Merge pull request #206 from ashawley/utility-escape-test

Add tests for Utility.escape

Author: ashawley

shared/src/test/scala/scala/xml/UtilityTest.scala

@@ -3,6 +3,7 @@ package scala.xml
 import org.junit.Test
 import org.junit.Assert.assertTrue
 import org.junit.Assert.assertEquals
+import org.junit.Assert.assertNotEquals
 
 class UtilityTest {
 
@@ -63,4 +64,133 @@ class UtilityTest {
     assertEquals("<node><!-- comment  --></node>", Utility.serialize(x, stripComments = false).toString)
   }
 
+  val printableAscii: Seq[Char] = {
+    (' ' to '/') ++ // Punctuation
+    ('0' to '9') ++ // Digits
+    (':' to '@') ++ // Punctuation (cont.)
+    ('A' to 'Z') ++ // Uppercase
+    ('[' to '`') ++ // Punctuation (cont.)
+    ('a' to 'z') ++ // Lowercase
+    ('{' to '~')    // Punctuation (cont.)
+  }
+
+  val escapedChars: Seq[Char] =
+    Utility.Escapes.escMap.keys.toSeq
+
+  @Test
+  def escapePrintablesTest: Unit = {
+    for {
+      char <- (printableAscii.diff(escapedChars))
+    } yield {
+      assertEquals(char.toString, Utility.escape(char.toString))
+    }
+  }
+
+  @Test
+  def escapeEscapablesTest: Unit = {
+    for {
+      char <- escapedChars
+    } yield {
+      assertNotEquals(char.toString, Utility.escape(char.toString))
+    }
+  }
+
+  @Test
+  def escapeAsciiControlCharsTest: Unit = {
+
+    /* Escapes that Scala (Java) doesn't support.
+     * \u0007 -> \a  (bell)
+     * \u001B -> \e  (escape)
+     * \u000B -> \v  (vertical tab)
+     * \u007F -> DEL (delete)
+     */
+    val input = " \u0007\b\u001B\f\n\r\t\u000B\u007F"
+
+    val expect = " \n\r\t\u007F"
+
+    val result = Utility.escape(input)
+
+    assertEquals(printfc(expect), printfc(result)) // Pretty,
+    assertEquals(expect, result)                   // but verify.
+  }
+
+  @Test
+  def escapeUnicodeExtendedControlCodesTest: Unit = {
+    for {
+      char <- ('\u0080' to '\u009f') // Extended control codes (C1)
+    } yield {
+      assertEquals(char.toString, Utility.escape(char.toString))
+    }
+  }
+
+  @Test
+  def escapeUnicodeTwoBytesTest: Unit = {
+    for {
+      char <- ('\u00A0' to '\u07FF') // Two bytes (cont.)
+    } yield {
+      assertEquals(char.toString, Utility.escape(char.toString))
+    }
+  }
+
+  @Test
+  def escapeUnicodeThreeBytesTest: Unit = {
+    for {
+      char <- ('\u0800' to '\uFFFF') // Three bytes
+    } yield {
+      assertEquals(char.toString, Utility.escape(char.toString))
+    }
+  }
+
+  /**
+   * Human-readable character printing
+   * 
+   * Think of `od -c` of unix od(1) command.
+   * 
+   * Or think of `printf("%c", i)` in C, but a little better.
+   */
+  def printfc(str: String) = {
+    str.map(prettyChar).mkString
+  }
+
+  /**
+   * Visual representation of characters that enhances output of
+   * failed test assertions.
+   */
+  val prettyChar: Map[Char,String] = Map(
+    '\u0000' -> "\\0", // Null
+    '\u0001' -> "^A",  // Start of header
+    '\u0002' -> "^B",  // Start of text
+    '\u0003' -> "^C",  // End of text
+    '\u0004' -> "^D",  // End of transmission
+    '\u0005' -> "^E",  // Enquiry
+    '\u0006' -> "^F",  // Acknowledgment
+    '\u0007' -> "\\a", // Bell (^G)
+    '\b'     -> "\\b", // Backspace (^H)
+    '\t'     -> "\\t", // Tab (^I)
+    '\n'     -> "\\n", // Newline (^J)
+    '\u000B' -> "\\v", // Vertical tab (^K)
+    '\f'     -> "\\f", // Form feed (^L)
+    '\r'     -> "\\r", // Carriage return (^M)
+    '\u000E' -> "^N",  // Shift out
+    '\u000F' -> "^O",  // Shift in
+    '\u0010' -> "^P",  // Data link escape
+    '\u0011' -> "^Q",  // DC1 (XON)
+    '\u0012' -> "^R",  // DC2
+    '\u0013' -> "^S",  // DC3 (XOFF)
+    '\u0014' -> "^T",  // DC4
+    '\u0015' -> "^U",  // Negative acknowledgment
+    '\u0016' -> "^V",  // Synchronous idle
+    '\u0017' -> "^W",  // End of transmission block
+    '\u0018' -> "^X",  // Cancel
+    '\u0019' -> "^Y",  // End of medium
+    '\u001A' -> "^Z",  // Substitute
+    '\u001B' -> "\\e", // Escape
+    '\u001C' -> "^\\", // File separator
+    '\u001D' -> "^]",  // Group separator
+    '\u001E' -> "^^",  // Record separator
+    '\u001F' -> "^_",  // Unit separator
+    '\u007F' -> "^?"   // Delete
+  ).toMap.withDefault {
+    key: Char => key.toString
+  }
 }