src/Pure/General/yxml.scala
author wenzelm
Thu, 06 May 2010 17:49:57 +0200
changeset 36684 943f1ca7b375
parent 34201 c95dcd12f48a
child 36685 2b3076cfd6dd
permissions -rw-r--r--
misc tuning -- accumulate body via ListBuffer;
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     1
/*  Title:      Pure/General/yxml.scala
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     2
    Author:     Makarius
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     3
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     4
Efficient text representation of XML trees.
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     5
*/
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     6
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     7
package isabelle
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     8
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
     9
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    10
import scala.collection.mutable.ListBuffer
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    11
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    12
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    13
object YXML
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    14
{
27943
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    15
  /* chunk markers */
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    16
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    17
  private val X = '\5'
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    18
  private val Y = '\6'
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
    19
  private val X_string = X.toString
27945
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
    20
  private val Y_string = Y.toString
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    21
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    22
27943
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    23
  /* iterate over chunks (resembles space_explode in ML) */
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    24
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    25
  private def chunks(sep: Char, source: CharSequence) = new Iterator[CharSequence]
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    26
  {
27943
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    27
    private val end = source.length
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    28
    private var state = if (end == 0) None else get_chunk(-1)
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    29
    private def get_chunk(i: Int) =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    30
    {
27943
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    31
      if (i < end) {
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    32
        var j = i; do j += 1 while (j < end && source.charAt(j) != sep)
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    33
        Some((source.subSequence(i + 1, j), j))
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    34
      }
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    35
      else None
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    36
    }
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    37
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    38
    def hasNext() = state.isDefined
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    39
    def next() = state match {
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    40
      case Some((s, i)) => { state = get_chunk(i); s }
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    41
      case None => throw new NoSuchElementException("next on empty iterator")
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    42
    }
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    43
  }
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    44
f34ff5e7728f replaced Pattern.split by chunks iterator (more efficient, resembles ML version more closely);
wenzelm
parents: 27930
diff changeset
    45
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    46
  /* decoding pseudo UTF-8 */
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    47
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    48
  private class Decode_Chars(decode: String => String,
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    49
    buffer: Array[Byte], start: Int, end: Int) extends CharSequence
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    50
  {
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    51
    def length: Int = end - start
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    52
    def charAt(i: Int): Char = (buffer(start + i).asInstanceOf[Int] & 0xFF).asInstanceOf[Char]
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    53
    def subSequence(i: Int, j: Int): CharSequence =
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    54
      new Decode_Chars(decode, buffer, start + i, start + j)
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    55
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    56
    // toString with adhoc decoding: abuse of CharSequence interface
34201
c95dcd12f48a separate Standard_System (Cygwin/Posix compatibility) vs. Isabelle_System (settings environment etc.);
wenzelm
parents: 34198
diff changeset
    57
    override def toString: String = decode(Standard_System.decode_permissive_utf8(this))
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    58
  }
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    59
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    60
  def decode_chars(decode: String => String,
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    61
    buffer: Array[Byte], start: Int, end: Int): CharSequence =
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    62
  {
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    63
    require(0 <= start && start <= end && end <= buffer.length)
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    64
    new Decode_Chars(decode, buffer, start, end)
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    65
  }
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    66
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    67
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    68
  /* parsing */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    69
27993
6dd90ef9f927 simplified exceptions: use plain error function / RuntimeException;
wenzelm
parents: 27971
diff changeset
    70
  private def err(msg: String) = error("Malformed YXML: " + msg)
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    71
  private def err_attribute() = err("bad attribute")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    72
  private def err_element() = err("bad element")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    73
  private def err_unbalanced(name: String) =
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    74
    if (name == "") err("unbalanced element")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    75
    else err("unbalanced element \"" + name + "\"")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    76
27944
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    77
  private def parse_attrib(source: CharSequence) = {
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    78
    val s = source.toString
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    79
    val i = s.indexOf('=')
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    80
    if (i <= 0) err_attribute()
29563
4773c5c994dc intern names of elements and attributes;
wenzelm
parents: 29521
diff changeset
    81
    (s.substring(0, i).intern, s.substring(i + 1))
27944
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    82
  }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    83
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    84
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    85
  def parse_body(source: CharSequence): List[XML.Tree] =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    86
  {
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    87
    /* stack operations */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    88
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    89
    def buffer(): ListBuffer[XML.Tree] = new ListBuffer[XML.Tree]
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    90
    var stack: List[((String, XML.Attributes), ListBuffer[XML.Tree])] = List((("", Nil), buffer()))
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    91
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    92
    def add(x: XML.Tree)
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    93
    {
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    94
      (stack: @unchecked) match { case ((_, body) :: _) => body += x }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    95
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    96
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    97
    def push(name: String, atts: XML.Attributes)
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    98
    {
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    99
      if (name == "") err_element()
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   100
      else stack = ((name, atts), buffer()) :: stack
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   101
    }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   102
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   103
    def pop()
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   104
    {
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   105
      (stack: @unchecked) match {
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   106
        case ((("", _), _) :: _) => err_unbalanced("")
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   107
        case (((name, atts), body) :: pending) =>
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   108
          stack = pending; add(XML.Elem(name, atts, body.toList))
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   109
      }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   110
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   111
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   112
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   113
    /* parse chunks */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   114
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
   115
    for (chunk <- chunks(X, source) if chunk.length != 0) {
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
   116
      if (chunk.length == 1 && chunk.charAt(0) == Y) pop()
27945
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
   117
      else {
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
   118
        chunks(Y, chunk).toList match {
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
   119
          case ch :: name :: atts if ch.length == 0 =>
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
   120
            push(name.toString.intern, atts.map(parse_attrib))
27945
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
   121
          case txts => for (txt <- txts) add(XML.Text(txt.toString))
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
   122
        }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   123
      }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   124
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   125
    stack match {
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
   126
      case List((("", _), body)) => body.toList
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   127
      case ((name, _), _) :: _ => err_unbalanced(name)
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   128
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   129
  }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   130
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   131
  def parse(source: CharSequence): XML.Tree =
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   132
    parse_body(source) match {
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   133
      case List(result) => result
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   134
      case Nil => XML.Text("")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   135
      case _ => err("multiple results")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   136
    }
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   137
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   138
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   139
  /* failsafe parsing */
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   140
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   141
  private def markup_failsafe(source: CharSequence) =
34119
ae92efb48784 markup bad YXML as malformed;
wenzelm
parents: 34118
diff changeset
   142
    XML.Elem (Markup.MALFORMED, Nil,
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   143
      List(XML.Text(source.toString.replace(X_string, "\\<^X>").replace(Y_string, "\\<^Y>"))))
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   144
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   145
  def parse_body_failsafe(source: CharSequence): List[XML.Tree] =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   146
  {
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   147
    try { parse_body(source) }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   148
    catch { case _: RuntimeException => List(markup_failsafe(source)) }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   149
  }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   150
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   151
  def parse_failsafe(source: CharSequence): XML.Tree =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   152
  {
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   153
    try { parse(source) }
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   154
    catch { case _: RuntimeException => markup_failsafe(source) }
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   155
  }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   156
}