fix(builder): strip UTF-8 BOM from .ino sources before preprocessing (#2983)

* fix(builder): strip UTF-8 BOM from .ino sources before preprocessing

When a sketch .ino is saved as UTF-8 *with BOM*, the BOM bytes (EF BB BF)
reach the compiler and cause:
  stray '\357' in program
  stray '\273' in program
  stray '\277' in program

This strips the BOM at read-time so the merged .cpp and copied sources are clean.

Refs: arduino/arduino-ide#2752

* test: add unit tests for stripUTF8BOM

Author: ritesh006

internal/arduino/builder/sketch.go

@@ -37,6 +37,14 @@ var (
 	includesArduinoH = regexp.MustCompile(`(?m)^\s*#\s*include\s*[<\"]Arduino\.h[>\"]`)
 )
 
+// stripUTF8BOM removes a UTF-8 BOM if present at the start of the file.
+func stripUTF8BOM(b []byte) []byte {
+	if len(b) >= 3 && b[0] == 0xEF && b[1] == 0xBB && b[2] == 0xBF {
+		return b[3:]
+	}
+	return b
+}
+
 // prepareSketchBuildPath copies the sketch source files in the build path.
 // The .ino files are merged together to create a .cpp file (by the way, the
 // .cpp file still needs to be Arduino-preprocessed to compile).
@@ -82,6 +90,7 @@ func (b *Builder) sketchMergeSources(overrides map[string]string) (int, string,
 		if err != nil {
 			return "", errors.New(i18n.Tr("reading file %[1]s: %[2]s", f, err))
 		}
+		data = stripUTF8BOM(data)
 		return string(data), nil
 	}
 
@@ -136,7 +145,9 @@ func (b *Builder) sketchCopyAdditionalFiles(buildPath *paths.Path, overrides map
 			if err != nil {
 				return fmt.Errorf("%s: %w", i18n.Tr("unable to read contents of the source item"), err)
 			}
+			s = stripUTF8BOM(s)
 			sourceBytes = s
+
 		}
 
 		// tag each additional file with the filename of the source it was copied from

internal/arduino/builder/sketch_test.go

@@ -106,3 +106,22 @@ func TestCopyAdditionalFiles(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, info1.ModTime(), info2.ModTime())
 }
+
+func TestStripUTF8BOM(t *testing.T) {
+	// Case 1: Input with BOM
+	inputWithBOM := []byte{0xEF, 0xBB, 0xBF, 'H', 'e', 'l', 'l', 'o'}
+	expected := []byte("Hello")
+
+	output := stripUTF8BOM(inputWithBOM)
+	require.Equal(t, expected, output, "BOM should be stripped")
+
+	// Case 2: Input without BOM
+	inputNoBOM := []byte("Hello")
+	outputNoBOM := stripUTF8BOM(inputNoBOM)
+	require.Equal(t, inputNoBOM, outputNoBOM, "Input without BOM should remain unchanged")
+
+	// Case 3: Input shorter than 3 bytes (edge case)
+	shortInput := []byte{0xEF}
+	outputShort := stripUTF8BOM(shortInput)
+	require.Equal(t, shortInput, outputShort, "Short input should remain unchanged")
+}