Skip to content

Commit

Permalink
Add a few more tests / improve code coverage
Browse files Browse the repository at this point in the history
  • Loading branch information
@stephentoub
stephentoub committed Jan 9, 2020
1 parent 3f270ff commit 95965c9
Show file tree
Hide file tree
Showing 12 changed files with 441 additions and 174 deletions.
114 changes: 59 additions & 55 deletions 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 @@ -222,62 +222,66 @@ public static int OpcodeSize(int opcode)
}

#if DEBUG
private static readonly Dictionary<int, string> s_codeStr = new Dictionary<int, string>()
{
{ Onerep, nameof(Onerep) },
{ Notonerep, nameof(Notonerep) },
{ Setrep, nameof(Setrep) },
{ Oneloop, nameof(Oneloop) },
{ Notoneloop, nameof(Notoneloop) },
{ Setloop, nameof(Setloop) },
{ Onelazy, nameof(Onelazy) },
{ Notonelazy, nameof(Notonelazy) },
{ Setlazy, nameof(Setlazy) },
{ One, nameof(One) },
{ Notone, nameof(Notone) },
{ Set, nameof(Set) },
{ Multi, nameof(Multi) },
{ Ref, nameof(Ref) },
{ Bol, nameof(Bol) },
{ Eol, nameof(Eol) },
{ Boundary, nameof(Boundary) },
{ Nonboundary, nameof(Nonboundary) },
{ Beginning, nameof(Beginning) },
{ Start, nameof(Start) },
{ EndZ, nameof(EndZ) },
{ End, nameof(End) },
{ Nothing, nameof(Nothing) },
{ Lazybranch, nameof(Lazybranch) },
{ Branchmark, nameof(Branchmark) },
{ Lazybranchmark, nameof(Lazybranchmark) },
{ Nullcount, nameof(Nullcount) },
{ Setcount, nameof(Setcount) },
{ Branchcount, nameof(Branchcount) },
{ Lazybranchcount, nameof(Lazybranchcount) },
{ Nullmark, nameof(Nullmark) },
{ Setmark, nameof(Setmark) },
{ Capturemark, nameof(Capturemark) },
{ Getmark, nameof(Getmark) },
{ Setjump, nameof(Setjump) },
{ Backjump, nameof(Backjump) },
{ Forejump, nameof(Forejump) },
{ Testref, nameof(Testref) },
{ Goto, nameof(Goto) },
{ Stop, nameof(Stop) },
{ ECMABoundary, nameof(ECMABoundary) },
{ NonECMABoundary, nameof(NonECMABoundary) },
{ Oneloopatomic, nameof(Oneloopatomic) },
{ Notoneloopatomic, nameof(Notoneloopatomic) },
{ Setloopatomic, nameof(Setloopatomic) },
};

[ExcludeFromCodeCoverage]
private static string OperatorDescription(int Opcode) =>
(s_codeStr.TryGetValue(Opcode & Mask, out string? codeStr) ? codeStr : "(unknown)") +
((Opcode & Ci) != 0 ? "-Ci" : "") +
((Opcode & Rtl) != 0 ? "-Rtl" : "") +
((Opcode & Back) != 0 ? "-Back" : "") +
((Opcode & Back2) != 0 ? "-Back2" : "");
private static string OperatorDescription(int Opcode)
{
string codeStr = (Opcode & Mask) switch
{
Onerep => nameof(Onerep),
Notonerep => nameof(Notonerep),
Setrep => nameof(Setrep),
Oneloop => nameof(Oneloop),
Notoneloop => nameof(Notoneloop),
Setloop => nameof(Setloop),
Onelazy => nameof(Onelazy),
Notonelazy => nameof(Notonelazy),
Setlazy => nameof(Setlazy),
One => nameof(One),
Notone => nameof(Notone),
Set => nameof(Set),
Multi => nameof(Multi),
Ref => nameof(Ref),
Bol => nameof(Bol),
Eol => nameof(Eol),
Boundary => nameof(Boundary),
Nonboundary => nameof(Nonboundary),
Beginning => nameof(Beginning),
Start => nameof(Start),
EndZ => nameof(EndZ),
End => nameof(End),
Nothing => nameof(Nothing),
Lazybranch => nameof(Lazybranch),
Branchmark => nameof(Branchmark),
Lazybranchmark => nameof(Lazybranchmark),
Nullcount => nameof(Nullcount),
Setcount => nameof(Setcount),
Branchcount => nameof(Branchcount),
Lazybranchcount => nameof(Lazybranchcount),
Nullmark => nameof(Nullmark),
Setmark => nameof(Setmark),
Capturemark => nameof(Capturemark),
Getmark => nameof(Getmark),
Setjump => nameof(Setjump),
Backjump => nameof(Backjump),
Forejump => nameof(Forejump),
Testref => nameof(Testref),
Goto => nameof(Goto),
Stop => nameof(Stop),
ECMABoundary => nameof(ECMABoundary),
NonECMABoundary => nameof(NonECMABoundary),
Oneloopatomic => nameof(Oneloopatomic),
Notoneloopatomic => nameof(Notoneloopatomic),
Setloopatomic => nameof(Setloopatomic),
_ => "(unknown)"
};

return
codeStr +
((Opcode & Ci) != 0 ? "-Ci" : "") +
((Opcode & Rtl) != 0 ? "-Rtl" : "") +
((Opcode & Back) != 0 ? "-Back" : "") +
((Opcode & Back2) != 0 ? "-Back2" : "");
}

[ExcludeFromCodeCoverage]
public string OpcodeDescription(int offset)
Expand Down
81 changes: 40 additions & 41 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 @@ -1661,7 +1661,7 @@ LocalBuilder RentInt32Local()
// Returns a rented Int32 local.
void ReturnInt32Local(LocalBuilder int32Local)
{
Debug.Assert(int32Local != null);
Debug.Assert(iterationLocals != null);
Debug.Assert(int32Local.LocalType == typeof(int));
iterationLocals.Push(int32Local);
}
Expand Down Expand Up @@ -2637,16 +2637,6 @@ private void InitializeCultureForGoIfNecessary()
}
}

#if DEBUG
/// <summary>Debug only: emit code to print out a message.</summary>
[ExcludeFromCodeCoverage]
private void Message(string str)
{
Ldstr(str);
Call(s_debugWriteLine!);
}
#endif

/// <summary>
/// The main translation function. It translates the logic for a single opcode at
/// the current position. The structure of this function exactly mirrors
Expand All @@ -2664,37 +2654,9 @@ private void GenerateOneCode()
{
#if DEBUG
if ((_options & RegexOptions.Debug) != 0)
{
Mvlocfld(_runtextposLocal!, s_runtextposField);
Mvlocfld(_runtrackposLocal!, s_runtrackposField);
Mvlocfld(_runstackposLocal!, s_runstackposField);
Ldthis();
Callvirt(s_dumpStateM);

var sb = new StringBuilder();
if (_backpos > 0)
{
sb.AppendFormat("{0:D6} ", _backpos);
}
else
{
sb.Append(" ");
}
sb.Append(_code!.OpcodeDescription(_codepos));

if ((_regexopcode & RegexCode.Back) != 0)
{
sb.Append(" Back");
}

if ((_regexopcode & RegexCode.Back2) != 0)
{
sb.Append(" Back2");
}

Message(sb.ToString());
}
DumpBacktracking();
#endif

LocalBuilder charInClassLocal;

// Before executing any RegEx code in the unrolled loop,
Expand Down Expand Up @@ -4399,5 +4361,42 @@ private void EmitTimeoutCheck()
Callvirt(s_checkTimeoutMethod);
MarkLabel(label);
}

#if DEBUG
/// <summary>Emit code to print out the current state of the runner.</summary>
[ExcludeFromCodeCoverage]
private void DumpBacktracking()
{
Mvlocfld(_runtextposLocal!, s_runtextposField);
Mvlocfld(_runtrackposLocal!, s_runtrackposField);
Mvlocfld(_runstackposLocal!, s_runstackposField);
Ldthis();
Callvirt(s_dumpStateM);

var sb = new StringBuilder();
if (_backpos > 0)
{
sb.AppendFormat("{0:D6} ", _backpos);
}
else
{
sb.Append(" ");
}
sb.Append(_code!.OpcodeDescription(_codepos));

if ((_regexopcode & RegexCode.Back) != 0)
{
sb.Append(" Back");
}

if ((_regexopcode & RegexCode.Back2) != 0)
{
sb.Append(" Back2");
}

Ldstr(sb.ToString());
Call(s_debugWriteLine!);
}
#endif
}
}
105 changes: 49 additions & 56 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 @@ -926,65 +926,58 @@ public int ChildCount()
}

#if DEBUG
private static readonly Dictionary<int, string> s_typeStr = new Dictionary<int, string>()
{
{ Oneloop, nameof(Oneloop) },
{ Notoneloop, nameof(Notoneloop) },
{ Setloop, nameof(Setloop) },
{ Onelazy, nameof(Onelazy) },
{ Notonelazy, nameof(Notonelazy) },
{ Setlazy, nameof(Setlazy) },
{ One, nameof(One) },
{ Notone, nameof(Notone) },
{ Set, nameof(Set) },
{ Multi, nameof(Multi) },
{ Ref, nameof(Ref) },
{ Bol, nameof(Bol) },
{ Eol, nameof(Eol) },
{ Boundary, nameof(Boundary) },
{ Nonboundary, nameof(Nonboundary) },
{ ECMABoundary, nameof(ECMABoundary) },
{ NonECMABoundary, nameof(NonECMABoundary) },
{ Beginning, nameof(Beginning) },
{ Start, nameof(Start) },
{ EndZ, nameof(EndZ) },
{ End, nameof(End) },
{ Oneloopatomic, nameof(Oneloopatomic) },
{ Notoneloopatomic, nameof(Notoneloopatomic) },
{ Setloopatomic, nameof(Setloopatomic) },
{ Nothing, nameof(Nothing) },
{ Empty, nameof(Empty) },
{ Lazyloop, nameof(Lazyloop) },
{ Capture, nameof(Capture) },
{ Oneloop, nameof(Group) },
{ Oneloop, nameof(Require) },
{ Oneloop, nameof(Prevent) },
{ Oneloop, nameof(Atomic) },
{ Oneloop, nameof(Testref) },
{ Oneloop, nameof(Testgroup) },
};

[ExcludeFromCodeCoverage]
public string Description()
{
StringBuilder argSb = new StringBuilder();

argSb.Append(s_typeStr.TryGetValue(Type, out string? typeStr) ? typeStr : "(unknown)");

if ((Options & RegexOptions.ExplicitCapture) != 0)
argSb.Append("-C");
if ((Options & RegexOptions.IgnoreCase) != 0)
argSb.Append("-I");
if ((Options & RegexOptions.RightToLeft) != 0)
argSb.Append("-L");
if ((Options & RegexOptions.Multiline) != 0)
argSb.Append("-M");
if ((Options & RegexOptions.Singleline) != 0)
argSb.Append("-S");
if ((Options & RegexOptions.IgnorePatternWhitespace) != 0)
argSb.Append("-X");
if ((Options & RegexOptions.ECMAScript) != 0)
argSb.Append("-E");

string typeStr = Type switch
{
Oneloop => nameof(Oneloop),
Notoneloop => nameof(Notoneloop),
Setloop => nameof(Setloop),
Onelazy => nameof(Onelazy),
Notonelazy => nameof(Notonelazy),
Setlazy => nameof(Setlazy),
One => nameof(One),
Notone => nameof(Notone),
Set => nameof(Set),
Multi => nameof(Multi),
Ref => nameof(Ref),
Bol => nameof(Bol),
Eol => nameof(Eol),
Boundary => nameof(Boundary),
Nonboundary => nameof(Nonboundary),
ECMABoundary => nameof(ECMABoundary),
NonECMABoundary => nameof(NonECMABoundary),
Beginning => nameof(Beginning),
Start => nameof(Start),
EndZ => nameof(EndZ),
End => nameof(End),
Oneloopatomic => nameof(Oneloopatomic),
Notoneloopatomic => nameof(Notoneloopatomic),
Setloopatomic => nameof(Setloopatomic),
Nothing => nameof(Nothing),
Empty => nameof(Empty),
Lazyloop => nameof(Lazyloop),
Capture => nameof(Capture),
Group => nameof(Group),
Require => nameof(Require),
Prevent => nameof(Prevent),
Atomic => nameof(Atomic),
Testref => nameof(Testref),
Testgroup => nameof(Testgroup),
_ => "(unknown)"
};

var argSb = new StringBuilder().Append(typeStr);

if ((Options & RegexOptions.ExplicitCapture) != 0) argSb.Append("-C");
if ((Options & RegexOptions.IgnoreCase) != 0) argSb.Append("-I");
if ((Options & RegexOptions.RightToLeft) != 0) argSb.Append("-L");
if ((Options & RegexOptions.Multiline) != 0) argSb.Append("-M");
if ((Options & RegexOptions.Singleline) != 0) argSb.Append("-S");
if ((Options & RegexOptions.IgnorePatternWhitespace) != 0) argSb.Append("-X");
if ((Options & RegexOptions.ECMAScript) != 0) argSb.Append("-E");

switch (Type)
{
Expand Down
20 changes: 3 additions & 17 deletions .../System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexParseException.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ namespace System.Text.RegularExpressions
[Serializable]
internal sealed class RegexParseException : ArgumentException
{
private readonly RegexParseError _error;
private readonly RegexParseError _error; // tests access this via private reflection

/// <summary>
/// The error that happened during parsing.
Expand All @@ -27,28 +27,14 @@ public RegexParseException(RegexParseError error, int offset, string message) :
Offset = offset;
}

public RegexParseException() : base()
{
}

public RegexParseException(string message) : base(message)
{
}

public RegexParseException(string message, Exception inner) : base(message, inner)
{
}

private RegexParseException(SerializationInfo info, StreamingContext context)
: base(info, context)
private RegexParseException(SerializationInfo info, StreamingContext context) : base(info, context)
{
}

public override void GetObjectData(SerializationInfo info, StreamingContext context)
{
base.GetObjectData(info, context);
// To maintain serialization support with netfx.
info.SetType(typeof(ArgumentException));
info.SetType(typeof(ArgumentException)); // To maintain serialization support with netfx.
}
}
}
Loading

0 comments on commit 95965c9

Please sign in to comment.