Re: [PATCH 3/4] merge: make merge.renormalize work for all uses of merge machinery

[Elijah Newren <newren@gmail.com>]

On Sun, Aug 2, 2020 at 12:23 PM Junio C Hamano <gitster@pobox.com> wrote:
>
> "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes:
>
> > From: Elijah Newren <newren@gmail.com>
> >
> > The 'merge' command is not the only one that does merges; other commands
> > like checkout -m or rebase do as well.  Unfortunately, the only area of
> > the code that checked for the "merge.renormalize" config setting was in
> > builtin/merge.c, meaning it could only affect merges performed by the
> > "merge" command.  Move the handling of this config setting to
> > merge_recursive_config() so that other commands can benefit from it as
> > well.  Fixes a few tests in t6038.
>
> Makes sense, but.
>
> > @@ -771,13 +771,6 @@ static int merge_working_tree(const struct checkout_opts *opts,
> >                        */
> >
> >                       add_files_to_cache(NULL, NULL, 0);
> > -                     /*
> > -                      * NEEDSWORK: carrying over local changes
> > -                      * when branches have different end-of-line
> > -                      * normalization (or clean+smudge rules) is
> > -                      * a pain; plumb in an option to set
> > -                      * o.renormalize?
> > -                      */
> >                       init_merge_options(&o, the_repository);
>
> Now init_merge_options() takes care of setting o.renormalize, which
> does exactly what the comment wanted to see happen.
>
> > diff --git a/builtin/merge.c b/builtin/merge.c
> > index 7da707bf55..52f03ea715 100644
> > --- a/builtin/merge.c
> > +++ b/builtin/merge.c
> > @@ -72,7 +72,7 @@ static const char **xopts;
> >  static size_t xopts_nr, xopts_alloc;
> >  static const char *branch;
> >  static char *branch_mergeoptions;
> > -static int option_renormalize;
> > +static int option_renormalize = -1;
>
> Do we still need this file-scope static variable at all?
>
> > @@ -621,8 +621,6 @@ static int git_merge_config(const char *k, const char *v, void *cb)
> >               return git_config_string(&pull_octopus, k, v);
> >       else if (!strcmp(k, "commit.cleanup"))
> >               return git_config_string(&cleanup_arg, k, v);
> > -     else if (!strcmp(k, "merge.renormalize"))
> > -             option_renormalize = git_config_bool(k, v);
>
> We no longer set value to option_renormalize in git_merge_config()
> that is used only from cmd_merge().
>
> > @@ -721,7 +719,8 @@ static int try_merge_strategy(const char *stra
> >               if (!strcmp(strategy, "subtree"))
> >                       o.subtree_shift = "";
> >
> > -             o.renormalize = option_renormalize;
> > +             if (option_renormalize != -1)
> > +                     o.renormalize = option_renormalize;
>
> And if somebody sets option_renormalize, then (and only then) we
> override o.renormalize.  One line before the precontext of this hunk
> is a call to init_merge_options() that would have set o.renormalize
> by reading merge.renormalize configuration.  So there must be another
> place where option_renormalize comes from other than that configuration
> variable.
>
> But I do not see the variable mentioned anywhere else in the
> original   The assignment to it in git_merge_config() you just got
> rid of was the only one that used to assign to it.
>
> > diff --git a/merge-recursive.c b/merge-recursive.c
> > index 36948eafb7..a1c8b36ddb 100644
> > --- a/merge-recursive.c
> > +++ b/merge-recursive.c
> > @@ -3791,9 +3791,12 @@ int merge_recursive_generic(struct merge_options *opt,
> >  static void merge_recursive_config(struct merge_options *opt)
> >  {
> >       char *value = NULL;
> > +     int renormalize = 0;
> >       git_config_get_int("merge.verbosity", &opt->verbosity);
> >       git_config_get_int("diff.renamelimit", &opt->rename_limit);
> >       git_config_get_int("merge.renamelimit", &opt->rename_limit);
> > +     git_config_get_bool("merge.renormalize", &renormalize);
> > +     opt->renormalize = renormalize;
> >       if (!git_config_get_string("diff.renames", &value)) {
> >               opt->detect_renames = git_config_rename("diff.renames", value);
> >               free(value);
>
> OK.
>
> IOW, wouldn't the attached be a no-op code clean-up on top?
>
>  builtin/merge.c | 3 ---
>  1 file changed, 3 deletions(-)
>
> diff --git a/builtin/merge.c b/builtin/merge.c
> index 52f03ea715..74829a838e 100644
> --- a/builtin/merge.c
> +++ b/builtin/merge.c
> @@ -72,7 +72,6 @@ static const char **xopts;
>  static size_t xopts_nr, xopts_alloc;
>  static const char *branch;
>  static char *branch_mergeoptions;
> -static int option_renormalize = -1;
>  static int verbosity;
>  static int allow_rerere_auto;
>  static int abort_current_merge;
> @@ -719,8 +718,6 @@ static int try_merge_strategy(const char *strategy, struct commit_list *common,
>                 if (!strcmp(strategy, "subtree"))
>                         o.subtree_shift = "";
>
> -               if (option_renormalize != -1)
> -                       o.renormalize = option_renormalize;
>                 o.show_rename_progress =
>                         show_progress == -1 ? isatty(2) : show_progress;
>

Indeed; I'll squash that in for my next iteration.