Re: doc elisp intro cross reference fixes

[Richard Stallman]

    I think calling such cases "broken" is alarmist.
    There might be cases where some defcustoms get placed in groups where
    they were not meant to go, but I can't think of any way this can break
    any code.

It would "break the code" in the sense that the unchanged defcustoms
would no longer do what they are intended to do.

Giving users a way to avoid repeated :group specifications could be
convenient, but it looks like this way of doing it is unreliable.
So let's consider another way: a special expression

  (defcustom-group FOO)

which would mean, "use group FOO by default for the rest of this file, or
until another defcustom-group."  This way, just changing the defgroups
would not alter the default for :group in any defcustom.

This would do basically the same job of defaulting for convenience, but in
a more reliable way.

Do you see any drawback in this change?