SI-7356 - Source.mkString performs painfully slow (...)

1. Wrote a custom mkString for BufferedSource.

2. Moved the logic for rescuing the iterator-buffered char out of
BufferedLineIterator and into a private method in BufferedSource.

Speed test based on the one in the issue tracker (but written correctly)
indicates that performance is equal or better to getLines.  This resolves
SI-7356 in a minimal fashion.

1 files changed, 32 insertions, 14 deletions

diff --git a/src/library/scala/io/BufferedSource.scala b/src/library/scala/io/BufferedSource.scala
index c170d28127..832c7b23f9 100644
--- a/src/library/scala/io/BufferedSource.scala
+++ b/src/library/scala/io/BufferedSource.scala
@@ -8,9 +8,11 @@
 
 package scala.io
 
+import java.util.Arrays
 import java.io.{ InputStream, BufferedReader, InputStreamReader, PushbackReader }
 import Source.DefaultBufSize
 import scala.collection.{ Iterator, AbstractIterator }
+import scala.collection.mutable.ArrayBuffer
 
 /** This object provides convenience methods to create an iterable
  *  representation of a source file.
@@ -39,8 +41,8 @@ class BufferedSource(inputStream: InputStream, bufferSize: Int)(implicit val cod
     takeWhile (_ != -1)
     map (_.toChar)
   )
-
-  class BufferedLineIterator extends AbstractIterator[String] with Iterator[String] {
+  
+  private def decachedReader: BufferedReader = {
     // Don't want to lose a buffered char sitting in iter either. Yes,
     // this is ridiculous, but if I can't get rid of Source, and all the
     // Iterator bits are designed into Source, and people create Sources
@@ -48,18 +50,21 @@ class BufferedSource(inputStream: InputStream, bufferSize: Int)(implicit val cod
     // that calls hasNext to find out if they're empty, and that leads
     // to chars being buffered, and no, I don't work here, they left a
     // door unlocked.
-    private val lineReader: BufferedReader = {
-      // To avoid inflicting this silliness indiscriminately, we can
-      // skip it if the char reader was never created: and almost always
-      // it will not have been created, since getLines will be called
-      // immediately on the source.
-      if (charReaderCreated && iter.hasNext) {
-        val pb = new PushbackReader(charReader)
-        pb unread iter.next().toInt
-        new BufferedReader(pb, bufferSize)
-      }
-      else charReader
+    // To avoid inflicting this silliness indiscriminately, we can
+    // skip it if the char reader was never created: and almost always
+    // it will not have been created, since getLines will be called
+    // immediately on the source.
+    if (charReaderCreated && iter.hasNext) {
+      val pb = new PushbackReader(charReader)
+      pb unread iter.next().toInt
+      new BufferedReader(pb, bufferSize)
     }
+    else charReader
+  }
+    
+
+  class BufferedLineIterator extends AbstractIterator[String] with Iterator[String] {
+    private val lineReader = decachedReader
     var nextLine: String = null
 
     override def hasNext = {
@@ -79,5 +84,18 @@ class BufferedSource(inputStream: InputStream, bufferSize: Int)(implicit val cod
   }
 
   override def getLines(): Iterator[String] = new BufferedLineIterator
+  
+  /** Efficiently converts the entire remaining input into a string. */
+  override def mkString = {
+    // Speed up slurping of whole data set in the simplest cases.
+    val allReader = decachedReader
+    val sb = new StringBuilder
+    val buf = new Array[Char](bufferSize)
+    var n = 0
+    while (n != -1) {
+      n = charReader.read(buf)
+      if (n>0) sb.appendAll(buf, 0, n)
+    }
+    sb.result
+  }
 }
-