Michael Schubert <msc...@elegosoft.com> writes:
> Currently, we allow refname components to start with a hyphen. There's
> no good reason to do so...
That is way too weak as a justification to potentially break
existing repositories.
Refusal upon attempted creation is probably OK, which is why the two
checks you removed in your patches are fine. I do not know if it is
justifiable to do that in check_refname_component() that is used in
the reading codepath.
> diff --git a/builtin/tag.c b/builtin/tag.c
> index 7b1be85..c99fb42 100644
> --- a/builtin/tag.c
> +++ b/builtin/tag.c
> @@ -403,9 +403,6 @@ static int parse_msg_arg(const struct option *opt, const
> char *arg, int unset)
>
> static int strbuf_check_tag_ref(struct strbuf *sb, const char *name)
> {
> - if (name[0] == '-')
> - return -1;
> -
> strbuf_reset(sb);
> strbuf_addf(sb, "refs/tags/%s", name);
>
> diff --git a/refs.c b/refs.c
> index da74a2b..5714681 100644
> --- a/refs.c
> +++ b/refs.c
> @@ -62,6 +62,8 @@ static int check_refname_component(const char *refname, int
> flags)
> if (refname[1] == '\0')
> return -1; /* Component equals ".". */
> }
> + if (refname[0] == '-')
> + return -1; /* Component starts with '-'. */
> if (cp - refname >= 5 && !memcmp(cp - 5, ".lock", 5))
> return -1; /* Refname ends with ".lock". */
> return cp - refname;
> diff --git a/sha1_name.c b/sha1_name.c
> index 5d81ea0..132d369 100644
> --- a/sha1_name.c
> +++ b/sha1_name.c
> @@ -892,8 +892,6 @@ int strbuf_branchname(struct strbuf *sb, const char *name)
> int strbuf_check_branch_ref(struct strbuf *sb, const char *name)
> {
> strbuf_branchname(sb, name);
> - if (name[0] == '-')
> - return -1;
> strbuf_splice(sb, 0, 0, "refs/heads/", 11);
> return check_refname_format(sb->buf, 0);
> }
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
