Eric Sunshine <sunshine@xxxxxxxxxxxxxx> writes: > On Thu, Dec 19, 2019 at 2:22 PM Junio C Hamano <gitster@xxxxxxxxx> wrote: >> In any case, here is what I tentatively have in my tree (with heavy >> rewrite to the proposed log message). >> >> +test_expect_success 'commit without staging files fails and displays hints' ' >> + echo "initial" >>file && > > The use of '>>' here rather than '>' feels wrong, especially when > "initial" is used for both the file body and the commit message, > causing a reader of the test to wonder if this test somehow depends > upon earlier tests. Yeah, makes sense. This was verbatim from v1 but I think starting the file from scratch like you suggest makes it clearer what is going on. > >> + git add file && >> + git commit -m initial && >> + echo "changes" >>file && >> + test_must_fail git commit -m update >actual && >> + test_i18ngrep "no changes added to commit (use \"git add\" and/or \"git commit -a\")" actual >> +'
