fix: diagnostics wrong position on unicode characters (#248)

server/src/services/validation/worker/build/solcCompile.ts

@@ -43,5 +43,28 @@ export async function solcCompile(
     });
   }
 
+  // Normalize errors' sourceLocation to use utf-8 offsets instead of byte offsets
+  for (const error of output.errors) {
+    const source = input.sources[error.sourceLocation?.file];
+
+    if (source === undefined) {
+      continue;
+    }
+
+    const normalizeOffset = (text: string, offset: number) => {
+      return Buffer.from(text, "utf-8").slice(0, offset).toString("utf-8")
+        .length;
+    };
+
+    error.sourceLocation.start = normalizeOffset(
+      source.content,
+      error.sourceLocation.start
+    );
+    error.sourceLocation.end = normalizeOffset(
+      source.content,
+      error.sourceLocation.end
+    );
+  }
+
   return { output, solcBuild };
 }

test/integration/tests/diagnostics/diagnostics.test.ts

@@ -44,4 +44,17 @@ suite("diagnostics", function () {
       'Contract "Counter" should be marked as abstract'
     );
   });
+
+  test.only("[diagnostics] unicode characters", async () => {
+    const uri = getDocUri(__dirname, "./UnicodeCharacters.sol");
+    await openFileInEditor(uri);
+
+    await checkOrWaitDiagnostic(
+      uri,
+      new vscode.Range(6, 4, 6, 14),
+      vscode.DiagnosticSeverity.Error,
+      "solidity",
+      "Different number of arguments in return statement"
+    );
+  });
 });

test/integration/tests/diagnostics/testdata/UnicodeCharacters.sol

@@ -0,0 +1,9 @@
+// SPDX-License-Identifier: MIT
+pragma solidity ^0.8.3;
+
+contract UnicodeCharacters {
+  //ñññññ
+  function bar() public {
+    return 123;
+  }
+}