Reading multibyte characters leads to corrupt FileStore and shows the wrong diagnostic. Using the built-in method String#readNBytes helps to prevent it. (georgewfraser#221)

Author: trett

src/main/java/org/javacs/debug/proto/DebugAdapter.java

@@ -52,21 +52,11 @@ private static char read(InputStream client) {
 
     private static String readLength(InputStream client, int byteLength) {
         // Eat whitespace
-        // Have observed problems with extra \r\n sequences from VSCode
-        var next = read(client);
-        while (Character.isWhitespace(next)) {
-            next = read(client);
-        }
-        // Append next
-        var result = new StringBuilder();
-        var i = 0;
-        while (true) {
-            result.append(next);
-            i++;
-            if (i == byteLength) break;
-            next = read(client);
+        try {
+            return new String(client.readNBytes(byteLength), StandardCharsets.UTF_8).stripLeading();
+        } catch (IOException e) {
+            throw new RuntimeException("An error occurred during the reading of client data", e);
         }
-        return result.toString();
     }
 
     static String nextToken(InputStream client) {

src/main/java/org/javacs/lsp/LSP.java

@@ -58,21 +58,11 @@ private static char read(InputStream client) {
 
     private static String readLength(InputStream client, int byteLength) {
         // Eat whitespace
-        // Have observed problems with extra \r\n sequences from VSCode
-        var next = read(client);
-        while (Character.isWhitespace(next)) {
-            next = read(client);
-        }
-        // Append next
-        var result = new StringBuilder();
-        var i = 0;
-        while (true) {
-            result.append(next);
-            i++;
-            if (i == byteLength) break;
-            next = read(client);
+        try {
+            return new String(client.readNBytes(byteLength), StandardCharsets.UTF_8).stripLeading();
+        } catch (IOException e) {
+            throw new RuntimeException("An error occurred during the reading of client data", e);
         }
-        return result.toString();
     }
 
     static String nextToken(InputStream client) {

src/test/java/org/javacs/lsp/LspTest.java

@@ -3,6 +3,7 @@
 import static org.hamcrest.Matchers.*;
 import static org.junit.Assert.*;
 
+import com.google.gson.Gson;
 import com.google.gson.JsonObject;
 import java.io.IOException;
 import java.io.PipedInputStream;
@@ -14,6 +15,8 @@
 import org.junit.Test;
 
 public class LspTest {
+
+    private static final Gson gson = new Gson();
     PipedInputStream buffer = new PipedInputStream(10 * 1024 * 1024); // 10 MB buffer
     PipedOutputStream writer = new PipedOutputStream();
 
@@ -93,6 +96,27 @@ public void readMessage() throws IOException {
         assertThat(parse.params, equalTo(new JsonObject()));
     }
 
+    @Test
+    public void readMessageWithMultiBytesCharacters() throws IOException {
+        var params = new ShowMessageParams();
+        params.message = "🔥";
+        var message =
+                String.format(
+                        "{\"jsonrpc\":\"2.0\",\"id\":1,\"method\":\"initialize\",\"params\": %s}", gson.toJson(params));
+        var header = String.format("Content-Length: %d\r\n\r\n", message.getBytes().length);
+        writer.write(header.getBytes());
+        writer.write(message.getBytes());
+
+        var token = LSP.nextToken(buffer);
+        assertThat(token, equalTo(message));
+
+        var parse = LSP.parseMessage(token);
+        assertThat(parse.jsonrpc, equalTo("2.0"));
+        assertThat(parse.id, equalTo(1));
+        assertThat(parse.method, equalTo("initialize"));
+        assertThat(parse.params, equalTo(gson.toJsonTree(params)));
+    }
+
     @Test
     public void excludeDefaults() {
         var item = new CompletionItem();