From c250d1924be18f2b5a426e48bb77bfa93ee6b57f Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 2 Jul 2025 08:36:10 +0200
Subject: [PATCH 1/6] Find PR base branch without GitHub API

---
 .gitlab/find-gh-base-ref.sh | 100 ++++++++++++------------------------
 1 file changed, 34 insertions(+), 66 deletions(-)

diff --git a/.gitlab/find-gh-base-ref.sh b/.gitlab/find-gh-base-ref.sh
index e007ab796a2..6d61efb51c1 100755
--- a/.gitlab/find-gh-base-ref.sh
+++ b/.gitlab/find-gh-base-ref.sh
@@ -2,6 +2,9 @@
 # Determines the base branch for the current PR (if we are running in a PR).
 set -euo pipefail
 
+# DEBUG: List all branches (to stderr)
+git branch -a --sort=committerdate --format='%(refname:short)' >&2
+
 CURRENT_HEAD_SHA="$(git rev-parse HEAD)"
 if [[ -z "${CURRENT_HEAD_SHA:-}" ]]; then
   echo "Failed to determine current HEAD SHA" >&2
@@ -40,74 +43,39 @@ if [[ $(git log --pretty=oneline origin/master..HEAD | wc -l) -eq 1 ]]; then
   exit 0
 fi
 
-# In GitLab: we have no reference to the base branch or even the PR number.
-# We have to find it from the current branch name, which is defined in
-# CI_COMMIT_REF_NAME.
-if [[ -z "${CI_COMMIT_REF_NAME}" ]]; then
-  echo "CI_COMMIT_REF_NAME is not set, not running in GitLab CI?" >&2
-  exit 1
-fi
-
-# In GitLab, CI_PROJECT_NAME is set, otherwise, set it for testing.
-export CI_PROJECT_NAME="${CI_PROJECT_NAME:-dd-trace-java}"
+get_distance_from_merge_base() {
+  local candidate_base="$1"
+  merge_base_sha=$(git merge-base "$candidate_base" HEAD)
+  distance=$(git log --pretty=oneline "$merge_base_sha".."$CURRENT_HEAD_SHA" | wc -l)
+  echo "Distance from $candidate_base is $distance" >&2
+  echo "$distance"
+}
 
-if [[ -z "${GITHUB_TOKEN:-}" ]]; then
-  echo "GITHUB_TOKEN is not set, fetching from AWS SSM" >&2
-  if ! command -v aws >/dev/null 2>&1; then
-    echo "aws is not installed, please install it" >&2
-    exit 1
+# Find the best base ref: the master/release branch whose merge base is closest to HEAD.
+# If there are multiple candidates (e.g. immediately after a release branch is created), we cannot
+# disambiguate and return an error.
+# NOTE: GitHub API is more robust for this task, but we hit rate limits.
+BEST_CANDIDATES=(origin/master)
+BEST_DISTANCE=$(get_distance_from_merge_base origin/master)
+mapfile -t CANDIDATE_BASES < <(git branch -a --sort=committerdate --format='%(refname:short)' --list 'origin/release/v*' | tac)
+for candidate_base in "${CANDIDATE_BASES[@]}"; do
+  distance=$(get_distance_from_merge_base "$candidate_base")
+  if [[ $distance -lt $BEST_DISTANCE ]]; then
+    BEST_DISTANCE=$distance
+    BEST_CANDIDATES=("$candidate_base")
+  elif [[ $distance -eq $BEST_DISTANCE ]]; then
+    BEST_CANDIDATES+=("$candidate_base")
   fi
-  set +e
-  GITHUB_TOKEN=$(aws ssm get-parameter --name "ci.$CI_PROJECT_NAME.gh_release_token" --with-decryption --query "Parameter.Value" --output text)
-  set -e
-  if [[ -z "${GITHUB_TOKEN:-}" ]]; then
-    echo "Failed to fetch GITHUB_TOKEN from AWS SSM" >&2
-    exit 1
-  fi
-  export GITHUB_TOKEN
-fi
-
-if ! command -v curl >/dev/null 2>&1; then
-  echo "curl is not installed, please install it" >&2
-  exit 1
-fi
+done
 
-if ! command -v jq >/dev/null 2>&1; then
-  echo "jq is not installed, please install it" >&2
-  exit 1
+if [[ ${#BEST_CANDIDATES[@]} -eq 1 ]]; then
+  # Remote the origin/ prefix
+  base_ref="${BEST_CANDIDATES[0]#origin/}"
+  echo "Base ref is ${base_ref}" >&2
+  save_cache "${base_ref}" "$CURRENT_HEAD_SHA"
+  echo "${base_ref}"
+  exit 0
 fi
 
-while true; do
-  set +e
-  PR_DATA=$(curl \
-    -XGET \
-    --silent \
-    --include \
-    --fail-with-body \
-    -H 'Accept: application/vnd.github+json' \
-    -H "Authorization: Bearer ${GITHUB_TOKEN}" \
-    -H "X-GitHub-Api-Version: 2022-11-28" \
-    "https://api.github.com/repos/datadog/dd-trace-java/pulls?head=DataDog:${CI_COMMIT_REF_NAME}&sort=updated&direction=desc")
-  exit_code=$?
-  set -e
-  if [[ ${exit_code} -eq 0 ]]; then
-    PR_NUMBER=$(echo "$PR_DATA" | sed '1,/^[[:space:]]*$/d' | jq -r '.[].number')
-    PR_BASE_REF=$(echo "$PR_DATA" | sed '1,/^[[:space:]]*$/d' | jq -r '.[].base.ref')
-    if [[ -n "${PR_BASE_REF:-}" ]]; then
-      echo "PR is https://github.com/datadog/dd-trace-java/pull/${PR_NUMBER} and base ref is ${PR_BASE_REF}">&2
-      save_cache "${PR_BASE_REF}" "$CURRENT_HEAD_SHA"
-      echo "${PR_BASE_REF}"
-      exit 0
-    fi
-  fi
-  if echo "$PR_DATA" | grep -q "^x-ratelimit-reset:"; then
-    reset_timestamp=$(echo -n "$PR_DATA" | grep "^x-ratelimit-reset:" | sed -e 's/^x-ratelimit-reset: //' -e 's/\r//')
-    now=$(date +%s)
-    sleep_time=$((reset_timestamp - now + 1))
-    echo "GitHub rate limit exceeded, sleeping for ${sleep_time} seconds" >&2
-    sleep "${sleep_time}"
-    continue
-  fi
-  echo -e "GitHub request failed for an unknown reason:\n$(echo "$PR_DATA" | sed '/^$/q')" >&2
-  exit 1
-done
+echo "Base ref is ambiguous, candidates are: ${BEST_CANDIDATES[*]}" >&2
+exit 1

From d405fdff1c465ffa645d97ce68a5447bdbde5517 Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 2 Jul 2025 09:15:52 +0200
Subject: [PATCH 2/6] Enable git-based skipping again

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

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index c01b08f7bcb..aadc0a593c5 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -119,8 +119,7 @@ default:
 
 .gitlab_base_ref_params: &gitlab_base_ref_params
   - |
-    # FIXME: Disabled until we find a way to not hit GitHub API rate limit
-    if false && [[ ! $CI_COMMIT_BRANCH =~ ^(master|release/.*)$ ]]; then
+    if [[ ! $CI_COMMIT_BRANCH =~ ^(master|release/.*)$ ]]; then
       export GIT_BASE_REF=$(.gitlab/find-gh-base-ref.sh)
       if [[ -n "$GIT_BASE_REF" ]]; then
         export GRADLE_PARAMS="$GRADLE_PARAMS -PgitBaseRef=origin/$GIT_BASE_REF"

From 3c91d4d2a983664a7ccc4a44ee3ad6aa1a1a42a0 Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 2 Jul 2025 09:38:37 +0200
Subject: [PATCH 3/6] Clarifying comments

---
 .gitlab/find-gh-base-ref.sh | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/.gitlab/find-gh-base-ref.sh b/.gitlab/find-gh-base-ref.sh
index 6d61efb51c1..7bb6f9dd0ca 100755
--- a/.gitlab/find-gh-base-ref.sh
+++ b/.gitlab/find-gh-base-ref.sh
@@ -2,9 +2,6 @@
 # Determines the base branch for the current PR (if we are running in a PR).
 set -euo pipefail
 
-# DEBUG: List all branches (to stderr)
-git branch -a --sort=committerdate --format='%(refname:short)' >&2
-
 CURRENT_HEAD_SHA="$(git rev-parse HEAD)"
 if [[ -z "${CURRENT_HEAD_SHA:-}" ]]; then
   echo "Failed to determine current HEAD SHA" >&2
@@ -77,5 +74,10 @@ if [[ ${#BEST_CANDIDATES[@]} -eq 1 ]]; then
   exit 0
 fi
 
+# If base ref is ambiguous, we cannot determine the correct one.
+# Example: a release branch is created, and a PR is opened starting from the
+# commit where the release branch was created. The distance to the merge base
+# for both master and the release branch is the same. In this case, we bail
+# out, and make no assumption on which is the correct base ref.
 echo "Base ref is ambiguous, candidates are: ${BEST_CANDIDATES[*]}" >&2
 exit 1

From b958dbc62585cec7a4e03eeae208da3c2bca7ddc Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 9 Jul 2025 09:43:34 +0200
Subject: [PATCH 4/6] Use rev-list instead if log

---
 .gitlab/find-gh-base-ref.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.gitlab/find-gh-base-ref.sh b/.gitlab/find-gh-base-ref.sh
index 7bb6f9dd0ca..f6a9fa9c7ee 100755
--- a/.gitlab/find-gh-base-ref.sh
+++ b/.gitlab/find-gh-base-ref.sh
@@ -33,7 +33,7 @@ if [[ -f $CACHE_PATH ]]; then
 fi
 
 # Happy path: if we're just one commit away from master, base ref is master.
-if [[ $(git log --pretty=oneline origin/master..HEAD | wc -l) -eq 1 ]]; then
+if [[ $(git rev-list --count origin/master..HEAD) -eq 1 ]]; then
   echo "We are just one commit away from master, base ref is master" >&2
   save_cache "master" "$CURRENT_HEAD_SHA"
   echo "master"
@@ -43,7 +43,7 @@ fi
 get_distance_from_merge_base() {
   local candidate_base="$1"
   merge_base_sha=$(git merge-base "$candidate_base" HEAD)
-  distance=$(git log --pretty=oneline "$merge_base_sha".."$CURRENT_HEAD_SHA" | wc -l)
+  distance=$(git rev-list --count "$merge_base_sha".."$CURRENT_HEAD_SHA")
   echo "Distance from $candidate_base is $distance" >&2
   echo "$distance"
 }

From a4c9d0bfacff8fc3accd18aecccc024b8177b1eb Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 9 Jul 2025 09:45:04 +0200
Subject: [PATCH 5/6] Use locals

---
 .gitlab/find-gh-base-ref.sh | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/.gitlab/find-gh-base-ref.sh b/.gitlab/find-gh-base-ref.sh
index f6a9fa9c7ee..69e43181018 100755
--- a/.gitlab/find-gh-base-ref.sh
+++ b/.gitlab/find-gh-base-ref.sh
@@ -42,6 +42,8 @@ fi
 
 get_distance_from_merge_base() {
   local candidate_base="$1"
+  local merge_base_sha
+  local distance
   merge_base_sha=$(git merge-base "$candidate_base" HEAD)
   distance=$(git rev-list --count "$merge_base_sha".."$CURRENT_HEAD_SHA")
   echo "Distance from $candidate_base is $distance" >&2

From 0ca758d26b00e7b9f464a185059d6a5bf4b071dd Mon Sep 17 00:00:00 2001
From: Santiago Mola <santiago.mola@datadoghq.com>
Date: Wed, 9 Jul 2025 10:08:48 +0200
Subject: [PATCH 6/6] Support project branches, move further logic within
 find-gh-base-ref.sh

---
 .gitlab-ci.yml              | 12 +++++-------
 .gitlab/find-gh-base-ref.sh | 22 +++++++++++++++++++++-
 2 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index aadc0a593c5..27840d96a49 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -119,13 +119,11 @@ default:
 
 .gitlab_base_ref_params: &gitlab_base_ref_params
   - |
-    if [[ ! $CI_COMMIT_BRANCH =~ ^(master|release/.*)$ ]]; then
-      export GIT_BASE_REF=$(.gitlab/find-gh-base-ref.sh)
-      if [[ -n "$GIT_BASE_REF" ]]; then
-        export GRADLE_PARAMS="$GRADLE_PARAMS -PgitBaseRef=origin/$GIT_BASE_REF"
-      else
-        echo "Failed to find base ref for PR" >&2
-      fi
+    export GIT_BASE_REF=$(.gitlab/find-gh-base-ref.sh)
+    if [[ -n "$GIT_BASE_REF" ]]; then
+      export GRADLE_PARAMS="$GRADLE_PARAMS -PgitBaseRef=origin/$GIT_BASE_REF"
+    else
+      echo "Failed to find base ref for PR" >&2
     fi
 
 .gradle_build: &gradle_build
diff --git a/.gitlab/find-gh-base-ref.sh b/.gitlab/find-gh-base-ref.sh
index 69e43181018..fe85ac0e4aa 100755
--- a/.gitlab/find-gh-base-ref.sh
+++ b/.gitlab/find-gh-base-ref.sh
@@ -2,6 +2,18 @@
 # Determines the base branch for the current PR (if we are running in a PR).
 set -euo pipefail
 
+if [[ -n "${CI_COMMIT_BRANCH:-}" ]]; then
+  echo "CI_COMMIT_BRANCH is set to $CI_COMMIT_BRANCH" >&2
+else
+  echo "CI_COMMIT_BRANCH is not set, skipping base ref detection" >&2
+  exit 1
+fi
+
+if [[ $CI_COMMIT_BRANCH =~ ^(master|release/.*)$ ]]; then
+  echo "CI_COMMIT_BRANCH is a master or release branch, skipping base ref detection" >&2
+  exit 1
+fi
+
 CURRENT_HEAD_SHA="$(git rev-parse HEAD)"
 if [[ -z "${CURRENT_HEAD_SHA:-}" ]]; then
   echo "Failed to determine current HEAD SHA" >&2
@@ -56,7 +68,15 @@ get_distance_from_merge_base() {
 # NOTE: GitHub API is more robust for this task, but we hit rate limits.
 BEST_CANDIDATES=(origin/master)
 BEST_DISTANCE=$(get_distance_from_merge_base origin/master)
-mapfile -t CANDIDATE_BASES < <(git branch -a --sort=committerdate --format='%(refname:short)' --list 'origin/release/v*' | tac)
+
+# If the current branch is not a project/ branch, project/ branches are candidates.
+# This accounts for the case when the project/ branch is being merged to master.
+if [[ ! "$CI_COMMIT_BRANCH" =~ ^project/.*$ ]]; then
+  mapfile -t CANDIDATE_BASES < <(git branch -a --sort=committerdate --format='%(refname:short)' --list 'origin/release/v*' --list 'origin/project/*' | tac)
+else
+  mapfile -t CANDIDATE_BASES < <(git branch -a --sort=committerdate --format='%(refname:short)' --list 'origin/release/v*' | tac)
+fi
+
 for candidate_base in "${CANDIDATE_BASES[@]}"; do
   distance=$(get_distance_from_merge_base "$candidate_base")
   if [[ $distance -lt $BEST_DISTANCE ]]; then