Some incremental steps on overhauling IO.

scala.io.Codec for encapsulating charset encoding/decoding info, and
then uses that class as the world's first default/implicit parameter:
see BufferedSource.fromInputStream.

4 files changed, 94 insertions, 72 deletions

diff --git a/src/dotnet-library/scala/io/Codec.scala b/src/dotnet-library/scala/io/Codec.scala
new file mode 100644
index 0000000000..e291d7a90d
--- /dev/null
+++ b/src/dotnet-library/scala/io/Codec.scala
@@ -0,0 +1 @@
+/* Codec.scala does not exist for the dotnet target */
\ No newline at end of file
diff --git a/src/library/scala/io/BufferedSource.scala b/src/library/scala/io/BufferedSource.scala
index 5926e9bd53..b6543688f8 100644
--- a/src/library/scala/io/BufferedSource.scala
+++ b/src/library/scala/io/BufferedSource.scala
@@ -11,29 +11,34 @@
 
 package scala.io
 
-import java.io.InputStream
-import java.nio.{ByteBuffer, CharBuffer}
-import java.nio.channels.{ByteChannel, Channels, ReadableByteChannel}
-import java.nio.charset.{Charset, CharsetDecoder}
+import java.io.{ InputStream, Reader, BufferedReader, InputStreamReader, IOException }
+import java.nio.charset.{ Charset, CharsetDecoder, CodingErrorAction, CharacterCodingException, MalformedInputException }
+import java.nio.channels.Channels
+import Source._
 
-object BufferedSource {
-
-  /** same as fromInputStream(inpStream, Charset.forName(enc), buffer_size, do_reset) */
-  def fromInputStream(inpStream: InputStream, enc: String, buffer_size: Int, do_reset: () => Source): BufferedSource =
-    fromInputStream(inpStream, Charset.forName(enc), buffer_size, do_reset)
-
-  /** same as fromInputStream(inpStream, charSet.newDecoder(), buffer_size, do_reset) */
-  def fromInputStream(inpStream: InputStream, charSet: Charset, buffer_size: Int, do_reset: () => Source): BufferedSource =
-    fromInputStream(inpStream, charSet.newDecoder(), buffer_size, do_reset)
-
-  /** constructs a BufferedSource instance from an input stream, using given decoder */
-  def fromInputStream(inpStream: InputStream, decoder: CharsetDecoder, buffer_size: Int, do_reset: () => Source): BufferedSource = {
-    val byteChannel = Channels.newChannel(inpStream)
-    return new {
-      val buf_size = buffer_size
-     } with BufferedSource(byteChannel, decoder) {
-      override def reset = do_reset()
-      def close { inpStream.close }
+object BufferedSource
+{
+  /** Reads data from <code>inputStream</code> with a buffered reader,
+   *  using encoding in implicit parameter <code>codec</code>.
+   *
+   *  @param  inputStream  the input stream from which to read
+   *  @param  bufferSize   buffer size (defaults to BufferedSource.DefaultBufSize)
+   *  @param  reset        a () => Source which resets the stream (defaults to Source.NoReset)
+   *  @param  codec        (implicit) a scala.io.Codec specifying behavior (defaults to BufferedSource.defaultCodec)
+   *  @return              the buffered source
+   */
+  def fromInputStream(
+    inputStream: InputStream,
+    bufferSize: Int = DefaultBufSize,
+    reset: () => Source = null
+  )(implicit codec: Codec = defaultCodec) =
+  {
+    if (reset == null) new BufferedSource(inputStream, bufferSize, codec)
+    else {
+      def _reset = reset
+      new BufferedSource(inputStream, bufferSize, codec) {
+        override def reset = _reset()
+      }
     }
   }
 }
@@ -44,49 +49,31 @@ object BufferedSource {
  *  @author  Burak Emir
  *  @version 1.0, 19/08/2004
  */
-abstract class BufferedSource(byteChannel: ReadableByteChannel, decoder: CharsetDecoder) extends Source {
-
-  val buf_size: Int
-
-  def close: Unit
-
-  val byteBuffer = ByteBuffer.allocate(buf_size)
-  var charBuffer = CharBuffer.allocate(buf_size)
-  byteBuffer.position(byteBuffer.limit())
-  charBuffer.position(charBuffer.limit())
-  decoder.reset()
-  var endOfInput = false
+class BufferedSource(
+  inputStream: InputStream,
+  bufferSize: Int,
+  codec: Codec)
+extends Source
+{
+  val decoder = codec.decoder
+  decoder.reset
+  decoder onMalformedInput codec.malformedAction
+  val reader = new BufferedReader(new InputStreamReader(inputStream, decoder), bufferSize)
 
-  def fillBuffer() = {
-    byteBuffer.compact()
-    charBuffer.compact()
-    var num_bytes = byteChannel.read(byteBuffer)
-    while (0 == num_bytes) {
-      Thread.sleep(1);  // wait 1 ms for new data
-      num_bytes = byteChannel.read(byteBuffer)
-    }
-    endOfInput = (num_bytes == -1)
-    byteBuffer.flip()
-    decoder.decode(byteBuffer, charBuffer, endOfInput)
-    if (endOfInput) decoder.flush(charBuffer)
-    charBuffer.flip()
-  }
   override val iter = new Iterator[Char] {
-    var buf_char = {
-      fillBuffer()
-      if (endOfInput) ' ' else charBuffer.get()
-    }
-    def hasNext = { charBuffer.remaining() > 0 || !endOfInput}
+    private def getc(): Char =
+      try     { reader.read().toChar }
+      catch   { case e: CharacterCodingException => codec receivedMalformedInput e }
+
+    var buf_char = getc
+    def hasNext = reader.ready()
     def next = {
       val c = buf_char
-      if (charBuffer.remaining() == 0) {
-        fillBuffer()
-      }
-      if (!endOfInput) {
-        buf_char = charBuffer.get()
-      }
+      if (hasNext) buf_char = getc
       c
     }
   }
+  def close: Unit     = reader.close
+  def reset(): Source = NoReset()
 }
 
diff --git a/src/library/scala/io/Codec.scala b/src/library/scala/io/Codec.scala
new file mode 100644
index 0000000000..9ecf3f0b1b
--- /dev/null
+++ b/src/library/scala/io/Codec.scala
@@ -0,0 +1,38 @@
+/*                     __                                               *\
+**     ________ ___   / /  ___     Scala API                            **
+**    / __/ __// _ | / /  / _ |    (c) 2003-2009, LAMP/EPFL             **
+**  __\ \/ /__/ __ |/ /__/ __ |    http://scala-lang.org/               **
+** /____/\___/_/ |_/____/_/ | |                                         **
+**                          |/                                          **
+\*                                                                      */
+
+// $Id$
+
+package scala.io
+
+import java.nio.charset.{ Charset, CharsetDecoder, CodingErrorAction }
+
+/** A class for character encoding/decoding preferences.
+ *
+ */
+class Codec(charSet: Charset) {
+  def decoder = charSet.newDecoder()
+
+  // by default we replace bad chars with the decoder's replacement value (e.g. "?")
+  // this behavior can be altered by overriding these two methods
+  def malformedAction(): CodingErrorAction = CodingErrorAction.IGNORE
+  def receivedMalformedInput(e: Exception): Char = decoder.replacement()(0)
+}
+
+object Codec {
+  def apply(encoding: String): Codec        = new Codec(Charset forName encoding)
+  def apply(charSet: Charset): Codec        = new Codec(charSet)
+  def apply(decoder: CharsetDecoder): Codec = {
+    val _decoder = decoder
+    new Codec(decoder.charset()) { override def decoder = _decoder }
+  }
+
+  implicit def string2codec(s: String) = apply(s)
+  implicit def charset2codec(c: Charset) = apply(c)
+  implicit def decoder2codec(cd: CharsetDecoder) = apply(cd)
+}
\ No newline at end of file
diff --git a/src/library/scala/io/Source.scala b/src/library/scala/io/Source.scala
index b79de507f6..756cb66e8b 100644
--- a/src/library/scala/io/Source.scala
+++ b/src/library/scala/io/Source.scala
@@ -24,12 +24,13 @@ import java.nio.charset.Charset
  *  @author  Burak Emir
  *  @version 1.0, 19/08/2004
  */
-object Source {
-
+object Source
+{
   val DefaultBufSize = 2048
-
   val NoReset: () => Source = () => throw new UnsupportedOperationException()
 
+  def defaultCodec = Codec(util.Properties.encodingString)
+
   /** Creates a <code>Source</code> instance from the given array of bytes,
    *  with empty description.
    *
@@ -129,7 +130,7 @@ object Source {
     val inpStream = new FileInputStream(file)
     val size = if (bufferSize > 0) bufferSize else Source.DefaultBufSize
     setFileDescriptor(file,
-      BufferedSource.fromInputStream(inpStream, enc, size, { () => fromFile(file, enc, size)}))
+      BufferedSource.fromInputStream(inpStream, size, () => fromFile(file, enc, size))(Codec(enc)))
   }
 
   /** This method sets the descr property of the given source to a string of the form "file:"+path
@@ -195,7 +196,7 @@ object Source {
   @deprecated
   def fromInputStream(istream: InputStream, enc: String, maxlen: Option[Int]): Source = {
     val limit = maxlen match { case Some(i) => i; case None => 0 }
-    val bi = new BufferedInputStream(istream, Source.DefaultBufSize)
+    val bi = new BufferedInputStream(istream, DefaultBufSize)
     val bytes = new collection.mutable.ArrayBuffer[Byte]()
     var b = 0
     var i = 0
@@ -206,15 +207,10 @@ object Source {
     fromBytes(bytes.toArray, enc)
   }
 
-  /** same as BufferedSource.fromInputStream(is, enc, Source.DefaultBufSize)
+  /** same as BufferedSource.fromInputStream(is)
    */
-  def fromInputStream(is: InputStream, enc: String): Source =
-    BufferedSource.fromInputStream(is, enc, Source.DefaultBufSize, { () => fromInputStream(is, enc) })
-
-  /** same as BufferedSource.fromInputStream(is, "utf-8", Source.DefaultBufSize) */
-  def fromInputStream(is: InputStream): Source =
-    BufferedSource.fromInputStream(is, "utf-8", Source.DefaultBufSize, { () => fromInputStream(is) })
-
+  def fromInputStream(is: InputStream, codec: Codec = defaultCodec): Source =
+    BufferedSource.fromInputStream(is, DefaultBufSize, () => fromInputStream(is, codec))(codec)
 }
 
 /** The class <code>Source</code> implements an iterable representation