From 74906f88b375a25c7ee711244072257325fc44fa Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ra=C3=BAl=20Cumplido?= <raulcumplido@gmail.com>
Date: Wed, 19 Oct 2022 15:51:41 +0200
Subject: [PATCH 01/48] Fake commit

---
 .github/workflows/comment_bot.yml | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/comment_bot.yml b/.github/workflows/comment_bot.yml
index 4a5461752b3..4cec82619d0 100644
--- a/.github/workflows/comment_bot.yml
+++ b/.github/workflows/comment_bot.yml
@@ -31,15 +31,29 @@ permissions:
 jobs:
   crossbow:
     name: Listen!
-    if: startsWith(github.event.comment.body, '@github-actions crossbow')
+    if: ${{ github.event.issue.pull_request && startsWith(github.event.comment.body, '@github-actions crossbow')}}
     runs-on: ubuntu-latest
     steps:
+      - name: Get PR SHA
+        id: sha
+        uses: actions/github-script@v4
+        with:
+          result-encoding: string
+          script: |
+            const { owner, repo, number } = context.issue;
+            const pr = await github.pulls.get({
+              owner,
+              repo,
+              pull_number: number,
+            });
+            return pr.data.head.sha
       - name: Checkout Arrow
         uses: actions/checkout@v3
         with:
           path: arrow
           # fetch the tags for version number generation
           fetch-depth: 0
+          ref: ${{ steps.sha.outputs.result }}
       - name: Set up Python
         uses: actions/setup-python@v4
         with:

From b0c8cbde00bac5f3007be74d37b220b6d4ca28ec Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 31 Aug 2022 10:53:27 -0400
Subject: [PATCH 02/48] Dask and Pandas repositories now use default branches
 named 'main'.

---
 dev/archery/archery/docker/tests/test_docker.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/dev/archery/archery/docker/tests/test_docker.py b/dev/archery/archery/docker/tests/test_docker.py
index bc25738becf..fa6b59f936d 100644
--- a/dev/archery/archery/docker/tests/test_docker.py
+++ b/dev/archery/archery/docker/tests/test_docker.py
@@ -259,12 +259,12 @@ def test_arrow_example_validation_passes(arrow_compose_path):
 def test_compose_default_params_and_env(arrow_compose_path):
     compose = DockerCompose(arrow_compose_path, params=dict(
         UBUNTU='18.04',
-        DASK='master'
+        DASK='main'
     ))
     assert compose.config.dotenv == arrow_compose_env
     assert compose.config.params == {
         'UBUNTU': '18.04',
-        'DASK': 'master',
+        'DASK': 'main',
     }
 
 
@@ -492,7 +492,7 @@ def test_compose_push(arrow_compose_path):
 def test_compose_error(arrow_compose_path):
     compose = DockerCompose(arrow_compose_path, params=dict(
         PYTHON='3.8',
-        PANDAS='master'
+        PANDAS='main'
     ))
 
     error = subprocess.CalledProcessError(99, [])
@@ -503,7 +503,7 @@ def test_compose_error(arrow_compose_path):
     exception_message = str(exc.value)
     assert "exited with a non-zero exit code 99" in exception_message
     assert "PANDAS: latest" in exception_message
-    assert "export PANDAS=master" in exception_message
+    assert "export PANDAS=main" in exception_message
 
 
 def test_image_with_gpu(arrow_compose_path):

From 09eb082dc0279f8481e86cd04b6e6ebcc318e9b1 Mon Sep 17 00:00:00 2001
From: Kevin Gurney <kgurney@mathworks.com>
Date: Wed, 31 Aug 2022 11:05:02 -0400
Subject: [PATCH 03/48] Replace 'master' with 'default' in comment about Travis
 CI default behavior.

---
 dev/archery/archery/crossbow/core.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index a83c190d121..b64a6f733ae 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -133,7 +133,7 @@ def format_all(items, pattern):
 
 # configurations for setting up branch skipping
 # - appveyor has a feature to skip builds without an appveyor.yml
-# - travis reads from the master branch and applies the rules
+# - travis reads from the default branch and applies the rules
 # - circle requires the configuration to be present on all branch, even ones
 #   that are configured to be skipped
 # - azure skips branches without azure-pipelines.yml by default

From 363c7b358a1dc7cf264ed7103037064576dd6870 Mon Sep 17 00:00:00 2001
From: Kevin Gurney <kgurney@mathworks.com>
Date: Wed, 31 Aug 2022 11:13:16 -0400
Subject: [PATCH 04/48] Remove mention of "master" from help text for
 --arrow-branch crossbow CLI option.

---
 dev/archery/archery/crossbow/cli.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index 456942173e7..55a2338a280 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -96,7 +96,7 @@ def check_config(obj, config_path):
                    'locally. Examples: https://github.com/apache/arrow or '
                    'https://github.com/kszucs/arrow.')
 @click.option('--arrow-branch', '-b', default=None,
-              help='Give the branch name explicitly, e.g. master, ARROW-1949.')
+              help='Give the branch name explicitly, e.g. ARROW-1949.')
 @click.option('--arrow-sha', '-t', default=None,
               help='Set commit SHA or Tag name explicitly, e.g. f67a515, '
                    'apache-arrow-0.11.1.')
@@ -225,7 +225,7 @@ def verify_release_candidate(obj, base_branch, create_pr,
                    'locally. Examples: https://github.com/apache/arrow or '
                    'https://github.com/kszucs/arrow.')
 @click.option('--arrow-branch', '-b', default=None,
-              help='Give the branch name explicitly, e.g. master, ARROW-1949.')
+              help='Give the branch name explicitly, e.g. ARROW-1949.')
 @click.option('--arrow-sha', '-t', default=None,
               help='Set commit SHA or Tag name explicitly, e.g. f67a515, '
                    'apache-arrow-0.11.1.')

From d5fa3d4d7615554a701d172e7fdba5d0ed9c40df Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 31 Aug 2022 11:30:52 -0400
Subject: [PATCH 05/48] Pandas repository uses 'main' as the default branch.

---
 dev/archery/archery/docker/cli.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dev/archery/archery/docker/cli.py b/dev/archery/archery/docker/cli.py
index c7b42c094f6..565efed05c5 100644
--- a/dev/archery/archery/docker/cli.py
+++ b/dev/archery/archery/docker/cli.py
@@ -217,7 +217,7 @@ def docker_run(obj, image, command, *, env, user, force_pull, force_build,
     PYTHON=3.8 archery docker run conda-python
 
     # disable the cache only for the leaf image
-    PANDAS=master archery docker run --no-leaf-cache conda-python-pandas
+    PANDAS=main archery docker run --no-leaf-cache conda-python-pandas
 
     # entirely skip building the image
     archery docker run --no-pull --no-build conda-python

From 4a4f987ca1bcd6336d56398ada5920c36cc8e3ce Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 31 Aug 2022 11:40:28 -0400
Subject: [PATCH 06/48] Add base_branch property to Release object and modify
 commits_to_pick to use the property.

---
 dev/archery/archery/release/core.py | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 2c775c7506b..00981bbae59 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -361,6 +361,10 @@ def commits(self):
         commit_range = f"{lower}..{upper}"
         return list(map(Commit, self.repo.iter_commits(commit_range)))
 
+    @property
+    def base_branch(self):
+        return "master"
+
     def curate(self, minimal=False):
         # handle commits with parquet issue key specially and query them from
         # jira and add it to the issues
@@ -422,9 +426,9 @@ def changelog(self):
         return JiraChangelog(release=self, categories=categories)
 
     def commits_to_pick(self, exclude_already_applied=True):
-        # collect commits applied on the main branch since the root of the
+        # collect commits applied on the default branch since the root of the
         # maintenance branch (the previous major release)
-        commit_range = f"{self.previous.tag}..master"
+        commit_range = f"{self.previous.tag}..{self.base_branch}"
 
         # keeping the original order of the commits helps to minimize the merge
         # conflicts during cherry-picks

From ef4342d129e816d96a8d788674070ce9ed280f99 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 1 Sep 2022 10:56:02 -0400
Subject: [PATCH 07/48] Modify 'archery' command line interface to reference
 the mainline development branch generically.

---
 dev/archery/archery/cli.py | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/dev/archery/archery/cli.py b/dev/archery/archery/cli.py
index f76e326a8e1..677d935da67 100644
--- a/dev/archery/archery/cli.py
+++ b/dev/archery/archery/cli.py
@@ -529,7 +529,7 @@ def benchmark_run(ctx, rev_or_path, src, preserve, output, cmake_extras,
               help="Hide counters field in diff report.")
 @click.argument("contender", metavar="[<contender>",
                 default=ArrowSources.WORKSPACE, required=False)
-@click.argument("baseline", metavar="[<baseline>]]", default="origin/master",
+@click.argument("baseline", metavar="[<baseline>]]", default="origin/HEAD",
                 required=False)
 @click.pass_context
 def benchmark_diff(ctx, src, preserve, output, language, cmake_extras,
@@ -542,7 +542,8 @@ def benchmark_diff(ctx, src, preserve, output, language, cmake_extras,
 
     The caller can optionally specify both the contender and the baseline. If
     unspecified, the contender will default to the current workspace (like git)
-    and the baseline will default to master.
+    and the baseline will default to the mainline development branch (i.e.
+    default git branch).
 
     Each target (contender or baseline) can either be a git revision
     (commit, tag, special values like HEAD) or a cmake build directory. This
@@ -559,12 +560,12 @@ def benchmark_diff(ctx, src, preserve, output, language, cmake_extras,
     Examples:
 
     \b
-    # Compare workspace (contender) with master (baseline)
+    # Compare workspace (contender) against the mainline development branch (baseline)
     \b
     archery benchmark diff
 
     \b
-    # Compare master (contender) with latest version (baseline)
+    # Compare the mainline development branch (contender) against the latest version (baseline)
     \b
     export LAST=$(git tag -l "apache-arrow-[0-9]*" | sort -rV | head -1)
     \b

From 0fcd9026eb8a8621ec5b0d949833c9d1bac7fdd3 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 8 Sep 2022 16:25:26 -0400
Subject: [PATCH 08/48] Dynamically compute the default branch name for archery
 and crossbow core/cli files

---
 dev/archery/archery/crossbow/cli.py  |  7 ++++++-
 dev/archery/archery/crossbow/core.py | 19 ++++++++++++++++---
 dev/archery/archery/release/core.py  | 27 +++++++++++++++++++++++++--
 3 files changed, 47 insertions(+), 6 deletions(-)

diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index 55a2338a280..eb12fad25ef 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -16,8 +16,10 @@
 # under the License.
 
 from pathlib import Path
+from pickle import TRUE
 import time
 import sys
+import pygit2
 
 import click
 
@@ -155,9 +157,12 @@ def submit(obj, tasks, groups, params, job_prefix, config_path, arrow_version,
         queue.push()
         click.echo('Pushed job identifier is: `{}`'.format(job.branch))
 
+# Get the default branch name from the repository
+arrow_source_dir = ArrowSources.find()
+repo = Repo(arrow_source_dir.path)
 
 @crossbow.command()
-@click.option('--base-branch', default="master",
+@click.option('--base-branch', default=repo.default_branch_name,
               help='Set base branch for the PR.')
 @click.option('--create-pr', is_flag=True, default=False,
               help='Create GitHub Pull Request')
diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index b64a6f733ae..ce03ac8795b 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -361,6 +361,13 @@ def signature(self):
         return pygit2.Signature(self.user_name, self.user_email,
                                 int(time.time()))
 
+    @property
+    def default_branch_name(self):
+        ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
+        target_name = ref_obj.target
+        target_name_tokenized = target_name.split("/")
+        return target_name_tokenized[-1]
+
     def create_tree(self, files):
         builder = self.repo.TreeBuilder()
 
@@ -382,7 +389,7 @@ def create_commit(self, files, parents=None, message='',
         if parents is None:
             # by default use the main branch as the base of the new branch
             # required to reuse github actions cache across crossbow tasks
-            commit, _ = self.repo.resolve_refish("master")
+            commit, _ = self.repo.resolve_refish(self.default_branch_name)
             parents = [commit.id]
         tree_id = self.create_tree(files)
 
@@ -546,8 +553,10 @@ def github_overwrite_release_assets(self, tag_name, target_commitish,
                         'Unsupported upload method {}'.format(method)
                     )
 
-    def github_pr(self, title, head=None, base="master", body=None,
+    def github_pr(self, title, head=None, base=None, body=None,
                   github_token=None, create=False):
+        # Default value for base is the default_branch name()
+        base = self.default_branch_name() if base is None else base
         github_token = github_token or self.github_token
         repo = self.as_github_repo(github_token=github_token)
         if create:
@@ -1289,11 +1298,15 @@ def validate(self):
                     'is: `{}`'.format(task_name, str(e))
                 )
 
+        # Get the default branch name from the repository
+        arrow_source_dir = ArrowSources.find()
+        repo = Repo(arrow_source_dir.path)
+
         # validate that the defined tasks are renderable, in order to to that
         # define the required object with dummy data
         target = Target(
             head='e279a7e06e61c14868ca7d71dea795420aea6539',
-            branch='master',
+            branch=repo.default_branch_name,
             remote='https://github.com/apache/arrow',
             version='1.0.0dev123',
             r_version='0.13.0.100000123',
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 00981bbae59..c15ec8ffbc3 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -226,6 +226,27 @@ def url(self):
     def title(self):
         return self._title
 
+class DefaultBranchName(object):
+    def __new__(self):
+        if not hasattr(self, 'instance'):
+            self.instance = super(DefaultBranchName, self).__new__(self)
+            arrow = ArrowSources.find()
+            arrow = ArrowSources.find()
+            repo = Repo(arrow.path)
+            remotes = repo.remotes
+            origin = repo.remotes["origin"]
+            origin_refs = origin.refs
+            origin_head = origin_refs["HEAD"] # git.RemoteReference object to origin/HEAD
+            origin_head_reference = origin_head.reference # git.RemoteReference object to origin/main
+            origin_head_name = origin_head_reference.name # Should return "origin/main" or "origin/master"
+            origin_head_name_tokenized = origin_head_name.split("/")
+            self.default_branch_name = origin_head_name_tokenized[-1]
+        return self.instance
+
+    @property
+    def value(self):
+        return self.default_branch_name
+
 
 class Release:
 
@@ -363,7 +384,8 @@ def commits(self):
 
     @property
     def base_branch(self):
-        return "master"
+        default_branch_name = DefaultBranchName()
+        return default_branch_name.value()
 
     def curate(self, minimal=False):
         # handle commits with parquet issue key specially and query them from
@@ -480,7 +502,8 @@ def branch(self):
 
     @property
     def base_branch(self):
-        return "master"
+        default_branch_name = DefaultBranchName()
+        return default_branch_name.value()
 
     @cached_property
     def siblings(self):

From 7cc70a74e2dd6ede659f30f89dcac9c294227ecd Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 8 Sep 2022 16:30:46 -0400
Subject: [PATCH 09/48] Performed python linting

---
 dev/archery/archery/cli.py          |  6 ++++--
 dev/archery/archery/crossbow/cli.py |  4 ++--
 dev/archery/archery/release/core.py | 11 +++++++----
 3 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/dev/archery/archery/cli.py b/dev/archery/archery/cli.py
index 677d935da67..3f608872741 100644
--- a/dev/archery/archery/cli.py
+++ b/dev/archery/archery/cli.py
@@ -560,12 +560,14 @@ def benchmark_diff(ctx, src, preserve, output, language, cmake_extras,
     Examples:
 
     \b
-    # Compare workspace (contender) against the mainline development branch (baseline)
+    # Compare workspace (contender) against the mainline development branch
+    # (baseline)
     \b
     archery benchmark diff
 
     \b
-    # Compare the mainline development branch (contender) against the latest version (baseline)
+    # Compare the mainline development branch (contender) against the latest
+    # version (baseline)
     \b
     export LAST=$(git tag -l "apache-arrow-[0-9]*" | sort -rV | head -1)
     \b
diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index eb12fad25ef..b7390ef1417 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -16,10 +16,8 @@
 # under the License.
 
 from pathlib import Path
-from pickle import TRUE
 import time
 import sys
-import pygit2
 
 import click
 
@@ -157,10 +155,12 @@ def submit(obj, tasks, groups, params, job_prefix, config_path, arrow_version,
         queue.push()
         click.echo('Pushed job identifier is: `{}`'.format(job.branch))
 
+
 # Get the default branch name from the repository
 arrow_source_dir = ArrowSources.find()
 repo = Repo(arrow_source_dir.path)
 
+
 @crossbow.command()
 @click.option('--base-branch', default=repo.default_branch_name,
               help='Set base branch for the PR.')
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index c15ec8ffbc3..b95a90b6d9c 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -226,6 +226,7 @@ def url(self):
     def title(self):
         return self._title
 
+
 class DefaultBranchName(object):
     def __new__(self):
         if not hasattr(self, 'instance'):
@@ -233,12 +234,14 @@ def __new__(self):
             arrow = ArrowSources.find()
             arrow = ArrowSources.find()
             repo = Repo(arrow.path)
-            remotes = repo.remotes
             origin = repo.remotes["origin"]
             origin_refs = origin.refs
-            origin_head = origin_refs["HEAD"] # git.RemoteReference object to origin/HEAD
-            origin_head_reference = origin_head.reference # git.RemoteReference object to origin/main
-            origin_head_name = origin_head_reference.name # Should return "origin/main" or "origin/master"
+            # git.RemoteReference object to origin/HEAD
+            origin_head = origin_refs["HEAD"]
+            # git.RemoteReference object to origin/main
+            origin_head_reference = origin_head.reference
+            # Should return "origin/main" or "origin/master"
+            origin_head_name = origin_head_reference.name
             origin_head_name_tokenized = origin_head_name.split("/")
             self.default_branch_name = origin_head_name_tokenized[-1]
         return self.instance

From 20d6c223aea6bc8fc349ab2720c8ebe5a74ee96e Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 14:55:29 -0400
Subject: [PATCH 10/48] remove duplicate code

---
 dev/archery/archery/release/core.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index b95a90b6d9c..997ed76417b 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -232,7 +232,6 @@ def __new__(self):
         if not hasattr(self, 'instance'):
             self.instance = super(DefaultBranchName, self).__new__(self)
             arrow = ArrowSources.find()
-            arrow = ArrowSources.find()
             repo = Repo(arrow.path)
             origin = repo.remotes["origin"]
             origin_refs = origin.refs

From f850d52160911a40c14563a1a6aa8db809581946 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 14:56:45 -0400
Subject: [PATCH 11/48] Print debugging info for default_branch_name Repo class
 function

---
 dev/archery/archery/crossbow/core.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index ce03ac8795b..175f59ea016 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -363,6 +363,10 @@ def signature(self):
 
     @property
     def default_branch_name(self):
+        for remote in self.repo.remotes:
+            print(remote.name)
+        for ref in self.repo.references.objects:
+            print(ref.target)
         ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
         target_name = ref_obj.target
         target_name_tokenized = target_name.split("/")

From 5ff60d757b7ceb963079eb7b8e96460d9ad55686 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 15:04:28 -0400
Subject: [PATCH 12/48] Print more debugging info

---
 dev/archery/archery/crossbow/core.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 175f59ea016..b144f16580c 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -365,8 +365,9 @@ def signature(self):
     def default_branch_name(self):
         for remote in self.repo.remotes:
             print(remote.name)
-        for ref in self.repo.references.objects:
-            print(ref.target)
+        # for ref in self.repo.references.objects:
+        #     print(ref.target)
+        print(self.repo.resolve_refish('origin/HEAD'))
         ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
         target_name = ref_obj.target
         target_name_tokenized = target_name.split("/")

From ec4bdf71ccb3eaf13bb3fb4969787b17328b5aa8 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 15:08:49 -0400
Subject: [PATCH 13/48] Remove resolve_refish print command

---
 dev/archery/archery/crossbow/core.py | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index b144f16580c..0e77bccc85d 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -363,11 +363,15 @@ def signature(self):
 
     @property
     def default_branch_name(self):
+        print("Remotes:")
         for remote in self.repo.remotes:
             print(remote.name)
-        # for ref in self.repo.references.objects:
-        #     print(ref.target)
-        print(self.repo.resolve_refish('origin/HEAD'))
+        for ref in self.repo.references.objects:
+            print("ref.target" + ref.target)
+            print("ref.raw_target" + ref.raw_target)
+            print("ref.shorthand" + ref.shorthand)
+            print("ref.raw_shorthand" + ref.raw_shorthand)
+            print("ref.name" + ref.name)
         ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
         target_name = ref_obj.target
         target_name_tokenized = target_name.split("/")

From 8dd439e8ab2906e8a617398aef231d2c4ff7ac51 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 15:11:31 -0400
Subject: [PATCH 14/48] Add new line between reference object details

---
 dev/archery/archery/crossbow/core.py | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 0e77bccc85d..86851ff7080 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -367,11 +367,12 @@ def default_branch_name(self):
         for remote in self.repo.remotes:
             print(remote.name)
         for ref in self.repo.references.objects:
-            print("ref.target" + ref.target)
-            print("ref.raw_target" + ref.raw_target)
-            print("ref.shorthand" + ref.shorthand)
-            print("ref.raw_shorthand" + ref.raw_shorthand)
-            print("ref.name" + ref.name)
+            print(ref.target)
+            print(ref.raw_target)
+            print(ref.shorthand)
+            print(ref.raw_shorthand)
+            print(ref.name)
+            print("\n")
         ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
         target_name = ref_obj.target
         target_name_tokenized = target_name.split("/")

From 7732c593c3ed2bb82263080795ee47ace2b5687d Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 13 Sep 2022 15:30:21 -0400
Subject: [PATCH 15/48] Print branches

---
 dev/archery/archery/crossbow/core.py | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 86851ff7080..ce110ae195a 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -366,13 +366,15 @@ def default_branch_name(self):
         print("Remotes:")
         for remote in self.repo.remotes:
             print(remote.name)
-        for ref in self.repo.references.objects:
-            print(ref.target)
-            print(ref.raw_target)
-            print(ref.shorthand)
-            print(ref.raw_shorthand)
-            print(ref.name)
-            print("\n")
+        # for ref in self.repo.references.objects:
+        #     print(ref.target)
+        #     print(ref.raw_target)
+        #     print(ref.shorthand)
+        #     print(ref.raw_shorthand)
+        #     print(ref.name)
+        #     print("\n")
+        for branch in self.repo.branches:
+            print(branch)
         ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
         target_name = ref_obj.target
         target_name_tokenized = target_name.split("/")

From 3081853f269c3d08be3a9d5890563cd85a55bbf9 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 10:09:00 -0400
Subject: [PATCH 16/48] Use environment variable, DEFAULT_BRANCH, that is set
 in the yml file.

---
 dev/archery/archery/crossbow/core.py | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index ce110ae195a..a1cedd2379d 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -363,9 +363,11 @@ def signature(self):
 
     @property
     def default_branch_name(self):
-        print("Remotes:")
-        for remote in self.repo.remotes:
-            print(remote.name)
+        default_branch_name = os.getenv("DEFAULT_BRANCH")
+        print("**********default_branch_name" + default_branch_name)
+        # print("Remotes:")
+        # for remote in self.repo.remotes:
+        #     print(remote.name)
         # for ref in self.repo.references.objects:
         #     print(ref.target)
         #     print(ref.raw_target)
@@ -373,12 +375,13 @@ def default_branch_name(self):
         #     print(ref.raw_shorthand)
         #     print(ref.name)
         #     print("\n")
-        for branch in self.repo.branches:
-            print(branch)
-        ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
-        target_name = ref_obj.target
-        target_name_tokenized = target_name.split("/")
-        return target_name_tokenized[-1]
+        # for branch in self.repo.branches:
+        #     print(branch)
+        # ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
+        # target_name = ref_obj.target
+        # target_name_tokenized = target_name.split("/")
+        # return target_name_tokenized[-1]
+        return default_branch_name
 
     def create_tree(self, files):
         builder = self.repo.TreeBuilder()

From fd6299689afdd25d88bee5bc60966122e8aa2b55 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 10:12:51 -0400
Subject: [PATCH 17/48] Remove string concatenation, types incompatible

---
 dev/archery/archery/crossbow/core.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index a1cedd2379d..1ad0bb8633c 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -364,7 +364,7 @@ def signature(self):
     @property
     def default_branch_name(self):
         default_branch_name = os.getenv("DEFAULT_BRANCH")
-        print("**********default_branch_name" + default_branch_name)
+        print(default_branch_name)
         # print("Remotes:")
         # for remote in self.repo.remotes:
         #     print(remote.name)

From eb36ffbe4e90fa44398398bba1be344ff00f616d Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:15:22 -0400
Subject: [PATCH 18/48] Enable both CI workflows and local repository workflows
 for getting default branch in archery/crossbow/core.py

---
 dev/archery/archery/crossbow/core.py | 27 ++++++++++-----------------
 1 file changed, 10 insertions(+), 17 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 1ad0bb8633c..688af5d55ff 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -364,23 +364,16 @@ def signature(self):
     @property
     def default_branch_name(self):
         default_branch_name = os.getenv("DEFAULT_BRANCH")
-        print(default_branch_name)
-        # print("Remotes:")
-        # for remote in self.repo.remotes:
-        #     print(remote.name)
-        # for ref in self.repo.references.objects:
-        #     print(ref.target)
-        #     print(ref.raw_target)
-        #     print(ref.shorthand)
-        #     print(ref.raw_shorthand)
-        #     print(ref.name)
-        #     print("\n")
-        # for branch in self.repo.branches:
-        #     print(branch)
-        # ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
-        # target_name = ref_obj.target
-        # target_name_tokenized = target_name.split("/")
-        # return target_name_tokenized[-1]
+
+        if default_branch_name == None:
+            try:
+                ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
+                target_name = ref_obj.target
+                target_name_tokenized = target_name.split("/")
+                default_branch_name = target_name_tokenized[-1]
+            except:
+                raise RuntimeError('DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+
         return default_branch_name
 
     def create_tree(self, files):

From 0e8c8ea95dd649793add1e3beef2d5b92475b726 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:26:43 -0400
Subject: [PATCH 19/48] Enable both CI workflows and local repository workflows
 for getting default branch in archery/release/core.py

---
 dev/archery/archery/release/core.py | 42 ++++++++++++++++++++---------
 1 file changed, 30 insertions(+), 12 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 997ed76417b..15a952d22f7 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -22,6 +22,7 @@
 import pathlib
 import shelve
 import warnings
+import os
 
 from git import Repo
 from jira import JIRA
@@ -231,18 +232,35 @@ class DefaultBranchName(object):
     def __new__(self):
         if not hasattr(self, 'instance'):
             self.instance = super(DefaultBranchName, self).__new__(self)
-            arrow = ArrowSources.find()
-            repo = Repo(arrow.path)
-            origin = repo.remotes["origin"]
-            origin_refs = origin.refs
-            # git.RemoteReference object to origin/HEAD
-            origin_head = origin_refs["HEAD"]
-            # git.RemoteReference object to origin/main
-            origin_head_reference = origin_head.reference
-            # Should return "origin/main" or "origin/master"
-            origin_head_name = origin_head_reference.name
-            origin_head_name_tokenized = origin_head_name.split("/")
-            self.default_branch_name = origin_head_name_tokenized[-1]
+
+            default_branch_name = os.getenv("DEFAULT_BRANCH")
+
+            if default_branch_name == None:
+                try:
+                    # Set up repo object 
+                    arrow = ArrowSources.find()
+                    repo = Repo(arrow.path)
+                    origin = repo.remotes["origin"]
+                    origin_refs = origin.refs
+
+                    # git.RemoteReference object to origin/HEAD
+                    origin_head = origin_refs["HEAD"]
+
+                    # git.RemoteReference object to origin/main or origin/master
+                    origin_head_reference = origin_head.reference
+
+                    # Should return "origin/main" or "origin/master"
+                    origin_head_name = origin_head_reference.name
+                    origin_head_name_tokenized = origin_head_name.split("/")
+
+                    # The last token is the default branch name
+                    default_branch_name = origin_head_name_tokenized[-1]
+                except:
+                    raise RuntimeError('DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+
+            # Set default branch as class property
+            self.default_branch_name = default_branch_name
+            
         return self.instance
 
     @property

From c0d5ea22bb08e1d5645161b77c0607da841e5ef9 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:39:40 -0400
Subject: [PATCH 20/48] Add DEFAULT_BRANCH environment variable to archery.yml
 test step for unittests

---
 .github/workflows/archery.yml | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 36146d46aa4..696c4242a55 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -65,7 +65,9 @@ jobs:
         run: pip install pytest responses -e dev/archery[all]
       - name: Archery Unittests
         working-directory: dev/archery
-        run: pytest -v archery
+        run: |
+          DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
+          pytest -v archery
       - name: Archery Docker Validation
         run: archery docker check-config
       - name: Crossbow Check Config

From 07226f9afb85652c331c2f8304076c3ad58bd56d Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:47:02 -0400
Subject: [PATCH 21/48] Set workflow-wide environment variable, DEFAULT_BRANCH,
 for archery.yml

---
 .github/workflows/archery.yml | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 696c4242a55..7e32a262281 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -31,6 +31,9 @@ on:
       - 'dev/tasks/**'
       - 'docker-compose.yml'
 
+env:
+  DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+
 concurrency:
   group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
   cancel-in-progress: true
@@ -65,9 +68,7 @@ jobs:
         run: pip install pytest responses -e dev/archery[all]
       - name: Archery Unittests
         working-directory: dev/archery
-        run: |
-          DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
-          pytest -v archery
+        run: pytest -v archery
       - name: Archery Docker Validation
         run: archery docker check-config
       - name: Crossbow Check Config

From 99938f70139fbe0c76f8ae843396719ec9fe9466 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:49:35 -0400
Subject: [PATCH 22/48] Print reason for skipping tests

---
 .github/workflows/archery.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 7e32a262281..5d300d5ccee 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -68,7 +68,7 @@ jobs:
         run: pip install pytest responses -e dev/archery[all]
       - name: Archery Unittests
         working-directory: dev/archery
-        run: pytest -v archery
+        run: pytest -rs -v archery
       - name: Archery Docker Validation
         run: archery docker check-config
       - name: Crossbow Check Config

From 63ecf94affc628fd9d23c0f42cb906d8a744ab3e Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 11:53:53 -0400
Subject: [PATCH 23/48] Add 'enable-integration' flag to ensure crossbowcli
 tests run

---
 .github/workflows/archery.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 5d300d5ccee..2dad44cd5bc 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -68,7 +68,7 @@ jobs:
         run: pip install pytest responses -e dev/archery[all]
       - name: Archery Unittests
         working-directory: dev/archery
-        run: pytest -rs -v archery
+        run: pytest -rs --enable-integration -v archery
       - name: Archery Docker Validation
         run: archery docker check-config
       - name: Crossbow Check Config

From 1f0138e290b9daa671cc83464a2912d62312b117 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 14 Sep 2022 12:09:20 -0400
Subject: [PATCH 24/48] Factor out GitFixup step DEFAULT_BRANCH value

---
 .github/workflows/archery.yml | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 2dad44cd5bc..5e4d1764a6a 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -55,9 +55,7 @@ jobs:
           fetch-depth: 0
       - name: Git Fixup
         shell: bash
-        run: |
-          DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
-          git branch $DEFAULT_BRANCH origin/$DEFAULT_BRANCH || true
+        run: git branch $DEFAULT_BRANCH origin/$DEFAULT_BRANCH || true
       - name: Setup Python
         uses: actions/setup-python@v4
         with:
@@ -68,7 +66,7 @@ jobs:
         run: pip install pytest responses -e dev/archery[all]
       - name: Archery Unittests
         working-directory: dev/archery
-        run: pytest -rs --enable-integration -v archery
+        run: pytest -v archery
       - name: Archery Docker Validation
         run: archery docker check-config
       - name: Crossbow Check Config

From 2e0740170d0cfea5576a6236e7cdee4bc75faabd Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 15 Sep 2022 10:04:30 -0400
Subject: [PATCH 25/48] Run python linting

---
 dev/archery/archery/cli.py           | 2 +-
 dev/archery/archery/crossbow/core.py | 3 ++-
 dev/archery/archery/release/core.py  | 7 ++++---
 3 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/dev/archery/archery/cli.py b/dev/archery/archery/cli.py
index 3f608872741..105a64c0603 100644
--- a/dev/archery/archery/cli.py
+++ b/dev/archery/archery/cli.py
@@ -571,7 +571,7 @@ def benchmark_diff(ctx, src, preserve, output, language, cmake_extras,
     \b
     export LAST=$(git tag -l "apache-arrow-[0-9]*" | sort -rV | head -1)
     \b
-    archery benchmark diff master "$LAST"
+    archery benchmark diff <default-branch> "$LAST"
 
     \b
     # Compare g++7 (contender) with clang++-8 (baseline) builds
diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 688af5d55ff..603dfdbd4f0 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -372,7 +372,8 @@ def default_branch_name(self):
                 target_name_tokenized = target_name.split("/")
                 default_branch_name = target_name_tokenized[-1]
             except:
-                raise RuntimeError('DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+                raise RuntimeError(
+                    'DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
 
         return default_branch_name
 
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 15a952d22f7..131be01a5b1 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -237,7 +237,7 @@ def __new__(self):
 
             if default_branch_name == None:
                 try:
-                    # Set up repo object 
+                    # Set up repo object
                     arrow = ArrowSources.find()
                     repo = Repo(arrow.path)
                     origin = repo.remotes["origin"]
@@ -256,11 +256,12 @@ def __new__(self):
                     # The last token is the default branch name
                     default_branch_name = origin_head_name_tokenized[-1]
                 except:
-                    raise RuntimeError('DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+                    raise RuntimeError(
+                        'DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
 
             # Set default branch as class property
             self.default_branch_name = default_branch_name
-            
+
         return self.instance
 
     @property

From 96e6535c9257f5778664d33e541d6f8082b87de9 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Mon, 19 Sep 2022 14:26:30 -0400
Subject: [PATCH 26/48] Address bare except and line lengths

---
 dev/archery/archery/crossbow/core.py |  8 +++++---
 dev/archery/archery/release/core.py  | 16 ++++++++++------
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 603dfdbd4f0..d7d54b146a4 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -365,15 +365,17 @@ def signature(self):
     def default_branch_name(self):
         default_branch_name = os.getenv("DEFAULT_BRANCH")
 
-        if default_branch_name == None:
+        if default_branch_name is None:
             try:
                 ref_obj = self.repo.references["refs/remotes/origin/HEAD"]
                 target_name = ref_obj.target
                 target_name_tokenized = target_name.split("/")
                 default_branch_name = target_name_tokenized[-1]
-            except:
+            except KeyError:
                 raise RuntimeError(
-                    'DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+                    'DEFAULT_BRANCH environment variable is not set. Git '
+                    'repository does not contain a '
+                    '\'refs/remotes/origin/HEAD\' reference.')
 
         return default_branch_name
 
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 131be01a5b1..83fe659baa0 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -235,7 +235,7 @@ def __new__(self):
 
             default_branch_name = os.getenv("DEFAULT_BRANCH")
 
-            if default_branch_name == None:
+            if default_branch_name is None:
                 try:
                     # Set up repo object
                     arrow = ArrowSources.find()
@@ -243,21 +243,25 @@ def __new__(self):
                     origin = repo.remotes["origin"]
                     origin_refs = origin.refs
 
-                    # git.RemoteReference object to origin/HEAD
+                    # Get git.RemoteReference object to origin/HEAD
                     origin_head = origin_refs["HEAD"]
 
-                    # git.RemoteReference object to origin/main or origin/master
+                    # Get git.RemoteReference object to origin/main or
+                    # origin/master
                     origin_head_reference = origin_head.reference
 
-                    # Should return "origin/main" or "origin/master"
+                    # Get string value of remote head reference, should return
+                    # "origin/main" or "origin/master"
                     origin_head_name = origin_head_reference.name
                     origin_head_name_tokenized = origin_head_name.split("/")
 
                     # The last token is the default branch name
                     default_branch_name = origin_head_name_tokenized[-1]
-                except:
+                except KeyError:
                     raise RuntimeError(
-                        'DEFAULT_BRANCH environment variable is not set. Git repository does not contain \'refs/remotes/origin/HEAD\' reference.')
+                        'DEFAULT_BRANCH environment variable is not set. Git '
+                        'repository does not contain a'
+                        '\'refs/remotes/origin/HEAD\' reference.')
 
             # Set default branch as class property
             self.default_branch_name = default_branch_name

From c66f3f10a988921dd20abf61c80ad4b3170a5391 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Mon, 19 Sep 2022 17:09:56 -0400
Subject: [PATCH 27/48] Add context to error message when obtaining default
 branch name.

---
 dev/archery/archery/crossbow/core.py | 6 +++---
 dev/archery/archery/release/core.py  | 6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index d7d54b146a4..02228cc66ee 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -373,9 +373,9 @@ def default_branch_name(self):
                 default_branch_name = target_name_tokenized[-1]
             except KeyError:
                 raise RuntimeError(
-                    'DEFAULT_BRANCH environment variable is not set. Git '
-                    'repository does not contain a '
-                    '\'refs/remotes/origin/HEAD\' reference.')
+                    'Unable to determine default branch name: DEFAULT_BRANCH '
+                    'environment variable is not set. Git repository does not '
+                    'contain a \'refs/remotes/origin/HEAD\' reference.')
 
         return default_branch_name
 
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 83fe659baa0..e242970bfcf 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -259,9 +259,9 @@ def __new__(self):
                     default_branch_name = origin_head_name_tokenized[-1]
                 except KeyError:
                     raise RuntimeError(
-                        'DEFAULT_BRANCH environment variable is not set. Git '
-                        'repository does not contain a'
-                        '\'refs/remotes/origin/HEAD\' reference.')
+                    'Unable to determine default branch name: DEFAULT_BRANCH '
+                    'environment variable is not set. Git repository does not '
+                    'contain a \'refs/remotes/origin/HEAD\' reference.')
 
             # Set default branch as class property
             self.default_branch_name = default_branch_name

From 09d898f62e3e10f992208172c2dd3f6a9b9a12af Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 20 Sep 2022 14:12:12 -0400
Subject: [PATCH 28/48] add debugging print statement in
 archery/archery/release/core.py commits_to_pick()

---
 dev/archery/archery/release/core.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index e242970bfcf..1ded212822b 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -475,6 +475,7 @@ def changelog(self):
     def commits_to_pick(self, exclude_already_applied=True):
         # collect commits applied on the default branch since the root of the
         # maintenance branch (the previous major release)
+        print(self.base_branch)
         commit_range = f"{self.previous.tag}..{self.base_branch}"
 
         # keeping the original order of the commits helps to minimize the merge

From 873cd8c312c8386570d0fc68652d3fd443b54abe Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 20 Sep 2022 14:37:28 -0400
Subject: [PATCH 29/48] Debugging statements for DefaultBranchName constructor

---
 dev/archery/archery/release/core.py | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 1ded212822b..e4578ee6679 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -236,12 +236,21 @@ def __new__(self):
             default_branch_name = os.getenv("DEFAULT_BRANCH")
 
             if default_branch_name is None:
+                print("default_branch_name could not be determined by the environment variable")
                 try:
                     # Set up repo object
                     arrow = ArrowSources.find()
                     repo = Repo(arrow.path)
                     origin = repo.remotes["origin"]
                     origin_refs = origin.refs
+                    
+                    print("repo.remotes:")
+                    for remote in repo.remotes:
+                        print(remote)
+
+                    print("origin.refs:")
+                    for ref in origin.refs:
+                        print(ref)
 
                     # Get git.RemoteReference object to origin/HEAD
                     origin_head = origin_refs["HEAD"]
@@ -250,6 +259,8 @@ def __new__(self):
                     # origin/master
                     origin_head_reference = origin_head.reference
 
+                    print(origin_head_reference)
+
                     # Get string value of remote head reference, should return
                     # "origin/main" or "origin/master"
                     origin_head_name = origin_head_reference.name

From 9b94745108388bde051ed8ceaa0a73887e06da76 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 20 Sep 2022 15:50:48 -0400
Subject: [PATCH 30/48] Remove base_branch property of Release, instead add
 default_branch_property

---
 dev/archery/archery/release/core.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index e4578ee6679..872fa91399e 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -243,7 +243,7 @@ def __new__(self):
                     repo = Repo(arrow.path)
                     origin = repo.remotes["origin"]
                     origin_refs = origin.refs
-                    
+
                     print("repo.remotes:")
                     for remote in repo.remotes:
                         print(remote)
@@ -419,7 +419,7 @@ def commits(self):
         return list(map(Commit, self.repo.iter_commits(commit_range)))
 
     @property
-    def base_branch(self):
+    def default_branch(self):
         default_branch_name = DefaultBranchName()
         return default_branch_name.value()
 
@@ -486,8 +486,8 @@ def changelog(self):
     def commits_to_pick(self, exclude_already_applied=True):
         # collect commits applied on the default branch since the root of the
         # maintenance branch (the previous major release)
-        print(self.base_branch)
-        commit_range = f"{self.previous.tag}..{self.base_branch}"
+        print(self.default_branch)
+        commit_range = f"{self.previous.tag}..{self.default_branch}"
 
         # keeping the original order of the commits helps to minimize the merge
         # conflicts during cherry-picks

From a6edc7f1d579ee2ca8e28e4b9f50b38d604b2e18 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 20 Sep 2022 17:09:32 -0400
Subject: [PATCH 31/48] Use separate function for computing default branch.

---
 dev/archery/archery/release/core.py | 16 ++--------------
 1 file changed, 2 insertions(+), 14 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 872fa91399e..8896a04161a 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -236,7 +236,6 @@ def __new__(self):
             default_branch_name = os.getenv("DEFAULT_BRANCH")
 
             if default_branch_name is None:
-                print("default_branch_name could not be determined by the environment variable")
                 try:
                     # Set up repo object
                     arrow = ArrowSources.find()
@@ -244,14 +243,6 @@ def __new__(self):
                     origin = repo.remotes["origin"]
                     origin_refs = origin.refs
 
-                    print("repo.remotes:")
-                    for remote in repo.remotes:
-                        print(remote)
-
-                    print("origin.refs:")
-                    for ref in origin.refs:
-                        print(ref)
-
                     # Get git.RemoteReference object to origin/HEAD
                     origin_head = origin_refs["HEAD"]
 
@@ -259,8 +250,6 @@ def __new__(self):
                     # origin/master
                     origin_head_reference = origin_head.reference
 
-                    print(origin_head_reference)
-
                     # Get string value of remote head reference, should return
                     # "origin/main" or "origin/master"
                     origin_head_name = origin_head_reference.name
@@ -420,8 +409,8 @@ def commits(self):
 
     @property
     def default_branch(self):
-        default_branch_name = DefaultBranchName()
-        return default_branch_name.value()
+        dbn = DefaultBranchName()
+        return dbn.value
 
     def curate(self, minimal=False):
         # handle commits with parquet issue key specially and query them from
@@ -486,7 +475,6 @@ def changelog(self):
     def commits_to_pick(self, exclude_already_applied=True):
         # collect commits applied on the default branch since the root of the
         # maintenance branch (the previous major release)
-        print(self.default_branch)
         commit_range = f"{self.previous.tag}..{self.default_branch}"
 
         # keeping the original order of the commits helps to minimize the merge

From ba7fc6f9c2e40e821173d8f501dcd1bc13b3e655 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Tue, 20 Sep 2022 17:29:17 -0400
Subject: [PATCH 32/48] Refactor the default branch code to be calculated
 within Release class

---
 dev/archery/archery/release/core.py | 83 ++++++++++++-----------------
 1 file changed, 33 insertions(+), 50 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 8896a04161a..20cb47a2eb4 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -228,51 +228,6 @@ def title(self):
         return self._title
 
 
-class DefaultBranchName(object):
-    def __new__(self):
-        if not hasattr(self, 'instance'):
-            self.instance = super(DefaultBranchName, self).__new__(self)
-
-            default_branch_name = os.getenv("DEFAULT_BRANCH")
-
-            if default_branch_name is None:
-                try:
-                    # Set up repo object
-                    arrow = ArrowSources.find()
-                    repo = Repo(arrow.path)
-                    origin = repo.remotes["origin"]
-                    origin_refs = origin.refs
-
-                    # Get git.RemoteReference object to origin/HEAD
-                    origin_head = origin_refs["HEAD"]
-
-                    # Get git.RemoteReference object to origin/main or
-                    # origin/master
-                    origin_head_reference = origin_head.reference
-
-                    # Get string value of remote head reference, should return
-                    # "origin/main" or "origin/master"
-                    origin_head_name = origin_head_reference.name
-                    origin_head_name_tokenized = origin_head_name.split("/")
-
-                    # The last token is the default branch name
-                    default_branch_name = origin_head_name_tokenized[-1]
-                except KeyError:
-                    raise RuntimeError(
-                    'Unable to determine default branch name: DEFAULT_BRANCH '
-                    'environment variable is not set. Git repository does not '
-                    'contain a \'refs/remotes/origin/HEAD\' reference.')
-
-            # Set default branch as class property
-            self.default_branch_name = default_branch_name
-
-        return self.instance
-
-    @property
-    def value(self):
-        return self.default_branch_name
-
-
 class Release:
 
     def __new__(self, version, jira=None, repo=None):
@@ -407,10 +362,39 @@ def commits(self):
         commit_range = f"{lower}..{upper}"
         return list(map(Commit, self.repo.iter_commits(commit_range)))
 
-    @property
+    @cached_property
     def default_branch(self):
-        dbn = DefaultBranchName()
-        return dbn.value
+        default_branch_name = os.getenv("DEFAULT_BRANCH")
+
+        if default_branch_name is None:
+            try:
+                # Set up repo object
+                arrow = ArrowSources.find()
+                repo = Repo(arrow.path)
+                origin = repo.remotes["origin"]
+                origin_refs = origin.refs
+
+                # Get git.RemoteReference object to origin/HEAD
+                origin_head = origin_refs["HEAD"]
+
+                # Get git.RemoteReference object to origin/main or
+                # origin/master
+                origin_head_reference = origin_head.reference
+
+                # Get string value of remote head reference, should return
+                # "origin/main" or "origin/master"
+                origin_head_name = origin_head_reference.name
+                origin_head_name_tokenized = origin_head_name.split("/")
+
+                # The last token is the default branch name
+                default_branch_name = origin_head_name_tokenized[-1]
+            except KeyError:
+                raise RuntimeError(
+                    'Unable to determine default branch name: DEFAULT_BRANCH '
+                    'environment variable is not set. Git repository does not '
+                    'contain a \'refs/remotes/origin/HEAD\' reference.')
+
+        return default_branch_name
 
     def curate(self, minimal=False):
         # handle commits with parquet issue key specially and query them from
@@ -527,8 +511,7 @@ def branch(self):
 
     @property
     def base_branch(self):
-        default_branch_name = DefaultBranchName()
-        return default_branch_name.value()
+        return self.default_branch
 
     @cached_property
     def siblings(self):

From c0979987e7f340a1101a69fe911fa3660dab8281 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 21 Sep 2022 16:28:00 -0400
Subject: [PATCH 33/48] Add DEFAULT_BRANCH environment variable to Execute
 Docker Build step in integration.yml

---
 .github/workflows/integration.yml | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/.github/workflows/integration.yml b/.github/workflows/integration.yml
index 2c5ffb084c0..4380dec782d 100644
--- a/.github/workflows/integration.yml
+++ b/.github/workflows/integration.yml
@@ -50,6 +50,11 @@ permissions:
 
 env:
   DOCKER_VOLUME_PREFIX: ".docker/"
+<<<<<<< HEAD
+=======
+  ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
+  ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
+>>>>>>> 9f0281356 (Only add DEFAULT_BRANCH environment variable to archery docker command in integration.yml)
 
 jobs:
 
@@ -85,6 +90,7 @@ jobs:
         env:
           ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
           ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
+          DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
         run: archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 conda-integration
       - name: Docker Push
         if: success() && github.event_name == 'push' && github.repository == 'apache/arrow'

From 4139d8225ad7c93a374b3a42fbbfaf0b0100321d Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 21 Sep 2022 16:31:08 -0400
Subject: [PATCH 34/48] In integration.yml, merge edits from default branch and
 current feature branch.

---
 .github/workflows/integration.yml | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/.github/workflows/integration.yml b/.github/workflows/integration.yml
index 4380dec782d..3e6d0d14c78 100644
--- a/.github/workflows/integration.yml
+++ b/.github/workflows/integration.yml
@@ -87,11 +87,18 @@ jobs:
       - name: Setup Archery
         run: pip install -e dev/archery[docker]
       - name: Execute Docker Build
+<<<<<<< HEAD
         env:
           ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
           ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
           DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
         run: archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 conda-integration
+=======
+        run: >
+          archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 -e
+          DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
+          conda-integration
+>>>>>>> 2600c98a1 (In integration.yml, fix multi-line symbol.)
       - name: Docker Push
         if: success() && github.event_name == 'push' && github.repository == 'apache/arrow'
         env:

From 7d6ab6201105e13e4294118f98bec075e5a9ecf1 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 22 Sep 2022 10:29:13 -0400
Subject: [PATCH 35/48] Add DEFAULT_BRANCH env var for archery docker run
 command in .travis.yml

---
 .travis.yml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/.travis.yml b/.travis.yml
index a15822b4a74..4a957b0ce78 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -187,6 +187,7 @@ script:
   - |
     archery docker run \
       ${DOCKER_RUN_ARGS} \
+      -e DEFAULT_BRANCH=${{ github.event.repository.default_branch }} \
       --volume ${PWD}/build:/build \
       ${DOCKER_IMAGE_ID}
 

From 7e1d68dd19f40b5be6d3ff642deeeb3cbce43f5c Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Fri, 23 Sep 2022 11:03:03 -0400
Subject: [PATCH 36/48] Use git command to get default branch name in
 .travis.yml

---
 .travis.yml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/.travis.yml b/.travis.yml
index 4a957b0ce78..b2caa04ba81 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -185,9 +185,10 @@ install:
 
 script:
   - |
+    GITHUB_DEFAULT_BRANCH_NAME=git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@
     archery docker run \
       ${DOCKER_RUN_ARGS} \
-      -e DEFAULT_BRANCH=${{ github.event.repository.default_branch }} \
+      -e DEFAULT_BRANCH=${GITHUB_DEFAULT_BRANCH_NAME} \
       --volume ${PWD}/build:/build \
       ${DOCKER_IMAGE_ID}
 

From d587402b83eedb67d55d92b4874a5ca52be8fbef Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Fri, 23 Sep 2022 12:31:07 -0400
Subject: [PATCH 37/48] Fix integration.yml merge

---
 .github/workflows/integration.yml | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/.github/workflows/integration.yml b/.github/workflows/integration.yml
index 3e6d0d14c78..e7cd861cd31 100644
--- a/.github/workflows/integration.yml
+++ b/.github/workflows/integration.yml
@@ -50,11 +50,6 @@ permissions:
 
 env:
   DOCKER_VOLUME_PREFIX: ".docker/"
-<<<<<<< HEAD
-=======
-  ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
-  ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
->>>>>>> 9f0281356 (Only add DEFAULT_BRANCH environment variable to archery docker command in integration.yml)
 
 jobs:
 
@@ -87,18 +82,13 @@ jobs:
       - name: Setup Archery
         run: pip install -e dev/archery[docker]
       - name: Execute Docker Build
-<<<<<<< HEAD
         env:
           ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
           ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
-          DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
-        run: archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 conda-integration
-=======
         run: >
           archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 -e
           DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
           conda-integration
->>>>>>> 2600c98a1 (In integration.yml, fix multi-line symbol.)
       - name: Docker Push
         if: success() && github.event_name == 'push' && github.repository == 'apache/arrow'
         env:

From d944930e96f86c6c691e5e9725ce930220510021 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Fri, 23 Sep 2022 15:22:16 -0400
Subject: [PATCH 38/48] Set and export the DEFAULT_BRANCH env var for the
 archery command.

---
 .travis.yml | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index b2caa04ba81..43fcc679066 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -184,11 +184,10 @@ install:
   - sudo -H pip3 install -e dev/archery[docker]
 
 script:
+  - export DEFAULT_BRANCH=$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)
   - |
-    GITHUB_DEFAULT_BRANCH_NAME=git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@
     archery docker run \
       ${DOCKER_RUN_ARGS} \
-      -e DEFAULT_BRANCH=${GITHUB_DEFAULT_BRANCH_NAME} \
       --volume ${PWD}/build:/build \
       ${DOCKER_IMAGE_ID}
 

From 6a2dccd0c48b820efcb50a2f1c14c589dd5b5304 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 12 Oct 2022 13:30:02 -0400
Subject: [PATCH 39/48] Remove computation for default branch name from module
 loading step in archery/archery/crossbow/cli.py

---
 dev/archery/archery/crossbow/cli.py | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index b7390ef1417..ba5136ca15b 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -155,14 +155,8 @@ def submit(obj, tasks, groups, params, job_prefix, config_path, arrow_version,
         queue.push()
         click.echo('Pushed job identifier is: `{}`'.format(job.branch))
 
-
-# Get the default branch name from the repository
-arrow_source_dir = ArrowSources.find()
-repo = Repo(arrow_source_dir.path)
-
-
 @crossbow.command()
-@click.option('--base-branch', default=repo.default_branch_name,
+@click.option('--base-branch', default=None,
               help='Set base branch for the PR.')
 @click.option('--create-pr', is_flag=True, default=False,
               help='Create GitHub Pull Request')
@@ -195,6 +189,13 @@ def verify_release_candidate(obj, base_branch, create_pr,
     # The verify-release-candidate command will create a PR (or find one)
     # and add the verify-rc* comment to trigger the verify tasks
 
+    # Default value for base_branch is the repository's default branch name
+    if base_branch is None:
+        # Get the default branch name from the repository
+        arrow_source_dir = ArrowSources.find()
+        repo = Repo(arrow_source_dir.path)
+        base_branch = repo.default_branch_name
+
     # Redefine Arrow repo to use the correct arrow remote.
     arrow = Repo(path=obj['arrow'].path, remote_url=remote)
     response = arrow.github_pr(title=pr_title, head=head_branch,

From dc6e939d3ee9d48dc5300bb1774b02dcedc5780a Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 12 Oct 2022 14:02:31 -0400
Subject: [PATCH 40/48] Removing error if default branch cannot be determined,
 default to 'master' for now in dev/archery/archery/crossbow/core.py

---
 dev/archery/archery/crossbow/core.py | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 02228cc66ee..13dbaebcc28 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -28,6 +28,7 @@
 from io import StringIO
 from pathlib import Path
 from datetime import date
+import warnings
 
 import jinja2
 from ruamel.yaml import YAML
@@ -372,10 +373,14 @@ def default_branch_name(self):
                 target_name_tokenized = target_name.split("/")
                 default_branch_name = target_name_tokenized[-1]
             except KeyError:
-                raise RuntimeError(
-                    'Unable to determine default branch name: DEFAULT_BRANCH '
-                    'environment variable is not set. Git repository does not '
-                    'contain a \'refs/remotes/origin/HEAD\' reference.')
+                # TODO: ARROW-18011 to track changing the hard coded default
+                # value from "master" to "main".
+                default_branch_name = "master"
+                warnings.warn('Unable to determine default branch name: '
+                    'DEFAULT_BRANCH environment variable is not set. Git '
+                    'repository does not contain a \'refs/remotes/origin/HEAD\''
+                    ' reference. Setting the default branch name to ' +
+                    default_branch_name, RuntimeWarning)
 
         return default_branch_name
 

From 3bd8a2e9649b6efd095b63d2af1fee8c9092abdf Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 12 Oct 2022 14:05:34 -0400
Subject: [PATCH 41/48] Alphabetize the standard library imports in
 dev/archery/archery/release/core.py

---
 dev/archery/archery/release/core.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 20cb47a2eb4..2786f96a663 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -18,11 +18,11 @@
 from abc import abstractmethod
 from collections import defaultdict
 import functools
-import re
+import os
 import pathlib
+import re
 import shelve
 import warnings
-import os
 
 from git import Repo
 from jira import JIRA

From c1670e6418a3b345b934bc4bcfd0eb6a8575eeec Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 12 Oct 2022 14:26:23 -0400
Subject: [PATCH 42/48] Remove error in the case that the default branch name
 could not be determined, default to 'master' in
 dev/archery/archery/release/core.py

---
 dev/archery/archery/release/core.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index 2786f96a663..a34a11cc3e6 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -389,10 +389,14 @@ def default_branch(self):
                 # The last token is the default branch name
                 default_branch_name = origin_head_name_tokenized[-1]
             except KeyError:
-                raise RuntimeError(
-                    'Unable to determine default branch name: DEFAULT_BRANCH '
-                    'environment variable is not set. Git repository does not '
-                    'contain a \'refs/remotes/origin/HEAD\' reference.')
+                # TODO: ARROW-18011 to track changing the hard coded default
+                # value from "master" to "main".
+                default_branch_name = "master"
+                warnings.warn('Unable to determine default branch name: '
+                    'DEFAULT_BRANCH environment variable is not set. Git '
+                    'repository does not contain a \'refs/remotes/origin/HEAD\''
+                    ' reference. Setting the default branch name to ' +
+                    default_branch_name, RuntimeWarning)
 
         return default_branch_name
 

From a7793a05dd547dce913877140ff6a0edfb47c472 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 12 Oct 2022 14:42:39 -0400
Subject: [PATCH 43/48] Remame DEFAULT_BRANCH env var to ARCHERY_DEFAULT_RBANCH

---
 .github/workflows/archery.yml        |  4 ++--
 .github/workflows/integration.yml    |  2 +-
 .travis.yml                          |  2 +-
 dev/archery/archery/crossbow/cli.py  |  1 +
 dev/archery/archery/crossbow/core.py | 11 ++++++-----
 dev/archery/archery/release/core.py  | 11 ++++++-----
 6 files changed, 17 insertions(+), 14 deletions(-)

diff --git a/.github/workflows/archery.yml b/.github/workflows/archery.yml
index 5e4d1764a6a..d337ec797cf 100644
--- a/.github/workflows/archery.yml
+++ b/.github/workflows/archery.yml
@@ -32,7 +32,7 @@ on:
       - 'docker-compose.yml'
 
 env:
-  DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
+  ARCHERY_DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
 
 concurrency:
   group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
@@ -55,7 +55,7 @@ jobs:
           fetch-depth: 0
       - name: Git Fixup
         shell: bash
-        run: git branch $DEFAULT_BRANCH origin/$DEFAULT_BRANCH || true
+        run: git branch $ARCHERY_DEFAULT_BRANCH origin/$ARCHERY_DEFAULT_BRANCH || true
       - name: Setup Python
         uses: actions/setup-python@v4
         with:
diff --git a/.github/workflows/integration.yml b/.github/workflows/integration.yml
index e7cd861cd31..0e6351c523e 100644
--- a/.github/workflows/integration.yml
+++ b/.github/workflows/integration.yml
@@ -87,7 +87,7 @@ jobs:
           ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
         run: >
           archery docker run -e ARCHERY_INTEGRATION_WITH_RUST=1 -e
-          DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
+          ARCHERY_DEFAULT_BRANCH=${{ github.event.repository.default_branch }}
           conda-integration
       - name: Docker Push
         if: success() && github.event_name == 'push' && github.repository == 'apache/arrow'
diff --git a/.travis.yml b/.travis.yml
index 43fcc679066..a96e07f0c43 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -184,7 +184,7 @@ install:
   - sudo -H pip3 install -e dev/archery[docker]
 
 script:
-  - export DEFAULT_BRANCH=$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)
+  - export ARCHERY_DEFAULT_BRANCH=$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)
   - |
     archery docker run \
       ${DOCKER_RUN_ARGS} \
diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index ba5136ca15b..ec8eab2a4b5 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -155,6 +155,7 @@ def submit(obj, tasks, groups, params, job_prefix, config_path, arrow_version,
         queue.push()
         click.echo('Pushed job identifier is: `{}`'.format(job.branch))
 
+
 @crossbow.command()
 @click.option('--base-branch', default=None,
               help='Set base branch for the PR.')
diff --git a/dev/archery/archery/crossbow/core.py b/dev/archery/archery/crossbow/core.py
index 13dbaebcc28..553b2d69216 100644
--- a/dev/archery/archery/crossbow/core.py
+++ b/dev/archery/archery/crossbow/core.py
@@ -364,7 +364,7 @@ def signature(self):
 
     @property
     def default_branch_name(self):
-        default_branch_name = os.getenv("DEFAULT_BRANCH")
+        default_branch_name = os.getenv("ARCHERY_DEFAULT_BRANCH")
 
         if default_branch_name is None:
             try:
@@ -377,10 +377,11 @@ def default_branch_name(self):
                 # value from "master" to "main".
                 default_branch_name = "master"
                 warnings.warn('Unable to determine default branch name: '
-                    'DEFAULT_BRANCH environment variable is not set. Git '
-                    'repository does not contain a \'refs/remotes/origin/HEAD\''
-                    ' reference. Setting the default branch name to ' +
-                    default_branch_name, RuntimeWarning)
+                              'ARCHERY_DEFAULT_BRANCH environment variable is '
+                              'not set. Git repository does not contain a '
+                              '\'refs/remotes/origin/HEAD\'reference. Setting '
+                              'the default branch name to' +
+                              default_branch_name, RuntimeWarning)
 
         return default_branch_name
 
diff --git a/dev/archery/archery/release/core.py b/dev/archery/archery/release/core.py
index a34a11cc3e6..d9c85740093 100644
--- a/dev/archery/archery/release/core.py
+++ b/dev/archery/archery/release/core.py
@@ -364,7 +364,7 @@ def commits(self):
 
     @cached_property
     def default_branch(self):
-        default_branch_name = os.getenv("DEFAULT_BRANCH")
+        default_branch_name = os.getenv("ARCHERY_DEFAULT_BRANCH")
 
         if default_branch_name is None:
             try:
@@ -393,10 +393,11 @@ def default_branch(self):
                 # value from "master" to "main".
                 default_branch_name = "master"
                 warnings.warn('Unable to determine default branch name: '
-                    'DEFAULT_BRANCH environment variable is not set. Git '
-                    'repository does not contain a \'refs/remotes/origin/HEAD\''
-                    ' reference. Setting the default branch name to ' +
-                    default_branch_name, RuntimeWarning)
+                              'ARCHERY_DEFAULT_BRANCH environment variable is '
+                              'not set. Git repository does not contain a '
+                              '\'refs/remotes/origin/HEAD\'reference. Setting '
+                              'the default branch name to' +
+                              default_branch_name, RuntimeWarning)
 
         return default_branch_name
 

From 40c6390e5229a32af5551f54dc07f99e95c17806 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 13 Oct 2022 15:40:05 -0400
Subject: [PATCH 44/48] Reuse arrow Repo object for getting the default branch
 name, if needed

---
 dev/archery/archery/crossbow/cli.py | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index ec8eab2a4b5..d9ab7a33841 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -190,15 +190,14 @@ def verify_release_candidate(obj, base_branch, create_pr,
     # The verify-release-candidate command will create a PR (or find one)
     # and add the verify-rc* comment to trigger the verify tasks
 
+    # Redefine Arrow repo to use the correct arrow remote.
+    arrow = Repo(path=obj['arrow'].path, remote_url=remote)
+
     # Default value for base_branch is the repository's default branch name
     if base_branch is None:
         # Get the default branch name from the repository
-        arrow_source_dir = ArrowSources.find()
-        repo = Repo(arrow_source_dir.path)
-        base_branch = repo.default_branch_name
-
-    # Redefine Arrow repo to use the correct arrow remote.
-    arrow = Repo(path=obj['arrow'].path, remote_url=remote)
+        base_branch = arrow.default_branch_name
+    
     response = arrow.github_pr(title=pr_title, head=head_branch,
                                base=base_branch, body=pr_body,
                                github_token=obj['queue'].github_token,

From 3775c65e308a9a24df45f4516dee5fca8c4f98e3 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Thu, 13 Oct 2022 16:34:10 -0400
Subject: [PATCH 45/48] Update the dask and pandas install scripts to use
 default branch computed by Git rather than hard-coded defaults

---
 ci/scripts/install_dask.sh   | 5 ++++-
 ci/scripts/install_pandas.sh | 5 ++++-
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/ci/scripts/install_dask.sh b/ci/scripts/install_dask.sh
index eb9c4e3dd42..50be7b38a6c 100755
--- a/ci/scripts/install_dask.sh
+++ b/ci/scripts/install_dask.sh
@@ -26,7 +26,10 @@ fi
 
 dask=$1
 
-if [ "${dask}" = "master" ]; then
+# Get Git default branch name
+DEFAULT_BRANCH="$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)"
+
+if [ "${dask}" = "${DEFAULT_BRANCH}" ]; then
   pip install https://github.com/dask/dask/archive/main.tar.gz#egg=dask[dataframe]
 elif [ "${dask}" = "latest" ]; then
   pip install dask[dataframe]
diff --git a/ci/scripts/install_pandas.sh b/ci/scripts/install_pandas.sh
index 5aca65f825a..135f60edf67 100755
--- a/ci/scripts/install_pandas.sh
+++ b/ci/scripts/install_pandas.sh
@@ -35,7 +35,10 @@ else
   pip install numpy==${numpy}
 fi
 
-if [ "${pandas}" = "master" ]; then
+# Get Git default branch name
+DEFAULT_BRANCH="$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)"
+
+if [ "${pandas}" = "${DEFAULT_BRANCH}" ]; then
   pip install git+https://github.com/pandas-dev/pandas.git --no-build-isolation
 elif [ "${pandas}" = "nightly" ]; then
   pip install --extra-index-url https://pypi.anaconda.org/scipy-wheels-nightly/simple --pre pandas

From 669a14eff38834515d866e2e3d772678bc4a8783 Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Fri, 14 Oct 2022 14:56:49 -0400
Subject: [PATCH 46/48] Change the flag for indicating upstream development
 version of Pandas and Dask to 'upstream_devel', and update the documentation.

---
 ci/scripts/install_dask.sh                               | 5 +----
 ci/scripts/install_pandas.sh                             | 5 +----
 dev/archery/archery/docker/tests/test_docker.py          | 8 ++++----
 docs/source/developers/continuous_integration/docker.rst | 7 ++++---
 4 files changed, 10 insertions(+), 15 deletions(-)

diff --git a/ci/scripts/install_dask.sh b/ci/scripts/install_dask.sh
index 50be7b38a6c..8d712a88a6a 100755
--- a/ci/scripts/install_dask.sh
+++ b/ci/scripts/install_dask.sh
@@ -26,10 +26,7 @@ fi
 
 dask=$1
 
-# Get Git default branch name
-DEFAULT_BRANCH="$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)"
-
-if [ "${dask}" = "${DEFAULT_BRANCH}" ]; then
+if [ "${dask}" = "upstream_devel" ]; then
   pip install https://github.com/dask/dask/archive/main.tar.gz#egg=dask[dataframe]
 elif [ "${dask}" = "latest" ]; then
   pip install dask[dataframe]
diff --git a/ci/scripts/install_pandas.sh b/ci/scripts/install_pandas.sh
index 135f60edf67..be29e5da554 100755
--- a/ci/scripts/install_pandas.sh
+++ b/ci/scripts/install_pandas.sh
@@ -35,10 +35,7 @@ else
   pip install numpy==${numpy}
 fi
 
-# Get Git default branch name
-DEFAULT_BRANCH="$(git rev-parse --abbrev-ref origin/HEAD | sed s@origin/@@)"
-
-if [ "${pandas}" = "${DEFAULT_BRANCH}" ]; then
+if [ "${pandas}" = "upstream_devel" ]; then
   pip install git+https://github.com/pandas-dev/pandas.git --no-build-isolation
 elif [ "${pandas}" = "nightly" ]; then
   pip install --extra-index-url https://pypi.anaconda.org/scipy-wheels-nightly/simple --pre pandas
diff --git a/dev/archery/archery/docker/tests/test_docker.py b/dev/archery/archery/docker/tests/test_docker.py
index fa6b59f936d..386b7c2bdae 100644
--- a/dev/archery/archery/docker/tests/test_docker.py
+++ b/dev/archery/archery/docker/tests/test_docker.py
@@ -259,12 +259,12 @@ def test_arrow_example_validation_passes(arrow_compose_path):
 def test_compose_default_params_and_env(arrow_compose_path):
     compose = DockerCompose(arrow_compose_path, params=dict(
         UBUNTU='18.04',
-        DASK='main'
+        DASK='upstream_devel'
     ))
     assert compose.config.dotenv == arrow_compose_env
     assert compose.config.params == {
         'UBUNTU': '18.04',
-        'DASK': 'main',
+        'DASK': 'upstream_devel',
     }
 
 
@@ -492,7 +492,7 @@ def test_compose_push(arrow_compose_path):
 def test_compose_error(arrow_compose_path):
     compose = DockerCompose(arrow_compose_path, params=dict(
         PYTHON='3.8',
-        PANDAS='main'
+        PANDAS='upstream_devel'
     ))
 
     error = subprocess.CalledProcessError(99, [])
@@ -503,7 +503,7 @@ def test_compose_error(arrow_compose_path):
     exception_message = str(exc.value)
     assert "exited with a non-zero exit code 99" in exception_message
     assert "PANDAS: latest" in exception_message
-    assert "export PANDAS=main" in exception_message
+    assert "export PANDAS=upstream_devel" in exception_message
 
 
 def test_image_with_gpu(arrow_compose_path):
diff --git a/docs/source/developers/continuous_integration/docker.rst b/docs/source/developers/continuous_integration/docker.rst
index 7035dfe9ec6..49061f5b847 100644
--- a/docs/source/developers/continuous_integration/docker.rst
+++ b/docs/source/developers/continuous_integration/docker.rst
@@ -85,13 +85,13 @@ where the leaf image is ``conda-python-pandas``.
 
 .. code:: bash
 
-    PANDAS=master archery docker run --no-leaf-cache conda-python-pandas
+    PANDAS=upstream_devel archery docker run --no-leaf-cache conda-python-pandas
 
 Which translates to:
 
 .. code:: bash
 
-    export PANDAS=master
+    export PANDAS=upstream_devel
     docker-compose pull --ignore-pull-failures conda-cpp
     docker-compose pull --ignore-pull-failures conda-python
     docker-compose build conda-cpp
@@ -102,7 +102,7 @@ Which translates to:
 Note that it doesn't pull the conda-python-pandas image and disable the cache
 when building it.
 
-``PANDAS`` is a `build parameter <Docker Build Parameters>`_, see the
+``PANDAS`` is a :ref:`build parameter <docker-build-parameters>`, see the
 defaults in the .env file.
 
 **To entirely skip building the image:**
@@ -178,6 +178,7 @@ image when building Glib, Ruby, R and Python bindings.
 This reduces duplication and streamlines maintenance, but makes the
 docker-compose configuration more complicated.
 
+.. _docker-build-parameters:
 Docker Build Parameters
 ~~~~~~~~~~~~~~~~~~~~~~~
 

From ddce92e596f2ca8180418c547be01da189bdcc2c Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Fri, 14 Oct 2022 15:09:38 -0400
Subject: [PATCH 47/48] Run python linting

---
 dev/archery/archery/crossbow/cli.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dev/archery/archery/crossbow/cli.py b/dev/archery/archery/crossbow/cli.py
index d9ab7a33841..58aed560329 100644
--- a/dev/archery/archery/crossbow/cli.py
+++ b/dev/archery/archery/crossbow/cli.py
@@ -197,7 +197,7 @@ def verify_release_candidate(obj, base_branch, create_pr,
     if base_branch is None:
         # Get the default branch name from the repository
         base_branch = arrow.default_branch_name
-    
+
     response = arrow.github_pr(title=pr_title, head=head_branch,
                                base=base_branch, body=pr_body,
                                github_token=obj['queue'].github_token,

From f884601631bee67b2d0926ba940a477e6358fcff Mon Sep 17 00:00:00 2001
From: Fiona La <fionala7@gmail.com>
Date: Wed, 19 Oct 2022 09:16:54 -0400
Subject: [PATCH 48/48] Update Dask and Pandas version flag in tasks.yml and
 dev/archery/archery/docker/cli.py

---
 dev/archery/archery/docker/cli.py | 3 ++-
 dev/tasks/tasks.yml               | 4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/dev/archery/archery/docker/cli.py b/dev/archery/archery/docker/cli.py
index 565efed05c5..6f571f27bff 100644
--- a/dev/archery/archery/docker/cli.py
+++ b/dev/archery/archery/docker/cli.py
@@ -217,7 +217,8 @@ def docker_run(obj, image, command, *, env, user, force_pull, force_build,
     PYTHON=3.8 archery docker run conda-python
 
     # disable the cache only for the leaf image
-    PANDAS=main archery docker run --no-leaf-cache conda-python-pandas
+    PANDAS=upstream_devel archery docker run --no-leaf-cache
+    conda-python-pandas
 
     # entirely skip building the image
     archery docker run --no-pull --no-build conda-python
diff --git a/dev/tasks/tasks.yml b/dev/tasks/tasks.yml
index bdf53ff1dac..8cbb7615b7b 100644
--- a/dev/tasks/tasks.yml
+++ b/dev/tasks/tasks.yml
@@ -1492,7 +1492,7 @@ tasks:
                                                                      ("3.7", "latest", "latest", False),
                                                                      ("3.8", "latest", "latest", False),
                                                                      ("3.8", "nightly", "nightly", False),
-                                                                     ("3.9", "master", "nightly", False)] %}
+                                                                     ("3.9", "upstream_devel", "nightly", False)] %}
   test-conda-python-{{ python_version }}-pandas-{{ pandas_version }}:
     ci: github
     template: docker-tests/github.linux.yml
@@ -1512,7 +1512,7 @@ tasks:
       image: conda-python-pandas
 {% endfor %}
 
-{% for dask_version in ["latest", "master"] %}
+{% for dask_version in ["latest", "upstream_devel"] %}
   test-conda-python-3.9-dask-{{ dask_version }}:
     ci: github
     template: docker-tests/github.linux.yml