"Philip Peterson via GitGitGadget" <gitgitgadget@xxxxxxxxx> writes: > From: Philip Peterson <philip.c.peterson@xxxxxxxxx> > > The imperative format was a little hard to read, so I rewrote the test cases > in a declarative style by defining a common structure for each test case and > its assertions. > > Signed-off-by: Philip Peterson <philip.c.peterson@xxxxxxxxx> > --- > t/unit-tests/t-apply.c | 123 ++++++++++++++++++++++++++--------------- > 1 file changed, 78 insertions(+), 45 deletions(-) In this project, we do not add a version of a known-to-be-bad file in patch [1/2], only to be immediately improved in patch [2/2]. Unless, of course, there is a good reason to do so. And "to preserve the true history of what happened in the developer's working tree" is not a good reason. We give our developers "rebase -i" and other means to rewrite their Git history, not just because we want them to be able to pretend that they are a better developer who never make such a mistake or misdesign in the first place, but because a polished history is easier to review in the shorter term and to maintain in the longer term. Thanks.
