Hi Paul, > > I claim that using _gl_<something> is the less risky compromise. > > Yes, thanks for explaining, and that sounds good to me too. Can we go in > that direction?
Moving from __gl_<something> to _gl_<something> is what I already did in the 4 patches. Do you mean, we should also move from gl_<something> to _gl_<something> for non-public API, such as gl_consolesafe_fwrite gl_isinfd gl_signbitl gl_LDBL_SNAN gl_MB_CUR_MAX gl_fd_to_handle ? That would be fine with me. Bruno
