Skip to content

Commit

Permalink
Added new regex option RegexOptions.AnyNewLine
Browse files Browse the repository at this point in the history
  • Loading branch information
@shishirchawla
shishirchawla committed Jan 10, 2020
1 parent 0bdab8a commit 55c713b
Show file tree
Hide file tree
Showing 13 changed files with 324 additions and 21 deletions.
1 change: 1 addition & 0 deletions src/libraries/System.Text.RegularExpressions/ref/System.Text.RegularExpressions.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -243,6 +243,7 @@ public enum RegexOptions
RightToLeft = 64,
ECMAScript = 256,
CultureInvariant = 512,
AnyNewLine = 1024,
}
public abstract partial class RegexRunner
{
Expand Down
5 changes: 3 additions & 2 deletions src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/Regex.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ namespace System.Text.RegularExpressions
/// </summary>
public partial class Regex : ISerializable
{
private const int MaxOptionShift = 10;
private const int MaxOptionShift = 11;

protected internal string? pattern; // The string pattern provided
protected internal RegexOptions roptions; // the top-level options from the options string
Expand Down Expand Up @@ -137,7 +137,8 @@ internal static void ValidateOptions(RegexOptions options)
#if DEBUG
RegexOptions.Debug |
#endif
RegexOptions.CultureInvariant)) != 0)
RegexOptions.CultureInvariant |
RegexOptions.AnyNewLine)) != 0)
{
throw new ArgumentOutOfRangeException(nameof(options));
}
Expand Down
7 changes: 6 additions & 1 deletion src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCode.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ internal sealed class RegexCode
public const int Beginning = 18; // \A
public const int Start = 19; // \G
public const int EndZ = 20; // \Z
public const int End = 21; // \Z
public const int End = 21; // \z

public const int Nothing = 22; // Reject!

Expand Down Expand Up @@ -88,6 +88,9 @@ internal sealed class RegexCode
public const int Notoneloopatomic = 44; // lef,back set,min,max (?> . {,n} )
public const int Setloopatomic = 45; // lef,back set,min,max (?> [\d]{,n} )

public const int AnyEndZ = 46; // \Z
public const int AnyEol = 47; // $

// Modifiers for alternate modes
public const int Mask = 63; // Mask to get unmodified ordinary operator
public const int Rtl = 64; // bit to indicate that we're reverse scanning.
Expand Down Expand Up @@ -168,13 +171,15 @@ public static int OpcodeSize(int opcode)
case Nothing:
case Bol:
case Eol:
case AnyEol:
case Boundary:
case Nonboundary:
case ECMABoundary:
case NonECMABoundary:
case Beginning:
case Start:
case EndZ:
case AnyEndZ:
case End:
case Nullmark:
case Setmark:
Expand Down
121 changes: 119 additions & 2 deletions ...raries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCompiler.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -876,7 +876,7 @@ protected void GenerateFindFirstChar()
}
}

if ((_anchors & (RegexFCD.Beginning | RegexFCD.Start | RegexFCD.EndZ | RegexFCD.End)) != 0)
if ((_anchors & (RegexFCD.Beginning | RegexFCD.Start | RegexFCD.EndZ | RegexFCD.AnyEndZ | RegexFCD.End)) != 0)
{
if (!_code!.RightToLeft)
{
Expand Down Expand Up @@ -908,7 +908,7 @@ protected void GenerateFindFirstChar()
MarkLabel(l1);
}

if ((_anchors & RegexFCD.EndZ) != 0)
if ((_anchors & (RegexFCD.EndZ | RegexFCD.AnyEndZ)) != 0)
{
Label l1 = DefineLabel();
Ldthisfld(s_runtextposField);
Expand Down Expand Up @@ -978,6 +978,60 @@ protected void GenerateFindFirstChar()
MarkLabel(l2);
}

if ((_anchors & RegexFCD.AnyEndZ) != 0)
{
LocalBuilder diff = _temp1Local;
Label l1 = DefineLabel();
Label l2 = DefineLabel();
Label l3 = DefineLabel();
Ldthisfld(s_runtextendField);
Ldthisfld(s_runtextposField);
Sub();
Stloc(diff);
Ldloc(diff);
Ldc(2);
Bgt(l1);
Ldloc(diff);
Ldc(2);
Blt(l2);
Ldthisfld(s_runtextField);
Ldthisfld(s_runtextposField);
Callvirt(s_stringGetCharsMethod);
Ldc('\r');
Bne(l1);
Ldthisfld(s_runtextField);
Ldthisfld(s_runtextposField);
Ldc(1);
Add();
Callvirt(s_stringGetCharsMethod);
Ldc('\n');
Bne(l1);
Br(l3);

MarkLabel(l2);
Ldloc(diff);
Ldc(1);
Blt(l3);
Ldthisfld(s_runtextField);
Ldthisfld(s_runtextposField);
Callvirt(s_stringGetCharsMethod);
Ldc('\n');
Beq(l3);
Ldthisfld(s_runtextField);
Ldthisfld(s_runtextposField);
Callvirt(s_stringGetCharsMethod);
Ldc('\r');
Beq(l3);

MarkLabel(l1);
Ldthis();
Ldthisfld(s_runtextbegField);
Stfld(s_runtextposField);
Ldc(0);
Ret();
MarkLabel(l3);
}

if ((_anchors & RegexFCD.Start) != 0)
{
Label l1 = DefineLabel();
Expand Down Expand Up @@ -3361,6 +3415,23 @@ private void GenerateOneCode()
break;
}

case RegexCode.AnyEol:
//: if (Rightchars() > 0 && CharAt(Textpos()) != '\n' && CharAt(Textpos()) != '\r')
//: break Backward;
{
Label l1 = _labels![NextCodepos()];
Ldloc(_runtextposLocal!);
Ldloc(_runtextendLocal!);
Bge(l1);
Rightchar();
Ldc('\n');
Beq(l1);
Rightchar();
Ldc('\r');
BneFar(_backtrack);
break;
}

case RegexCode.Boundary:
case RegexCode.Nonboundary:
//: if (!IsBoundary(Textpos(), _textbeg, _textend))
Expand Down Expand Up @@ -3431,6 +3502,52 @@ private void GenerateOneCode()
BneFar(_backtrack);
break;

case RegexCode.AnyEndZ:
//: if (rightChars > 2)
//: break Backward;
//: if (rightChars == 1 && CharAt(Textpos()) != '\r' && CharAt(Textpos()) != '\n')
//: break Backward;
//: if (rightChars == 2 && (CharAt(Textpos()) != '\r' || CharAt(Textpos()+1) != '\n'))
//: break Backward;
{
LocalBuilder diff = _temp1Local!;
Label l1 = DefineLabel();

Ldloc(_runtextendLocal!);
Ldloc(_runtextposLocal!);
Sub();
Stloc(diff);
Ldloc(diff);
Ldc(2);
BgtFar(_backtrack);
Ldloc(diff);
Ldc(2);
Blt(l1);
Rightchar();
Ldc('\r');
BneFar(_backtrack);
Ldloc(_runtextLocal!);
Ldloc(_runtextposLocal!);
Ldc(1);
Add();
Callvirt(s_stringGetCharsMethod);
Ldc('\n');
BneFar(_backtrack);
Br(_labels![NextCodepos()]);

MarkLabel(l1);
Ldloc(diff);
Ldc(1);
Blt(_labels![NextCodepos()]);
Rightchar();
Ldc('\n');
Beq(_labels![NextCodepos()]);
Rightchar();
Ldc('\r');
BneFar(_backtrack);
break;
}

case RegexCode.End:
//: if (Rightchars() > 0)
//: break Backward;
Expand Down
15 changes: 15 additions & 0 deletions src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexFCD.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,9 @@ internal ref struct RegexFCD
public const int Boundary = 0x0040;
public const int ECMABoundary = 0x0080;

public const int AnyEndZ = 0x0100;
public const int AnyEol = 0x0200;

private readonly List<RegexFC> _fcStack;
private ValueListBuilder<int> _intStack; // must not be readonly
private bool _skipAllChildren; // don't process any more children at the current level
Expand Down Expand Up @@ -130,11 +133,13 @@ public static RegexPrefix Prefix(RegexTree tree)

case RegexNode.Bol:
case RegexNode.Eol:
case RegexNode.AnyEol:
case RegexNode.Boundary:
case RegexNode.ECMABoundary:
case RegexNode.Beginning:
case RegexNode.Start:
case RegexNode.EndZ:
case RegexNode.AnyEndZ:
case RegexNode.End:
case RegexNode.Empty:
case RegexNode.Require:
Expand Down Expand Up @@ -185,11 +190,13 @@ public static int Anchors(RegexTree tree)

case RegexNode.Bol:
case RegexNode.Eol:
case RegexNode.AnyEol:
case RegexNode.Boundary:
case RegexNode.ECMABoundary:
case RegexNode.Beginning:
case RegexNode.Start:
case RegexNode.EndZ:
case RegexNode.AnyEndZ:
case RegexNode.End:
return result | AnchorFromType(curNode.Type);

Expand Down Expand Up @@ -217,11 +224,13 @@ private static int AnchorFromType(int type) =>
{
RegexNode.Bol => Bol,
RegexNode.Eol => Eol,
RegexNode.AnyEol => AnyEol,
RegexNode.Boundary => Boundary,
RegexNode.ECMABoundary => ECMABoundary,
RegexNode.Beginning => Beginning,
RegexNode.Start => Start,
RegexNode.EndZ => EndZ,
RegexNode.AnyEndZ => AnyEndZ,
RegexNode.End => End,
_ => 0,
};
Expand All @@ -244,10 +253,14 @@ public static string AnchorDescription(int anchors)
sb.Append(", ECMABoundary");
if (0 != (anchors & Eol))
sb.Append(", Eol");
if (0 != (anchors & AnyEol))
sb.Append(", AnyEol");
if (0 != (anchors & End))
sb.Append(", End");
if (0 != (anchors & EndZ))
sb.Append(", EndZ");
if (0 != (anchors & AnyEndZ))
sb.Append(", AnyEndZ");

if (sb.Length >= 2)
return (sb.ToString(2, sb.Length - 2));
Expand Down Expand Up @@ -479,13 +492,15 @@ private void CalculateFC(int NodeType, RegexNode node, int CurIndex)
case RegexNode.Nothing:
case RegexNode.Bol:
case RegexNode.Eol:
case RegexNode.AnyEol:
case RegexNode.Boundary:
case RegexNode.Nonboundary:
case RegexNode.ECMABoundary:
case RegexNode.NonECMABoundary:
case RegexNode.Beginning:
case RegexNode.Start:
case RegexNode.EndZ:
case RegexNode.AnyEndZ:
case RegexNode.End:
PushFC(new RegexFC(true));
break;
Expand Down
27 changes: 25 additions & 2 deletions ...ies/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexInterpreter.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -387,7 +387,8 @@ private char CharAt(int j)

protected override bool FindFirstChar()
{
if (0 != (_code.Anchors & (RegexFCD.Beginning | RegexFCD.Start | RegexFCD.EndZ | RegexFCD.End)))
if (0 != (_code.Anchors & (RegexFCD.Beginning | RegexFCD.Start | RegexFCD.EndZ |
RegexFCD.AnyEndZ | RegexFCD.End)))
{
if (!_code.RightToLeft)
{
Expand All @@ -397,7 +398,7 @@ protected override bool FindFirstChar()
runtextpos = runtextend;
return false;
}
if (0 != (_code.Anchors & RegexFCD.EndZ) && runtextpos < runtextend - 1)
if (0 != (_code.Anchors & (RegexFCD.EndZ | RegexFCD.AnyEndZ)) && runtextpos < runtextend - 1)
{
runtextpos = runtextend - 1;
}
Expand All @@ -411,6 +412,11 @@ protected override bool FindFirstChar()
if ((0 != (_code.Anchors & RegexFCD.End) && runtextpos < runtextend) ||
(0 != (_code.Anchors & RegexFCD.EndZ) && (runtextpos < runtextend - 1 ||
(runtextpos == runtextend - 1 && CharAt(runtextpos) != '\n'))) ||
(0 != (_code.Anchors & RegexFCD.AnyEndZ) && (runtextpos < runtextend - 2 ||
(runtextpos == runtextend - 2 && (CharAt(runtextpos) != '\r'
|| CharAt(runtextpos+1) != '\n')) ||
(runtextpos == runtextend - 1 && CharAt(runtextpos) != '\n'
&& CharAt(runtextpos) != '\r'))) ||
(0 != (_code.Anchors & RegexFCD.Start) && runtextpos < runtextstart))
{
runtextpos = runtextbeg;
Expand Down Expand Up @@ -967,6 +973,12 @@ protected override void Go()
advance = 0;
continue;

case RegexCode.AnyEol:
if (Rightchars() > 0 && CharAt(Textpos()) != '\n' && CharAt(Textpos()) != '\r')
break;
advance = 0;
continue;

case RegexCode.Boundary:
if (!IsBoundary(Textpos(), runtextbeg, runtextend))
break;
Expand Down Expand Up @@ -1009,6 +1021,17 @@ protected override void Go()
advance = 0;
continue;

case RegexCode.AnyEndZ:
int rightChars = Rightchars();
if (rightChars > 2)
break;
if (rightChars == 1 && CharAt(Textpos()) != '\r' && CharAt(Textpos()) != '\n')
break;
if (rightChars == 2 && (CharAt(Textpos()) != '\r' || CharAt(Textpos() + 1) != '\n'))
break;
advance = 0;
continue;

case RegexCode.End:
if (Rightchars() > 0)
break;
Expand Down
3 changes: 3 additions & 0 deletions src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexNode.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -69,13 +69,15 @@ internal sealed class RegexNode

public const int Bol = RegexCode.Bol; // ^
public const int Eol = RegexCode.Eol; // $
public const int AnyEol = RegexCode.AnyEol; // $
public const int Boundary = RegexCode.Boundary; // \b
public const int Nonboundary = RegexCode.Nonboundary; // \B
public const int ECMABoundary = RegexCode.ECMABoundary; // \b
public const int NonECMABoundary = RegexCode.NonECMABoundary; // \B
public const int Beginning = RegexCode.Beginning; // \A
public const int Start = RegexCode.Start; // \G
public const int EndZ = RegexCode.EndZ; // \Z
public const int AnyEndZ = RegexCode.AnyEndZ; // \Z
public const int End = RegexCode.End; // \z

public const int Oneloopatomic = RegexCode.Oneloopatomic; // c,n (?> a*)
Expand Down Expand Up @@ -978,6 +980,7 @@ public string Description()
if ((Options & RegexOptions.Singleline) != 0) argSb.Append("-S");
if ((Options & RegexOptions.IgnorePatternWhitespace) != 0) argSb.Append("-X");
if ((Options & RegexOptions.ECMAScript) != 0) argSb.Append("-E");
if ((Options & RegexOptions.AnyNewLine) != 0) argSb.Append("-A");

switch (Type)
{
Expand Down
1 change: 1 addition & 0 deletions ...braries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexOptions.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,5 +22,6 @@ public enum RegexOptions

ECMAScript = 0x0100, // "e"
CultureInvariant = 0x0200,
AnyNewLine = 0x0400, // "a", Treat "$" as (?=\r\z|\n\z|\r\n\z|\z)
}
}
Loading

0 comments on commit 55c713b

Please sign in to comment.