diff mbox series

checkout: special case error messages during noop switching

Message ID xmqqikxnqzz4.fsf@gitster.g (mailing list archive)
State Accepted
Commit d1e6c61272737926cd049a6126431fe9c12e4a54
Headers show
Series checkout: special case error messages during noop switching | expand

Commit Message

Junio C Hamano July 2, 2024, 8:51 p.m. UTC
"git checkout" ran with no branch and no pathspec behaves like
switching the branch to the current branch (in other words, a
no-op, except that it gives a side-effect "here are the modified
paths" report).  But unlike "git checkout HEAD" or "git checkout
main" (when you are on the 'main' branch), the user is much less
conscious that they are "switching" to the current branch.

This twists end-user expectation in a strange way.  There are
options (like "--ours") that make sense only when we are checking
out paths out of either the tree-ish or out of the index.  So the
error message the command below gives

    $ git checkout --ours
    fatal: '--ours/theirs' cannot be used with switching branches

is technically correct, but because the end-user may not even be
aware of the fact that the command they are issuing is about no-op
branch switching [*], they may find the error confusing.

Let's refactor the code to make it easier to special case the "no-op
branch switching" situation, and then customize the exact error
message for "--ours/--theirs".  Since it is more likely that the
end-user forgot to give pathspec that is required by the option,
let's make it say

    $ git checkout --ours
    fatal: '--ours/theirs' needs the paths to check out

instead.

Among the other options that are incompatible with branch switching,
there may be some that benefit by having messages tweaked when a
no-op branch switching is done, but I'll leave them as #leftoverbits
material.

[Footnote]

 * Yes, the end-users are irrational.  When they did not give
   "--ours", they take it granted that "git checkout" gives a short
   status, e.g..

    $ git checkout
    M	builtin/checkout.c
    M	t/t7201-co.sh

   exactly as a branch switching command.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 builtin/checkout.c | 21 ++++++++++++++-------
 t/t7201-co.sh      | 13 +++++++++++++
 2 files changed, 27 insertions(+), 7 deletions(-)

Comments

Martin von Zweigbergk July 17, 2024, 4:05 p.m. UTC | #1
Thanks! (This is a fix for a bug I reported internally at work.)

On Tue, Jul 2, 2024 at 10:52 PM Junio C Hamano <gitster@pobox.com> wrote:
>
> "git checkout" ran with no branch and no pathspec behaves like
> switching the branch to the current branch (in other words, a
> no-op, except that it gives a side-effect "here are the modified
> paths" report).  But unlike "git checkout HEAD" or "git checkout
> main" (when you are on the 'main' branch), the user is much less
> conscious that they are "switching" to the current branch.

Yes, that's exactly what happened to me. I should have used `git
restore` instead. I know that's the modern way of updating paths, so I
don't know why I didn't think of it. I just verified that `git restore
--ours <path>` works for restoring a conflicted file to my side.

> [Footnote]
>
>  * Yes, the end-users are irrational.  When they did not give
>    "--ours", they take it granted that "git checkout" gives a short
>    status, e.g..

I actually did not even know that it does that :) I'm a bit surprised
that it does, especially since `git checkout <non-HEAD>` doesn't seem
to do that. But that's off topic.
Junio C Hamano July 17, 2024, 4:15 p.m. UTC | #2
Martin von Zweigbergk <martinvonz@gmail.com> writes:

>>  * Yes, the end-users are irrational.  When they did not give
>>    "--ours", they take it granted that "git checkout" gives a short
>>    status, e.g..
>
> I actually did not even know that it does that :) I'm a bit surprised
> that it does, especially since `git checkout <non-HEAD>` doesn't seem
> to do that. But that's off topic.

It does but you wouldn't know unless you have modified paths.

Also checkout and switch would only allow checking out another
branch when these modified paths are the same between the original
HEAD and the other branch.

        $ git reset --hard
        $ echo >>GIT-VERSION-GEN
        $ git checkout
	M       GIT-VERSION-GEN

With a modified file, no-op checkout, report only.

	$ git checkout next
	M       GIT-VERSION-GEN
	Switched to branch 'next'

With a modified file, checkout another branch, with report.

	$ git checkout maint
	error: Your local changes to the following files would be overwritten...
		GIT-VERSION-GEN
	Please commit your changes or stash them...

With a modified file, that is different between next and maint, fail
to checkout the other branch.
Junio C Hamano July 17, 2024, 11:38 p.m. UTC | #3
Martin von Zweigbergk <martinvonz@gmail.com> writes:

> Thanks! (This is a fix for a bug I reported internally at work.)

Glad if this worked for you.

I'll mark the topic for 'next', but being so late in the cycle, one
day before 2.46-rc1 gets tagged, it is unlikely that it will become
a part of the upcoming release.

Thanks.
diff mbox series

Patch

diff --git a/builtin/checkout.c b/builtin/checkout.c
index 3cf44b4683..1748d68c96 100644
--- a/builtin/checkout.c
+++ b/builtin/checkout.c
@@ -1572,6 +1572,10 @@  static void die_if_switching_to_a_branch_in_use(struct checkout_opts *opts,
 static int checkout_branch(struct checkout_opts *opts,
 			   struct branch_info *new_branch_info)
 {
+	int noop_switch = (!new_branch_info->name &&
+			   !opts->new_branch &&
+			   !opts->force_detach);
+
 	if (opts->pathspec.nr)
 		die(_("paths cannot be used with switching branches"));
 
@@ -1583,9 +1587,14 @@  static int checkout_branch(struct checkout_opts *opts,
 		die(_("'%s' cannot be used with switching branches"),
 		    "--[no]-overlay");
 
-	if (opts->writeout_stage)
-		die(_("'%s' cannot be used with switching branches"),
-		    "--ours/--theirs");
+	if (opts->writeout_stage) {
+		const char *msg;
+		if (noop_switch)
+			msg = _("'%s' needs the paths to check out");
+		else
+			msg = _("'%s' cannot be used with switching branches");
+		die(msg, "--ours/--theirs");
+	}
 
 	if (opts->force && opts->merge)
 		die(_("'%s' cannot be used with '%s'"), "-f", "-m");
@@ -1612,10 +1621,8 @@  static int checkout_branch(struct checkout_opts *opts,
 		die(_("Cannot switch branch to a non-commit '%s'"),
 		    new_branch_info->name);
 
-	if (!opts->switch_branch_doing_nothing_is_ok &&
-	    !new_branch_info->name &&
-	    !opts->new_branch &&
-	    !opts->force_detach)
+	if (noop_switch &&
+	    !opts->switch_branch_doing_nothing_is_ok)
 		die(_("missing branch or commit argument"));
 
 	if (!opts->implicit_detach &&
diff --git a/t/t7201-co.sh b/t/t7201-co.sh
index 42352dc0db..793da6e64e 100755
--- a/t/t7201-co.sh
+++ b/t/t7201-co.sh
@@ -497,6 +497,19 @@  test_expect_success 'checkout unmerged stage' '
 	test ztheirside = "z$(cat file)"
 '
 
+test_expect_success 'checkout --ours is incompatible with switching' '
+	test_must_fail git checkout --ours 2>error &&
+	test_grep "needs the paths to check out" error &&
+
+	test_must_fail git checkout --ours HEAD 2>error &&
+	test_grep "cannot be used with switching" error &&
+
+	test_must_fail git checkout --ours main 2>error &&
+	test_grep "cannot be used with switching" error &&
+
+	git checkout --ours file
+'
+
 test_expect_success 'checkout path with --merge from tree-ish is a no-no' '
 	setup_conflicting_index &&
 	test_must_fail git checkout -m HEAD -- file