dev: Add a script to auto fix all lint violations

ci/scripts/doc_prettier_check.sh

@@ -17,41 +17,68 @@
 # specific language governing permissions and limitations
 # under the License.
 
-SCRIPT_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/$(basename "${BASH_SOURCE[0]}")"
-
-MODE="--check"
-ACTION="Checking"
-if [ $# -gt 0 ]; then
-  if [ "$1" = "--write" ]; then
-    MODE="--write"
-    ACTION="Formatting"
-  else
-    echo "Usage: $0 [--write]" >&2
-    exit 1
-  fi
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT_NAME="$(basename "${BASH_SOURCE[0]}")"
+PRETTIER_VERSION="2.7.1"
+PRETTIER_TARGETS=(
+  '{datafusion,datafusion-cli,datafusion-examples,dev,docs}/**/*.md'
+  '!datafusion/CHANGELOG.md'
+  README.md
+  CONTRIBUTING.md
+)
+
+source "${SCRIPT_DIR}/utils/git.sh"
+
+MODE="check"
+ALLOW_DIRTY=0
+
+usage() {
+  cat >&2 <<EOF
+Usage: $SCRIPT_NAME [--write] [--allow-dirty]
+
+Runs prettier@${PRETTIER_VERSION} over markdown docs.
+--write         Run with \`--write\` to format files (requires a clean git worktree, no uncommitted changes).
+--allow-dirty   Allow \`--write\` to run even when the git worktree has uncommitted changes.
+EOF
+  exit 1
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --write)
+      MODE="write"
+      ;;
+    --allow-dirty)
+      ALLOW_DIRTY=1
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+  shift
+done
+
+if [[ "$MODE" == "write" && $ALLOW_DIRTY -eq 0 ]]; then
+  require_clean_work_tree "$SCRIPT_NAME" || exit 1
 fi
 
-echo "$SCRIPT_PATH: $ACTION documents with prettier"
+echo "[${SCRIPT_NAME}] prettier@${PRETTIER_VERSION} ${MODE}"
 
 # Ensure `npx` is available
 if ! command -v npx >/dev/null 2>&1; then
   echo "npx is required to run the prettier check. Install Node.js (e.g., brew install node) and re-run." >&2
   exit 1
 fi
-
-# Ignore subproject CHANGELOG.md because it is machine generated
-npx [email protected] $MODE \
-  '{datafusion,datafusion-cli,datafusion-examples,dev,docs}/**/*.md' \
-  '!datafusion/CHANGELOG.md' \
-  README.md \
-  CONTRIBUTING.md
-status=$?
-
-if [ $status -ne 0 ]; then
-  if [ "$MODE" = "--check" ]; then
-    echo "Prettier check failed. Re-run with --write (e.g., ./ci/scripts/doc_prettier_check.sh --write) to format files, commit the changes, and re-run the check." >&2
-  else
-    echo "Prettier format failed. Files may have been modified; commit any changes and re-run." >&2
-  fi
-  exit $status
+
+PRETTIER_MODE=(--check)
+if [[ "$MODE" == "write" ]]; then
+  PRETTIER_MODE=(--write)
 fi
+
+# Ignore subproject CHANGELOG.md because it is machine generated
+npx "prettier@${PRETTIER_VERSION}" "${PRETTIER_MODE[@]}" "${PRETTIER_TARGETS[@]}"

ci/scripts/license_header.sh

@@ -17,6 +17,62 @@
 # specific language governing permissions and limitations
 # under the License.
 
-# Check Apache license header
-set -ex
-hawkeye check --config licenserc.toml
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT_NAME="$(basename "${BASH_SOURCE[0]}")"
+
+source "${SCRIPT_DIR}/utils/git.sh"
+
+MODE="check"
+ALLOW_DIRTY=0
+HAWKEYE_CONFIG="licenserc.toml"
+
+usage() {
+  cat >&2 <<EOF
+Usage: $SCRIPT_NAME [--write] [--allow-dirty]
+
+Checks Apache license headers with \`hawkeye check --config $HAWKEYE_CONFIG\`.
+--write         Run \`hawkeye format --config $HAWKEYE_CONFIG\` to auto-add/fix headers (requires a clean git worktree, no uncommitted changes).
+--allow-dirty   Allow \`--write\` to run even when the git worktree has uncommitted changes.
+EOF
+  exit 1
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --write)
+      MODE="write"
+      ;;
+    --allow-dirty)
+      ALLOW_DIRTY=1
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+  shift
+done
+
+if [[ "$MODE" == "write" && $ALLOW_DIRTY -eq 0 ]]; then
+  require_clean_work_tree "$SCRIPT_NAME" || exit 1
+fi
+
+if [[ "$MODE" == "write" ]]; then
+  echo "[${SCRIPT_NAME}] \`hawkeye format --config ${HAWKEYE_CONFIG}\`"
+  if ! hawkeye format --config "${HAWKEYE_CONFIG}"; then
+    status=$?
+    # hawkeye returns exit code 1 when it applies fixes; treat that as success.
+    if [[ $status -eq 1 ]]; then
+      echo "[${SCRIPT_NAME}] hawkeye format applied fixes (exit 1 treated as success)"
+    else
+      exit $status
+    fi
+  fi
+else
+  echo "[${SCRIPT_NAME}] \`hawkeye check --config ${HAWKEYE_CONFIG}\`"
+  hawkeye check --config "${HAWKEYE_CONFIG}"
+fi

ci/scripts/rust_clippy.sh

@@ -17,5 +17,60 @@
 # specific language governing permissions and limitations
 # under the License.
 
-set -ex
-cargo clippy --all-targets --workspace --features avro,integration-tests,extended_tests -- -D warnings
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT_NAME="$(basename "${BASH_SOURCE[0]}")"
+CLIPPY_FEATURES="avro,integration-tests,extended_tests"
+CLIPPY_ARGS=(--all-targets --workspace --features "$CLIPPY_FEATURES")
+CLIPPY_LINT_ARGS=(-- -D warnings)
+
+source "${SCRIPT_DIR}/utils/git.sh"
+
+MODE="check"
+ALLOW_DIRTY=0
+
+usage() {
+  cat >&2 <<EOF
+Usage: $SCRIPT_NAME [--write] [--allow-dirty]
+
+Runs \`cargo clippy\` to lint.
+--write         Run \`cargo clippy --fix\` to apply fixes for clippy lints (requires a clean git worktree, no uncommitted changes).
+--allow-dirty   Allow \`--write\` to run even when the git worktree has uncommitted or staged changes.
+EOF
+  exit 1
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --write)
+      MODE="write"
+      ;;
+    --allow-dirty)
+      ALLOW_DIRTY=1
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+  shift
+done
+
+if [[ "$MODE" == "write" && $ALLOW_DIRTY -eq 0 ]]; then
+  require_clean_work_tree "$SCRIPT_NAME" || exit 1
+fi
+
+CLIPPY_CMD=(cargo clippy)
+if [[ "$MODE" == "write" ]]; then
+  CLIPPY_CMD+=(--fix)
+  if [[ $ALLOW_DIRTY -eq 1 ]]; then
+    CLIPPY_CMD+=(--allow-dirty --allow-staged)
+  fi
+fi
+CLIPPY_CMD+=("${CLIPPY_ARGS[@]}" "${CLIPPY_LINT_ARGS[@]}")
+
+echo "[${SCRIPT_NAME}] \`${CLIPPY_CMD[*]}\`"
+"${CLIPPY_CMD[@]}"

ci/scripts/rust_docs.sh

@@ -17,6 +17,7 @@
 # specific language governing permissions and limitations
 # under the License.
 
+# Note: cargo doc does not support an auto-fix mode; this script runs the check-only build.
 set -ex
 export RUSTDOCFLAGS="-D warnings"
 cargo doc --document-private-items --no-deps --workspace

ci/scripts/rust_fmt.sh

@@ -17,5 +17,52 @@
 # specific language governing permissions and limitations
 # under the License.
 
-set -ex
-cargo fmt --all -- --check
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT_NAME="$(basename "${BASH_SOURCE[0]}")"
+source "${SCRIPT_DIR}/utils/git.sh"
+
+MODE="check"
+ALLOW_DIRTY=0
+
+usage() {
+  cat >&2 <<EOF
+Usage: $0 [--write] [--allow-dirty]
+
+Runs \`cargo fmt --all -- --check\` by default to verify Rust formatting.
+--write        Run \`cargo fmt --all\` to auto-fix formatting (requires a clean git worktree, no uncommitted changes).
+--allow-dirty  Allow \`--write\` to run even when the git worktree has uncommitted changes.
+EOF
+  exit 1
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --write)
+      MODE="write"
+      ;;
+    --allow-dirty)
+      ALLOW_DIRTY=1
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+  shift
+done
+
+if [[ "$MODE" == "write" && $ALLOW_DIRTY -eq 0 ]]; then
+  require_clean_work_tree "$SCRIPT_NAME" || exit 1
+fi
+
+if [[ "$MODE" == "write" ]]; then
+  echo "[${SCRIPT_NAME}] \`cargo fmt --all\`"
+  cargo fmt --all
+else
+  echo "[${SCRIPT_NAME}] \`cargo fmt --all -- --check\`"
+  cargo fmt --all -- --check
+fi

ci/scripts/rust_toml_fmt.sh

@@ -17,8 +17,53 @@
 # specific language governing permissions and limitations
 # under the License.
 
-# Run `taplo format` with flag `--check` in dry run to check formatting
-# without overwritng the file. If any error occur, you may want to
-# rerun `taplo format` to fix the formatting automatically.
-set -ex
-taplo format --check
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPT_NAME="$(basename "${BASH_SOURCE[0]}")"
+
+source "${SCRIPT_DIR}/utils/git.sh"
+
+MODE="check"
+ALLOW_DIRTY=0
+
+usage() {
+  cat >&2 <<EOF
+Usage: $0 [--write] [--allow-dirty]
+
+Runs \`taplo format --check\` by default to verify TOML formatting.
+--write        Run \`taplo format\` to auto-fix formatting (best-effort; requires a clean git worktree, no uncommitted changes).
+--allow-dirty  Allow \`--write\` to run even when the git worktree has uncommitted changes.
+EOF
+  exit 1
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --write)
+      MODE="write"
+      ;;
+    --allow-dirty)
+      ALLOW_DIRTY=1
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+  shift
+done
+
+if [[ "$MODE" == "write" && $ALLOW_DIRTY -eq 0 ]]; then
+  require_clean_work_tree "$SCRIPT_NAME" || exit 1
+fi
+
+if [[ "$MODE" == "write" ]]; then
+  echo "[${SCRIPT_NAME}] \`taplo format\`"
+  taplo format
+else
+  echo "[${SCRIPT_NAME}] \`taplo format --check\`"
+  taplo format --check
+fi