Re: ediff and coding systems

[Eli Zaretskii <eliz@gnu.org>]

> From: Kenichi Handa <handa@m17n.org>
> CC: kifer@cs.sunysb.edu, eliz@gnu.org, emacs-devel@gnu.org
> Date: Mon, 22 Oct 2007 12:42:47 +0900
> 
>   read-coding-system("Select coding system (default mule-utf-8): " mule-utf-8)
>   select-safe-coding-system-interactively(1 5557 (utf-8 ....))
>   select-safe-coding-system(1 5557 nil nil "/tmp/fineDiffA155544eB")
>   write-region(1 5557 "/tmp/fineDiffA155544eB" nil no-message)
>   ediff-make-temp-file(#<buffer  *ediff-tmp*> "fineDiffA" "/tmp/fineDiffA155544eB")
>   ediff-make-fine-diffs(5 noforce)
>   ediff-install-fine-diff-if-necessary(5)
>   ediff-next-difference(1)
>   call-interactively(ediff-next-difference)
> 
> I don't know why ediff-make-temp-file is called, but perhaps
> it should call write-region while binding
> coding-system-for-write to `no-conversion' or `emacs-mule'.

Not a good idea, I'm afraid: those temporary files (or parts thereof)
will need to be read back, when `diff' finishes, so each one needs to
be in the original encoding, or else the diffs will be incorrect.
E.g., imagine two files with the same text, but different encodings: I
don't think we want ediff say that they compare identical.