Christian Couder wrote:
> --- a/t/t3508-cherry-pick-many-commits.sh
> +++ b/t/t3508-cherry-pick-many-commits.sh
> @@ -92,4 +92,14 @@ test_expect_failure 'cherry-pick -3 fourth works' '
> test "$(git rev-parse --verify HEAD)" != "$(git rev-parse --verify fourth)"
> '
>
> +test_expect_success 'cherry-pick --stdin works' '
> + git checkout master &&
> + git reset --hard first &&
[...]
This test fails for me as written, since the previous test leaves some
files in an unmerged state. Patch 1 below works around that.
> --- a/builtin/revert.c
> +++ b/builtin/revert.c
> @@ -79,7 +80,7 @@ static void parse_args(int argc, const char **argv)
> }
>
> commit_argc = parse_options(argc, argv, NULL, options, usage_str, 0);
[...]
> @@ -527,10 +528,12 @@ static void prepare_revs(struct rev_info *revs)
> {
> int argc = 0;
> int i;
> - const char **argv = xmalloc((commit_argc + 4) * sizeof(*argv));
> + const char **argv = xmalloc((commit_argc + 5) * sizeof(*argv));
>
> argv[argc++] = NULL;
> argv[argc++] = "--no-walk";
> + if (read_stdin)
> + argv[argc++] = "--stdin";
Ah, I see the problem now. But it would be even nicer to allow arbitrary
rev-list options, so a person could ‘git cherry-pick --reverse a..b’,
for example.
In other words, how about something like patch 2 below?
Patch 3 is a small cleanup, as a bonus.
Christian Couder (1):
revert: accept arbitrary rev-list options
Jonathan Nieder (2):
t3508 (cherry-pick): futureproof against unmerged files
revert: do not rebuild argv on heap
Documentation/git-cherry-pick.txt | 7 ++++++
builtin/revert.c | 37 ++++++++++++++++++-----------------
t/t3501-revert-cherry-pick.sh | 18 +++++++++++++++++
t/t3508-cherry-pick-many-commits.sh | 24 ++++++++++++++++------
4 files changed, 61 insertions(+), 25 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html