Hi Christian,
On Wed, 23 Sep 2020, Christian Couder wrote:
> In 06f5608c14 (bisect--helper: `bisect_start` shell function
> partially in C, 2019-01-02), we changed the following shell
> code:
>
> - rev=$(git rev-parse -q --verify "$arg^{commit}") || {
> - test $has_double_dash -eq 1 &&
> - die "$(eval_gettext "'\$arg' does not appear to be a valid revision")"
> - break
> - }
> - revs="$revs $rev"
>
> into:
>
> + char *commit_id = xstrfmt("%s^{commit}", arg);
> + if (get_oid(commit_id, &oid) && has_double_dash)
> + die(_("'%s' does not appear to be a valid "
> + "revision"), arg);
> +
> + string_list_append(&revs, oid_to_hex(&oid));
> + free(commit_id);
>
> In case of an invalid "arg" when "has_double_dash" is false, the old
> code would "break" out of the argument loop.
>
> In the new C code though, `oid_to_hex(&oid)` is unconditonally
> appended to "revs". This is wrong first because "oid" is junk as
> `get_oid(commit_id, &oid)` failed and second because it doesn't break
> out of the argument loop.
>
> Not breaking out of the argument loop means that "arg" is then not
> treated as a path restriction (which is wrong).
Good catch!
> This is a bug fix for the bug Miriam talks about in:
>
> https://lore.kernel.org/git/20200923072740.20772-1-mirucam@xxxxxxxxx/
>
> and:
>
> https://lore.kernel.org/git/CAN7CjDDVp_i7dhpbAq5zrGW69nE6+SfivJQ-dembmu+WyqKiQQ@xxxxxxxxxxxxxx/
Thank you for clarifying.
> diff --git a/builtin/bisect--helper.c b/builtin/bisect--helper.c
> index 7dcc1b5188..538fa6f16b 100644
> --- a/builtin/bisect--helper.c
> +++ b/builtin/bisect--helper.c
> @@ -486,12 +486,16 @@ static int bisect_start(struct bisect_terms *terms, const char **argv, int argc)
> return error(_("unrecognized option: '%s'"), arg);
> } else {
> char *commit_id = xstrfmt("%s^{commit}", arg);
> - if (get_oid(commit_id, &oid) && has_double_dash)
> - die(_("'%s' does not appear to be a valid "
> - "revision"), arg);
> -
> - string_list_append(&revs, oid_to_hex(&oid));
> + int res = get_oid(commit_id, &oid);
> free(commit_id);
> + if (res) {
> + if (has_double_dash)
> + die(_("'%s' does not appear to be a valid "
> + "revision"), arg);
> + break;
> + } else {
> + string_list_append(&revs, oid_to_hex(&oid));
> + }
I would find that a lot easier to read if it was reordered thusly:
if (!get_oidf(&oid, "%s^{commit}", arg))
string_list_append(&revs, oid_to_hex(&oid));
else if (!has_double_dash)
break;
else
die(_("'%s' does not appear to be a valid "
revision"), arg);
And it would actually probably make sense to replace the `get_oid()` by
`get_oid_committish()` in the first place.
> }
> }
> pathspec_pos = i;
> diff --git a/t/t6030-bisect-porcelain.sh b/t/t6030-bisect-porcelain.sh
> index b886529e59..cb645cf8c8 100755
> --- a/t/t6030-bisect-porcelain.sh
> +++ b/t/t6030-bisect-porcelain.sh
> @@ -82,6 +82,13 @@ test_expect_success 'bisect fails if given any junk instead of revs' '
> git bisect bad $HASH4
> '
>
> +test_expect_success 'bisect start without -- uses unknown stuff as path restriction' '
s/stuff/arg/ maybe?
The rest of the patch looks good to me.
Thank you,
Dscho
> + git bisect reset &&
> + git bisect start foo bar &&
> + grep foo ".git/BISECT_NAMES" &&
> + grep bar ".git/BISECT_NAMES"
> +'
> +
> test_expect_success 'bisect reset: back in the master branch' '
> git bisect reset &&
> echo "* master" > branch.expect &&
> --
> 2.28.0.587.g1c7fdf1d8b
>
>
