#107: Simplify UnicodeDecoder.decode()

pathikrit · Feb 14, 2017 · 14276ac · 14276ac
1 parent 4307856
commit 14276ac
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 25 deletions.
diff --git a/core/src/main/scala/better/files/UnicodeDecoder.scala b/core/src/main/scala/better/files/UnicodeDecoder.scala
@@ -14,42 +14,33 @@ import java.nio.{ByteBuffer, CharBuffer}
 class UnicodeDecoder(defaultCharset: Charset) extends CharsetDecoder(null, 1, 1) {
   import UnicodeDecoder._
 
-  private var inferredCharset: Option[Charset] = None
+  private[this] var inferredCharset: Option[Charset] = None
 
   override def decodeLoop(in: ByteBuffer, out: CharBuffer) =
-    decode(
-      in = in,
-      out = out,
-      candidates = bomTable.keys.toList,
-      header = ByteBuffer.allocate(4).mark().asInstanceOf[ByteBuffer]
-    )
+    decode(in = in, out = out, candidates = bomTable.keys.toList)
 
   @annotation.tailrec
-  private def decode(in: ByteBuffer, out: CharBuffer, candidates: List[Charset], header: ByteBuffer): CoderResult = {
+  private[this] def decode(in: ByteBuffer, out: CharBuffer, candidates: List[Charset]): CoderResult = {
     if (isCharsetDetected) {
-      val decoder = detectedCharset().newDecoder()
-      val hasMore = in.remaining() > 0
-      if (header.remaining() > 0) decoder.decode(header, out, !hasMore)
-      decoder.decode(in, out, true)
+      detectedCharset().newDecoder().decode(in, out, true)
     } else if (candidates.isEmpty || in.remaining() <= 0) {
       inferredCharset = Some(defaultCharset)
-      header.reset()
-      decode(in, out, candidates, header)
+      in.position(0)
+      decode(in, out, Nil)
     } else {
+      val idx = in.position()
       val byte = in.get()
-      val idx = header.position()
-      header.put(byte)
       val newCandidates = candidates filter {charset =>
         val bom = bomTable(charset)
         bom.isDefinedAt(idx) && bom(idx) == byte
       }
       newCandidates match {
         case charset :: Nil if bomTable(charset).length == idx + 1 =>
           inferredCharset = Some(charset)
-          header.limit(idx)
+          in.position(idx + 1)
         case _ =>
       }
-      decode(in, out, newCandidates, header)
+      decode(in, out, newCandidates)
     }
   }
 

diff --git a/core/src/test/scala/better/files/FileSpec.scala b/core/src/test/scala/better/files/FileSpec.scala
@@ -158,6 +158,13 @@ class FileSpec extends FlatSpec with BeforeAndAfterEach with Matchers {
       .lines.toSeq should contain theSameElementsInOrderAs Seq("", "My name is", "Inigo Montoya", "x", "1")
   }
 
+  it should "handle BOM" in {
+    val file = File.resource("file_with_bom.txt")
+    val expectedContent = "I contain an offending UTF-8 BOM\n"
+    file.contentAsString should not equal expectedContent
+    file.contentAsString(charset = UnicodeDecoder("UTF-8")) shouldEqual expectedContent
+  }
+
   it should "glob" in {
     a1.glob("**/*.txt").map(_.name).toSeq.sorted shouldEqual Seq("t1.txt", "t2.txt")
     //a1.glob("*.txt").map(_.name).toSeq shouldEqual Seq("t1.txt", "t2.txt")
@@ -495,13 +502,6 @@ class FileSpec extends FlatSpec with BeforeAndAfterEach with Matchers {
     Seq.fill(2)(scanner.next[Animal]) should contain theSameElementsInOrderAs Seq(Cat("Garfield"), Dog("Woofer"))
   }
 
-  it should "handle BOM" in {
-    val file = File.resource("file_with_bom.txt")
-    val expectedContent = "I contain an offending UTF-8 BOM\n"
-    file.contentAsString should not equal expectedContent
-    file.contentAsString(charset = UnicodeDecoder("UTF-8")) shouldEqual expectedContent
-  }
-
   "file watcher" should "watch single files" in {
     assume(isCI)
     val file = File.newTemporaryFile(suffix = ".txt").writeText("Hello world")