checkout: die() on ambiguous tracking branches #477
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
SyntevoAlex
changed the title
|
/submit |
|
Submitted as [email protected] |
![gitgitgadget[bot]](https://avatars.githubusercontent.com/in/12836?s=60&v=4){kind=small}
| @@ -1113,12 +1113,43 @@ static void setup_new_branch_info_and_source_tree( | |||
| } | |||
There was a problem hiding this comment.
On the Git mailing list, Ævar Arnfjörð Bjarmason wrote (reply to this):
On Wed, Nov 27 2019, Alexandr Miloslavskiy via GitGitGadget wrote:
> From: Alexandr Miloslavskiy <[email protected]>
>
> Before this patch, when there were multiple DWIM candidates for remote
> branch, git decided to try the argument as pathspec instead. I believe
> that such behavior is a surprise: adding another remote suddenly causes
> git to discard file contents, because it was unsure which branch to
> pick. There was an incomplete attempt to prevent that in [3].
>
> I understand that this was never intended:
>
> [1] introduces the unexpected behavior. Before, there was fallback
> from not-a-ref to pathspec. This is reasonable DWIM. After, there is
> another fallback from ambiguous-remote to pathspec. I understand that
> it was kind of copy&paste oversight.
>
> [2] noticed the unexpected behavior but chose to semi-document it
> instead of forbidding, because the goal of the patch series was
> focused on something else.
>
> [3] adds `die()` when there is ambiguity between branch and file. The
> case of multiple tracking branches is seemingly overlooked.
>
> Change to complain about ambiguity instead of doing unexpected things.
>
> [1] Commit 70c9ac2f ("DWIM "git checkout frotz" to "git checkout -b frotz origin/frotz"" 2009-10-18)
> https://public-inbox.org/git/[email protected]/
> [2] Commit ad8d5104 ("checkout: add advice for ambiguous "checkout <branch>"", 2018-06-05)
> https://public-inbox.org/git/[email protected]/
> [3] Commit be4908f1 ("checkout: disambiguate dwim tracking branches and local files", 2018-11-13)
> https://public-inbox.org/git/[email protected]/
I'll reserve judgement on whether we really should do this for now, my
current opinion on the matter is undefined as I haven't re-paged this
behavior of checkout into my brain.
But a giant red flag here for me is that you say "I understand that this
was never intended".
Just from a cursory look at this that's not true, for better or worse it
*is* intended behavior. Most of the code you're moving around here is
what I added in ad8d5104b4 ("checkout: add advice for ambiguous
"checkout <branch>"", 2018-06-05), and the very start of that commit
message refers to the checkout documentation we have that explicitly
documents this edge case.
Digging a bit further reveals that we've had this behavior (again,
intended, not emergent) since 70c9ac2f19 ("DWIM "git checkout frotz" to
"git checkout -b frotz origin/frotz"", 2009-10-18), and had it
documented since 00bb4378c7 ("Documentation/git-checkout.txt: document
70c9ac2 behavior", 2012-12-17).
So at the very least I'd say you need a v2 where you amend the relevant
docs & commit message to make a case to the effect of "we've had this
since 2009, but it was never really all that important etc.".
Such a change should also be changing the docs etc. added in 8d7b558bae
("checkout & worktree: introduce checkout.defaultRemote",
2018-06-05). With this series our docs don't make a lot of sense anymore
& don't describe the behavior with the patches applied.
> Signed-off-by: Alexandr Miloslavskiy <[email protected]>
> ---
> builtin/checkout.c | 56 ++++++++++++++++++----------------------
> t/t2024-checkout-dwim.sh | 23 +++++++++++++++--
> 2 files changed, 46 insertions(+), 33 deletions(-)
>
> diff --git a/builtin/checkout.c b/builtin/checkout.c
> index e1b9df1543..6fb427990f 100644
> --- a/builtin/checkout.c
> +++ b/builtin/checkout.c
> @@ -1115,10 +1115,10 @@ static void setup_new_branch_info_and_source_tree(
>
> static const char *parse_remote_branch(const char *arg,
> struct object_id *rev,
> - int could_be_checkout_paths,
> - int *dwim_remotes_matched)
> + int could_be_checkout_paths)
> {
> - const char *remote = unique_tracking_name(arg, rev, dwim_remotes_matched);
> + int num_matches = 0;
> + const char *remote = unique_tracking_name(arg, rev, &num_matches);
>
> if (remote && could_be_checkout_paths) {
> die(_("'%s' could be both a local file and a tracking branch.\n"
> @@ -1126,6 +1126,22 @@ static const char *parse_remote_branch(const char *arg,
> arg);
> }
>
> + if (!remote && (num_matches > 1)) {
> + if (advice_checkout_ambiguous_remote_branch_name) {
> + advise(_("If you meant to check out a remote tracking branch on, e.g. 'origin',\n"
> + "you can do so by fully qualifying the name with the --track option:\n"
> + "\n"
> + " git checkout --track origin/<name>\n"
> + "\n"
> + "If you'd like to always have checkouts of an ambiguous <name> prefer\n"
> + "one remote, e.g. the 'origin' remote, consider setting\n"
> + "checkout.defaultRemote=origin in your config."));
> + }
> +
> + die(_("'%s' matched multiple (%d) remote tracking branches"),
> + arg, num_matches);
> + }
> +
> return remote;
> }
>
> @@ -1133,8 +1149,7 @@ static int parse_branchname_arg(int argc, const char **argv,
> int dwim_new_local_branch_ok,
> struct branch_info *new_branch_info,
> struct checkout_opts *opts,
> - struct object_id *rev,
> - int *dwim_remotes_matched)
> + struct object_id *rev)
> {
> const char **new_branch = &opts->new_branch;
> int argcount = 0;
> @@ -1240,8 +1255,7 @@ static int parse_branchname_arg(int argc, const char **argv,
>
> if (recover_with_dwim) {
> const char *remote = parse_remote_branch(arg, rev,
> - could_be_checkout_paths,
> - dwim_remotes_matched);
> + could_be_checkout_paths);
> if (remote) {
> *new_branch = arg;
> arg = remote;
> @@ -1505,7 +1519,6 @@ static int checkout_main(int argc, const char **argv, const char *prefix,
> const char * const usagestr[])
> {
> struct branch_info new_branch_info;
> - int dwim_remotes_matched = 0;
> int parseopt_flags = 0;
>
> memset(&new_branch_info, 0, sizeof(new_branch_info));
> @@ -1613,8 +1626,7 @@ static int checkout_main(int argc, const char **argv, const char *prefix,
> opts->track == BRANCH_TRACK_UNSPECIFIED &&
> !opts->new_branch;
> int n = parse_branchname_arg(argc, argv, dwim_ok,
> - &new_branch_info, opts, &rev,
> - &dwim_remotes_matched);
> + &new_branch_info, opts, &rev);
> argv += n;
> argc -= n;
> } else if (!opts->accept_ref && opts->from_treeish) {
> @@ -1672,28 +1684,10 @@ static int checkout_main(int argc, const char **argv, const char *prefix,
> }
>
> UNLEAK(opts);
> - if (opts->patch_mode || opts->pathspec.nr) {
> - int ret = checkout_paths(opts, new_branch_info.name);
> - if (ret && dwim_remotes_matched > 1 &&
> - advice_checkout_ambiguous_remote_branch_name)
> - advise(_("'%s' matched more than one remote tracking branch.\n"
> - "We found %d remotes with a reference that matched. So we fell back\n"
> - "on trying to resolve the argument as a path, but failed there too!\n"
> - "\n"
> - "If you meant to check out a remote tracking branch on, e.g. 'origin',\n"
> - "you can do so by fully qualifying the name with the --track option:\n"
> - "\n"
> - " git checkout --track origin/<name>\n"
> - "\n"
> - "If you'd like to always have checkouts of an ambiguous <name> prefer\n"
> - "one remote, e.g. the 'origin' remote, consider setting\n"
> - "checkout.defaultRemote=origin in your config."),
> - argv[0],
> - dwim_remotes_matched);
> - return ret;
> - } else {
> + if (opts->patch_mode || opts->pathspec.nr)
> + return checkout_paths(opts, new_branch_info.name);
> + else
> return checkout_branch(opts, &new_branch_info);
> - }
> }
>
> int cmd_checkout(int argc, const char **argv, const char *prefix)
> diff --git a/t/t2024-checkout-dwim.sh b/t/t2024-checkout-dwim.sh
> index fa0718c730..707c88ceba 100755
> --- a/t/t2024-checkout-dwim.sh
> +++ b/t/t2024-checkout-dwim.sh
> @@ -37,7 +37,9 @@ test_expect_success 'setup' '
> git checkout -b foo &&
> test_commit a_foo &&
> git checkout -b bar &&
> - test_commit a_bar
> + test_commit a_bar &&
> + git checkout -b ambiguous_branch_and_file &&
> + test_commit a_ambiguous_branch_and_file
> ) &&
> git init repo_b &&
> (
> @@ -46,7 +48,9 @@ test_expect_success 'setup' '
> git checkout -b foo &&
> test_commit b_foo &&
> git checkout -b baz &&
> - test_commit b_baz
> + test_commit b_baz &&
> + git checkout -b ambiguous_branch_and_file &&
> + test_commit b_ambiguous_branch_and_file
> ) &&
> git remote add repo_a repo_a &&
> git remote add repo_b repo_b &&
> @@ -75,6 +79,21 @@ test_expect_success 'checkout of branch from multiple remotes fails #1' '
> test_branch master
> '
>
> +test_expect_success 'when arg matches multiple remotes, do not fallback to interpreting as pathspec' '
> + git checkout -b t_ambiguous_branch_and_file &&
> + >ambiguous_branch_and_file &&
> + git add ambiguous_branch_and_file &&
> + git commit -m "ambiguous_branch_and_file" &&
> +
> + test_when_finished "git checkout -- ambiguous_branch_and_file" &&
> + echo "file contents" >ambiguous_branch_and_file &&
> + cp ambiguous_branch_and_file expect &&
> +
> + test_must_fail git checkout ambiguous_branch_and_file &&
> +
> + test_cmp expect ambiguous_branch_and_file
> +'
> +
> test_expect_success 'checkout of branch from multiple remotes fails with advice' '
> git checkout -B master &&
> test_might_fail git branch -D foo &&
There was a problem hiding this comment.
On the Git mailing list, Alexandr Miloslavskiy wrote (reply to this):
On 27.11.2019 16:43, Ævar Arnfjörð Bjarmason wrote:
> I'll reserve judgement on whether we really should do this for now, my
> current opinion on the matter is undefined as I haven't re-paged this
> behavior of checkout into my brain.
>
> But a giant red flag here for me is that you say "I understand that this
> was never intended".
>
> Just from a cursory look at this that's not true, for better or worse it
> *is* intended behavior. Most of the code you're moving around here is
> what I added in ad8d5104b4 ("checkout: add advice for ambiguous
> "checkout <branch>"", 2018-06-05), and the very start of that commit
> message refers to the checkout documentation we have that explicitly
> documents this edge case.
>
> Digging a bit further reveals that we've had this behavior (again,
> intended, not emergent) since 70c9ac2f19 ("DWIM "git checkout frotz" to
> "git checkout -b frotz origin/frotz"", 2009-10-18), and had it
> documented since 00bb4378c7 ("Documentation/git-checkout.txt: document
> 70c9ac2 behavior", 2012-12-17).
>
> So at the very least I'd say you need a v2 where you amend the relevant
> docs & commit message to make a case to the effect of "we've had this
> since 2009, but it was never really all that important etc.".
>
I'm sorry, can you please clarify?
My patch addresses the situation where there are *multiple* remote
candidates *and* a file with the same name.
My feeling is that in this case, reverting a file is an unintended surprise.
I understand previous patches this way:
ad8d5104 - patch series is mostly for "if there are *multiple* remotes,
disambiguate via checkout.defaultRemote". This essentially
converts the case of multiple remotes into a single remote.
However, this also semi-documents what I'm now preventing.
Was that really intended?
70c9ac2f - if there is *one* remote, DWIM it.
This isn't what I'm changing.
be4908f1 - if there is *one* remote *and* file, die().
This is what I'm extending further.
If the prevented behavior is documented, could you please quote it
explicitly?
> Such a change should also be changing the docs etc. added in 8d7b558bae
> ("checkout & worktree: introduce checkout.defaultRemote",
> 2018-06-05). With this series our docs don't make a lot of sense anymore
> & don't describe the behavior with the patches applied.
To my understanding, my patch doesn't affect those pieces of docs?
Please correct me if I'm wrong.
There was a problem hiding this comment.
On the Git mailing list, Alexandr Miloslavskiy wrote (reply to this):
Ævar, please let me know if your concern is resolved?
On 27.11.2019 17:09, Alexandr Miloslavskiy wrote:
>
>
> On 27.11.2019 16:43, Ævar Arnfjörð Bjarmason wrote:
>> I'll reserve judgement on whether we really should do this for now, my
>> current opinion on the matter is undefined as I haven't re-paged this
>> behavior of checkout into my brain.
>>
>> But a giant red flag here for me is that you say "I understand that this
>> was never intended".
>>
>> Just from a cursory look at this that's not true, for better or worse it
>> *is* intended behavior. Most of the code you're moving around here is
>> what I added in ad8d5104b4 ("checkout: add advice for ambiguous
>> "checkout <branch>"", 2018-06-05), and the very start of that commit
>> message refers to the checkout documentation we have that explicitly
>> documents this edge case.
>>
>> Digging a bit further reveals that we've had this behavior (again,
>> intended, not emergent) since 70c9ac2f19 ("DWIM "git checkout frotz" to
>> "git checkout -b frotz origin/frotz"", 2009-10-18), and had it
>> documented since 00bb4378c7 ("Documentation/git-checkout.txt: document
>> 70c9ac2 behavior", 2012-12-17).
>>
>> So at the very least I'd say you need a v2 where you amend the relevant
>> docs & commit message to make a case to the effect of "we've had this
>> since 2009, but it was never really all that important etc.".
>>
>
> I'm sorry, can you please clarify?
>
> My patch addresses the situation where there are *multiple* remote
> candidates *and* a file with the same name.
>
> My feeling is that in this case, reverting a file is an unintended
> surprise.
>
> I understand previous patches this way:
> ad8d5104 - patch series is mostly for "if there are *multiple* remotes,
> disambiguate via checkout.defaultRemote". This essentially
> converts the case of multiple remotes into a single remote.
> However, this also semi-documents what I'm now preventing.
> Was that really intended?
> 70c9ac2f - if there is *one* remote, DWIM it.
> This isn't what I'm changing.
> be4908f1 - if there is *one* remote *and* file, die().
> This is what I'm extending further.
>
> If the prevented behavior is documented, could you please quote it
> explicitly?
>
>> Such a change should also be changing the docs etc. added in 8d7b558bae
>> ("checkout & worktree: introduce checkout.defaultRemote",
>> 2018-06-05). With this series our docs don't make a lot of sense anymore
>> & don't describe the behavior with the patches applied.
>
> To my understanding, my patch doesn't affect those pieces of docs?
> Please correct me if I'm wrong.
This is done for the next commit to avoid crazy 7x tab code padding. Signed-off-by: Alexandr Miloslavskiy <[email protected]>
Before this patch, when there were multiple DWIM candidates for remote branch, git decided to try the argument as pathspec instead. I believe that such behavior is a surprise: adding another remote suddenly causes git to discard file contents, because it was unsure which branch to pick. There was an incomplete attempt to prevent that in [3]. I understand that this was never intended: [1] introduces the unexpected behavior. Before, there was fallback from not-a-ref to pathspec. This is reasonable DWIM. After, there is another fallback from ambiguous-remote to pathspec. I understand that it was kind of copy&paste oversight. [2] noticed the unexpected behavior but chose to semi-document it instead of forbidding, because the goal of the patch series was focused on something else. [3] adds `die()` when there is ambiguity between branch and file. The case of multiple tracking branches is seemingly overlooked. Change to complain about ambiguity instead of doing unexpected things. [1] Commit 70c9ac2 ("DWIM "git checkout frotz" to "git checkout -b frotz origin/frotz"" 2009-10-18) https://public-inbox.org/git/[email protected]/ [2] Commit ad8d510 ("checkout: add advice for ambiguous "checkout <branch>"", 2018-06-05) https://public-inbox.org/git/[email protected]/ [3] Commit be4908f ("checkout: disambiguate dwim tracking branches and local files", 2018-11-13) https://public-inbox.org/git/[email protected]/ Signed-off-by: Alexandr Miloslavskiy <[email protected]>
SyntevoAlex
force-pushed
the
#0224(git)_deny_ambiguous_checkout
branch
from
7dde1a3 to
575eeb9
Compare
|
/submit |
|
Submitted as [email protected] |
|
On the Git mailing list, Junio C Hamano wrote (reply to this): |
|
On the Git mailing list, Alexandr Miloslavskiy wrote (reply to this): |
|
This branch is now known as |
|
This patch series was integrated into pu via git@1ed3ace. |
|
See #504 for continuation |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Please refer to commit messages for details.
Rationale: this issue was standing in my way while working on my
--pathspec-from-filepatches.Since this looks like an oversight, I decided to fix it.
Changes from v1: