Pin charsets to UTF-8 in two test sites uncovered by charset audit

Audit of byte/char boundaries across all production and test source
revealed two test-only sites with implicit charset dependencies:

  - ContentPartTest.imageFileRejectsUnknownExtension: replace
    Files.write(file, "hello".getBytes()) with Files.writeString(file,
    "hello"). The string is ASCII so byte output was identical
    everywhere today, but Files.writeString uses UTF-8 by contract
    (per JDK docs since Java 11) and the new form is shorter.

  - LlamaModelTest.completeAndReadStdOut: pin the PrintStream that
    captures System.out and the matching ByteArrayOutputStream.toString
    to StandardCharsets.UTF_8. Model output is multi-byte UTF-8 token
    text; on a non-UTF-8 default host (e.g. Windows CP-1252) the
    captured assertion strings would mangle non-ASCII tokens. Removes
    the stale @SuppressWarnings('ImplicitDefaultCharsetUsage').

Spotbugs count unchanged at 0. Tests pass cleanly through the
charset-pinned capture path.

Author: claude

src/test/java/net/ladenthin/llama/ContentPartTest.java

@@ -116,7 +116,7 @@ public void imageFileDetectsGif() throws IOException {
     @Test
     public void imageFileRejectsUnknownExtension() throws IOException {
         Path file = tmp.resolve("doc.txt");
-        Files.write(file, "hello".getBytes());
+        Files.writeString(file, "hello");
         try {
             ContentPart.imageFile(file);
             fail("expected IllegalArgumentException for unknown extension");

src/test/java/net/ladenthin/llama/LlamaModelTest.java

@@ -8,6 +8,7 @@
 import static org.junit.jupiter.api.Assertions.*;
 
 import java.io.*;
+import java.nio.charset.StandardCharsets;
 import java.util.*;
 import java.util.regex.Pattern;
 import net.ladenthin.llama.args.LogFormat;
@@ -608,8 +609,7 @@ public void testLogStdout() {
     private String completeAndReadStdOut() {
         PrintStream stdOut = System.out;
         ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
-        @SuppressWarnings("ImplicitDefaultCharsetUsage")
-        PrintStream printStream = new PrintStream(outputStream);
+        PrintStream printStream = new PrintStream(outputStream, false, StandardCharsets.UTF_8);
         System.setOut(printStream);
 
         try {
@@ -622,7 +622,7 @@ private String completeAndReadStdOut() {
             printStream.close();
         }
 
-        return outputStream.toString();
+        return outputStream.toString(StandardCharsets.UTF_8);
     }
 
     private List<String> splitLines(String text) {