From d1d17255794d7193cdd0cf400cd5c0451ddc487f Mon Sep 17 00:00:00 2001
From: Fabian Meumertzheim <fabian@meumertzhe.im>
Date: Fri, 1 Apr 2022 17:32:08 +0200
Subject: [PATCH] Default to a UTF-8 locale in Java stub template

On non-macOS Unix, without any locale variable set, the OpenJDK defaults
to using ASCII rather than UTF-8 as the encoding for file system paths
(i.e., the value of the `sun.jnu.encoding` property).
---
 .../bazel/rules/java/java_stub_template.txt   |  9 +++++
 .../shell/bazel/unicode_filenames_test.sh     | 39 +++++++++++++++++++
 2 files changed, 48 insertions(+)

diff --git a/src/main/java/com/google/devtools/build/lib/bazel/rules/java/java_stub_template.txt b/src/main/java/com/google/devtools/build/lib/bazel/rules/java/java_stub_template.txt
index 2aabedcdf942f2..873f99e3c82dbd 100644
--- a/src/main/java/com/google/devtools/build/lib/bazel/rules/java/java_stub_template.txt
+++ b/src/main/java/com/google/devtools/build/lib/bazel/rules/java/java_stub_template.txt
@@ -362,6 +362,15 @@ if [ -z "$CLASSPATH_LIMIT" ]; then
   is_windows && CLASSPATH_LIMIT=7000 || CLASSPATH_LIMIT=120000
 fi
 
+# On non-macOS Unix, without any locale variable set, the OpenJDK defaults to
+# using ASCII rather than UTF-8 as the encoding for file system paths (i.e., the
+# value of the sun.jnu.encoding property).
+if ! is_macos && ! is_windows && [ -z "$LC_ALL" ] && [ -z "$LC_CTYPE" ] && [ -z "$LANG" ]; then
+  if locale -a | grep -q C.UTF-8; then
+    export LC_CTYPE=C.UTF-8
+  fi
+fi
+
 if (("${#CLASSPATH}" > ${CLASSPATH_LIMIT})); then
   export JACOCO_IS_JAR_WRAPPED=1
   create_and_run_classpath_jar
diff --git a/src/test/shell/bazel/unicode_filenames_test.sh b/src/test/shell/bazel/unicode_filenames_test.sh
index 7e38cca62e2bac..0bd0a331db9702 100755
--- a/src/test/shell/bazel/unicode_filenames_test.sh
+++ b/src/test/shell/bazel/unicode_filenames_test.sh
@@ -100,6 +100,11 @@ function has_utf8_locale() {
   [[ "${charmap}" == "UTF-8" ]]
 }
 
+function has_c_utf8_locale() {
+  charmap="$(LC_ALL=C.UTF-8 locale charmap 2>/dev/null)"
+  [[ "${charmap}" == "UTF-8" ]]
+}
+
 function test_utf8_source_artifact() {
   # Bazel relies on the JVM for filename encoding, and can only support
   # UTF-8 if either a UTF-8 or ISO-8859-1 locale is available.
@@ -206,4 +211,38 @@ function test_utf8_source_artifact_in_bep() {
   expect_log '"name":"pkg/srcs/ünïcödë fïlë.txt"'
 }
 
+function test_utf8_filename_in_java_test() {
+  if ! has_c_utf8_locale; then
+    echo "Skipping test (no C.UTF-8 locale)."
+    echo "Available locales (need C.UTF-8):"
+    locale -a
+    return
+  fi
+
+  touch WORKSPACE
+  mkdir pkg
+
+  cat >pkg/BUILD <<'EOF'
+java_test(
+    name = "Test",
+    srcs = ["Test.java"],
+    main_class = "Test",
+    use_testrunner = False,
+)
+EOF
+
+  cat >pkg/Test.java <<'EOF'
+import java.nio.file.Files;
+import java.io.IOException;
+
+class Test {
+    public static void main(String[] args) throws IOException {
+        Files.createTempFile("æøå", null);
+    }
+}
+EOF
+
+  bazel test //pkg:Test --test_output=errors 2>$TEST_log || fail "Test should pass"
+}
+
 run_suite "Tests for handling of Unicode filenames"