Skip to contentSkip to conversation

Improving advise messages from "switch"

PATCH Superseded Junio C Hamano 3 participants 11 messages Jan 27 – Feb 19

lore
Share & tools
lore
#0 lore raw

cover Improving advise messages from "switch"

Simon Cheng noticed [*] that "git switch <name>", in an attempt to create a local branch <name> after a remote tracking branch of the same name, fails correctly when multiple remotes have branches with that name, but gives an advise message to disambiguate using "git checkout".

Obviously, the user wanted to use "git switch", and "git switch" is equipped to do so, just like "git checkout" can, so it makes more sense to give hint based on "git switch".

The first patch of this series restructures the call chain of checkout, switch, and restore commands a bit, so that a helper deep in the callchain can find out what end-user command it is working for. The second patch takes advantage of the new structure to tweak the advise message in question, given from a helper function in a fairly deep part of the call chain.

[Reference]

* https://lore.kernel.org/git/CA+itcS0iyqNyzOP0cueLg7B3yadoEr_VWJ-QoL+YPFUPJiE2RQ@mail.gmail.com

Junio C Hamano (2):
checkout: pass program-readable token to unified "main"
checkout: tell "parse_remote_branch" which command is calling it

 builtin/checkout.c        | 92 ++++++++++++++++++++++++++++-----------
 t/t2027-checkout-track.sh | 21 +++++++++
 2 files changed, 88 insertions(+), 25 deletions(-)
Signature · 1 line
2.53.0-rc2-135-gb1217c0133
#1 lore raw

PATCH 1/2 checkout: pass program-readable token to unified "main"

The "git checkout", "git switch", and "git restore" commands share a single implementation, checkout_main(), which switches error message it gives using the usage string passed by each of these three front-ends.

In order to be able to tweak behaviours of the commands based on which one we are executing, invent an enum that denotes which one of these three commands is currently executing, and pass that to checkout_main() instead. With this step, there is no externally visible behaviour change, as this enum parameter is only used to choose among the three usage strings.

---
 builtin/checkout.c | 63 +++++++++++++++++++++++++++++++---------------
 1 file changed, 43 insertions(+), 20 deletions(-)
builtin/checkout.c +43 -20
index f9453473fe..4f189fde48 100644
@@ -43,22 +43,6 @@
 #include "parallel-checkout.h"
 #include "add-interactive.h"
 
-static const char * const checkout_usage[] = {
-	N_("git checkout [<options>] <branch>"),
-	N_("git checkout [<options>] [<branch>] -- <file>..."),
-	NULL,
-};
-
-static const char * const switch_branch_usage[] = {
-	N_("git switch [<options>] [<branch>]"),
-	NULL,
-};
-
-static const char * const restore_usage[] = {
-	N_("git restore [<options>] [--source=<branch>] <file>..."),
-	NULL,
-};
-
 struct checkout_opts {
 	int patch_mode;
 	int patch_context;
@@ -1293,6 +1277,13 @@ static void setup_new_branch_info_and_source_tree(
 	}
 }
 
+
+enum checkout_command {
+	CHECKOUT_CHECKOUT = 1,
+	CHECKOUT_SWITCH = 2,
+	CHECKOUT_RESTORE = 3,
+};
+
 static char *parse_remote_branch(const char *arg,
 				 struct object_id *rev,
 				 int could_be_checkout_paths)
@@ -1767,12 +1758,44 @@ static char cb_option = 'b';
 
 static int checkout_main(int argc, const char **argv, const char *prefix,
 			 struct checkout_opts *opts, struct option *options,
-			 const char * const usagestr[])
+			 enum checkout_command which_command)
 {
 	int parseopt_flags = 0;
 	struct branch_info new_branch_info = { 0 };
 	int ret;
 
+	static const char * const checkout_usage[] = {
+		N_("git checkout [<options>] <branch>"),
+		N_("git checkout [<options>] [<branch>] -- <file>..."),
+		NULL,
+	};
+
+	static const char * const switch_branch_usage[] = {
+		N_("git switch [<options>] [<branch>]"),
+		NULL,
+	};
+
+	static const char * const restore_usage[] = {
+		N_("git restore [<options>] [--source=<branch>] <file>..."),
+		NULL,
+	};
+
+	const char * const *usagestr;
+
+	switch (which_command) {
+	case CHECKOUT_CHECKOUT:
+		usagestr = checkout_usage;
+		break;
+	case CHECKOUT_SWITCH:
+		usagestr = switch_branch_usage;
+		break;
+	case CHECKOUT_RESTORE:
+		usagestr = restore_usage;
+		break;
+	default:
+		BUG("No such checkout variant %d", which_command);
+	}
+
 	opts->overwrite_ignore = 1;
 	opts->prefix = prefix;
 	opts->show_progress = -1;
@@ -2032,7 +2055,7 @@ int cmd_checkout(int argc,
 	options = add_checkout_path_options(&opts, options);
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     checkout_usage);
+			     CHECKOUT_CHECKOUT);
 }
 
 int cmd_switch(int argc,
@@ -2071,7 +2094,7 @@ int cmd_switch(int argc,
 	cb_option = 'c';
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     switch_branch_usage);
+			     CHECKOUT_SWITCH);
 }
 
 int cmd_restore(int argc,
@@ -2107,5 +2130,5 @@ int cmd_restore(int argc,
 	options = add_checkout_path_options(&opts, options);
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     restore_usage);
+			     CHECKOUT_RESTORE);
 }
Signature · 1 line
2.53.0-rc2-135-gb1217c0133
#2 lore raw

PATCH 2/2 checkout: tell "parse_remote_branch" which command is calling it

In reply to Junio C Hamano

When "git checkout <dwim>" and "git switch <dwim>" need to error out due to ambiguity of the branch name <dwim>, these command give an advise message that tells a sample command to show how to disambiguate from the parse_remote_branch() function. The sample command hardcodes "git checkout", since this feature predates "git switch" by a large margin. To a user who said "git switch <dwim>" and got this message, it is confusing.

Pass the "enum checkout_command", which was invented in the previous step for this exact purpose, down the call chain leading to parse_remote_branch() function to change the sample command shown to the user in this advise message.

---
 builtin/checkout.c        | 29 ++++++++++++++++++++++++-----
 t/t2027-checkout-track.sh | 21 +++++++++++++++++++++
 2 files changed, 45 insertions(+), 5 deletions(-)
builtin/checkout.c +24 -5
index 4f189fde48..17f31c30b2 100644
@@ -1286,7 +1286,8 @@ enum checkout_command {
 
 static char *parse_remote_branch(const char *arg,
 				 struct object_id *rev,
-				 int could_be_checkout_paths)
+				 int could_be_checkout_paths,
+				 enum checkout_command which_command)
 {
 	int num_matches = 0;
 	char *remote = unique_tracking_name(arg, rev, &num_matches);
@@ -1299,14 +1300,30 @@ static char *parse_remote_branch(const char *arg,
 
 	if (!remote && num_matches > 1) {
 	    if (advice_enabled(ADVICE_CHECKOUT_AMBIGUOUS_REMOTE_BRANCH_NAME)) {
+		    const char *cmdname;
+
+		    switch (which_command) {
+		    case CHECKOUT_CHECKOUT:
+			    cmdname = "checkout";
+			    break;
+		    case CHECKOUT_SWITCH:
+			    cmdname = "switch";
+			    break;
+		    default:
+			    BUG("command <%d> should not reach parse_remote_branch",
+				which_command);
+			    break;
+		    }
+
 		    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"
+			     "    git %s --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."));
+			     "checkout.defaultRemote=origin in your config."),
+			   cmdname);
 	    }
 
 	    die(_("'%s' matched multiple (%d) remote tracking branches"),
@@ -1318,6 +1335,7 @@ static char *parse_remote_branch(const char *arg,
 
 static int parse_branchname_arg(int argc, const char **argv,
 				int dwim_new_local_branch_ok,
+				enum checkout_command which_command,
 				struct branch_info *new_branch_info,
 				struct checkout_opts *opts,
 				struct object_id *rev)
@@ -1427,7 +1445,8 @@ static int parse_branchname_arg(int argc, const char **argv,
 
 		if (recover_with_dwim) {
 			remote = parse_remote_branch(arg, rev,
-						     could_be_checkout_paths);
+						     could_be_checkout_paths,
+						     which_command);
 			if (remote) {
 				*new_branch = arg;
 				arg = remote;
@@ -1916,7 +1935,7 @@ static int checkout_main(int argc, const char **argv, const char *prefix,
 			opts->dwim_new_local_branch &&
 			opts->track == BRANCH_TRACK_UNSPECIFIED &&
 			!opts->new_branch;
-		int n = parse_branchname_arg(argc, argv, dwim_ok,
+		int n = parse_branchname_arg(argc, argv, dwim_ok, which_command,
 					     &new_branch_info, opts, &rev);
 		argv += n;
 		argc -= n;
t/t2027-checkout-track.sh +21 -0
index a397790df5..e9f8d8ec48 100755
@@ -47,4 +47,25 @@ test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' '
 	test_cmp_config refs/heads/main branch.b4.merge
 '
 
+test_expect_success 'ambiguous tracking info' '
+	# Set up a few remote repositories
+	git init --bare --initial-branch=trunk src1 &&
+	git init --bare --initial-branch=trunk src2 &&
+	git push src1 one:refs/heads/trunk &&
+	git push src2 two:refs/heads/trunk &&
+
+	git remote add -f src1 "file://$PWD/src1" &&
+	git remote add -f src2 "file://$PWD/src2" &&
+
+	# DWIM
+	test_must_fail git checkout trunk 2>hint &&
+	test_grep "hint: *git checkout --track" hint &&
+	test_grep ! "hint: *git switch --track" hint &&
+
+	{ git update-ref -d refs/heads/trunk || :; } &&
+	test_must_fail git switch trunk 2>hint &&
+	test_grep ! "hint: *git checkout --track" hint &&
+	test_grep "hint: *git switch --track" hint
+'
+
 test_done
Signature · 1 line
2.53.0-rc2-135-gb1217c0133
#3 lore raw

Re: [PATCH 2/2] checkout: tell "parse_remote_branch" which command is calling it

Quoted 2 lines · Junio C Hamano

When "git checkout <dwim>" and "git switch <dwim>" need to error out due to ambiguity of the branch name <dwim>, these command give an

s/these command/these two commands/ or something.

Quoted 10 lines

advise message that tells a sample command to show how to disambiguate from the parse_remote_branch() function. The sample command hardcodes "git checkout", since this feature predates "git switch" by a large margin. To a user who said "git switch <dwim>" and got this message, it is confusing.

Pass the "enum checkout_command", which was invented in the previous step for this exact purpose, down the call chain leading to parse_remote_branch() function to change the sample command shown to the user in this advise message.

The commit message could also say that it adds a regression test for this advice (for the first time).

Quoted 23 lines
---
 builtin/checkout.c        | 29 ++++++++++++++++++++++++-----
 t/t2027-checkout-track.sh | 21 +++++++++++++++++++++
 2 files changed, 45 insertions(+), 5 deletions(-)

[snip]

t/t2027-checkout-track.sh +0 -0
index a397790df5..e9f8d8ec48 100755
@@ -47,4 +47,25 @@ test_expect_success 'checkout --track -b overrides

autoSetupMerge=inherit' '

t/t2027-checkout-track.sh +6 -0
 	test_cmp_config refs/heads/main branch.b4.merge
 '
 
+test_expect_success 'ambiguous tracking info' '
+	# Set up a few remote repositories
+	git init --bare --initial-branch=trunk src1 &&
+	git init --bare --initial-branch=trunk src2 &&
+	git push src1 one:refs/heads/trunk &&
+	git push src2 two:refs/heads/trunk &&

t/t2027-checkout-track.sh test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' ' · line 65

Quoted 10 lines
t/t2027-checkout-track.sh +10 -0
+
+	git remote add -f src1 "file://$PWD/src1" &&
+	git remote add -f src2 "file://$PWD/src2" &&
+
+	# DWIM
+	test_must_fail git checkout trunk 2>hint &&
+	test_grep "hint: *git checkout --track" hint &&
+	test_grep ! "hint: *git switch --track" hint &&
+
+	{ git update-ref -d refs/heads/trunk || :; } &&

I don’t understand what the purpose of this is after git checkout but before git switch. I can delete it and the test still passes. Is it post-test cleanup?

t/t2027-checkout-track.sh test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' ' · line 69

Quoted 4 lines
t/t2027-checkout-track.sh +4 -0
+	test_must_fail git switch trunk 2>hint &&
+	test_grep ! "hint: *git checkout --track" hint &&
+	test_grep "hint: *git switch --track" hint
+'

Maybe just the positive greps are enough. I read these a few times because I thought the order was wrong, i.e. that hint was overwritten before it got tested. The regression that they test are unlikely and these negative greps might not make immediate sense for future readers. I dunno.

I was about to suggest parameterizing the command but that makes it harder to grep for that part of the advice.

    checkout_with_cmd () {
            cmd="$1"
            test_must_fail git $cmd trunk 2>hint &&
            test_grep --fixed-string \
                      "hint:     git $cmd --track origin/<name>" hint
    }
Quoted 4 lines

+
test_done
--
2.53.0-rc2-135-gb1217c0133

#4 lore raw

Re: [PATCH 2/2] checkout: tell "parse_remote_branch" which command is calling it

Quoted 10 lines · Kristoffer Haugsbakk

t/t2027-checkout-track.sh test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' ' · line 65

Quoted 6 lines
t/t2027-checkout-track.sh +6 -0
+	# DWIM
+	test_must_fail git checkout trunk 2>hint &&
+	test_grep "hint: *git checkout --track" hint &&
+	test_grep ! "hint: *git switch --track" hint &&
+
+	{ git update-ref -d refs/heads/trunk || :; } &&

I don’t understand what the purpose of this is after git checkout but before git switch. I can delete it and the test still passes. Is it post-test cleanup?

Just in case "git checkout trunk" that was expected to fail still creates the 'trunk' branch by a bug. I do not want the failure of the next "git switch trunk" to be due to "hey, you already have a local branch of that name", and make sure the failure is from "you have two remotes with trunk, and I cannot tell which one you meant".

Quoted 10 lines

t/t2027-checkout-track.sh test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' ' · line 69

Quoted 4 lines
t/t2027-checkout-track.sh +4 -0
+	test_must_fail git switch trunk 2>hint &&
+	test_grep ! "hint: *git checkout --track" hint &&
+	test_grep "hint: *git switch --track" hint
+'

Maybe just the positive greps are enough. I read these a few times because I thought the order was wrong, i.e. that hint was overwritten before it got tested. The regression that they test are unlikely and these negative greps might not make immediate sense for future readers. I dunno.

Possibly. These tests to expect concrete strings in the output are already familiar with how these output strings are built, so they should know that when 'git checkout --track' appears, it is very unlikely that 'git switch --track' would appear there, for example.

Thanks.

#5 lore raw

v2 cover Improving advise messages from "switch"

In reply to Junio C Hamano

Simon Cheng noticed [*] that "git switch <name>", in an attempt to create a local branch <name> after a remote tracking branch of the same name, fails correctly when multiple remotes have branches with that name, but gives an advise message to disambiguate using "git checkout".

Obviously, the user wanted to use "git switch", and "git switch" is equipped to do so, just like "git checkout" can, so it makes more sense to give hint based on "git switch".

The first patch of this series restructures the call chain of checkout, switch, and restore commands a bit, so that a helper deep in the callchain can find out what end-user command it is working for. The second patch takes advantage of the new structure to tweak the advise message in question, given from a helper function in a fairly deep part of the call chain.

[Reference]

* https://lore.kernel.org/git/CA+itcS0iyqNyzOP0cueLg7B3yadoEr_VWJ-QoL+YPFUPJiE2RQ@mail.gmail.com

Junio C Hamano (2):
checkout: pass program-readable token to unified "main"
checkout: tell "parse_remote_branch" which command is calling it

 builtin/checkout.c        | 92 ++++++++++++++++++++++++++++-----------
 t/t2027-checkout-track.sh | 18 ++++++++
 2 files changed, 85 insertions(+), 25 deletions(-)

---

This second iteration simplifies the tests in the second patch somewhat, and rewords its proposed log message, as suggested by Kristoffer Haugsbakk.

#6 lore raw

PATCH v2 1/2 checkout: pass program-readable token to unified "main"

The "git checkout", "git switch", and "git restore" commands share a single implementation, checkout_main(), which switches error message it gives using the usage string passed by each of these three front-ends.

In order to be able to tweak behaviours of the commands based on which one we are executing, invent an enum that denotes which one of these three commands is currently executing, and pass that to checkout_main() instead. With this step, there is no externally visible behaviour change, as this enum parameter is only used to choose among the three usage strings.

---
* No change relative to v1

 builtin/checkout.c | 63 +++++++++++++++++++++++++++++++---------------
 1 file changed, 43 insertions(+), 20 deletions(-)
builtin/checkout.c +43 -20
index f9453473fe..4f189fde48 100644
@@ -43,22 +43,6 @@
 #include "parallel-checkout.h"
 #include "add-interactive.h"
 
-static const char * const checkout_usage[] = {
-	N_("git checkout [<options>] <branch>"),
-	N_("git checkout [<options>] [<branch>] -- <file>..."),
-	NULL,
-};
-
-static const char * const switch_branch_usage[] = {
-	N_("git switch [<options>] [<branch>]"),
-	NULL,
-};
-
-static const char * const restore_usage[] = {
-	N_("git restore [<options>] [--source=<branch>] <file>..."),
-	NULL,
-};
-
 struct checkout_opts {
 	int patch_mode;
 	int patch_context;
@@ -1293,6 +1277,13 @@ static void setup_new_branch_info_and_source_tree(
 	}
 }
 
+
+enum checkout_command {
+	CHECKOUT_CHECKOUT = 1,
+	CHECKOUT_SWITCH = 2,
+	CHECKOUT_RESTORE = 3,
+};
+
 static char *parse_remote_branch(const char *arg,
 				 struct object_id *rev,
 				 int could_be_checkout_paths)
@@ -1767,12 +1758,44 @@ static char cb_option = 'b';
 
 static int checkout_main(int argc, const char **argv, const char *prefix,
 			 struct checkout_opts *opts, struct option *options,
-			 const char * const usagestr[])
+			 enum checkout_command which_command)
 {
 	int parseopt_flags = 0;
 	struct branch_info new_branch_info = { 0 };
 	int ret;
 
+	static const char * const checkout_usage[] = {
+		N_("git checkout [<options>] <branch>"),
+		N_("git checkout [<options>] [<branch>] -- <file>..."),
+		NULL,
+	};
+
+	static const char * const switch_branch_usage[] = {
+		N_("git switch [<options>] [<branch>]"),
+		NULL,
+	};
+
+	static const char * const restore_usage[] = {
+		N_("git restore [<options>] [--source=<branch>] <file>..."),
+		NULL,
+	};
+
+	const char * const *usagestr;
+
+	switch (which_command) {
+	case CHECKOUT_CHECKOUT:
+		usagestr = checkout_usage;
+		break;
+	case CHECKOUT_SWITCH:
+		usagestr = switch_branch_usage;
+		break;
+	case CHECKOUT_RESTORE:
+		usagestr = restore_usage;
+		break;
+	default:
+		BUG("No such checkout variant %d", which_command);
+	}
+
 	opts->overwrite_ignore = 1;
 	opts->prefix = prefix;
 	opts->show_progress = -1;
@@ -2032,7 +2055,7 @@ int cmd_checkout(int argc,
 	options = add_checkout_path_options(&opts, options);
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     checkout_usage);
+			     CHECKOUT_CHECKOUT);
 }
 
 int cmd_switch(int argc,
@@ -2071,7 +2094,7 @@ int cmd_switch(int argc,
 	cb_option = 'c';
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     switch_branch_usage);
+			     CHECKOUT_SWITCH);
 }
 
 int cmd_restore(int argc,
@@ -2107,5 +2130,5 @@ int cmd_restore(int argc,
 	options = add_checkout_path_options(&opts, options);
 
 	return checkout_main(argc, argv, prefix, &opts, options,
-			     restore_usage);
+			     CHECKOUT_RESTORE);
 }
Signature · 1 line
2.53.0-rc2-135-gb1217c0133
#7 lore raw

Re: [PATCH v2 1/2] checkout: pass program-readable token to unified "main"

Quoted 4 lines · Junio C Hamano
builtin/checkout.c +0 -0
index f9453473fe..4f189fde48 100644

[snip]

builtin/checkout.c static char cb_option = 'b'; · line 1797

Quoted 13 lines
builtin/checkout.c +13 -0
+	switch (which_command) {
+	case CHECKOUT_CHECKOUT:
+		usagestr = checkout_usage;
+		break;
+	case CHECKOUT_SWITCH:
+		usagestr = switch_branch_usage;
+		break;
+	case CHECKOUT_RESTORE:
+		usagestr = restore_usage;
+		break;
+	default:
+		BUG("No such checkout variant %d", which_command);
+	}

Tiniest nit, really not worth addressing on its own: BUG messages typically start with a lower-case letter.

Other than that I like that we have less global constants with this change.

Patrick

#8 lore raw

Re: [PATCH v2 1/2] checkout: pass program-readable token to unified "main"

Quoted 23 lines · Patrick Steinhardt
Quoted 4 lines · Junio C Hamano
builtin/checkout.c +0 -0
index f9453473fe..4f189fde48 100644

[snip]

builtin/checkout.c static char cb_option = 'b'; · line 1797

Quoted 13 lines
builtin/checkout.c +13 -0
+	switch (which_command) {
+	case CHECKOUT_CHECKOUT:
+		usagestr = checkout_usage;
+		break;
+	case CHECKOUT_SWITCH:
+		usagestr = switch_branch_usage;
+		break;
+	case CHECKOUT_RESTORE:
+		usagestr = restore_usage;
+		break;
+	default:
+		BUG("No such checkout variant %d", which_command);
+	}

Tiniest nit, really not worth addressing on its own: BUG messages typically start with a lower-case letter.

Quoted 2 lines

Other than that I like that we have less global constants with this change.

Thanks.

As I locally amended the nit away while queuing, I guess I should merge this down to 'next'.

#9 lore raw

PATCH v2 2/2 checkout: tell "parse_remote_branch" which command is calling it

In reply to Junio C Hamano

When "git checkout <dwim>" and "git switch <dwim>" need to error out due to ambiguity of the branch name <dwim>, these two commands give an advise message with a sample command that tells the user how to disambiguate from the parse_remote_branch() function. The sample command hardcodes "git checkout", since this feature predates "git switch" by a large margin. To a user who said "git switch <dwim>" and got this message, it is confusing.

Pass the "enum checkout_command", which was invented in the previous step for this exact purpose, down the call chain leading to parse_remote_branch() function to change the sample command shown to the user in this advise message.

Also add a bit more test coverage for this "fail to DWIM under ambiguity" that we lack, as well as the message we produce when we fail.

---
 builtin/checkout.c        | 29 ++++++++++++++++++++++++-----
 t/t2027-checkout-track.sh | 18 ++++++++++++++++++
 2 files changed, 42 insertions(+), 5 deletions(-)
builtin/checkout.c +24 -5
index 4f189fde48..17f31c30b2 100644
@@ -1286,7 +1286,8 @@ enum checkout_command {
 
 static char *parse_remote_branch(const char *arg,
 				 struct object_id *rev,
-				 int could_be_checkout_paths)
+				 int could_be_checkout_paths,
+				 enum checkout_command which_command)
 {
 	int num_matches = 0;
 	char *remote = unique_tracking_name(arg, rev, &num_matches);
@@ -1299,14 +1300,30 @@ static char *parse_remote_branch(const char *arg,
 
 	if (!remote && num_matches > 1) {
 	    if (advice_enabled(ADVICE_CHECKOUT_AMBIGUOUS_REMOTE_BRANCH_NAME)) {
+		    const char *cmdname;
+
+		    switch (which_command) {
+		    case CHECKOUT_CHECKOUT:
+			    cmdname = "checkout";
+			    break;
+		    case CHECKOUT_SWITCH:
+			    cmdname = "switch";
+			    break;
+		    default:
+			    BUG("command <%d> should not reach parse_remote_branch",
+				which_command);
+			    break;
+		    }
+
 		    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"
+			     "    git %s --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."));
+			     "checkout.defaultRemote=origin in your config."),
+			   cmdname);
 	    }
 
 	    die(_("'%s' matched multiple (%d) remote tracking branches"),
@@ -1318,6 +1335,7 @@ static char *parse_remote_branch(const char *arg,
 
 static int parse_branchname_arg(int argc, const char **argv,
 				int dwim_new_local_branch_ok,
+				enum checkout_command which_command,
 				struct branch_info *new_branch_info,
 				struct checkout_opts *opts,
 				struct object_id *rev)
@@ -1427,7 +1445,8 @@ static int parse_branchname_arg(int argc, const char **argv,
 
 		if (recover_with_dwim) {
 			remote = parse_remote_branch(arg, rev,
-						     could_be_checkout_paths);
+						     could_be_checkout_paths,
+						     which_command);
 			if (remote) {
 				*new_branch = arg;
 				arg = remote;
@@ -1916,7 +1935,7 @@ static int checkout_main(int argc, const char **argv, const char *prefix,
 			opts->dwim_new_local_branch &&
 			opts->track == BRANCH_TRACK_UNSPECIFIED &&
 			!opts->new_branch;
-		int n = parse_branchname_arg(argc, argv, dwim_ok,
+		int n = parse_branchname_arg(argc, argv, dwim_ok, which_command,
 					     &new_branch_info, opts, &rev);
 		argv += n;
 		argc -= n;
t/t2027-checkout-track.sh +18 -0
index a397790df5..c01f1cd617 100755
@@ -47,4 +47,22 @@ test_expect_success 'checkout --track -b overrides autoSetupMerge=inherit' '
 	test_cmp_config refs/heads/main branch.b4.merge
 '
 
+test_expect_success 'ambiguous tracking info' '
+	# Set up a few remote repositories
+	git init --bare --initial-branch=trunk src1 &&
+	git init --bare --initial-branch=trunk src2 &&
+	git push src1 one:refs/heads/trunk &&
+	git push src2 two:refs/heads/trunk &&
+
+	git remote add -f src1 "file://$PWD/src1" &&
+	git remote add -f src2 "file://$PWD/src2" &&
+
+	# DWIM
+	test_must_fail git checkout trunk 2>hint.checkout &&
+	test_grep "hint: *git checkout --track" hint.checkout &&
+
+	test_must_fail git switch trunk 2>hint.switch &&
+	test_grep "hint: *git switch --track" hint.switch
+'
+
 test_done
Signature · 1 line
2.53.0-rc2-135-gb1217c0133
#10 lore raw

Re: [PATCH v2 2/2] checkout: tell "parse_remote_branch" which command is calling it

Quoted 12 lines · Junio C Hamano
builtin/checkout.c +4 -0
index 4f189fde48..17f31c30b2 100644
@@ -1299,14 +1300,30 @@ static char *parse_remote_branch(const char *arg,
 
 	if (!remote && num_matches > 1) {
 	    if (advice_enabled(ADVICE_CHECKOUT_AMBIGUOUS_REMOTE_BRANCH_NAME)) {
+		    const char *cmdname;
+
+		    switch (which_command) {
+		    case CHECKOUT_CHECKOUT:

builtin/checkout.c static char *parse_remote_branch(const char *arg, · line 1316

Quoted 10 lines
builtin/checkout.c +10 -0
+			    cmdname = "checkout";
+			    break;
+		    case CHECKOUT_SWITCH:
+			    cmdname = "switch";
+			    break;
+		    default:
+			    BUG("command <%d> should not reach parse_remote_branch",
+				which_command);
+			    break;
+		    }

parse_remote_branch() is only called by parse_branchname_arg(), which in turn is only called if opts->accept_ref is truish. And as that value only gets set to 1 in git-checkout(1) and git-switch(1) we know that we indeed don't have to care about git-restore(1).

Quoted 4 lines
builtin/checkout.c +0 -1
 		    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"

builtin/checkout.c static char *parse_remote_branch(const char *arg, · line 1329

Quoted 10 lines
builtin/checkout.c +3 -1
+			     "    git %s --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."));
+			     "checkout.defaultRemote=origin in your config."),
+			   cmdname);
 	    }
 
 	    die(_("'%s' matched multiple (%d) remote tracking branches"),

Yup, makes sense.

Overall this patch series looks good to me. There's been the one micronit in the preceding commit, but I really don't think it warrants a new version.

Thanks!

Patrick