Skip to contentSkip to conversation

rebase: fix bug in --fork-point

PATCH Superseded Alex Torok 5 participants 23 messages Dec 5, 2019 – Feb 11, 2020

lore
Share & tools
lore

Alex Torok

author
#0 lore raw

cover rebase: fix bug in --fork-point

git rebase --fork-point with a short branch ref name (without refs/heads/ prefix) for the upstream will not use the fork point between the upstream and the branch.

If git rebase --forkpoint is used with a full ref on the upstream (refs/heads/branchName), the fork point between the branch and upstream is correctly used.

This bug was introduced in 2.20.0 with the c implementation of rebase. I was able to reproduce it on next and master.

The code path for rebase with --fork-point and a user-provided upstream does not use a DWIM method to look up the full ref name. This leads to get_fork_point returning a null pointer and rebase using the upstream itself instead of the fork point between the upstream and branch.

I looked in other places that call get_fork_point and found that in handle_fork_point of builtin/merge_base.c, dwim_ref is used to find the full ref name before calling get_fork_point.

I initially based this change off of maint, but then saw that there were some tests added for rebase --fork-point on master, so I rebased onto there to build off of those tests, but I'm not sure if that was the "right" thing to do.

Alex Torok (3):
rebase: add test for rebase --fork-point with short upstream
rebase: refactor dwim_ref_or_die from merge-base.c
rebase: fix rebase to use full ref to find fork-point

 builtin/merge-base.c         |  9 +--------
 builtin/rebase.c             |  4 +++-
 refs.c                       | 14 ++++++++++++++
 refs.h                       |  1 +
 t/t3431-rebase-fork-point.sh |  1 +
 5 files changed, 20 insertions(+), 9 deletions(-)
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH 1/3 rebase: add test for rebase --fork-point with short upstream

This proves the bug of "rebase --fork-point upstream branch" not using the fork point of upstream and branch if upstream is not the full refname of the upstream branch.

---
 t/t3431-rebase-fork-point.sh | 1 +
 1 file changed, 1 insertion(+)
t/t3431-rebase-fork-point.sh +1 -0
index 78851b9a2a..6ecdae918e 100755
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
 test_rebase 'G F E D B A' --fork-point refs/heads/master
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
 test_rebase 'G F C E D B A' refs/heads/master
 test_rebase 'G F C D B A' --onto D refs/heads/master
Signature · 1 line
2.17.1
#2 lore raw

Re: [PATCH 1/3] rebase: add test for rebase --fork-point with short upstream

Quoted 8 lines · Alex Torok

This proves the bug of "rebase --fork-point upstream branch" not using the fork point of upstream and branch if upstream is not the full refname of the upstream branch.

---
 t/t3431-rebase-fork-point.sh | 1 +
 1 file changed, 1 insertion(+)

Is this new test expected to fail after applying only 1/3 and then starts working after applying all 3 patches?

If so, it probably makes a lot mroe sense to reorder the series to have 2/3 as a single preparatory patch, with 1/3 + 3/3 combined into a single patch "rebase: find --fork-point with full ref" to fix the code and protect the fix with the test at the same time.

Quoted 12 lines
t/t3431-rebase-fork-point.sh +1 -0
index 78851b9a2a..6ecdae918e 100755
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
 test_rebase 'G F E D B A' --fork-point refs/heads/master
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
 test_rebase 'G F C E D B A' refs/heads/master
 test_rebase 'G F C D B A' --onto D refs/heads/master
author
#3 lore raw

Re: [PATCH 1/3] rebase: add test for rebase --fork-point with short upstream

That was the intention. Your idea for ordering makes more sense.

I'll reorder the commits and submit new patches.

Quoted 32 lines · Junio C Hamano
Quoted 8 lines · Alex Torok

This proves the bug of "rebase --fork-point upstream branch" not using the fork point of upstream and branch if upstream is not the full refname of the upstream branch.

---
 t/t3431-rebase-fork-point.sh | 1 +
 1 file changed, 1 insertion(+)

Is this new test expected to fail after applying only 1/3 and then starts working after applying all 3 patches?

If so, it probably makes a lot mroe sense to reorder the series to have 2/3 as a single preparatory patch, with 1/3 + 3/3 combined into a single patch "rebase: find --fork-point with full ref" to fix the code and protect the fix with the test at the same time.

Quoted 12 lines
t/t3431-rebase-fork-point.sh +1 -0
index 78851b9a2a..6ecdae918e 100755
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
 test_rebase 'G F E D B A' --fork-point refs/heads/master
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
 test_rebase 'G F C E D B A' refs/heads/master
 test_rebase 'G F C D B A' --onto D refs/heads/master
author
#4 lore raw

PATCH 2/3 rebase: refactor dwim_ref_or_die from merge-base.c

In reply to Alex Torok

Pull logic for getting the full dwim ref name from a passed in ref name out of the handle_fork_point function of merge-base.c. This will allow git-rebase --fork-point to use the same method for getting the full ref name before calling get_fork_point.

I saw other *_or_die methods in other places and using that pattern seemed sane here.

I'm not 100% sure about the name or signature of dwim_ref_or_die. I feel like it should be named something like dwim_ref_name_or_die, unique_dwim_ref_or_die, or should be including the object_id argument even though it isn't used by the calling merge_base code, and won't be used in rebase.

This is my first patch submission for git, and I'd appreciate some feedback on naming/style wrt to this.

---
 builtin/merge-base.c |  9 +--------
 refs.c               | 14 ++++++++++++++
 refs.h               |  1 +
 3 files changed, 16 insertions(+), 8 deletions(-)
builtin/merge-base.c +1 -8
index e3f8da13b6..edd16f9fcd 100644
@@ -118,14 +118,7 @@ static int handle_fork_point(int argc, const char **argv)
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
+	dwim_ref_or_die(argv[0], strlen(argv[0]), &refname);
 
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
refs.c +14 -0
index 1ab0bb54d3..3b778f2df9 100644
@@ -639,6 +639,20 @@ int dwim_ref(const char *str, int len, struct object_id *oid, char **ref)
 	return repo_dwim_ref(the_repository, str, len, oid, ref);
 }
 
+void dwim_ref_or_die(const char *str, int len, char **ref)
+{
+	struct object_id discard;
+	switch (dwim_ref(str, len, &discard, ref)) {
+	case 0:
+		die("No such ref: '%s'", str);
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", str);
+	}
+}
+
+
 int expand_ref(struct repository *repo, const char *str, int len,
 	       struct object_id *oid, char **ref)
 {
refs.h +1 -0
index 730d05ad91..a961662382 100644
@@ -154,6 +154,7 @@ int repo_dwim_log(struct repository *r, const char *str, int len, struct object_
 int dwim_ref(const char *str, int len, struct object_id *oid, char **ref);
 int dwim_log(const char *str, int len, struct object_id *oid, char **ref);
 
+void dwim_ref_or_die(const char *str, int len, char **ref);
 /*
  * A ref_transaction represents a collection of reference updates that
  * should succeed or fail together.
Signature · 1 line
2.17.1
author
#5 lore raw

PATCH 3/3 rebase: fix rebase to use full ref to find fork-point

In reply to Alex Torok

rebase --fork-point needs to look up the full ref name before calling get_fork_point in the same manner that merge-base --fork-point does.

---
 builtin/rebase.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
builtin/rebase.c +3 -1
index e755087b0f..821994f676 100644
@@ -1980,8 +1980,10 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
 		struct commit *head =
 			lookup_commit_reference(the_repository,
 						&options.orig_head);
+		char * full_name;
+		dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);
 		options.restrict_revision =
-			get_fork_point(options.upstream_name, head);
+			get_fork_point(full_name, head);
 	}
 
 	if (repo_read_index(the_repository) < 0)
Signature · 1 line
2.17.1
author
#6 lore raw

v2 cover rebase: fix bug in --fork-point

In reply to Alex Torok

Reorder commits to include the fix and test in the same commit.

Alex Torok (2):
rebase: refactor dwim_ref_or_die from merge-base.c
rebase: find --fork-point with full ref

 builtin/merge-base.c         |  9 +--------
 builtin/rebase.c             |  4 +++-
 refs.c                       | 14 ++++++++++++++
 refs.h                       |  1 +
 t/t3431-rebase-fork-point.sh |  1 +
 5 files changed, 20 insertions(+), 9 deletions(-)
Signature · 1 line
2.17.1
author
#7 lore raw

PATCH v2 1/2 rebase: refactor dwim_ref_or_die from merge-base.c

Pull logic for getting the full dwim ref name from a passed in ref name out of the handle_fork_point function of merge-base.c. This will allow git-rebase --fork-point to use the same method for getting the full ref name before calling get_fork_point.

I saw other *_or_die methods in other places and using that pattern seemed sane here.

I'm not 100% sure about the name or signature of dwim_ref_or_die. I feel like it should be named something like dwim_ref_name_or_die, unique_dwim_ref_or_die, or should be including the object_id argument even though it isn't used by the calling merge_base code, and won't be used in rebase.

This is my first patch submission for git, and I'd appreciate some feedback on naming/style wrt to this.

---
 builtin/merge-base.c |  9 +--------
 refs.c               | 14 ++++++++++++++
 refs.h               |  1 +
 3 files changed, 16 insertions(+), 8 deletions(-)
builtin/merge-base.c +1 -8
index e3f8da13b6..edd16f9fcd 100644
@@ -118,14 +118,7 @@ static int handle_fork_point(int argc, const char **argv)
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
+	dwim_ref_or_die(argv[0], strlen(argv[0]), &refname);
 
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
refs.c +14 -0
index 1ab0bb54d3..3b778f2df9 100644
@@ -639,6 +639,20 @@ int dwim_ref(const char *str, int len, struct object_id *oid, char **ref)
 	return repo_dwim_ref(the_repository, str, len, oid, ref);
 }
 
+void dwim_ref_or_die(const char *str, int len, char **ref)
+{
+	struct object_id discard;
+	switch (dwim_ref(str, len, &discard, ref)) {
+	case 0:
+		die("No such ref: '%s'", str);
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", str);
+	}
+}
+
+
 int expand_ref(struct repository *repo, const char *str, int len,
 	       struct object_id *oid, char **ref)
 {
refs.h +1 -0
index 730d05ad91..a961662382 100644
@@ -154,6 +154,7 @@ int repo_dwim_log(struct repository *r, const char *str, int len, struct object_
 int dwim_ref(const char *str, int len, struct object_id *oid, char **ref);
 int dwim_log(const char *str, int len, struct object_id *oid, char **ref);
 
+void dwim_ref_or_die(const char *str, int len, char **ref);
 /*
  * A ref_transaction represents a collection of reference updates that
  * should succeed or fail together.
Signature · 1 line
2.17.1
#8 lore raw

Re: [PATCH v2 1/2] rebase: refactor dwim_ref_or_die from merge-base.c

Hi Alex,

Thanks for the contribution.

Quoted 5 lines · Alex Torok

Pull logic for getting the full dwim ref name from a passed in ref name out of the handle_fork_point function of merge-base.c. This will allow git-rebase --fork-point to use the same method for getting the full ref name before calling get_fork_point.

The remaining paragraphs shouldn't end up going into the commit message since they're more like "notes to reviewers". You should place these notes under the --- below so that they won't end up in the final commit message.

Quoted 44 lines

I saw other *_or_die methods in other places and using that pattern seemed sane here.

I'm not 100% sure about the name or signature of dwim_ref_or_die. I feel like it should be named something like dwim_ref_name_or_die, unique_dwim_ref_or_die, or should be including the object_id argument even though it isn't used by the calling merge_base code, and won't be used in rebase.

This is my first patch submission for git, and I'd appreciate some feedback on naming/style wrt to this.

---
 builtin/merge-base.c |  9 +--------
 refs.c               | 14 ++++++++++++++
 refs.h               |  1 +
 3 files changed, 16 insertions(+), 8 deletions(-)
builtin/merge-base.c +1 -8
index e3f8da13b6..edd16f9fcd 100644
@@ -118,14 +118,7 @@ static int handle_fork_point(int argc, const char **argv)
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
+	dwim_ref_or_die(argv[0], strlen(argv[0]), &refname);
 
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
refs.c +0 -0
index 1ab0bb54d3..3b778f2df9 100644

refs.c int dwim_ref(const char *str, int len, struct object_id *oid, char **ref) · line 647

Quoted 9 lines
refs.c +6 -0
@@ -639,6 +639,20 @@ int dwim_ref(const char *str, int len, struct object_id *oid, char **ref)
 	return repo_dwim_ref(the_repository, str, len, oid, ref);
 }
 
+void dwim_ref_or_die(const char *str, int len, char **ref)
+{
+	struct object_id discard;
+	switch (dwim_ref(str, len, &discard, ref)) {
+	case 0:
+		die("No such ref: '%s'", str);

I see that this code is copied so I'm fine with leaving it as is. But if you need to do any rerolls, it would be nice to, as a prepatory step, introduce a patch which 1. marks these strings for translation and 2. lowercases the first letter of the sentence which is the convention for messages that are seen by the end-user.

Perhaps:

	die(_("no such ref: '%s'"), str);

refs.c int dwim_ref(const char *str, int len, struct object_id *oid, char **ref) · line 651

Quoted 4 lines
refs.c +4 -0
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", str);

Same comment here.

Thanks,

Denton

Quoted 22 lines
refs.c +4 -0
+	}
+}
+
+
 int expand_ref(struct repository *repo, const char *str, int len,
 	       struct object_id *oid, char **ref)
 {
refs.h +1 -0
index 730d05ad91..a961662382 100644
@@ -154,6 +154,7 @@ int repo_dwim_log(struct repository *r, const char *str, int len, struct object_
 int dwim_ref(const char *str, int len, struct object_id *oid, char **ref);
 int dwim_log(const char *str, int len, struct object_id *oid, char **ref);
 
+void dwim_ref_or_die(const char *str, int len, char **ref);
 /*
  * A ref_transaction represents a collection of reference updates that
  * should succeed or fail together.
Signature · 1 line
2.17.1
author
#9 lore raw

Re: [PATCH v2 1/2] rebase: refactor dwim_ref_or_die from merge-base.c

Quoted 4 lines · Denton Liu

The remaining paragraphs shouldn't end up going into the commit message since they're more like "notes to reviewers". You should place these notes under the --- below so that they won't end up in the final commit message.

Thank you for the info! I'll keep that in mind for my next patch submission.

Quoted 13 lines

refs.c int dwim_ref(const char *str, int len, struct object_id *oid, char **ref) · line 647

Quoted 6 lines
refs.c +6 -0
+void dwim_ref_or_die(const char *str, int len, char **ref)
+{
+     struct object_id discard;
+     switch (dwim_ref(str, len, &discard, ref)) {
+     case 0:
+             die("No such ref: '%s'", str);

I see that this code is copied so I'm fine with leaving it as is. But if you need to do any rerolls, it would be nice to, as a prepatory step, introduce a patch which 1. marks these strings for translation and 2. lowercases the first letter of the sentence which is the convention for messages that are seen by the end-user.

Quoted 4 lines

Perhaps:

        die(_("no such ref: '%s'"), str);

Sounds good! I'll add a patch to the end of this series that addresses these strings not being translated.

Thank you for your feedback.
~Alex

Quoted 33 lines

refs.c int dwim_ref(const char *str, int len, struct object_id *oid, char **ref) · line 651

Quoted 4 lines
refs.c +4 -0
+     case 1:
+             break; /* good */
+     default:
+             die("Ambiguous refname: '%s'", str);

Same comment here.

Thanks,

Denton

Quoted 22 lines
refs.c +4 -0
+     }
+}
+
+
 int expand_ref(struct repository *repo, const char *str, int len,
             struct object_id *oid, char **ref)
 {
refs.h +1 -1
index 730d05ad91..a961662382 100644
@@ -154,6 +154,7 @@ int repo_dwim_log(struct repository *r, const char *str, int len, struct object_
 int dwim_ref(const char *str, int len, struct object_id *oid, char **ref);
 int dwim_log(const char *str, int len, struct object_id *oid, char **ref);
 
+void dwim_ref_or_die(const char *str, int len, char **ref);
 /*
  * A ref_transaction represents a collection of reference updates that
  * should succeed or fail together.
--

2.17.1

author
#10 lore raw

PATCH v2 2/2 rebase: find --fork-point with full ref

In reply to Alex Torok

rebase --fork-point needs to look up the full ref name before calling get_fork_point in the same manner that merge-base --fork-point does.

---
 builtin/rebase.c             | 4 +++-
 t/t3431-rebase-fork-point.sh | 1 +
 2 files changed, 4 insertions(+), 1 deletion(-)
builtin/rebase.c +3 -1
index e755087b0f..821994f676 100644
@@ -1980,8 +1980,10 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
 		struct commit *head =
 			lookup_commit_reference(the_repository,
 						&options.orig_head);
+		char * full_name;
+		dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);
 		options.restrict_revision =
-			get_fork_point(options.upstream_name, head);
+			get_fork_point(full_name, head);
 	}
 
 	if (repo_read_index(the_repository) < 0)
t/t3431-rebase-fork-point.sh +1 -0
index 78851b9a2a..6ecdae918e 100755
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
 test_rebase 'G F E D B A' --fork-point refs/heads/master
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
 test_rebase 'G F C E D B A' refs/heads/master
 test_rebase 'G F C D B A' --onto D refs/heads/master
Signature · 1 line
2.17.1
#11 lore raw

Re: [PATCH v2 2/2] rebase: find --fork-point with full ref

Hi Alex,

Quoted 14 lines · Alex Torok

rebase --fork-point needs to look up the full ref name before calling get_fork_point in the same manner that merge-base --fork-point does.

---
 builtin/rebase.c             | 4 +++-
 t/t3431-rebase-fork-point.sh | 1 +
 2 files changed, 4 insertions(+), 1 deletion(-)
builtin/rebase.c +0 -0
index e755087b0f..821994f676 100644

builtin/rebase.c int cmd_rebase(int argc, const char **argv, const char *prefix) · line 1983

Quoted 4 lines
builtin/rebase.c +1 -0
@@ -1980,8 +1980,10 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
 		struct commit *head =
 			lookup_commit_reference(the_repository,
 						&options.orig_head);
+		char * full_name;

nit: * should be attached to the variable name.

builtin/rebase.c int cmd_rebase(int argc, const char **argv, const char *prefix) · line 1984

Quoted 1 line
builtin/rebase.c +1 -0
+		dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);

Also, thinking about this more, would it be possible to put the dwim_ref logic into get_fork_point() directly? There are currently only these two callers so I suspect it should be fine and it'll result in cleaner logic.

We could also squash it down into one patch.

Quoted 11 lines
builtin/rebase.c +1 -1
 		options.restrict_revision =
-			get_fork_point(options.upstream_name, head);
+			get_fork_point(full_name, head);
 	}
 
 	if (repo_read_index(the_repository) < 0)
t/t3431-rebase-fork-point.sh +0 -0
index 78851b9a2a..6ecdae918e 100755

t/t3431-rebase-fork-point.sh test_rebase 'G F C D B A' --no-fork-point --onto D · line 52

Quoted 4 lines
t/t3431-rebase-fork-point.sh +1 -0
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
 test_rebase 'G F E D B A' --fork-point refs/heads/master
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master

It's not obvious why this was failing in the first place. Perhaps we could document it better in the commit message?

Maybe something like:

We used to pass in the upstream_name directly into the get_fork_point() machinery. However, get_fork_point() was expecting a fully qualified ref name even though most users use the short name for branches. This resulted in --fork-point not working as expected since, without the full ref name, the reflog lookup would fail and it would behave as if we weren't passing in --fork-point at all.

Also, I'm not why this test case in particular that was duplicated (and not the one above) given that the first three --fork-point test cases fail without the change to rebase. Perhaps we want to duplicate all "refs/heads/master" tests with a corresponding "master" test?

Thanks,

Denton

Quoted 6 lines
t/t3431-rebase-fork-point.sh +0 -0
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
 test_rebase 'G F C E D B A' refs/heads/master
 test_rebase 'G F C D B A' --onto D refs/heads/master
Signature · 1 line
2.17.1
#12 lore raw

Re: [PATCH v2 2/2] rebase: find --fork-point with full ref

Hi Alex

Thanks for working on this

Quoted 12 lines · Denton Liu

Hi Alex,

Quoted 9 lines · Alex Torok

rebase --fork-point needs to look up the full ref name before calling get_fork_point in the same manner that merge-base --fork-point does.

---
  builtin/rebase.c             | 4 +++-
  t/t3431-rebase-fork-point.sh | 1 +
  2 files changed, 4 insertions(+), 1 deletion(-)
Quoted 11 lines

builtin/rebase.c int cmd_rebase(int argc, const char **argv, const char *prefix) · line 1983

Quoted 9 lines
builtin/rebase.c +1 -0
index e755087b0f..821994f676 100644
@@ -1980,8 +1980,10 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
  		struct commit *head =
  			lookup_commit_reference(the_repository,
  						&options.orig_head);
+		char * full_name;

nit: * should be attached to the variable name.

I think you also need to free it once you've called get_fork_point() as well.

Quoted 7 lines

builtin/rebase.c int cmd_rebase(int argc, const char **argv, const char *prefix) · line 1984

Quoted 2 lines
builtin/rebase.c +1 -0
+		dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);

Also, thinking about this more, would it be possible to put the dwim_ref logic into get_fork_point() directly? There are currently only these two callers so I suspect it should be fine and it'll result in cleaner logic.

If you do that then it would be better to use error() rather than die() in get_fork_point() and return an error to the caller as we try to avoid adding code to libgit that dies. This lets the caller handle any cleanup that they need to before exiting.

Best Wishes

Phillip

Quoted 47 lines

We could also squash it down into one patch.

Quoted 11 lines
builtin/rebase.c +1 -1
  		options.restrict_revision =
-			get_fork_point(options.upstream_name, head);
+			get_fork_point(full_name, head);
  	}
 
  	if (repo_read_index(the_repository) < 0)
t/t3431-rebase-fork-point.sh +0 -0
index 78851b9a2a..6ecdae918e 100755

t/t3431-rebase-fork-point.sh test_rebase 'G F C D B A' --no-fork-point --onto D · line 52

Quoted 4 lines
t/t3431-rebase-fork-point.sh +1 -0
@@ -49,6 +49,7 @@ test_rebase 'G F C D B A' --no-fork-point --onto D
  test_rebase 'G F C B A' --no-fork-point --keep-base
  test_rebase 'G F E D B A' --fork-point refs/heads/master
  test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master

It's not obvious why this was failing in the first place. Perhaps we could document it better in the commit message?

Maybe something like:

We used to pass in the upstream_name directly into the get_fork_point() machinery. However, get_fork_point() was expecting a fully qualified ref name even though most users use the short name for branches. This resulted in --fork-point not working as expected since, without the full ref name, the reflog lookup would fail and it would behave as if we weren't passing in --fork-point at all.

Also, I'm not why this test case in particular that was duplicated (and not the one above) given that the first three --fork-point test cases fail without the change to rebase. Perhaps we want to duplicate all "refs/heads/master" tests with a corresponding "master" test?

Thanks,

Denton

Quoted 6 lines
t/t3431-rebase-fork-point.sh +0 -0
  test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
  test_rebase 'G F C E D B A' refs/heads/master
  test_rebase 'G F C D B A' --onto D refs/heads/master
Signature · 1 line
2.17.1
author
#13 lore raw

Re: [PATCH v2 2/2] rebase: find --fork-point with full ref

Thank you for the feedback Denton & Phillip!

Quoted 5 lines · Phillip Wood
Quoted 1 line · Denton Liu

nit: * should be attached to the variable name.

I think you also need to free it once you've called get_fork_point() as well.

Yup. Got it.

Quoted 9 lines · Denton Liu
Quoted 7 lines
Quoted 2 lines

+ dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);

Also, thinking about this more, would it be possible to put the dwim_ref logic into get_fork_point() directly? There are currently only these two callers so I suspect it should be fine and it'll result in cleaner logic.

Quoted 4 lines

If you do that then it would be better to use error() rather than die() in get_fork_point() and return an error to the caller as we try to avoid adding code to libgit that dies. This lets the caller handle any cleanup that they need to before exiting.

Would the signature of get_fork_point change to be something like:
int get_fork_point(const char *refname, struct commit *commit,
struct commit **fork_point, struct strbuf *err)

If not, could you point me to an example of some existing code that does what you're talking about?

Quoted 6 lines · Denton Liu
Quoted 5 lines

It's not obvious why this was failing in the first place. Perhaps we could document it better in the commit message?

Maybe something like:

Quoted 7 lines
Quoted 7 lines

We used to pass in the upstream_name directly into the get_fork_point() machinery. However, get_fork_point() was expecting a fully qualified ref name even though most users use the short name for branches. This resulted in --fork-point not working as expected since, without the full ref name, the reflog lookup would fail and it would behave as if we weren't passing in --fork-point at all.

Sounds good.

Quoted 4 lines
Quoted 4 lines

Also, I'm not why this test case in particular that was duplicated (and not the one above) given that the first three --fork-point test cases fail without the change to rebase. Perhaps we want to duplicate all "refs/heads/master" tests with a corresponding "master" test?

I only duplicated one so that there would only be one test case that would fail if a regression around getting the fork point with a short ref name was introduced.

I just happened to pick that one because it was closest to the rebase command I was running when I found the bug :)

I can include some of the above reasoning in the commit message. Alternatively:
* I could duplicate all of tests
* I could change all of the tests to use the short ref name

I'm leaning towards just leaving one test (maybe with a comment?) for the short ref name --fork-point so that there is more resolution around where a bug could be on test failure.

Let me know what you think,
Alex

#14 lore raw

Re: [PATCH v2 2/2] rebase: find --fork-point with full refgg

Hi Alex,

Quoted 20 lines · Alex Torok

Thank you for the feedback Denton & Phillip!

Quoted 5 lines · Phillip Wood
Quoted 1 line · Denton Liu

nit: * should be attached to the variable name.

I think you also need to free it once you've called get_fork_point() as well.

Yup. Got it.

Quoted 9 lines · Denton Liu
Quoted 7 lines
Quoted 2 lines

+ dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);

Also, thinking about this more, would it be possible to put the dwim_ref logic into get_fork_point() directly? There are currently only these two callers so I suspect it should be fine and it'll result in cleaner logic.

Quoted 8 lines
Quoted 4 lines

If you do that then it would be better to use error() rather than die() in get_fork_point() and return an error to the caller as we try to avoid adding code to libgit that dies. This lets the caller handle any cleanup that they need to before exiting.

Would the signature of get_fork_point change to be something like:
int get_fork_point(const char *refname, struct commit *commit,
struct commit **fork_point, struct strbuf *err)

I would drop the last parameter. If an error is detected, you could just do

	return error(_("oh no, something bad happened"));

Even though we try and avoid dying in the middle of libgit, we print errors out very often so it should be fine here.

Quoted 12 lines
Quoted 4 lines
Quoted 4 lines

Also, I'm not why this test case in particular that was duplicated (and not the one above) given that the first three --fork-point test cases fail without the change to rebase. Perhaps we want to duplicate all "refs/heads/master" tests with a corresponding "master" test?

I only duplicated one so that there would only be one test case that would fail if a regression around getting the fork point with a short ref name was introduced.

I just happened to pick that one because it was closest to the rebase command I was running when I found the bug :)

Quoted 8 lines

I can include some of the above reasoning in the commit message. Alternatively:
* I could duplicate all of tests
* I could change all of the tests to use the short ref name

I'm leaning towards just leaving one test (maybe with a comment?) for the short ref name --fork-point so that there is more resolution around where a bug could be on test failure.

I would just duplicate all of the tests. When the tests are pretty cheap to run (as they are in this case), I tend to err on the side of adding more tests since they might catch more odd edge-cases but, in this case, all of the fork point logic goes through one common block so the duplicate logic doesn't really buy us anything.

I'm pretty impartial so I'll leave it up to you ;)

Thanks,

Denton

Quoted 3 lines

Let me know what you think,
Alex

#15 lore raw

Re: [PATCH v2 2/2] rebase: find --fork-point with full refgg

Quoted 32 lines · Denton Liu

Hi Alex,

Quoted 20 lines · Alex Torok

Thank you for the feedback Denton & Phillip!

Quoted 5 lines · Phillip Wood
Quoted 1 line · Denton Liu

nit: * should be attached to the variable name.

I think you also need to free it once you've called get_fork_point() as well.

Yup. Got it.

Quoted 9 lines · Denton Liu
Quoted 7 lines
Quoted 2 lines

+ dwim_ref_or_die(options.upstream_name, strlen(options.upstream_name), &full_name);

Also, thinking about this more, would it be possible to put the dwim_ref logic into get_fork_point() directly? There are currently only these two callers so I suspect it should be fine and it'll result in cleaner logic.

Quoted 8 lines
Quoted 4 lines

If you do that then it would be better to use error() rather than die() in get_fork_point() and return an error to the caller as we try to avoid adding code to libgit that dies. This lets the caller handle any cleanup that they need to before exiting.

Would the signature of get_fork_point change to be something like:
int get_fork_point(const char *refname, struct commit *commit,
struct commit **fork_point, struct strbuf *err)

Quoted 7 lines

I would drop the last parameter. If an error is detected, you could just do

	return error(_("oh no, something bad happened"));

Even though we try and avoid dying in the middle of libgit, we print errors out very often so it should be fine here.

Yes that was what I was thinking of

Best Wishes

Phillip

Quoted 37 lines
Quoted 13 lines
Quoted 5 lines
Quoted 5 lines

Also, I'm not why this test case in particular that was duplicated (and not the one above) given that the first three --fork-point test cases fail without the change to rebase. Perhaps we want to duplicate all "refs/heads/master" tests with a corresponding "master" test?

I only duplicated one so that there would only be one test case that would fail if a regression around getting the fork point with a short ref name was introduced.

I just happened to pick that one because it was closest to the rebase command I was running when I found the bug :)

Quoted 8 lines

I can include some of the above reasoning in the commit message. Alternatively:
* I could duplicate all of tests
* I could change all of the tests to use the short ref name

I'm leaning towards just leaving one test (maybe with a comment?) for the short ref name --fork-point so that there is more resolution around where a bug could be on test failure.

I would just duplicate all of the tests. When the tests are pretty cheap to run (as they are in this case), I tend to err on the side of adding more tests since they might catch more odd edge-cases but, in this case, all of the fork point logic goes through one common block so the duplicate logic doesn't really buy us anything.

I'm pretty impartial so I'll leave it up to you ;)

Thanks,

Denton

Quoted 3 lines

Let me know what you think,
Alex

author
#16 lore raw

v3 cover rebase: fix --fork-point with short ref upstream

In reply to Alex Torok

Move the dwim ref logic into get_fork_point() and return an error code so that callers of get_fork_point() know if something went wrong.

Alex Torok (1):
rebase: fix --fork-point with short refname

 builtin/merge-base.c         | 14 +++-----------
 builtin/rebase.c             |  5 +++--
 commit.c                     | 19 +++++++++++++------
 commit.h                     |  2 +-
 refs.c                       | 14 ++++++++++++++
 refs.h                       |  2 ++
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 7 files changed, 56 insertions(+), 20 deletions(-)
Signature · 1 line
2.17.1
author
#17 lore raw

PATCH v3 1/1 rebase: fix --fork-point with short refname

We were directly passing in the user-provided upstream_name into get_fork_point(), but get_fork_point() was expecting a fully qualified ref name. This resulted in --fork-point not working as expected. Without the full ref name, get_fork_point could not find the fork point, and rebase behaved as if no --fork-point flag was passed in.

Pull logic for getting the full dwim ref name out of merge-base.c handle_fork_point into get_fork_point. This allows both of the locations that call get_fork_point to not need to handle doing the dwim ref lookup before trying to get the fork point.

Duplicate all of the rebase tests with a short refname to ensure that they work with short and long refnames.

---

One thing that I'm not super sure about is the error messaging that gets printed when a short refname is given to rebase. In the ambigous refname test that I added, the command output is:

warning: refname 'one' is ambiguous.
error: ambiguous refname: 'one'
fatal: could not get fork point

Which seems a bit odd ot have a warning, error, fatal stacked on top of each other. From a user-facing perspective it feels a bit odd to me, but I'm not sure what the general rules for error messaging are in git.

 builtin/merge-base.c         | 14 +++-----------
 builtin/rebase.c             |  5 +++--
 commit.c                     | 19 +++++++++++++------
 commit.h                     |  2 +-
 refs.c                       | 14 ++++++++++++++
 refs.h                       |  2 ++
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 7 files changed, 56 insertions(+), 20 deletions(-)
builtin/merge-base.c +3 -11
index e3f8da13b6..ecc4268d43 100644
@@ -114,26 +114,18 @@ static int handle_is_ancestor(int argc, const char **argv)
 static int handle_fork_point(int argc, const char **argv)
 {
 	struct object_id oid;
-	char *refname;
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
-
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
 		die("Not a valid object name: '%s'", commitname);
 
 	derived = lookup_commit_reference(the_repository, &oid);
 
-	fork_point = get_fork_point(refname, derived);
+	fork_point = NULL;
+	if (get_fork_point(argv[0], derived, &fork_point) < 0)
+		return 1;
 
 	if (!fork_point)
 		return 1;
builtin/rebase.c +3 -2
index e755087b0f..782cf5a890 100644
@@ -1980,8 +1980,9 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
 		struct commit *head =
 			lookup_commit_reference(the_repository,
 						&options.orig_head);
-		options.restrict_revision =
-			get_fork_point(options.upstream_name, head);
+		options.restrict_revision = NULL;
+		if (get_fork_point(options.upstream_name, head, &options.restrict_revision) < 0)
+			die(_("could not get fork point"));
 	}
 
 	if (repo_read_index(the_repository) < 0)
commit.c +13 -6
index 434ec030d6..874bc0cdf1 100644
@@ -920,19 +920,25 @@ static int collect_one_reflog_ent(struct object_id *ooid, struct object_id *noid
 	return 0;
 }
 
-struct commit *get_fork_point(const char *refname, struct commit *commit)
+int get_fork_point(const char *refname, struct commit *commit, struct commit **fork_point)
 {
 	struct object_id oid;
 	struct rev_collect revs;
 	struct commit_list *bases;
 	int i;
-	struct commit *ret = NULL;
+	char *full_refname;
+
+	if (dwim_unique_ref(refname, strlen(refname), &oid, &full_refname) < 0) {
+		free(full_refname);
+		return -1;
+	}
+
 
 	memset(&revs, 0, sizeof(revs));
 	revs.initial = 1;
-	for_each_reflog_ent(refname, collect_one_reflog_ent, &revs);
+	for_each_reflog_ent(full_refname, collect_one_reflog_ent, &revs);
 
-	if (!revs.nr && !get_oid(refname, &oid))
+	if (!revs.nr && !get_oid(full_refname, &oid))
 		add_one_commit(&oid, &revs);
 
 	for (i = 0; i < revs.nr; i++)
@@ -954,11 +960,12 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 	if (revs.nr <= i)
 		goto cleanup_return;
 
-	ret = bases->item;
+	*fork_point = bases->item;
 
 cleanup_return:
+	free(full_refname);
 	free_commit_list(bases);
-	return ret;
+	return 0;
 }
 
 static const char gpg_sig_header[] = "gpgsig";
commit.h +1 -1
index 221cdaa34b..d79a8eab48 100644
@@ -240,7 +240,7 @@ int register_commit_graft(struct repository *r, struct commit_graft *, int);
 void prepare_commit_graft(struct repository *r);
 struct commit_graft *lookup_commit_graft(struct repository *r, const struct object_id *oid);
 
-struct commit *get_fork_point(const char *refname, struct commit *commit);
+int get_fork_point(const char *refname, struct commit *commit, struct commit **fork_point);
 
 /* largest positive number a signed 32-bit integer can contain */
 #define INFINITE_DEPTH 0x7fffffff
refs.c +14 -0
index 1ab0bb54d3..853e45a6a3 100644
@@ -639,6 +639,20 @@ int dwim_ref(const char *str, int len, struct object_id *oid, char **ref)
 	return repo_dwim_ref(the_repository, str, len, oid, ref);
 }
 
+int dwim_unique_ref(const char* str, int len, struct object_id *oid, char **ref)
+{
+	struct object_id discard;
+	switch (dwim_ref(str, len, &discard, ref)) {
+	case 0:
+		return error(_("no such ref: '%s'"), str);
+	case 1:
+		break; /* good */
+	default:
+		return error(_("ambiguous refname: '%s'"), str);
+	}
+	return 0;
+}
+
 int expand_ref(struct repository *repo, const char *str, int len,
 	       struct object_id *oid, char **ref)
 {
refs.h +2 -0
index 730d05ad91..7aa348052b 100644
@@ -154,6 +154,8 @@ int repo_dwim_log(struct repository *r, const char *str, int len, struct object_
 int dwim_ref(const char *str, int len, struct object_id *oid, char **ref);
 int dwim_log(const char *str, int len, struct object_id *oid, char **ref);
 
+int dwim_unique_ref(const char* str, int len, struct object_id *oid, char **ref);
+
 /*
  * A ref_transaction represents a collection of reference updates that
  * should succeed or fail together.
t/t3431-rebase-fork-point.sh +20 -0
index 78851b9a2a..5b09aecd13 100755
@@ -47,11 +47,31 @@ test_rebase 'G F B A' --keep-base
 test_rebase 'G F C E D B A' --no-fork-point
 test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
+
 test_rebase 'G F E D B A' --fork-point refs/heads/master
+test_rebase 'G F E D B A' --fork-point master
+
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
+
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
+test_rebase 'G F B A' --fork-point --keep-base master
+
 test_rebase 'G F C E D B A' refs/heads/master
+test_rebase 'G F C E D B A' master
+
 test_rebase 'G F C D B A' --onto D refs/heads/master
+test_rebase 'G F C D B A' --onto D master
+
 test_rebase 'G F C B A' --keep-base refs/heads/master
+test_rebase 'G F C B A' --keep-base master
+
+test_expect_success "git rebase --fork-point with ambigous refname" "
+	git checkout master &&
+	git checkout -b one &&
+	git checkout side &&
+	git tag one &&
+	test_must_fail git rebase --fork-point --onto D one
+"
 
 test_done
Signature · 1 line
2.17.1
#18 lore raw

Re: [PATCH v3 1/1] rebase: fix --fork-point with short refname

Quoted 8 lines · Alex Torok

We were directly passing in the user-provided upstream_name into get_fork_point(), but get_fork_point() was expecting a fully qualified ref name. This resulted in --fork-point not working as expected. Without the full ref name, get_fork_point could not find the fork point, and rebase behaved as if no --fork-point flag was passed in.

Pull logic for getting the full dwim ref name out of merge-base.c handle_fork_point into get_fork_point. This allows both of the locations

"handle_fork_point() into get_fork_point()" to match the second line of the first paragraph, perhaps (similarly for "get_fork_point()" on the next line)?

Quoted 13 lines

that call get_fork_point to not need to handle doing the dwim ref lookup before trying to get the fork point.

Duplicate all of the rebase tests with a short refname to ensure that they work with short and long refnames.

---

One thing that I'm not super sure about is the error messaging that gets printed when a short refname is given to rebase. In the ambigous refname test that I added, the command output is:

Quoted 6 lines

warning: refname 'one' is ambiguous.
error: ambiguous refname: 'one'
fatal: could not get fork point

Which seems a bit odd ot have a warning, error, fatal stacked on top of each other.

Yes, it does look strange.

I think that the new "dwim_unique_ref()" helper is misdesigned as a generic helper function. It makes it hard for the callers to handle errors on their own for this function to return error().

A helper at the right level for this kind of thing would have been a function that does DWIM ref and returns the number of candidate refs the given short name would expand to, and do all that silently. The caller would then react to the returned number and say "no such" if there is no candidate, or "ambigous" if there are more than one.

Which is exactly what dwim_ref() is. In other words, dwim_unique_ref() helper, unless it is useful by multiple callers who want it to die() itself, is not all that useful as an abstraction.

If I were doing this, probably I would not add dwim_unique_ref() at all, and then I'd invent an extra variant of get_fork_point(), that the caller can choose to make it die or just silently return an error.

This caller would want it to die() without giving control back to it, but there may be some other callers that would want a finer control.

On the other hand, if the other caller(s) all want get_fork_point() to die, then that would also be fine. A quick glance tells me that the only other caller is in builtin/rebase.c and does this:

	if (fork_point > 0) {
		struct commit *head =
			lookup_commit_reference(the_repository,
						&options.orig_head);
		options.restrict_revision =
			get_fork_point(options.upstream_name, head);
	}

and there are other calls to die(_("...")) in the caller everywhere, so I'd say you are over-engineering a simple bugfix.

Wouldn't it be sufficient for this fix to be more like this?

-- >8 --

Subject: rebase: --fork-point regression fix

"git rebase --fork-point master" used to work OK, as it internally called "git merge-base --fork-point" that knew how to handle short refname and dwim it to the full refname before calling the underlying get_fork_point() function.

This is no longer true after the command was rewritten in C, as its internall call made directly to get_fork_point() does not dwim a short ref.

Move the "dwim the refname argument to the full refname" logic that is used in "git merge-base" to the underlying get_fork_point() function, so that the other caller of the function in the implementation of "git rebase" behaves the same way to fix this regression.

---
 builtin/merge-base.c         | 12 +-----------
 commit.c                     | 15 +++++++++++++--
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 3 files changed, 34 insertions(+), 13 deletions(-)
builtin/merge-base.c +1 -11
index e3f8da13b6..6719ac198d 100644
@@ -114,26 +114,16 @@ static int handle_is_ancestor(int argc, const char **argv)
 static int handle_fork_point(int argc, const char **argv)
 {
 	struct object_id oid;
-	char *refname;
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
-
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
 		die("Not a valid object name: '%s'", commitname);
 
 	derived = lookup_commit_reference(the_repository, &oid);
 
-	fork_point = get_fork_point(refname, derived);
+	fork_point = get_fork_point(argv[0], derived);
 
 	if (!fork_point)
 		return 1;
commit.c +13 -2
index 40890ae7ce..016f14fe95 100644
@@ -903,12 +903,22 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 	struct commit_list *bases;
 	int i;
 	struct commit *ret = NULL;
+	char *full_refname;
+
+	switch (dwim_ref(refname, strlen(refname), &oid, &full_refname)) {
+	case 0:
+		die("No such ref: '%s'", refname);
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", refname);
+	}
 
 	memset(&revs, 0, sizeof(revs));
 	revs.initial = 1;
-	for_each_reflog_ent(refname, collect_one_reflog_ent, &revs);
+	for_each_reflog_ent(full_refname, collect_one_reflog_ent, &revs);
 
-	if (!revs.nr && !get_oid(refname, &oid))
+	if (!revs.nr)
 		add_one_commit(&oid, &revs);
 
 	for (i = 0; i < revs.nr; i++)
@@ -934,6 +944,7 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 
 cleanup_return:
 	free_commit_list(bases);
+	free(full_refname);
 	return ret;
 }
 
t/t3431-rebase-fork-point.sh +20 -0
index 78851b9a2a..5b09aecd13 100755
@@ -47,11 +47,31 @@ test_rebase 'G F B A' --keep-base
 test_rebase 'G F C E D B A' --no-fork-point
 test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
+
 test_rebase 'G F E D B A' --fork-point refs/heads/master
+test_rebase 'G F E D B A' --fork-point master
+
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
+
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
+test_rebase 'G F B A' --fork-point --keep-base master
+
 test_rebase 'G F C E D B A' refs/heads/master
+test_rebase 'G F C E D B A' master
+
 test_rebase 'G F C D B A' --onto D refs/heads/master
+test_rebase 'G F C D B A' --onto D master
+
 test_rebase 'G F C B A' --keep-base refs/heads/master
+test_rebase 'G F C B A' --keep-base master
+
+test_expect_success "git rebase --fork-point with ambigous refname" "
+	git checkout master &&
+	git checkout -b one &&
+	git checkout side &&
+	git tag one &&
+	test_must_fail git rebase --fork-point --onto D one
+"
 
 test_done
author
#19 lore raw

Re: [PATCH v3 1/1] rebase: fix --fork-point with short refname

Quoted 4 lines · Junio C Hamano

and there are other calls to die(_("...")) in the caller everywhere, so I'd say you are over-engineering a simple bugfix.

Wouldn't it be sufficient for this fix to be more like this?

This is essentially what I had done in the v2 version of my patch with the only "big" difference being pulling out the dwim_ref() switching and dying logic into a dwim_ref_or_die() function.

Let me know if you want me to adjust my patch at all (my v2 patch is missing a call to free()...).

If the patch that you replied here with is sufficient for how you would fix it, I'm fine with you just using that.

Quoted 130 lines

-- >8 --

Subject: rebase: --fork-point regression fix

"git rebase --fork-point master" used to work OK, as it internally called "git merge-base --fork-point" that knew how to handle short refname and dwim it to the full refname before calling the underlying get_fork_point() function.

This is no longer true after the command was rewritten in C, as its internall call made directly to get_fork_point() does not dwim a short ref.

Move the "dwim the refname argument to the full refname" logic that is used in "git merge-base" to the underlying get_fork_point() function, so that the other caller of the function in the implementation of "git rebase" behaves the same way to fix this regression.

---
 builtin/merge-base.c         | 12 +-----------
 commit.c                     | 15 +++++++++++++--
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 3 files changed, 34 insertions(+), 13 deletions(-)
builtin/merge-base.c +1 -11
index e3f8da13b6..6719ac198d 100644
@@ -114,26 +114,16 @@ static int handle_is_ancestor(int argc, const char **argv)
 static int handle_fork_point(int argc, const char **argv)
 {
        struct object_id oid;
-       char *refname;
        struct commit *derived, *fork_point;
        const char *commitname;
 
-       switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-       case 0:
-               die("No such ref: '%s'", argv[0]);
-       case 1:
-               break; /* good */
-       default:
-               die("Ambiguous refname: '%s'", argv[0]);
-       }
-
        commitname = (argc == 2) ? argv[1] : "HEAD";
        if (get_oid(commitname, &oid))
                die("Not a valid object name: '%s'", commitname);
 
        derived = lookup_commit_reference(the_repository, &oid);
 
-       fork_point = get_fork_point(refname, derived);
+       fork_point = get_fork_point(argv[0], derived);
 
        if (!fork_point)
                return 1;
commit.c +13 -2
index 40890ae7ce..016f14fe95 100644
@@ -903,12 +903,22 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
        struct commit_list *bases;
        int i;
        struct commit *ret = NULL;
+       char *full_refname;
+
+       switch (dwim_ref(refname, strlen(refname), &oid, &full_refname)) {
+       case 0:
+               die("No such ref: '%s'", refname);
+       case 1:
+               break; /* good */
+       default:
+               die("Ambiguous refname: '%s'", refname);
+       }
 
        memset(&revs, 0, sizeof(revs));
        revs.initial = 1;
-       for_each_reflog_ent(refname, collect_one_reflog_ent, &revs);
+       for_each_reflog_ent(full_refname, collect_one_reflog_ent, &revs);
 
-       if (!revs.nr && !get_oid(refname, &oid))
+       if (!revs.nr)
                add_one_commit(&oid, &revs);
 
        for (i = 0; i < revs.nr; i++)
@@ -934,6 +944,7 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 
 cleanup_return:
        free_commit_list(bases);
+       free(full_refname);
        return ret;
 }
 
t/t3431-rebase-fork-point.sh +20 -0
index 78851b9a2a..5b09aecd13 100755
@@ -47,11 +47,31 @@ test_rebase 'G F B A' --keep-base
 test_rebase 'G F C E D B A' --no-fork-point
 test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
+
 test_rebase 'G F E D B A' --fork-point refs/heads/master
+test_rebase 'G F E D B A' --fork-point master
+
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
+
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
+test_rebase 'G F B A' --fork-point --keep-base master
+
 test_rebase 'G F C E D B A' refs/heads/master
+test_rebase 'G F C E D B A' master
+
 test_rebase 'G F C D B A' --onto D refs/heads/master
+test_rebase 'G F C D B A' --onto D master
+
 test_rebase 'G F C B A' --keep-base refs/heads/master
+test_rebase 'G F C B A' --keep-base master
+
+test_expect_success "git rebase --fork-point with ambigous refname" "
+       git checkout master &&
+       git checkout -b one &&
+       git checkout side &&
+       git tag one &&
+       test_must_fail git rebase --fork-point --onto D one
+"
 
 test_done
#20 lore raw

Re: [PATCH v3 1/1] rebase: fix --fork-point with short refname

In reply to Junio C Hamano

Hi Alex,

There are a couple of small issues with the patch below. If you iterate on it, here are some small suggestions:

Quoted 34 lines · Junio C Hamano

Subject: rebase: --fork-point regression fix

"git rebase --fork-point master" used to work OK, as it internally called "git merge-base --fork-point" that knew how to handle short refname and dwim it to the full refname before calling the underlying get_fork_point() function.

This is no longer true after the command was rewritten in C, as its internall call made directly to get_fork_point() does not dwim a short ref.

Move the "dwim the refname argument to the full refname" logic that is used in "git merge-base" to the underlying get_fork_point() function, so that the other caller of the function in the implementation of "git rebase" behaves the same way to fix this regression.

---
 builtin/merge-base.c         | 12 +-----------
 commit.c                     | 15 +++++++++++++--
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 3 files changed, 34 insertions(+), 13 deletions(-)
builtin/merge-base.c +0 -1
index e3f8da13b6..6719ac198d 100644
@@ -114,26 +114,16 @@ static int handle_is_ancestor(int argc, const char **argv)
 static int handle_fork_point(int argc, const char **argv)
 {
 	struct object_id oid;
-	char *refname;
 	struct commit *derived, *fork_point;
 	const char *commitname;

builtin/merge-base.c static int handle_is_ancestor(int argc, const char **argv) · line 129

Quoted 10 lines
builtin/merge-base.c +0 -9
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
-

In the unabridged v3, you cleaned this up by marking it for translation and also lowercasing the first letter of the sentence. That was good.

If you iterate on this, could you create a preparatory patch before this one that gives the same treatment to the other strings in merge-base? I count 8 instances of die() being called in merge-base so I think that the cleanup shouldn't be too arduous.

If we have a preparatory patch before this one, this patch will mostly be pure code movement which would be nice.

Quoted 19 lines
builtin/merge-base.c +1 -1
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
 		die("Not a valid object name: '%s'", commitname);
 
 	derived = lookup_commit_reference(the_repository, &oid);
 
-	fork_point = get_fork_point(refname, derived);
+	fork_point = get_fork_point(argv[0], derived);
 
 	if (!fork_point)
 		return 1;
commit.c +0 -0
index 40890ae7ce..016f14fe95 100644
@@ -903,12 +903,22 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 	struct commit_list *bases;
 	int i;
 	struct commit *ret = NULL;

commit.c struct commit *get_fork_point(const char *refname, struct commit *commit) · line 915

Quoted 10 lines
commit.c +10 -0
+	char *full_refname;
+
+	switch (dwim_ref(refname, strlen(refname), &oid, &full_refname)) {
+	case 0:
+		die("No such ref: '%s'", refname);
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", refname);
+	}

Yeah, we should fix these strings in the next iteration.

Quoted 43 lines
commit.c +3 -2
 	memset(&revs, 0, sizeof(revs));
 	revs.initial = 1;
-	for_each_reflog_ent(refname, collect_one_reflog_ent, &revs);
+	for_each_reflog_ent(full_refname, collect_one_reflog_ent, &revs);
 
-	if (!revs.nr && !get_oid(refname, &oid))
+	if (!revs.nr)
 		add_one_commit(&oid, &revs);
 
 	for (i = 0; i < revs.nr; i++)
@@ -934,6 +944,7 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 
 cleanup_return:
 	free_commit_list(bases);
+	free(full_refname);
 	return ret;
 }
 
t/t3431-rebase-fork-point.sh +11 -0
index 78851b9a2a..5b09aecd13 100755
@@ -47,11 +47,31 @@ test_rebase 'G F B A' --keep-base
 test_rebase 'G F C E D B A' --no-fork-point
 test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
+
 test_rebase 'G F E D B A' --fork-point refs/heads/master
+test_rebase 'G F E D B A' --fork-point master
+
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
+
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
+test_rebase 'G F B A' --fork-point --keep-base master
+
 test_rebase 'G F C E D B A' refs/heads/master
+test_rebase 'G F C E D B A' master
+
 test_rebase 'G F C D B A' --onto D refs/heads/master
+test_rebase 'G F C D B A' --onto D master
+

t/t3431-rebase-fork-point.sh test_rebase 'G F B A' --keep-base · line 75

Quoted 10 lines
t/t3431-rebase-fork-point.sh +9 -0
 test_rebase 'G F C B A' --keep-base refs/heads/master
+test_rebase 'G F C B A' --keep-base master
+
+test_expect_success "git rebase --fork-point with ambigous refname" "
+	git checkout master &&
+	git checkout -b one &&
+	git checkout side &&
+	git tag one &&
+	test_must_fail git rebase --fork-point --onto D one
+"

nit: use double-quotes instead of single-quotes to surround both the test case name and the actual code itself.

Thanks,

Denton

Quoted 2 lines
t/t3431-rebase-fork-point.sh +0 -0
 test_done
#21 lore raw

Re: [PATCH v3 1/1] rebase: fix --fork-point with short refname

Quoted 10 lines · Denton Liu

t/t3431-rebase-fork-point.sh test_rebase 'G F B A' --keep-base · line 75

Quoted 7 lines
t/t3431-rebase-fork-point.sh +7 -0
+test_expect_success "git rebase --fork-point with ambigous refname" "
+     git checkout master &&
+     git checkout -b one &&
+     git checkout side &&
+     git tag one &&
+     test_must_fail git rebase --fork-point --onto D one
+"

nit: use double-quotes instead of single-quotes to surround both the test case name and the actual code itself.

Denton meant to say that you should single-quote the test title and test body (not double-quote). In this particular test, the distinction doesn't matter presently, though it could matter if the test is later changed. In particular, if the test does any sort of variable interpolation (such as $foo), then the single quotes ensure that the interpolation will occur when the test is actually run (which is almost always what is desired) rather than at the time the test is defined (which is almost never wanted).

#22 lore raw

Re: [PATCH v4 1/1] rebase: --fork-point regression fix

"git rebase --fork-point master" used to work OK, as it internally called "git merge-base --fork-point" that knew how to handle short refname and dwim it to the full refname before calling the underlying get_fork_point() function.

This is no longer true after the command was rewritten in C, as its internall call made directly to get_fork_point() does not dwim a short ref.

Move the "dwim the refname argument to the full refname" logic that is used in "git merge-base" to the underlying get_fork_point() function, so that the other caller of the function in the implementation of "git rebase" behaves the same way to fix this regression.

[jc: revamped the fix and used Alex's tests]

---

* I was about to discard stalled topics and this caught my eyes.

 builtin/merge-base.c         | 12 +-----------
 commit.c                     | 15 +++++++++++++--
 t/t3431-rebase-fork-point.sh | 20 ++++++++++++++++++++
 3 files changed, 34 insertions(+), 13 deletions(-)
builtin/merge-base.c +1 -11
index e3f8da13b6..6719ac198d 100644
@@ -114,26 +114,16 @@ static int handle_is_ancestor(int argc, const char **argv)
 static int handle_fork_point(int argc, const char **argv)
 {
 	struct object_id oid;
-	char *refname;
 	struct commit *derived, *fork_point;
 	const char *commitname;
 
-	switch (dwim_ref(argv[0], strlen(argv[0]), &oid, &refname)) {
-	case 0:
-		die("No such ref: '%s'", argv[0]);
-	case 1:
-		break; /* good */
-	default:
-		die("Ambiguous refname: '%s'", argv[0]);
-	}
-
 	commitname = (argc == 2) ? argv[1] : "HEAD";
 	if (get_oid(commitname, &oid))
 		die("Not a valid object name: '%s'", commitname);
 
 	derived = lookup_commit_reference(the_repository, &oid);
 
-	fork_point = get_fork_point(refname, derived);
+	fork_point = get_fork_point(argv[0], derived);
 
 	if (!fork_point)
 		return 1;
commit.c +13 -2
index 40890ae7ce..016f14fe95 100644
@@ -903,12 +903,22 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 	struct commit_list *bases;
 	int i;
 	struct commit *ret = NULL;
+	char *full_refname;
+
+	switch (dwim_ref(refname, strlen(refname), &oid, &full_refname)) {
+	case 0:
+		die("No such ref: '%s'", refname);
+	case 1:
+		break; /* good */
+	default:
+		die("Ambiguous refname: '%s'", refname);
+	}
 
 	memset(&revs, 0, sizeof(revs));
 	revs.initial = 1;
-	for_each_reflog_ent(refname, collect_one_reflog_ent, &revs);
+	for_each_reflog_ent(full_refname, collect_one_reflog_ent, &revs);
 
-	if (!revs.nr && !get_oid(refname, &oid))
+	if (!revs.nr)
 		add_one_commit(&oid, &revs);
 
 	for (i = 0; i < revs.nr; i++)
@@ -934,6 +944,7 @@ struct commit *get_fork_point(const char *refname, struct commit *commit)
 
 cleanup_return:
 	free_commit_list(bases);
+	free(full_refname);
 	return ret;
 }
 
t/t3431-rebase-fork-point.sh +20 -0
index 78851b9a2a..172562789e 100755
@@ -47,11 +47,31 @@ test_rebase 'G F B A' --keep-base
 test_rebase 'G F C E D B A' --no-fork-point
 test_rebase 'G F C D B A' --no-fork-point --onto D
 test_rebase 'G F C B A' --no-fork-point --keep-base
+
 test_rebase 'G F E D B A' --fork-point refs/heads/master
+test_rebase 'G F E D B A' --fork-point master
+
 test_rebase 'G F D B A' --fork-point --onto D refs/heads/master
+test_rebase 'G F D B A' --fork-point --onto D master
+
 test_rebase 'G F B A' --fork-point --keep-base refs/heads/master
+test_rebase 'G F B A' --fork-point --keep-base master
+
 test_rebase 'G F C E D B A' refs/heads/master
+test_rebase 'G F C E D B A' master
+
 test_rebase 'G F C D B A' --onto D refs/heads/master
+test_rebase 'G F C D B A' --onto D master
+
 test_rebase 'G F C B A' --keep-base refs/heads/master
+test_rebase 'G F C B A' --keep-base master
+
+test_expect_success 'git rebase --fork-point with ambigous refname' '
+	git checkout master &&
+	git checkout -b one &&
+	git checkout side &&
+	git tag one &&
+	test_must_fail git rebase --fork-point --onto D one
+'
 
 test_done