From 9a7e6722ccb47562939fae4a457bfde4d5745065 Mon Sep 17 00:00:00 2001
From: Trevor Guidry <trevor.guidry@mongodb.com>
Date: Fri, 1 Nov 2024 12:21:34 -0400
Subject: [PATCH] SERVER-95851 Fix 0% coverage reports and pull requests have
 separate coveralls builds (#28737)

GitOrigin-RevId: e44387453918d42d722da105670efccf61efe51a
---
 etc/evergreen_yml_components/definitions.yml  | 14 ++++++++++
 .../functions/code_coverage_data_process.py   | 28 +++++++++++++++++++
 2 files changed, 42 insertions(+)

diff --git a/etc/evergreen_yml_components/definitions.yml b/etc/evergreen_yml_components/definitions.yml
index 54c8551bd74..27542d496f3 100644
--- a/etc/evergreen_yml_components/definitions.yml
+++ b/etc/evergreen_yml_components/definitions.yml
@@ -584,6 +584,14 @@ functions:
       exec_timeout_secs: ${exec_timeout_secs}
       timeout_secs: ${timeout_secs}
 
+  "set code coverage expansion": &set_code_coverage_expansion
+    command: expansions.update
+    display_name: "set code coverage expansion"
+    params:
+      updates:
+        - key: gather_code_coverage_results
+          value: "true"
+
   ### Set expansion macros used in each task.
   "set task expansion macros": &set_task_expansion_macros
     command: expansions.update
@@ -1160,6 +1168,8 @@ functions:
           - key: aws_secret_remote
             value: ${mongodatafiles_aws_secret}
     - *f_expansions_write
+    - *set_code_coverage_expansion
+    - *f_expansions_write
     - *set_up_remote_credentials
     - *f_expansions_write
     - *determine_resmoke_jobs
@@ -1188,6 +1198,8 @@ functions:
     - *f_expansions_write
     - *update_task_timeout
     - *f_expansions_write
+    - *set_code_coverage_expansion
+    - *f_expansions_write
     - command: expansions.update
       params:
         env:
@@ -1412,6 +1424,8 @@ functions:
     - *get_version_expansions
     - *apply_version_expansions
     - *f_expansions_write
+    - *set_code_coverage_expansion
+    - *f_expansions_write
     - *bazel_coverage_sh
 
   "ninja compile sh": &ninja_compile_sh
diff --git a/evergreen/functions/code_coverage_data_process.py b/evergreen/functions/code_coverage_data_process.py
index 1fa4a249ff4..9341331d971 100755
--- a/evergreen/functions/code_coverage_data_process.py
+++ b/evergreen/functions/code_coverage_data_process.py
@@ -7,6 +7,7 @@ import sys
 import tarfile
 from urllib.request import urlretrieve
 
+import git
 import requests
 import retry
 
@@ -72,6 +73,31 @@ def main():
         )
         return 0
 
+    should_gather_code_coverage = expansions.get("gather_code_coverage_results", False)
+    if not should_gather_code_coverage:
+        print("Missing 'gather_code_coverage_results' expansion, skipping code coverage.")
+        return 0
+
+    if not os.path.exists(".git"):
+        print("No git repo found in working directory. Code coverage needs git repo to function.")
+        return 1
+
+    repo = git.Repo()
+    head_sha = repo.head.object.hexsha
+    evergreen_revision = expansions.get("revision")
+    # We need all code coverage runs from the same patch to have the same commit on HEAD
+    # When different runs have different git SHAs they are grouped in different builds in coveralls.
+    #
+    # Because of the weird cloning we do to have a git repo present during code coverage tasks
+    # and the inconsistent behaviors of the local git state during evergreen patches and
+    # commit-queue runs. We need to always ensure that the HEAD matches the evergreen revision
+    # for consistency in coveralls.
+    if head_sha != evergreen_revision:
+        print("test not equal")
+        print(f"head_sha: {head_sha}")
+        print(f"evergreen revision: {evergreen_revision}")
+        repo.git.reset("--mixed", evergreen_revision)
+
     disallowed_arches = {"s390x", "s390", "ppc64le", "ppc64", "ppc", "ppcle"}
     arch = platform.uname().machine.lower()
     print(f"Detected arch: {arch}")
@@ -123,6 +149,8 @@ def main():
                 )
             args.append(bazel_coverage_report_location)
             retry_coveralls_report(args)
+            # no gcda files are generated from bazel coverage so we can exit early here
+            return 0
 
     scons_build_dir = os.path.join(workdir, "src", "build", "debug")
     if os.path.exists(scons_build_dir):