In message <[EMAIL PROTECTED]
rosoft.com>you write:
>
> At this point we have the following situation:
> The current makefile/configure combination doesn't
> work on systems without symbolic links because of
> the nested quoting problem discussed in this thread.
>
> No one has yet come up with a "decent" solution to
> the problem (when all the constraints are considered)
> that doesn't involve $(). (We can discuss whether
> $() is "decent" separately, but that's an issue of
> availablity, not understandability.)
>
> I'd submit a patch (a) if FSF could accept it (and
> I'm short a signature yet on that, but working on it)
> and (b) if I knew what would be accepted. In the
> meantime, since $() works for me, I'm keeping a private
> patch that solves the problem (although it's not what
> I'd do if $() were the permanent solution, what I have
> is more "rot proof", but not clean).
>
> However, that leaves other systems without symlinks out
> in the cold. Do the maintainers want to pick one of the
> ugly solutions and declare it "the best we can do", or ???
> (If they were to pick one, the variant that uses a
> generated script to generate a piece of a makefile that's
> included using multiple -f options seems the least onerous
> to me.)
I think we need to pick one of the ugly solutions and declare it
the best we can do. Whatever code we install needs to have some
comments explaining what it's doing.
jeff
