Make sure Bazel patches files with line ending '\n' on all platforms

This is to ensure all Bazel patched files have the exact same content and hash value. For example, for patched bzl files, their hash could end up in Bzlmod lockfile. Without this change, the lockfile will be different on Windows due to line ending being `\r\n`.

RELNOTES:
PiperOrigin-RevId: 564504118
Change-Id: Ia61493dca0c034341279d0d5e613354d4c6b9e19

src/main/java/com/google/devtools/build/lib/bazel/repository/PatchUtil.java

@@ -248,7 +248,7 @@ private static ImmutableList<String> readFile(Path file) throws IOException {
   }
 
   private static void writeFile(Path file, List<String> content) throws IOException {
-    FileSystemUtils.writeLinesAs(file, UTF_8, content);
+    FileSystemUtils.writeLinesAs(file, UTF_8, content, "\n");
   }
 
   private static boolean getReadPermission(int permission) {

src/main/java/com/google/devtools/build/lib/vfs/FileSystemUtils.java

@@ -708,17 +708,17 @@ public static void writeContent(Path outputFile, byte[] content) throws IOExcept
   }
 
   /**
-   * Writes lines to file using the given encoding, ending every line with a line break '\n'
-   * character.
+   * Writes lines to file using the given encoding, ending every line with a system specific line
+   * break character.
    */
   @ThreadSafe // but not atomic
   public static void writeLinesAs(Path file, Charset charset, String... lines) throws IOException {
     writeLinesAs(file, charset, Arrays.asList(lines));
   }
 
   /**
-   * Writes lines to file using the given encoding, ending every line with a line break '\n'
-   * character.
+   * Writes lines to file using the given encoding, ending every line with a system specific line
+   * break character.
    */
   @ThreadSafe // but not atomic
   public static void writeLinesAs(Path file, Charset charset, Iterable<String> lines)
@@ -728,17 +728,28 @@ public static void writeLinesAs(Path file, Charset charset, Iterable<String> lin
   }
 
   /**
-   * Appends lines to file using the given encoding, ending every line with a line break '\n'
+   * Writes lines to file using the given encoding, ending every line with a given line break
    * character.
    */
   @ThreadSafe // but not atomic
+  public static void writeLinesAs(
+      Path file, Charset charset, Iterable<String> lines, String lineBreak) throws IOException {
+    file.getParentDirectory().createDirectoryAndParents();
+    asByteSink(file).asCharSink(charset).writeLines(lines, lineBreak);
+  }
+
+  /**
+   * Appends lines to file using the given encoding, ending every line with a system specific line
+   * break character.
+   */
+  @ThreadSafe // but not atomic
   public static void appendLinesAs(Path file, Charset charset, String... lines) throws IOException {
     appendLinesAs(file, charset, Arrays.asList(lines));
   }
 
   /**
-   * Appends lines to file using the given encoding, ending every line with a line break '\n'
-   * character.
+   * Appends lines to file using the given encoding, ending every line with a system specific line
+   * break character.
    */
   @ThreadSafe // but not atomic
   public static void appendLinesAs(Path file, Charset charset, Iterable<String> lines)

src/test/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryContextTest.java

@@ -317,7 +317,7 @@ public void testPatch() throws Exception {
     context.createFile(
         context.path("my.patch"), "--- foo\n+++ foo\n" + ONE_LINE_PATCH, false, true, thread);
     context.patch(patchFile, StarlarkInt.of(0), thread);
-    testOutputFile(foo.getPath(), String.format("line one%nline two%n"));
+    testOutputFile(foo.getPath(), "line one\nline two\n");
   }
 
   @Test