On Thu, Nov 19, 2020 at 11:50:26AM -0800, Junio C Hamano wrote: > Jeff King <peff@xxxxxxxx> writes: > > > I took a look at converting some of the existing tests. This seems to > > work. It's a bit longer to read, perhaps, but I kind of like that the > > expected outcome is all laid out. It also pollutes the test output less > > (e.g., if you wanted to count up skipped tests in the whole suite, you'd > > get a bunch of noise from t0000 for these uninteresting skips). > > > > Thoughts? I think this is something I'd do on top of your patch. > > Yes, it looks nice as the expectation is expressed much clearly. OK, then here's the whole thing. I ended up with a few more cleanups, too. This is all on top of Gábor's patches. It's conceptually independent, but the textual wrangling was annoying enough it didn't make any sense to require you to do it again during merging. ;) Plus I do not think either topic is high-risk nor urgent enough to worry too much about one blocking the other. The diffstat is scary, but it's mostly the final patch, which is pretty mechanical. [1/4]: t0000: keep clean-up tests together [2/4]: t0000: run prereq tests inside sub-test [3/4]: t0000: run cleaning test inside sub-test [4/4]: t0000: consistently use single quotes for outer tests t/t0000-basic.sh | 570 +++++++++++++++++++++++------------------------ 1 file changed, 284 insertions(+), 286 deletions(-) -Peff
