Re: [GSoC][PATCH] lib-read-tree-m-3way: modernize a test script (style)

> On Fri, Jan 28, 2022 at 4:34 PM Eric Sunshine <sunshine@sunshineco.com> wrote:
> > In this case, the indentation of the entire body of the for-loop needs
> > to be fixed to use tabs rather than spaces, however, fixing all the
> > indentation problems together with the other problems can make for a
> > too-noisy patch for reviewers to really see what is going on. As such,
> > it may be better to make this a multi-patch series in which one patch
> > fixes indentation problems only.
>
> > As mentioned above, changing the body of the test from double- to
> > single-quoted string should (presumably) be okay since the body gets
> > eval'd and `$p` will be visible at the time of `eval`, however, mixing
> > this sort of change in with all the other changes being made makes it
> > hard for reviewers to spot such little details, let alone reason about
> > correctness. As such, switching of quote types is also probably the
> > sort of change which should be split out into its own patch. The same
> > comment applies to other changes following this one.
>
> I think so. I was thinking fixing all the general styling issues in one
> patch (since they are all style related), but now I realize that the general
> style patch can be divided into sub-patches for clearer review experience.
>
> And my question is, should I do this "multi-patch series" thing in the form of
> "-v<n>" (all under this thread), e.g. "v2" or "v3"? Or I just submit
> multiple patches separately (a new thread for each one)?

> > Overall, with the exception of the test title which needs to
> > interpolate `$p`, the rest of the changes are probably reasonable and
> > benign. It's important to understand that lengthy patches like this
> > full of mechanical changes tend to be quite taxing on reviewers, so
> > it's a good idea to help in any way you can to ease the review burden.
> > This can be done, for instance, by making only a single type of change
> > per patch (i.e. indentation fixes), or by limiting the scope or
> > breadth of the changes, which is especially important for this sort of
>
> I'm not quite sure what this means, and I quote, "limiting the scope or
> breadth of the changes". Are you suggesting, for example,
> fixing fewer occurrences of tab indentation issue in a patch; or, for
> example, limiting the
> fix to a specific "test_expect_success" block, and do multiple patches
> sequentially?