Re: Issue: REQUIRE-PATHNAME-DEFAULTS (Version 3)

[Dan L. Pierson <pierson%mist@multimax.ARPA>]

    This looks better, but I feel really uncomfortable with the idea that
    an implementation would be allowed to decide on its own that modules
    which haven't been PROVIDE'd have been loaded anyway.  If I say
    (REQUIRE "foo") and if I haven't done a (PROVIDE "foo"), I want to get
    an error.  I don't want the implementation to decide on its own that
    module "foo" has been loaded just because I've loaded a file named
    "foo", for example.  The file named "foo" might contain only part of
    the module "foo", or it might contain the system definition (defined
    using some variant of DEFSYSTEM) for module "foo", or it might contain
    something else entirely.
    
I was thinking similar thoughts while writing the rationale.  If the
person who requested this feature (JonL?) doesn't come up with a
convincing defense soon, I'll probably remove it.