On Wed, May 27, 2020 at 11:03 AM Junio C Hamano <gitster@xxxxxxxxx> wrote: > > Jacob Keller <jacob.e.keller@xxxxxxxxx> writes: > > > This new logic is more robust, as we will correctly identify superseded > > options, and ensure that both _git_switch and _git_checkout enable DWIM > > in similar ways. > > > > Add several tests which demonstrate the new expected behavior. Note that > > some of the git switch tests are marked as failures because the default > > git switch completion with --guess is sub-par as discussed in a previous > > commit. This will be fixed in a future change. > > "as discussed in a previous commit"? 1/9 does not mention --guess > at all. Perhaps this is an unintended damage due to rebasing? > Yes I originally had a patch earlier in the series which just added test cases, but I later moved that into the patch that fixed the tests. > In any case, up to this point, the log messages are a bit > frustrating read, primarily because the author has spent too much > time on DWIM and expect everybody understands exactly what he means > when he says "DWIM mode is enabled", while a reader in me keeps > asking: Right, I thought I had an explanation of this somewhere but it appears to have been lost while re-ordering and re-wording. > > - what refs get included when I type <TAB> without DWIM mode? > > - what extra refs get included when DWIM mode is enabled? > > - under DWIM mode, do certain refs (that would be included without > DWIM mode) stop appearing in the completion? > > and without getting any answer to them. Perhaps the proposed log > message for [1/9] can be made a bit more robust to cover it? > Sure, I'll look at what I can do to make this more logical. > Thanks. >
