src/Pure/General/yxml.scala
author wenzelm
Sun, 08 Aug 2010 14:00:59 +0200
changeset 38234 4b610fbb2d83
parent 38231 968844caaff9
child 38267 e50c283dd125
permissions -rw-r--r--
YXML.parse: refrain from interning, let XML.Cache do it (partially);
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
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    23
  /* 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
    24
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    25
  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
    26
    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
    27
  {
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    28
    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
    29
    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
    30
    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
    31
      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
    32
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    33
    // 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
    34
    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
    35
  }
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    36
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    37
  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
    38
    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
    39
  {
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    40
    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
    41
    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
    42
  }
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    43
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    44
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    45
  /* parsing */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    46
27993
6dd90ef9f927 simplified exceptions: use plain error function / RuntimeException;
wenzelm
parents: 27971
diff changeset
    47
  private def err(msg: String) = error("Malformed YXML: " + msg)
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    48
  private def err_attribute() = err("bad attribute")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    49
  private def err_element() = err("bad element")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    50
  private def err_unbalanced(name: String) =
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    51
    if (name == "") err("unbalanced element")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    52
    else err("unbalanced element \"" + name + "\"")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    53
27944
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    54
  private def parse_attrib(source: CharSequence) = {
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    55
    val s = source.toString
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    56
    val i = s.indexOf('=')
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    57
    if (i <= 0) err_attribute()
38234
4b610fbb2d83 YXML.parse: refrain from interning, let XML.Cache do it (partially);
wenzelm
parents: 38231
diff changeset
    58
    (s.substring(0, i), s.substring(i + 1))
27944
2bf3f30558ed parse_attrib: more efficient due to indexOf('=');
wenzelm
parents: 27943
diff changeset
    59
  }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    60
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    61
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    62
  def parse_body(source: CharSequence): List[XML.Tree] =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
    63
  {
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    64
    /* stack operations */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    65
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    66
    def buffer(): ListBuffer[XML.Tree] = new ListBuffer[XML.Tree]
38230
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    67
    var stack: List[(Markup, ListBuffer[XML.Tree])] = List((Markup("", Nil), buffer()))
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    68
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    69
    def add(x: XML.Tree)
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    70
    {
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    71
      (stack: @unchecked) match { case ((_, body) :: _) => body += x }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    72
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    73
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    74
    def push(name: String, atts: XML.Attributes)
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    75
    {
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    76
      if (name == "") err_element()
38230
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    77
      else stack = (Markup(name, atts), buffer()) :: stack
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    78
    }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    79
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    80
    def pop()
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    81
    {
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    82
      (stack: @unchecked) match {
38230
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    83
        case ((Markup("", _), _) :: _) => err_unbalanced("")
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    84
        case ((markup, body) :: pending) =>
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    85
          stack = pending
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
    86
          add(XML.Elem(markup, body.toList))
36684
943f1ca7b375 misc tuning -- accumulate body via ListBuffer;
wenzelm
parents: 34201
diff changeset
    87
      }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    88
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    89
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    90
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    91
    /* parse chunks */
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
    92
36685
2b3076cfd6dd slightly more general Library.chunks;
wenzelm
parents: 36684
diff changeset
    93
    for (chunk <- Library.chunks(source, X) if chunk.length != 0) {
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    94
      if (chunk.length == 1 && chunk.charAt(0) == Y) pop()
27945
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
    95
      else {
36685
2b3076cfd6dd slightly more general Library.chunks;
wenzelm
parents: 36684
diff changeset
    96
        Library.chunks(chunk, Y).toList match {
34099
2541de190d92 added decode_chars, with raw character view on byte buffer and adhoc decoding via toString;
wenzelm
parents: 32450
diff changeset
    97
          case ch :: name :: atts if ch.length == 0 =>
38234
4b610fbb2d83 YXML.parse: refrain from interning, let XML.Cache do it (partially);
wenzelm
parents: 38231
diff changeset
    98
            push(name.toString, atts.map(parse_attrib))
27945
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
    99
          case txts => for (txt <- txts) add(XML.Text(txt.toString))
d2dc5a1903e8 tuned parse performance: avoid splitting terminal Y chunk;
wenzelm
parents: 27944
diff changeset
   100
        }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   101
      }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   102
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   103
    stack match {
38230
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
   104
      case List((Markup("", _), body)) => body.toList
ed147003de4b simplified type XML.Tree: embed Markup directly, avoid slightly odd triple;
wenzelm
parents: 36685
diff changeset
   105
      case (Markup(name, _), _) :: _ => err_unbalanced(name)
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   106
    }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   107
  }
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   108
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   109
  def parse(source: CharSequence): XML.Tree =
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   110
    parse_body(source) match {
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   111
      case List(result) => result
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   112
      case Nil => XML.Text("")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   113
      case _ => err("multiple results")
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   114
    }
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   115
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   116
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   117
  /* failsafe parsing */
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   118
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   119
  private def markup_failsafe(source: CharSequence) =
38231
968844caaff9 simplified some Markup;
wenzelm
parents: 38230
diff changeset
   120
    XML.elem (Markup.MALFORMED,
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   121
      List(XML.Text(source.toString.replace(X_string, "\\<^X>").replace(Y_string, "\\<^Y>"))))
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   122
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   123
  def parse_body_failsafe(source: CharSequence): List[XML.Tree] =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   124
  {
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   125
    try { parse_body(source) }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   126
    catch { case _: RuntimeException => List(markup_failsafe(source)) }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   127
  }
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   128
32450
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   129
  def parse_failsafe(source: CharSequence): XML.Tree =
375db037f4d2 misc tuning;
wenzelm
parents: 31521
diff changeset
   130
  {
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   131
    try { parse(source) }
29521
736bf7117153 added parse_body_failsafe;
wenzelm
parents: 29180
diff changeset
   132
    catch { case _: RuntimeException => markup_failsafe(source) }
27960
65b10d8ef0c6 added parse_failsafe;
wenzelm
parents: 27946
diff changeset
   133
  }
27930
2b44df907cc2 Efficient text representation of XML trees.
wenzelm
parents:
diff changeset
   134
}