LGTM.  Alternatively, the test could be
  (null (notmuch-wash....))
with the correct answer being 't'.  That would avoid the awkward
detour through a string, but either way is good as long as this test
passes.

Quoth david@tethera.net on Aug 30 at 10:09 pm:
> From: David Bremner <bremner@debian.org>
> 
> The behaviour of "emacsclient --eval nil" changed from emacs23 to
> emacs24, and in emacs24 it prints 'nil' rather than an empty string.
> 
> (format "%S" foo) produces a sexpr form of foo, and is consistent
> between the two versions.
> ---
> 
> This fixes another test failure on emacs24. 
> 
> I guess maybe all test_emacs output could be canonicalized this way,
> but I suspect that would be pretty disruptive.
> 
>  test/emacs-subject-to-filename |    6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/test/emacs-subject-to-filename b/test/emacs-subject-to-filename
> index 176e685..a0ffdfe 100755
> --- a/test/emacs-subject-to-filename
> +++ b/test/emacs-subject-to-filename
> @@ -8,10 +8,10 @@ test_emacs '(ignore)'
>  
>  # test notmuch-wash-subject-to-patch-sequence-number (subject)
>  test_begin_subtest "no patch sequence number"
> -output=$(test_emacs '(notmuch-wash-subject-to-patch-sequence-number
> -      "[PATCH] A normal patch subject without numbers")'
> +output=$(test_emacs '(format "%S" (notmuch-wash-subject-to-patch-sequence-number
> +      "[PATCH] A normal patch subject without numbers"))'
>  )
> -test_expect_equal "$output" ""
> +test_expect_equal "$output" '"nil"'
>  
>  test_begin_subtest "patch sequence number #1"
>  output=$(test_emacs '(notmuch-wash-subject-to-patch-sequence-number