On 05/06/2020 19:03, Han-Wen Nienhuys via GitGitGadget wrote:
> From: Han-Wen Nienhuys <hanwen@xxxxxxxxxx>
>
> Signed-off-by: Han-Wen Nienhuys <hanwen@xxxxxxxxxx>
> ---
> refs.c | 2 +-
> refs.h | 2 ++
> 2 files changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/refs.c b/refs.c
> index 12908066b13..812fee47108 100644
> --- a/refs.c
> +++ b/refs.c
> @@ -311,7 +311,7 @@ int read_ref(const char *refname, struct object_id *oid)
> return read_ref_full(refname, RESOLVE_REF_READING, oid, NULL);
> }
>
> -static int refs_ref_exists(struct ref_store *refs, const char *refname)
> +int refs_ref_exists(struct ref_store *refs, const char *refname)
> {
> return !!refs_resolve_ref_unsafe(refs, refname, RESOLVE_REF_READING, NULL, NULL);
> }
It is a shame that ref_exists() does not take a struct repository. The
code in the following patches would be pleasanter if we could write
ref_exists(r, ref)
rather than
refs_ref_exists(get_main_ref_store(r), ref)
all the time
Maybe we should think about updating the existing callers to
ref_exists() to pass a struct repository.
The existing code is inconsistent about its repository handling - we
create the refs with update_ref() which operates on the main repository
but when checking their existence and deleting them we use a path which
depends on the repository. I've realized now the answer to my question
about using delete_ref() in my reply to the previous patch - it does not
take a repository - maybe it should along with update_ref() but that
might be more work than you want to do though.
Best Wishes
Phillip
> diff --git a/refs.h b/refs.h
> index 4dad8f24914..7aaa1226551 100644
> --- a/refs.h
> +++ b/refs.h
> @@ -105,6 +105,8 @@ int refs_verify_refname_available(struct ref_store *refs,
> const struct string_list *skip,
> struct strbuf *err);
>
> +int refs_ref_exists(struct ref_store *refs, const char *refname);
> +
> int ref_exists(const char *refname);
>
> int should_autocreate_reflog(const char *refname);
>
