sg/rebase-progress, was Re: What's cooking in git.git (Jun 2019, #04; Fri, 14)

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Jun 17, 2019, 18:25 UTC
Message-ID: <nycvar.QRO.7.76.6.1906171944170.44@tvgsbejvaqbjf.bet>
In-Reply-To: <xmqqh88ruc1b.fsf@gitster-ct.c.googlers.com>

Hi Junio,

On Fri, 14 Jun 2019, Junio C Hamano wrote:

Show 11 quoted lines

> * sg/rebase-progress (2019-06-11) 4 commits
>  - progress: use term_clear_line()
>  - rebase: fix garbled progress display with '-x'
>  - pager: add a helper function to clear the last line in the terminal
>  - t3404-rebase-interactive: use the 'q_to_cr' helper
>
>  Use "Erase in Line" CSI sequence that is already used in the editor
>  support to clear cruft in the progress output.
>
>  Its test may have to be updated to work around glitch in test setup
>  on Windows.

Two thoughts on that:

- the problem is an existing test that had to be adjusted for this patch
  series,

- the existing test has the underlying issue that it casts too wide a net,
  and catches *non*-regressions more easily than regressions (read: it is
  false-positive-happy).

And most importantly, the "glitch in the test setup" was actually a full-blown bug in Git for Windows' MSYS2 runtime (not very likely to hit anyone, though), and I fixed it in the meantime. The build for `pu` passes now:

https://dev.azure.com/gitgitgadget/git/_build/results?buildId=10882

and also the build for `sg/rebase-progress`:

https://dev.azure.com/gitgitgadget/git/_build/results?buildId=10885

Given that my remaining gripe with the tests was hardly introduced by this patch series (although I would have loved to see it fixed in the course of a next iteration), I tend to be in favor of advancing this to `next` as-is. I can always offer a patch on top to clean up as I had suggested.

Ciao, Dscho