`fatal: cannot force update the branch ... checkout out at ...` is confusing when it isn't active
Consider this series of unfortunate commands:
```sh % (cd $(mktemp -d); git init; (touch a; git add a; git commit -m a; touch b; git add b; git commit -m b; git bisect start; git branch next HEAD~; git checkout next ) 2>/dev/null >/dev/null; git log --oneline -1 HEAD; git branch -f main HEAD;) Initialized empty Git repository in /private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf/.git/ accc238 (HEAD -> next) a fatal: cannot force update the branch 'main' checked out at '/private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf' ```
It's true git in the repository does have a pin of sorts for the main branch and that git-rebase would be very upset if the branch's location were changed, but as an end user, the main branch is not checked out, the next branch is checked out.
Quoted 13 lines · Josh Soref
Consider this series of unfortunate commands:
```sh % (cd $(mktemp -d); git init; (touch a; git add a; git commit -m a; touch b; git add b; git commit -m b; git bisect start; git branch next HEAD~; git checkout next ) 2>/dev/null >/dev/null; git log --oneline -1 HEAD; git branch -f main HEAD;) Initialized empty Git repository in /private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf/.git/ accc238 (HEAD -> next) a fatal: cannot force update the branch 'main' checked out at '/private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf' ```
Quoted 4 lines
It's true git in the repository does have a pin of sorts for the
mainbranch and thatgit-rebasewould be very upset if the branch's location were changed, but as an end user, themainbranch is not checked out, thenextbranch is checked out.
Fair enough. Perhaps "cannot force update the branch 'main' in use" would alleviate the confusion?
Quoted 7 lines · Josh Soref
fatal: cannot force update the branch 'main' checked out at '/private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf'
It's true git in the repository does have a pin of sorts for the
mainbranch and thatgit-rebasewould be very upset if the branch's location were changed, but as an end user, themainbranch is not checked out, thenextbranch is checked out.
Quoted 2 lines · Junio C Hamano
Fair enough. Perhaps "cannot force update the branch 'main' in use" would alleviate the confusion?
That would be better.
Even better would be to tell me what's using it (rebase). I'm not sure how possible that is.
Quoted 9 lines · Josh Soref
Quoted 7 lines · Josh Soref
fatal: cannot force update the branch 'main' checked out at '/private/var/folders/r3/n29fz25x72x191fdv6mhhr3m0000gp/T/tmp.fGD64HAf'
It's true git in the repository does have a pin of sorts for the
mainbranch and thatgit-rebasewould be very upset if the branch's location were changed, but as an end user, themainbranch is not checked out, thenextbranch is checked out.
Quoted 8 lines
Quoted 2 lines · Junio C Hamano
Fair enough. Perhaps "cannot force update the branch 'main' in use" would alleviate the confusion?
That would be better.
Even better would be to tell me what's using it (rebase). I'm not sure how possible that is.
When branch.c:prepare_checked_out() notices which branch is in use in what way in which worktree, it does check many conditions (e.g. "in this worktree, is a rebase in progress and if so which branch is being affected?") but the information is lost there to only record a mapping from branch names to worktree location. So "it is in use over there" with exact location is given, but "how it is used" is not recorded.
It shouldn't be a rocket surgery to teach the function to leave that information, but I am not offhand sure how valuable it would be to do so. After all, once you learn which worktree of yours is using the branch you wanted to touch, you'll either know already or it would be easy for you to find out what you have been doing in that other worktree anyway, no?
Re* [rebase] `fatal: cannot force update the branch ... checkout out at ...` is confusing when it isn't active
Quoted 6 lines · Junio C Hamano
It shouldn't be a rocket surgery to teach the function to leave that information, but I am not offhand sure how valuable it would be to do so. After all, once you learn which worktree of yours is using the branch you wanted to touch, you'll either know already or it would be easy for you to find out what you have been doing in that other worktree anyway, no?
So, before we forget, let's make a minimum patch and let the list archive it.
----- >8 --------- >8 --------- >8 --------- >8 --------- >8 -----
Subject: [PATCH] branch: update the message to refuse touching a branch in-use
The "git branch -f" command can refuse to force-update a branch that is in use in another worktree. The original rationale for this behaviour was that updating a branch that is checked out in another worktree, without making a matching change to the index and the working tree files in that worktree, will lead to a very confused user. "git diff HEAD" will no longer give a useful patch, because HEAD is a commit unrelated to what the index and the working tree in the worktree were based on, for example.
These days, however, the same mechanism also protects branches that are being rebased or bisected, and the same machanism is expected to be the right place to add more checks, when we decide to protect branches undergoing other kinds of operations. We forgot to rethink the messaging, which originally said that we are refusing to touch the branch because it is "checked out" elsewhere, when d2ba271a (branch: check for bisects and rebases, 2022-06-14) started to protect branches that are being rebased or bisected.
The spirit of the check is that we do not want to disrupt the use of the same branch in other worktrees. Let's reword the message slightly to say that the branch is "in use", instead of "checked out".
We could teach the branch.c:prepare_checked_out_branches() function to remember why it decided that a particular branch needs protecting (i.e. was it because it was checked out? being bisected? something else?) in addition to which worktree the branch was in use, and use that in the error message to say "you cannot force update this branch because it is being bisected in the worktree X", etc., but it is dubious that such extra complexity is worth it. The message already tells which directory the worktree in question is, and it should be just a "chdir" away for the user to find out what state it is in, if the user felt curious enough. So let's not go there yet.
--- branch.c | 2 +- t/t2407-worktree-heads.sh | 10 +++++----- 2 files changed, 6 insertions(+), 6 deletions(-)
branch.c +1 -1
@@ -471,7 +471,7 @@ int validate_new_branchname(const char *name, struct strbuf *ref, int force)if ((path =branch_checked_out (ref->buf)))die (_ ("cannot force update the branch '%s' " -"checked out at '%s'" ), +"in use at '%s'" ), ref->buf +strlen ("refs/heads/" ), path);return 1 ;
t/t2407-worktree-heads.sh +5 -5
@@ -58,7 +58,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in bisect' ' git -C wt-4 bisect good wt-1 && test_must_fail git branch -f wt-4 HEAD 2>err && - grep "cannot force update the branch '\''wt-4'\'' checked out at.*wt-4" err + grep "cannot force update the branch '\''wt-4'\'' in use at.*wt-4" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (apply)' '
@@ -68,7 +68,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (app test_must_fail git -C wt-2 rebase --apply conflict-2 && test_must_fail git branch -f wt-2 HEAD 2>err && - grep "cannot force update the branch '\''wt-2'\'' checked out at.*wt-2" err + grep "cannot force update the branch '\''wt-2'\'' in use at.*wt-2" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (merge)' '
@@ -78,7 +78,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (mer test_must_fail git -C wt-2 rebase conflict-2 && test_must_fail git branch -f wt-2 HEAD 2>err && - grep "cannot force update the branch '\''wt-2'\'' checked out at.*wt-2" err + grep "cannot force update the branch '\''wt-2'\'' in use at.*wt-2" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase with --update-refs' '
@@ -90,7 +90,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase with for i in 3 4 do test_must_fail git branch -f can-be-updated HEAD 2>err && - grep "cannot force update the branch '\''can-be-updated'\'' checked out at.*wt-3" err || + grep "cannot force update the branch '\''can-be-updated'\'' in use at.*wt-3" err || return 1 done '
@@ -150,7 +150,7 @@ test_expect_success 'refuse to overwrite when in error states' ' for i in 1 2 do test_must_fail git branch -f fake-$i HEAD 2>err && - grep "cannot force update the branch '\''fake-$i'\'' checked out at" err || + grep "cannot force update the branch '\''fake-$i'\'' in use at" err || return 1 done '
Signature · 1 line
2.41.0-327-gaa9166bcc0
PATCH v2 1/1 branch: update the message to refuse touching a branch in-use
The "git branch -f" command can refuse to force-update a branch that is used by another worktree. The original rationale for this behaviour was that updating a branch that is checked out in another worktree, without making a matching change to the index and the working tree files in that worktree, will lead to a very confused user. "git diff HEAD" will no longer give a useful patch, because HEAD is a commit unrelated to what the index and the working tree in the worktree were based on, for example.
These days, the same mechanism also protects branches that are being rebased or bisected, and the same machanism is expected to be the right place to add more checks, when we decide to protect branches undergoing other kinds of operations. We however forgot to rethink the messaging, which originally said that we are refusing to touch the branch because it is "checked out" elsewhere, when d2ba271a (branch: check for bisects and rebases, 2022-06-14) started to protect branches that are being rebased or bisected.
The spirit of the check has always been that we do not want to disrupt the use of the same branch in other worktrees. Let's reword the message slightly to say that the branch is "used by" another worktree, instead of "checked out".
We could teach the branch.c:prepare_checked_out_branches() function to remember why it decided that a particular branch needs protecting (i.e. was it because it was checked out? being bisected? something else?) in addition to which worktree the branch was in use, and use that in the error message to say "you cannot force update this branch because it is being bisected in the worktree X", etc., but it is dubious that such extra complexity is worth it. The message already tells which directory the worktree in question is, and it should be just a "chdir" away for the user to find out what state it is in, if the user felt curious enough. So let's not go there yet.
---
* Instead of saying "branch in use", use a bit more natural "branch is used by ...".
branch.c | 2 +- t/t2407-worktree-heads.sh | 10 +++++----- 2 files changed, 6 insertions(+), 6 deletions(-)
branch.c +1 -1
@@ -471,7 +471,7 @@ int validate_new_branchname(const char *name, struct strbuf *ref, int force)if ((path =branch_checked_out (ref->buf)))die (_ ("cannot force update the branch '%s' " -"checked out at '%s'" ), +"used by worktree at '%s'" ), ref->buf +strlen ("refs/heads/" ), path);return 1 ;
t/t2407-worktree-heads.sh +5 -5
@@ -58,7 +58,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in bisect' ' git -C wt-4 bisect good wt-1 && test_must_fail git branch -f wt-4 HEAD 2>err && - grep "cannot force update the branch '\''wt-4'\'' checked out at.*wt-4" err + grep "cannot force update the branch '\''wt-4'\'' used by worktree at.*wt-4" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (apply)' '
@@ -68,7 +68,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (app test_must_fail git -C wt-2 rebase --apply conflict-2 && test_must_fail git branch -f wt-2 HEAD 2>err && - grep "cannot force update the branch '\''wt-2'\'' checked out at.*wt-2" err + grep "cannot force update the branch '\''wt-2'\'' used by worktree at.*wt-2" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (merge)' '
@@ -78,7 +78,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase (mer test_must_fail git -C wt-2 rebase conflict-2 && test_must_fail git branch -f wt-2 HEAD 2>err && - grep "cannot force update the branch '\''wt-2'\'' checked out at.*wt-2" err + grep "cannot force update the branch '\''wt-2'\'' used by worktree at.*wt-2" err ' test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase with --update-refs' '
@@ -90,7 +90,7 @@ test_expect_success !SANITIZE_LEAK 'refuse to overwrite: worktree in rebase with for i in 3 4 do test_must_fail git branch -f can-be-updated HEAD 2>err && - grep "cannot force update the branch '\''can-be-updated'\'' checked out at.*wt-3" err || + grep "cannot force update the branch '\''can-be-updated'\'' used by worktree at.*wt-3" err || return 1 done '
@@ -150,7 +150,7 @@ test_expect_success 'refuse to overwrite when in error states' ' for i in 1 2 do test_must_fail git branch -f fake-$i HEAD 2>err && - grep "cannot force update the branch '\''fake-$i'\'' checked out at" err || + grep "cannot force update the branch '\''fake-$i'\'' used by worktree at" err || return 1 done '
Signature · 1 line
2.41.0-376-gcba07a324d