namespace System.Text.RegularExpressions {
class Interpreter : IMachine {
+ private int ReadProgramCount (int ptr)
+ {
+ int ret = program [ptr + 1];
+ ret <<= 16;
+ ret += program [ptr];
+ return ret;
+ }
+
public Interpreter (ushort[] program) {
this.program = program;
this.qs = null;
Debug.Assert ((OpCode)program[0] == OpCode.Info, "Regex", "Cant' find info block");
- this.group_count = program[1] + 1;
- this.match_min = program[2];
- //this.match_max = program[3];
+ this.group_count = ReadProgramCount (1) + 1;
+ this.match_min = ReadProgramCount (3);
+ //this.match_max = ReadProgramCount (5);
// setup
- this.program_start = 4;
+ this.program_start = 7;
this.groups = new int [group_count];
}
// True
switch ((Position)program[pc + 4]) {
- case Position.StartOfString:
+ case Position.StartOfString:
if (anch_reverse || anch_offset == 0) {
- ptr = anch_offset;
+ if (anch_reverse)
+ ptr = anch_offset;
if (TryMatch (ref ptr, pc + skip))
goto Pass;
}
break;
case Position.StartOfLine:
-
if (anch_ptr == 0) {
ptr = 0;
if (TryMatch (ref ptr, pc + skip))
break;
}
- case OpCode.Character: case OpCode.Category:
+ case OpCode.Character: case OpCode.Category: case OpCode.NotCategory:
case OpCode.Range: case OpCode.Set: {
if (!EvalChar (mode, ref ptr, ref pc, false))
goto Fail;
case OpCode.Repeat: {
this.repeat = new RepeatContext (
this.repeat, // previous context
- program[pc + 2], // minimum
- program[pc + 3], // maximum
+ ReadProgramCount (pc + 2), // minimum
+ ReadProgramCount (pc + 4), // maximum
(flags & OpFlags.Lazy) != 0, // lazy
- pc + 4 // subexpression
+ pc + 6 // subexpression
);
if (Eval (Mode.Match, ref ptr, pc + program[pc + 1]))
++ current.Count;
current.Start = ptr;
deep = current;
- if (!Eval (Mode.Match, ref ptr, repeat.Expression)) {
+ if (!Eval (Mode.Match, ref ptr, current.Expression)) {
current.Start = start;
current.Count = start_count;
goto Fail;
goto Pass;
}
- DegenerateMatch:
if (ptr == current.Start) {
// degenerate match ... match tail or fail
this.repeat = current.Previous;
+ deep = null;
if (Eval (Mode.Match, ref ptr, pc + 1))
goto Pass;
for (;;) {
// match tail first ...
this.repeat = current.Previous;
+ deep = null;
int cp = Checkpoint ();
if (Eval (Mode.Match, ref ptr, pc + 1))
goto Pass;
// ... then match more
this.repeat = current;
if (current.IsMaximum)
- return false;
+ goto Fail;
++ current.Count;
current.Start = ptr;
deep = current;
}
if (deep != current) // recursive mode
goto Pass;
- // Degenerate match: no point retrying current.Expression if tail match fails
+ // Degenerate match: ptr has not moved since the last (failed) tail match.
+ // So, next and subsequent tail matches will fail.
if (ptr == current.Start)
- goto DegenerateMatch;
+ goto Fail;
}
} else {
int stack_size = stack.Count;
// then, match the tail, backtracking as necessary.
this.repeat = current.Previous;
for (;;) {
+ deep = null;
if (Eval (Mode.Match, ref ptr, pc + 1)) {
stack.Count = stack_size;
goto Pass;
case OpCode.FastRepeat: {
this.fast = new RepeatContext (
fast,
- program[pc + 2], // minimum
- program[pc + 3], // maximum
+ ReadProgramCount (pc + 2), // minimum
+ ReadProgramCount (pc + 4), // maximum
(flags & OpFlags.Lazy) != 0, // lazy
- pc + 4 // subexpression
+ pc + 6 // subexpression
);
- deep = fast;
-
fast.Start = ptr;
int cp = Checkpoint ();
pc += program[pc + 1]; // tail expression
ushort tail_word = program[pc];
- int c1, c2; // first character of tail operator
- int coff; // 0 or -1 depending on direction
+ int c1 = -1; // first character of tail operator
+ int c2 = -1; // ... and the same character, in upper case if ignoring case
+ int coff = 0; // 0 or -1 depending on direction
OpCode tail_op = (OpCode)(tail_word & 0xff);
if (tail_op == OpCode.Character || tail_op == OpCode.String) {
OpFlags tail_flags = (OpFlags)(tail_word & 0xff00);
+ if ((tail_flags & OpFlags.Negate) != 0)
+ goto skip;
+
if (tail_op == OpCode.String)
{
int offset = 0;
else
coff = 0;
}
- else {
- c1 = c2 = -1;
- coff = 0;
- }
+ skip:
if (fast.IsLazy) {
if (!fast.IsMinimum && !Eval (Mode.Count, ref ptr, fast.Expression)) {
//Console.WriteLine ("lazy fast: failed mininum.");
while (true) {
int p = ptr + coff;
- if ((c1 < 0 || (p >= 0 && p < text_end && (c1 == text[p] || c2 == text[p]))) &&
- Eval (Mode.Match, ref ptr, pc))
- break;
+ if (c1 < 0 || (p >= 0 && p < text_end && (c1 == text[p] || c2 == text[p]))) {
+ deep = null;
+ if (Eval (Mode.Match, ref ptr, pc))
+ break;
+ }
if (fast.IsMaximum) {
//Console.WriteLine ("lazy fast: failed with maximum.");
while (true) {
int p = ptr + coff;
- if ((c1 < 0 || (p >= 0 && p < text_end && (c1 == text[p] || c2 == text[p]))) &&
- Eval (Mode.Match, ref ptr, pc))
- break;
+ if (c1 < 0 || (p >= 0 && p < text_end && (c1 == text[p] || c2 == text[p]))) {
+ deep = null;
+ if (Eval (Mode.Match, ref ptr, pc))
+ break;
+ }
-- fast.Count;
if (!fast.IsMinimum) {
case OpCode.Category: {
if (CategoryUtils.IsCategory ((Category)program[pc ++], c))
return !negate;
+ break;
+ }
+ case OpCode.NotCategory: {
+ if (!CategoryUtils.IsCategory ((Category)program[pc ++], c))
+ return !negate;
break;
}