Re: [GSoC][PATCHl 1/6] rebase -i: add --ignore-whitespace flag

[Rohit Ashiwal <rohit.ashiwal265@gmail.com>]

Hi Junio

On Wed, Aug 7, 2019 at 11:08 AM Junio C Hamano <gitster@pobox.com> wrote:
>
> Rohit Ashiwal <rohit.ashiwal265@gmail.com> writes:
>
> > diff --git a/builtin/rebase.c b/builtin/rebase.c
> > index db6ca9bd7d..3c195ddc73 100644
> > --- a/builtin/rebase.c
> > +++ b/builtin/rebase.c
> > ...
> > @@ -511,6 +523,8 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
> >       argc = parse_options(argc, argv, NULL, options,
> >                       builtin_rebase_interactive_usage, PARSE_OPT_KEEP_ARGV0);
>
> Curious.  Did you rebase onto an older codebase?  I think the round
> currently queued in my tree already has c0e78f7e46 which merged the
> jk/unused-params-final-batch topic that updated this call to
> parse_options() to pass prefix.  Perhaps you want to fetch from me
> what has been on 'pu' (it should be "log master..10827432") and
> compare what you had?  Unless there is no compelling reason not to,
> it would be a good idea to base the reroll on the same commit as
> the commit on which the previous version has been queued, which in
> this case is 9c9b961d ("The sixth batch", 2019-07-19).

Oh~ I didn't notice it. I'll rebase and re send the patch. Thanks for
pointing out.

> > diff --git a/t/t3433-rebase-options-compatibility.sh b/t/t3433-rebase-options-compatibility.sh
> > new file mode 100755
> > index 0000000000..e617d3150e
> > --- /dev/null
> > +++ b/t/t3433-rebase-options-compatibility.sh
> > @@ -0,0 +1,65 @@
> > ...
> > +
> > +     git checkout --orphan master &&
> > +     cat >file <<-EOF &&
> > +     line 1
> > +             line 2
> > +     line 3
> > +     EOF
>
> The second line triggers "indent with spaces" whitespace error; you
> can protect these spaces in the leading indent like so:
>
>         sed -e "s/^|//" <<-\EOF &&
>         |line 1
>         |        line 2
>         |line 3
>         EOF
>
> Also make it a habit to always quote the EOF token that begins the
> here document (i.e. <<-\EOF) when the here document does not need
> variable interpolation.  I suspect (but I did not read all of the
> tests) that there may be many other instances of here document in
> this patch that can use the same improvement.

Will keep this in mind.

Thanks
Rohit