Adapted indentation in scaladoc code blocks

and fixed a pesky crash in the syntax highlighting caused by invalid
chars (0x0E) in MarkupParser.scala.

3 files changed, 76 insertions, 18 deletions

diff --git a/src/compiler/scala/tools/nsc/doc/html/SyntaxHigh.scala b/src/compiler/scala/tools/nsc/doc/html/SyntaxHigh.scala
index f19f449d2c..f67abc58da 100644
--- a/src/compiler/scala/tools/nsc/doc/html/SyntaxHigh.scala
+++ b/src/compiler/scala/tools/nsc/doc/html/SyntaxHigh.scala
@@ -219,24 +219,24 @@ private[html] object SyntaxHigh {
           parse(" ", i+1)
         case '&' =>
           parse("&", i+1)
-        case '<' =>
+        case '<' if i+1 < buf.length =>
           val ch = buf(i+1).toChar
           if (ch == '-' || ch == ':' || ch == '%')
             parse("<span class=\"kw\">&lt;"+ch+"</span>", i+2)
           else
             parse("&lt;", i+1)
         case '>' =>
-          if (buf(i+1) == ':')
+          if (i+1 < buf.length && buf(i+1) == ':')
             parse("<span class=\"kw\">&gt;:</span>", i+2)
           else
             parse("&gt;", i+1)
         case '=' =>
-          if (buf(i+1) == '>')
+          if (i+1 < buf.length && buf(i+1) == '>')
             parse("<span class=\"kw\">=&gt;</span>", i+2)
           else
             parse(buf(i).toChar.toString, i+1)
         case '/' =>
-          if (buf(i+1) == '/' || buf(i+1) == '*') {
+          if (i+1 < buf.length && (buf(i+1) == '/' || buf(i+1) == '*')) {
             val c = comment(i+1)
             parse("<span class=\"cmt\">"+c+"</span>", i+c.length)
           } else
@@ -257,9 +257,9 @@ private[html] object SyntaxHigh {
           else
             parse(buf(i).toChar.toString, i+1)
         case _ =>
-          if (i == 0 || !Character.isJavaIdentifierPart(buf(i-1).toChar)) {
+          if (i == 0 || (i >= 1 && !Character.isJavaIdentifierPart(buf(i-1).toChar))) {
             if (Character.isDigit(buf(i)) ||
-                (buf(i) == '.' && Character.isDigit(buf(i+1)))) {
+                (buf(i) == '.' && i + 1 < buf.length && Character.isDigit(buf(i+1)))) {
               val s = numlit(i)
               parse("<span class=\"num\">"+s+"</span>", i+s.length)
             } else {
diff --git a/src/compiler/scala/tools/nsc/doc/model/comment/CommentFactory.scala b/src/compiler/scala/tools/nsc/doc/model/comment/CommentFactory.scala
index efa524503c..dbbd6ab432 100644
--- a/src/compiler/scala/tools/nsc/doc/model/comment/CommentFactory.scala
+++ b/src/compiler/scala/tools/nsc/doc/model/comment/CommentFactory.scala
@@ -491,7 +491,7 @@ trait CommentFactory { thisFactory: ModelFactory with CommentFactory =>
       else
         jump("}}}")
       blockEnded("code block")
-      Code(getRead)
+      Code(normalizeIndentation(getRead))
     }
 
     /** {{{ title ::= ('=' inline '=' | "==" inline "==" | ...) '\n' }}} */
@@ -732,6 +732,64 @@ trait CommentFactory { thisFactory: ModelFactory with CommentFactory =>
         nextChar()
     }
 
+    /** 
+     *  Eliminates the (common) leading spaces in all lines, based on the first line 
+     *  For indented pieces of code, it reduces the indent to the least whitespace prefix:
+     *    {{{
+     *       indented example
+     *       another indented line
+     *       if (condition)
+     *         then do something;
+     *       ^ this is the least whitespace prefix
+     *    }}}
+     */
+    def normalizeIndentation(_code: String): String = {
+
+      var code = _code.trim
+      var maxSkip = Integer.MAX_VALUE
+      var crtSkip = 0
+      var wsArea = true
+      var index = 0
+      var firstLine = true
+      var emptyLine = true
+
+      while (index < code.length) {
+        code(index) match {
+          case ' ' =>
+            if (wsArea) 
+              crtSkip += 1
+          case c =>
+            wsArea = (c == '\n')
+            maxSkip = if (firstLine || emptyLine) maxSkip else if (maxSkip <= crtSkip) maxSkip else crtSkip 
+            crtSkip = if (c == '\n') 0 else crtSkip
+            firstLine = if (c == '\n') false else firstLine
+            emptyLine = if (c == '\n') true else false
+        }
+        index += 1
+      }
+
+      if (maxSkip == 0)
+        code
+      else {
+        index = 0
+        val builder = new StringBuilder
+        while (index < code.length) {
+          builder.append(code(index))
+          if (code(index) == '\n') {
+            // we want to skip as many spaces are available, if there are less spaces (like on empty lines, do not
+            // over-consume them)
+            index += 1
+            val limit = index + maxSkip
+            while ((index < code.length) && (code(index) == ' ') && index < limit)
+              index += 1
+          }
+          else
+            index += 1
+        }
+        builder.toString
+      }
+    }
+
     def checkParaEnded(): Boolean = {
       (char == endOfText) ||
       ((char == endOfLine) && {
diff --git a/src/library/scala/xml/parsing/MarkupParser.scala b/src/library/scala/xml/parsing/MarkupParser.scala
index 1de08b3025..74781914e3 100644
--- a/src/library/scala/xml/parsing/MarkupParser.scala
+++ b/src/library/scala/xml/parsing/MarkupParser.scala
@@ -134,7 +134,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   //
 
   /** {{{
-   *  &lt;? prolog ::= xml S ... ?&gt;
+   *  <? prolog ::= xml S ... ?>
    *  }}} */
   def xmlProcInstr(): MetaData = {
     xToken("xml")
@@ -195,7 +195,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  &lt;? prolog ::= xml S?
+   *  <? prolog ::= xml S?
    *  // this is a bit more lenient than necessary...
    *  }}} */
   def prolog(): (Option[String], Option[String], Option[Boolean]) =
@@ -355,7 +355,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  '&lt;! CharData ::= [CDATA[ ( {char} - {char}"]]&gt;"{char} ) ']]&gt;'
+   *  '<! CharData ::= [CDATA[ ( {char} - {char}"]]>"{char} ) ']]>'
    *
    *  see [15]
    *  }}} */
@@ -369,7 +369,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  Comment ::= '&lt;!--' ((Char - '-') | ('-' (Char - '-')))* '--&gt;'
+   *  Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
    *
    * see [15]
    *  }}} */
@@ -399,7 +399,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  '&lt;' content1 ::=  ...
+   *  '<' content1 ::=  ...
    *  }}} */
   def content1(pscope: NamespaceBinding, ts: NodeBuffer) {
     ch match {
@@ -420,7 +420,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  content1 ::=  '&lt;' content1 | '&amp;' charref ...
+   *  content1 ::=  '<' content1 | '&' charref ...
    *  }}} */
   def content(pscope: NamespaceBinding): NodeSeq = {
     var ts = new NodeBuffer
@@ -490,7 +490,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   /** parses document type declaration and assigns it to instance variable
    *  dtd.
    *  {{{
-   *  &lt;! parseDTD ::= DOCTYPE name ... >
+   *  <! parseDTD ::= DOCTYPE name ... >
    *  }}} */
   def parseDTD() { // dirty but fast
     var extID: ExternalID = null
@@ -545,8 +545,8 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  '&lt;' element ::= xmlTag1 '&gt;'  { xmlExpr | '{' simpleExpr '}' } ETag
-   *               | xmlTag1 '/' '&gt;'
+   *  '<' element ::= xmlTag1 '>'  { xmlExpr | '{' simpleExpr '}' } ETag
+   *               | xmlTag1 '/' '>'
    *  }}} */
   def element1(pscope: NamespaceBinding): NodeSeq = {
     val pos = this.pos
@@ -778,7 +778,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  &lt;! attlist := ATTLIST
+   *  <! attlist := ATTLIST
    *  }}} */
   def attrDecl() = {
     xToken("TTLIST")
@@ -824,7 +824,7 @@ trait MarkupParser extends MarkupParserCommon with TokenTests
   }
 
   /** {{{
-   *  &lt;! element := ELEMENT
+   *  <! element := ELEMENT
    *  }}} */
   def entityDecl() = {
     var isParameterEntity = false