Re: [PATCH 03/15] add, reset: ensure submodules can be added or reset

[Brandon Williams <bmwill@google.com>]

On 07/25, Stefan Beller wrote:
> On Tue, Jul 25, 2017 at 2:39 PM, Brandon Williams <bmwill@google.com> wrote:
> > Commit aee9c7d65 (Submodules: Add the new "ignore" config option for
> > diff and status) ...
> 
> introduced in 2010, so quite widely spread.
> 
> > ...  introduced the ignore configuration option for
> > submodules so that configured submodules could be omitted from the
> > status and diff commands.  Because this flag is respected in the diff
> > machinery it has the unintended consequence of potentially prohibiting
> > users from adding or resetting a submodule, even when a path to the
> > submodule is explicitly given.
> >
> > Ensure that submodules can be added or set, even if they are configured
> > to be ignored, by setting the `DIFF_OPT_OVERRIDE_SUBMODULE_CONFIG` diff
> > flag.
> >
> > Signed-off-by: Brandon Williams <bmwill@google.com>
> > ---
> >  builtin/add.c   | 1 +
> >  builtin/reset.c | 1 +
> >  2 files changed, 2 insertions(+)
> >
> > diff --git a/builtin/add.c b/builtin/add.c
> > index e888fb8c5..6f271512f 100644
> > --- a/builtin/add.c
> > +++ b/builtin/add.c
> > @@ -116,6 +116,7 @@ int add_files_to_cache(const char *prefix,
> >         rev.diffopt.output_format = DIFF_FORMAT_CALLBACK;
> >         rev.diffopt.format_callback = update_callback;
> >         rev.diffopt.format_callback_data = &data;
> > +       rev.diffopt.flags |= DIFF_OPT_OVERRIDE_SUBMODULE_CONFIG;
> 
> 
> This flag occurs once in the code base, with the comment:
>     /*
>      * Unless the user did explicitly request a submodule
>      * ignore mode by passing a command line option we do
>      * not ignore any changed submodule SHA-1s when
>      * comparing index and parent, no matter what is
>      * configured. Otherwise we won't commit any
>      * submodules which were manually staged, which would
>      * be really confusing.
>      */
>     int diff_flags = DIFF_OPT_OVERRIDE_SUBMODULE_CONFIG;
> 
> in prepare_commit, so commit ignores the .gitmodules file.
> 
> This allows git-add to add ignored submodules, currently ignored submodules
> would have to be added using the plumbing
>     git update-index --add --cacheinfo 160000,$SHA1,<gitlink>
> 
> This makes sense, though a test demonstrating the change in behavior
> would be nice, but git-add doesn't seem to change as it doesn't even load
> the git modules config?

I can add a comment to the code but its already being tested in the
submodule test suite.  The only reason this doesn't cause any changes
now is that the gitmodules config is never loaded, but that may
change if we decide to allow lazy-loading of the gitmodules file (like
the last couple patches in this series do).

-- 
Brandon Williams