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

ritesh006 · ritesh006 · commit b36b87ae9bce · 2025-08-24T11:50:27.000+05:30
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
diff --git a/internal/arduino/builder/sketch.go b/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

Original file line number	Diff line number	Diff line change
`@@ -37,6 +37,14 @@ var (`
`37`	`37`	includesArduinoH = regexp.MustCompile(`(?m)^\s#\sinclude\s*[<\"]Arduino\.h[>\"]`)
`38`	`38`	`)`
`39`	`39`
	`40`	`+// stripUTF8BOM removes a UTF-8 BOM if present at the start of the file.`
	`41`	`+func stripUTF8BOM(b []byte) []byte {`
	`42`	`+ if len(b) >= 3 && b[0] == 0xEF && b[1] == 0xBB && b[2] == 0xBF {`
	`43`	`+ return b[3:]`
	`44`	`+ }`
	`45`	`+ return b`
	`46`	`+}`
	`47`	`+`
`40`	`48`	`// prepareSketchBuildPath copies the sketch source files in the build path.`
`41`	`49`	`// The .ino files are merged together to create a .cpp file (by the way, the`
`42`	`50`	`// .cpp file still needs to be Arduino-preprocessed to compile).`
`@@ -82,6 +90,7 @@ func (b *Builder) sketchMergeSources(overrides map[string]string) (int, string,`
`82`	`90`	`if err != nil {`
`83`	`91`	`return "", errors.New(i18n.Tr("reading file %[1]s: %[2]s", f, err))`
`84`	`92`	`}`
	`93`	`+ data = stripUTF8BOM(data)`
`85`	`94`	`return string(data), nil`
`86`	`95`	`}`
`87`	`96`
`@@ -136,7 +145,9 @@ func (b Builder) sketchCopyAdditionalFiles(buildPath paths.Path, overrides map`
`136`	`145`	`if err != nil {`
`137`	`146`	`return fmt.Errorf("%s: %w", i18n.Tr("unable to read contents of the source item"), err)`
`138`	`147`	`}`
	`148`	`+ s = stripUTF8BOM(s)`
`139`	`149`	`sourceBytes = s`
	`150`	`+`
`140`	`151`	`}`
`141`	`152`
`142`	`153`	`// tag each additional file with the filename of the source it was copied from`