On Thu, 02 Feb 2012 10:00:59 -0800, Jameson Graef Rollins <jrollins@finestructure.net> wrote: > Hey, Dmitry. I'm so sorry I sent my last email on your original patch > before I saw this new series. I do now like your original proposal > better, since it shows the diff based the names the caller provides, > which I now agree is probably the clearest and most robust solution. > The second patch in this series could still go through, though, no > matter what version of the change to test_expect_equal_file we go with. > Actually, we can do both: check file name for consistent diff order (from expected to actual) and use file names that the caller provides. What do you think? Regards, Dmitry > jamie.