Re: [PATCH 2/4] t4011: illustrate "diff-index -p" on stat-dirty paths

Junio C Hamano [off-list ref] writes:

The plumbing that looks at the working tree, i.e. "diff-index" and
"diff-files", always emit the "diff --git a/path b/path" header lines
without anything else for paths that are only stat-dirty (i.e. different
only because the cached stat information in the index no longer matches
that of the working tree, but the real contents are the same), when
these commands are run with "-p" option to produce patches.

Illustrate this current behaviour.  The new part that uses "-w" option
demonstrates that we do not show any "diff --git" header for blobs whose
true contents are different but compares the same when whitespaces are
ignored, which is inconsistent with the behaviour for stat-dirty paths.

[...]

-test_expect_success SYMLINKS 'diff identical, but newly created symlink' '
+test_expect_success SYMLINKS 'diff identical, but newly created symlink and file' '
 	cat >expected <<-\EOF &&
 	diff --git a/frotz b/frotz
+	diff --git a/nitfol b/nitfol
 	EOF
+	sleep 3 &&
+	rm -f frotz &&
+	echo xyzzy >nitfol3 &&
+	mv nitfol3 nitfol &&
 	ln -s xyzzy frotz &&
 	git diff-index -M -p $tree >current &&
+	compare_diff_patch expected current &&
+
+	>expected &&
+	git diff-index -M -p -w $tree >current &&
 	compare_diff_patch expected current
 '

I find the last bit of the commit message rather confusing.  You appear
to be using -w here to diff the stat-dirty worktree nitfol 'xyzzy\n'
against the $tree:nitfol which is also 'xyzzy\n'.

If that analysis is correct, then

  we do not show any "diff --git" header for blobs whose true contents
  are different but compares the same when whitespaces are ignored

is not what is going on here; the blobs have exactly the same content.
The difference is that

* without -w, the code "knows" from the lstat() data that the files are
  different, prints a header, and then fails to find any differences;

* with -w, the code correctly holds off on printing anything since it
  will invariably have to inspect the contents beforehand.

So perhaps you can say

  Illustrate this current behaviour.  Also demonstrate that with the
  "-w" option, we (correctly) hold off showing a "diff --git" header
  until actual differences have been found.  This also suppresses the
  header for merely stat-dirty files, which is inconsistent.

-- 
Thomas Rast
trast@{inf,student}.ethz.ch