Re: [PATCH] diff-files: treat "i-t-a" files as "not-in-index"

> > +     content="foo" &&
> > +     echo $content >not-empty &&
>
> The quoting decision is backwards in these two lines.  It is OK not
> to quote when the right hand side literal is clearly a single word
> without $IFS.  On the other hand, it is a good practice to always
> quote when using what is in a "$variable".

> > > +  touch empty &&
> >
> > Use of "touch" gives a wrong impression that you care about the file
> > timestamp; use something like ": >empty &&" instead when you care
> > about the presence of the file and do not care about its timestamp.
>
> I just realized that this is even more important in this case not to
> use "touch".
>
> The test that uses this file cares not just the presence, but it
> deeply cares that its contents is empty.  The thing it least cares
> about is its timestamp.
>
> The purpose of using "touch" is to update the timestamp, to keep the
> current contents if it exists, and to ensure it exists (as a side
> effect), in the decreasing order of importance.  Use of the command
> here misleads the readers.

> > +     git add -N empty not-empty &&
> > +     git diff-files -p >actual &&
> > +     hash_e=$(git hash-object empty) &&
> > +     hash_n=$(git hash-object not-empty) &&
> > +     cat >expect <<-EOF &&
> > +     diff --git a/empty b/empty
> > +     new file mode 100644
> > +     index 0000000..$(git rev-parse --short $hash_e)
> > +     diff --git a/not-empty b/not-empty
> > +     new file mode 100644
> > +     index 0000000..$(git rev-parse --short $hash_n)
> > +     --- /dev/null
> > +     +++ b/not-empty
> > +     @@ -0,0 +1 @@
> > +     +$content
> > +     EOF
> > +     test_cmp expect actual
> > +'
>
> OK.  Do we want to show what happens when "diff" and "diff --cached"
> are run with these two "added but not quite added yet" paths to
> contrast with this new case?