git@vger.kernel.org list mirror (unofficial, one of many)
 help / color / mirror / code / Atom feed
From: Benno Evers <benno@bmevers.de>
To: git@vger.kernel.org
Cc: spearce@spearce.org, Benno Evers <benno@bmevers.de>
Subject: [PATCH] describe: dont abort too early when searching tags
Date: Sun, 23 Feb 2020 13:51:02 +0100	[thread overview]
Message-ID: <20200223125102.6697-1-benno@bmevers.de> (raw)

When searching the commit graph for tag candidates, `git-describe`
will stop as soon as there is only one active branch left and
it already found an annotated tag as a candidate.

This works well as long as all branches eventually connect back
to a common root, but if the tags are found across branches
with no common ancestor

                  B
                  o----.
                        \
          o-----o---o----x
          A

it can happen that the search on one branch terminates prematurely
because a tag was found on another, independent branch. This scenario
isn't quite as obscure as it sounds, since cloning with a limited
depth often introduces many independent "dead ends" into the commit
graph.

The help text of `git-describe` states pretty clearly that when
describing a commit D, the number appended to the emitted tag X should
correspond to the number of commits found by `git log X..D`.

Thus, this commit modifies the stopping condition to only abort
the search when only one branch is left to search *and* all current
best candidates are descendants from that branch.

Signed-off-by: Benno Evers <benno@bmevers.de>
---
We originally found this issue in one of our internal CI jobs,
which relied on `git-describe` to assign filenames to the generated
artifacts.
---
 builtin/describe.c  | 22 +++++++++++++++----
 t/t6120-describe.sh | 53 ++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 70 insertions(+), 5 deletions(-)

diff --git a/builtin/describe.c b/builtin/describe.c
index b6df81d8d0..420f4c6401 100644
--- a/builtin/describe.c
+++ b/builtin/describe.c
@@ -376,11 +376,25 @@ static void describe_commit(struct object_id *oid, struct strbuf *dst)
 			if (!(c->object.flags & t->flag_within))
 				t->depth++;
 		}
+		/* Stop if last remaining path already covered by best candidate(s) */
 		if (annotated_cnt && !list) {
-			if (debug)
-				fprintf(stderr, _("finished search at %s\n"),
-					oid_to_hex(&c->object.oid));
-			break;
+			int best_depth = INT_MAX;
+			unsigned best_within = 0;
+			for (cur_match = 0; cur_match < match_cnt; cur_match++) {
+				struct possible_tag *t = &all_matches[cur_match];
+				if (t->depth < best_depth) {
+					best_depth = t->depth;
+					best_within = t->flag_within;
+				} else if (t->depth == best_depth) {
+					best_within |= t->flag_within;
+				}
+			}
+			if ((c->object.flags & best_within) == best_within) {
+				if (debug)
+					fprintf(stderr, _("finished search at %s\n"),
+						oid_to_hex(&c->object.oid));
+				break;
+			}
 		}
 		while (parents) {
 			struct commit *p = parents->item;
diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
index 09c50f3f04..d8cc08258e 100755
--- a/t/t6120-describe.sh
+++ b/t/t6120-describe.sh
@@ -479,4 +479,55 @@ test_expect_success 'name-rev covers all conditions while looking at parents' '
 	)
 '
 
-test_done
+#               B
+#               o
+#                \
+#  o-----o---o----x
+#        A
+#
+test_expect_success 'describe commits with disjoint bases' '
+	git init disjoint1 &&
+	(
+		cd disjoint1 &&
+
+		echo o >> file && git add file && git commit -m o &&
+		echo A >> file && git add file && git commit -m A &&
+		git tag A -a -m A &&
+		echo o >> file && git add file && git commit -m o &&
+
+		git checkout --orphan branch && rm file &&
+		echo B > file2 && git add file2 && git commit -m B &&
+		git tag B -a -m B &&
+		git merge --no-ff --allow-unrelated-histories master -m x &&
+
+		check_describe "A-3-*" HEAD
+	)
+'
+
+#           B
+#   o---o---o------------.
+#                         \
+#                  o---o---x
+#                  A
+#
+test_expect_success 'describe commits with disjoint bases 2' '
+	git init disjoint2 &&
+	(
+		cd disjoint2 &&
+
+		echo A >> file && git add file && GIT_COMMITTER_DATE="2020-01-01 18:00" git commit -m A &&
+		git tag A -a -m A &&
+		echo o >> file && git add file && GIT_COMMITTER_DATE="2020-01-01 18:01" git commit -m o &&
+
+		git checkout --orphan branch &&
+		echo o >> file2 && git add file2 && GIT_COMMITTER_DATE="2020-01-01 15:00" git commit -m o &&
+		echo o >> file2 && git add file2 && GIT_COMMITTER_DATE="2020-01-01 15:01" git commit -m o &&
+		echo B >> file2 && git add file2 && GIT_COMMITTER_DATE="2020-01-01 15:02" git commit -m B &&
+		git tag B -a -m B &&
+		git merge --no-ff --allow-unrelated-histories master -m x &&
+
+		check_describe "B-3-*" HEAD
+	)
+'
+
+test_done
\ No newline at end of file
-- 
2.20.1


             reply	other threads:[~2020-02-23 12:58 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-23 12:51 Benno Evers [this message]
2020-02-24 20:52 ` Junio C Hamano
2020-02-25 19:07   ` benno
2020-02-25 20:10     ` Junio C Hamano
2021-02-28 19:54 ` [PATCH 00/10] describe tests: refactor & fix recent broken tests Ævar Arnfjörð Bjarmason
2021-03-09  0:47   ` Junio C Hamano
2021-03-12 23:59   ` Junio C Hamano
2021-04-12 11:21   ` [PATCH v2 0/5] describe test fixes Ævar Arnfjörð Bjarmason
2021-04-12 11:21     ` [PATCH v2 1/5] describe tests: improve test for --work-tree & --dirty Ævar Arnfjörð Bjarmason
2021-04-12 11:21     ` [PATCH v2 2/5] describe tests: refactor away from glob matching Ævar Arnfjörð Bjarmason
2021-04-12 11:21     ` [PATCH v2 3/5] describe tests: don't rely on err.actual from "check_describe" Ævar Arnfjörð Bjarmason
2021-04-12 11:21     ` [PATCH v2 4/5] describe tests: fix nested "test_expect_success" call Ævar Arnfjörð Bjarmason
2021-04-12 11:21     ` [PATCH v2 5/5] describe tests: support -C in "check_describe" Ævar Arnfjörð Bjarmason
2021-04-12 11:33   ` [PATCH v2 0/2] svn tests: trivial "set -e" in main body of test fixes Ævar Arnfjörð Bjarmason
2021-04-12 11:33     ` [PATCH v2 1/2] svn tests: remove legacy re-setup from init-clone test Ævar Arnfjörð Bjarmason
2021-04-12 11:33     ` [PATCH v2 2/2] svn tests: refactor away a "set -e" in test body Ævar Arnfjörð Bjarmason
2021-02-28 19:54 ` [PATCH 01/10] describe tests: improve test for --work-tree & --dirty Ævar Arnfjörð Bjarmason
2021-02-28 19:54 ` [PATCH 02/10] describe tests: refactor away from glob matching Ævar Arnfjörð Bjarmason
2021-03-01 21:26   ` Junio C Hamano
2021-02-28 19:54 ` [PATCH 03/10] describe tests: always assert empty stderr from "describe" Ævar Arnfjörð Bjarmason
2021-03-01 21:32   ` Junio C Hamano
2021-02-28 19:54 ` [PATCH 04/10] test-lib functions: add an --annotated-tag option to "test_commit" Ævar Arnfjörð Bjarmason
2021-03-01 21:41   ` Junio C Hamano
2021-03-02  9:34     ` Ævar Arnfjörð Bjarmason
2021-03-03  6:35       ` Junio C Hamano
2021-02-28 19:54 ` [PATCH 05/10] describe tests: convert setup to use test_commit Ævar Arnfjörð Bjarmason
2021-03-01 21:42   ` Junio C Hamano
2021-02-28 19:54 ` [PATCH 06/10] describe tests: fix nested "test_expect_success" call Ævar Arnfjörð Bjarmason
2021-02-28 19:54 ` [PATCH 07/10] describe tests: support -C in "check_describe" Ævar Arnfjörð Bjarmason
2021-02-28 19:54 ` [PATCH 08/10] svn tests: remove legacy re-setup from init-clone test Ævar Arnfjörð Bjarmason
2021-02-28 19:54 ` [PATCH 09/10] svn tests: refactor away a "set -e" in test body Ævar Arnfjörð Bjarmason
2021-02-28 21:14   ` Eric Wong
2021-02-28 19:54 ` [PATCH 10/10] test-lib: return 1 from test_expect_{success,failure} Ævar Arnfjörð Bjarmason
2021-03-01 21:43   ` Junio C Hamano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200223125102.6697-1-benno@bmevers.de \
    --to=benno@bmevers.de \
    --cc=git@vger.kernel.org \
    --cc=spearce@spearce.org \
    --subject='Re: [PATCH] describe: dont abort too early when searching tags' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Code repositories for project(s) associated with this inbox:

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).