Skip UTF-8 BOM bytes from Dockerignore if exist

This fix tries to address issues related to #23221 where Dockerignore
may consists of UTF-8 BOM. This likely happens when Notepad
tries to save a file as UTF-8 in Windows.

This fix skips the UTF-8 BOM bytes from the beginning of the
Dockerignore if exists.

Additional tests has been added to cover the changes in this fix.

This fix is related to #23221 (UTF-8 BOM in Dockerfile).

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>

builder/dockerignore/dockerignore.go

@@ -2,6 +2,7 @@ package dockerignore
 
 import (
 	"bufio"
+	"bytes"
 	"fmt"
 	"io"
 	"path/filepath"
@@ -18,10 +19,18 @@ func ReadAll(reader io.ReadCloser) ([]string, error) {
 	defer reader.Close()
 	scanner := bufio.NewScanner(reader)
 	var excludes []string
+	currentLine := 0
 
+	utf8bom := []byte{0xEF, 0xBB, 0xBF}
 	for scanner.Scan() {
+		scannedBytes := scanner.Bytes()
+		// We trim UTF8 BOM
+		if currentLine == 0 {
+			scannedBytes = bytes.TrimPrefix(scannedBytes, utf8bom)
+		}
+		pattern := string(scannedBytes)
+		currentLine++
 		// Lines starting with # (comments) are ignored before processing
-		pattern := scanner.Text()
 		if strings.HasPrefix(pattern, "#") {
 			continue
 		}