Re: Colon-array variables can crash "sh" emulation

[Bart Schaefer <schaefer@brasslantern.com>]

On Apr 20, 11:21am, Peter Stephenson wrote:
}
} The only robust solution looks like being not let PATH's ename element
} point at path.  Otherwise the code following ename has to jump through
} hoops to see if the parameters are "really" linked, which is far too
} late.

Yes.

} The fly in the ointment here is that colonarrsetfn() refuses to fix up
} the environment unless the ename element is set.  It's not clear to me
} why since I don't think ename can be unset for special tied variables
} which are the only use of colonarrsetfn().

This is probably not be true any longer, but at one time I think "ename"
was doing double duty as a flag that the parameter was in fact exported.
My understanding was that it's called "ename" because it's the name in
the environment to which the parameter refers.  This is probably only
the case for the array version of the parameter and not for the scalar,
or something like that.