/// <summary>
/// Add a transition from NState "this"
/// to NState "nxt", for each character
/// value in the leaf range list.
/// If the characters are packed, transform
/// from character ordinal to equivalence class
/// ordinal.
/// </summary>
/// <param name="leaf">The regex leaf node</param>
/// <param name="nxt">The destination state</param>
public void AddClsTrans(Leaf leaf, NState nxt)
{
BitArray cls = new BitArray(myNfaInst.MaxSym);
if (myNfaInst.Pack)
{
foreach (int ord in leaf.rangeLit.equivClasses)
cls[ord] = true;
}
else
{
foreach (CharRange rng in leaf.rangeLit.list.Ranges)
for (int i = rng.minChr; i <= rng.maxChr; i++)
cls[i] = true;
if (leaf.rangeLit.list.IsInverted)
cls = cls.Not();
}
AddClsTrans(cls, nxt);
}
/// <summary>
/// Add a transition from NState "this"
/// to NState "nxt", for each character
/// value in the leaf range list.
/// If the characters are packed, transform
/// from character ordinal to equivalence class
/// ordinal.
/// </summary>
/// <param name="leaf">The regex leaf node</param>
/// <param name="nxt">The destination state</param>
public void AddClsTrans(Leaf leaf, NState nxt)
{
if (myNfaInst.parent.task.CaseAgnostic) {
leaf.rangeLit.list = leaf.rangeLit.list.MakeCaseAgnosticList();
leaf.rangeLit.list.Canonicalize();
}
BitArray cls = new BitArray(myNfaInst.MaxSym);
if (myNfaInst.Pack)
{
foreach (int ord in leaf.rangeLit.equivClasses)
cls[ord] = true;
}
else
{
foreach (CharRange rng in leaf.rangeLit.list.Ranges)
for (int i = rng.minChr; i <= rng.maxChr; i++)
cls[i] = true;
if (leaf.rangeLit.list.IsInverted)
cls = cls.Not();
}
AddClsTrans(cls, nxt);
}
//
// For version 1.0.1 recognize any line-end character if /unicode
//
static void AddAnchorContext(NfsaInstance nInst, NState endS, RuleDesc rule)
{
NState nEnd = nInst.MkState();
Leaf temp = new Leaf(RegOp.charClass);
temp.rangeLit = RangeLiteral.RightAnchors;
nInst.MakePath(temp, endS, nEnd);
nInst.MarkAccept(nEnd, rule);
nEnd.rhCntx = 1;
}
// CharClass : "[" ["^"] {code | code "-" code | FilteredClass}+ "]" ;
internal Leaf CharClass() {
// Assert chr == '['
// Need to build a new string taking into account char escapes
Leaf leaf = new Leaf( RegOp.charClass );
bool invert = false;
scan(); // read past '['
if (!esc && chr == '^') {
invert = true;
scan(); // read past '^'
}
leaf.rangeLit = new RangeLiteral( invert );
// Special case of '-' at start, taken as ordinary class member.
// This is correct for LEX specification, but is undocumented
// behavior for FLEX. GPLEX gives a friendly warning, just in
// case this is actually a typographical error.
if (!esc && chr == '-') {
Warn( 113, index - 1, 1, "-" );
leaf.rangeLit.list.Add( new CharRange( '-' ) );
scan(); // read past -'
}
while (chr != NUL && (esc || chr != ']')) {
int lhCodePoint;
int startIx = index - 1; // save starting index for error reporting
lhCodePoint = (esc ? EscapedChar() : CodePoint());
if (!esc && lhCodePoint == (int)'-')
Error( 82, startIx, index - startIx, null );
//
// There are three possible elements here:
// * a singleton character
// * a character range
// * a filtered class like [:IsLetter:]
//
if (chr == '[' && !esc && peek() == ':') // character category
{
Leaf rslt = FilteredClass();
leaf.Merge( rslt );
}
else {
scan();
if (!esc && chr == '-') // character range
{
scan();
if (!esc && chr == ']') {
// Special case of '-' at end, taken as ordinary class member.
// This is correct for LEX specification, but is undocumented
// behavior for FLEX. GPLEX gives a friendly warning, just in
// case this is actually a typographical error.
leaf.rangeLit.list.Add( new CharRange( lhCodePoint ) );
leaf.rangeLit.list.Add( new CharRange( '-' ) );
//Error(81, idx, index - idx - 1);
Warn( 114, startIx, index - startIx - 1, String.Format(
CultureInfo.InvariantCulture,
"'{0}','{1}'",
CharacterUtilities.Map( lhCodePoint ),
'-' ) );
}
else {
int rhCodePoint = (esc ? EscapedChar() : CodePoint());
if (rhCodePoint < lhCodePoint)
Error( 54, startIx, index - startIx, null );
scan();
leaf.rangeLit.list.Add( new CharRange( lhCodePoint, rhCodePoint ) );
}
}
else // character singleton
{
leaf.rangeLit.list.Add( new CharRange( lhCodePoint ) );
}
}
}
checkAndScan( ']' );
leaf.rangeLit.list.Canonicalize();
return leaf;
}
// Primitive : CharClassExpr | NamedRegexReference | "." | escapedChar | char ;
internal RegExTree Primitive() {
RegExTree tmp;
if (!esc && chr == '[')
tmp = CharClassExpr();
else if (!esc && chr == '{' && !Char.IsDigit( peek() ))
tmp = UseRegexRef();
else if (!esc && chr == '.') {
Leaf leaf = new Leaf( RegOp.charClass );
leaf.rangeLit = new RangeLiteral( true );
scan();
leaf.rangeLit.list.Add( new CharRange( '\n' ) );
tmp = leaf;
}
// Remaining cases are:
// 1. escaped character (maybe beyond ffff limit)
// 2. ordinary unicode character
// 3. maybe a surrogate pair in future
else if (esc) {
tmp = new Leaf( EscapedChar() );
scan();
}
else {
tmp = new Leaf( (int)chr );
scan();
}
return tmp;
}
internal void Intersect( Leaf rhOperand ) {
this.rangeLit.list = this.rangeLit.list.AND( rhOperand.rangeLit.list );
}
internal void Subtract( Leaf subtrahend ) {
this.rangeLit.list = this.rangeLit.list.SUB( subtrahend.rangeLit.list );
}
internal void Merge( Leaf addend ) {
foreach (CharRange rng in addend.rangeLit.list.Ranges)
this.rangeLit.list.Add( rng );
}