# Re: \[PATCH v5 2/2\] difftool.c: learn a new way start at specified file

- From: ZheNing Hu \<hidden\>
- Date: 2021-02-17 16:19:33 
- List: git
- Page: https://ratatoskr.run/git/2021/02/7876720
- Thread: https://ratatoskr.run/git/2021/02/7875304/t.md

```text
Junio C Hamano [ref] 于2021年2月17日周三 下午6:31写道：
>
> "ZheNing Hu via GitGitGadget" [ref] writes:
>
> > From: ZheNing Hu <redacted>
> >
> > `git difftool` only allow us to select file to view in turn.
> > If there is a commit with many files and we exit in the search,
>
> I am not sure what "in the search" refers to.  "in the middle" I
> would understand, though.
>
> > We will have to traverse list again to get the file diff which
>
> Let's downcase this "We".
>
> > we want to see. Therefore, here is a new method: user can use
> > `git difftool --rotate-to=<filename>` or `git difftool --skip-to=<filename>`
> > to start viewing from the specified file, This will improve the
> > user experience.
>
> Do we need both?  I'd rather not to give end-user-facing commands
> too many knobs that would do similar things.  Too many choices to
> choose from without clear answer to "which one should I prefer to
> use?" is a bad combination for end-users.
>
So users will not need to use `git difftool --skip-to`? Then I am confused
about the meaning of `git difftool --skip-to`.
> > diff --git a/Documentation/diff-options.txt b/Documentation/diff-options.txt
> > index 7c5b3cf42bcc..aa2b5c11f20b 100644
> > --- a/Documentation/diff-options.txt
> > +++ b/Documentation/diff-options.txt
> > @@ -701,7 +701,7 @@ components matches the pattern.  For example, the pattern "`foo*bar`"
> >  matches "`fooasdfbar`" and "`foo/bar/baz/asdf`" but not "`foobarx`".
> >
> >  --skip-to=<file>::
> > ---rotate-to=<file::
> > +--rotate-to=<file>::
> >       Discard the files before the named <file> from the output
> >       (i.e. 'skip to'), or move them to the end of the output
> >       (i.e. 'rotate to').  These were invented primarily for use
>
> Thanks for correcting, but this change should not be a part of this
> patch.  Instead, you help the other's topic by giving a review (and
> you could just have said "there there is closing '>' missing").
>
Okay, I will pay attention later.
> > diff --git a/Documentation/git-difftool.txt b/Documentation/git-difftool.txt
> > index 484c485fd06c..c64dff69c976 100644
> > --- a/Documentation/git-difftool.txt
> > +++ b/Documentation/git-difftool.txt
> > @@ -34,6 +34,16 @@ OPTIONS
> >       This is the default behaviour; the option is provided to
> >       override any configuration settings.
> >
> > +--rotate-to=<file>::
> > +     Internally call `git diff --rotate-to=<file>`,
> > +     show the change in the specified path first.
> > +     Files before the specified path will be moved to the last output.
> > +
> > +--skip-to=<file>::
> > +     Internally call `git diff --skip-to=<file>`,
> > +     skip the output to the specified path.
> > +     Files before the specified path will not output.
> > +
>
> This, unlike the "diffcore" stuff, is end-user facing, and it is
> better not to force the readers even know what --skip-to option
> to the diff does (after all, difftool users are using 'git difftool'
> and they are not necessarily 'git diff' users).
>
>     --skip-to=<file>::
>             Start showing the diff for the given path, skipping all
>             the paths before it.
>
> or something, perhaps.
>
> > +test_expect_success 'difftool --skip-to' '
> > +     difftool_test_setup &&
> > +     test_when_finished git reset --hard &&
> > +     git difftool --no-prompt --extcmd=cat --skip-to="2" HEAD^ >output &&
> > +     cat >expect <<-\EOF &&
> > +     2
> > +     4
> > +     EOF
> > +     test_cmp output expect &&
> > +     test_must_fail git difftool --no-prompt --extcmd=cat --skip-to="3" HEAD^
> > +'
>
> This probably should be split into two independent tests.  One to
> check that the non-failing case works as expected, the other to
> check that a bogus command line option errors out as expected.
>
I will finish it.
> Thanks.

Besides, I have some curiosity about one place in the code in your
patch:
> int cmd_diff_tree(...)
> ...
> if (read_stdin) {
> ...
> opt->diffopt.rotate_to_strict = 0;
> ...
}
This is the only place where rotate_to_strict is set to zero,
So the "git log -p" you mentioned earlier is here Called this code
to avoid exiting  the program because of the wrong path, right?

Thanks.
```