Re: [PATCH 1/3] diff.c: add --output-indicator-{new, old, context}

> On Mon, Aug 20, 2018 at 12:31 PM Johannes Schindelin
> <Johannes.Schindelin@gmx.de> wrote:
> >
> > On Fri, 17 Aug 2018, Stefan Beller wrote:
> >
> > > This will prove useful in range-diff in a later patch as we will be able
> > > to differentiate between adding a new file (that line is starting with
> > > +++ and then the file name) and regular new lines.
> > >
> > > It could also be useful for experimentation in new patch formats, i.e.
> > > we could teach git to emit moved lines with lines other than +/-.
> >
> > Thanks.
> >
> > > diff --git a/diff.c b/diff.c
> > > index c5c7739ce34..03486c35b75 100644
> > > --- a/diff.c
> > > +++ b/diff.c
> > > @@ -1281,7 +1281,9 @@ static void emit_diff_symbol_from_struct(struct diff_options *o,
> > >                       else if (c == '-')
> > >                               set = diff_get_color_opt(o, DIFF_FILE_OLD);
> > >               }
> > > -             emit_line_ws_markup(o, set_sign, set, reset, ' ', line, len,
> >                                     ^
> > Here we already pass `o`... so...
> >
> > > +             emit_line_ws_markup(o, set_sign, set, reset,
> > > +                                 o->output_indicators[OUTPUT_INDICATOR_CONTEXT],
> >                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > ... here, we could simply pass `OUTPUT_INDICATOR_CONTEXT` and let the
> > callee look it up in`o->output_indicators[]`...
> >
> > I read all three patches and did not see a reason why we could not
> > simplify the code that way.
> >
> > Other than that: great!
> 
> Thanks!
> 
> I considered it, but was put off by the (small) effort of yet another
> diff refactoring.
> 
> I'll include it in a resend if a resend is needed, otherwise
> I would suggest to make it a patch on top?