Something I haven't heard explained on this thread yet is why the change would be worthwhile in terms of cost (to e.g. the GNU project) vs. benefit (to e.g. users).
The cost is, I hope, reasonably obvious: - the maintenance burden of the extra code and maintaining both fork/exec and posix_spawn implementations. - the cost to users of the additional bugs introduced by human error during this maintenance Benefits on the other hand probably accrue mainly to users (which is fine): - but actually, what are those benefits? - are those benefits enjoyed by users other than Threos users? Who? Would this e.g. work better for Cygwin? James.