced users -- this I accept. But given that support was previously absent and NIST's own recommendation to stop using it immediately, this is a decision I don't truly understand.