Roland McGrath, le Thu 06 Aug 2015 01:01:55 -0700, a écrit : > then a header that defines away the functions as > no-op always-fail or no-op pretend-to-succeed or something else distinct > from "implement as a faithful emulation of Linux usage regardless of the > sensibility of such usage on the Hurd" could well be the most useful thing.
Or maybe not: providing such a failing call makes those programs enable portions of code it expects to succeed, and thus getting new failures. It would indeed probably be useful to actually check these usage before including mount(). Samuel
