[PATCH RESEND] zic, various tests: use LFS I/O functions explicitly where needed

[Paul Eggert]

Nick Alcock wrote:

> we've seen zic fail with bizarre error messages in this situation
> already.

I think those failures were fixed by the EOVERFLOW check that is in zic already. 
That being said, the patch shouldn't hurt correctness and should help 
performance a bit in some cases, so I just now installed a more-portable patch 
upstream. Please see:

http://mm.icann.org/pipermail/tz/2017-June/025164.html

So, an alternative to your patch would be to install this upstream patch (it's 
OK to have both patches, of course).