public void RegexStateMachine()
{
// Initialize unicode-related classes to avoid having their cost in the time computation
var sw = Stopwatch.StartNew();
var provider = new UnicodeCharSetProvider();
provider.GetClassSet(CharSetClass.Digit);
provider.GetClassSet(CharSetClass.Dot);
provider.GetClassSet(CharSetClass.Space);
provider.GetClassSet(CharSetClass.Word);
UnicodeRanges.FromUnicodeName("InCombining_Diacritical_Marks");
sw.Stop();
this.output.WriteLine("Unicode init ms: " + sw.ElapsedMilliseconds);
sw = Stopwatch.StartNew();
RegexLexer.CreateStateMachine(out var stateMachine, out var startStateId);
var lapBuild = sw.ElapsedTicks;
var sm = stateMachine.Compile();
var lapCompile = sw.ElapsedTicks;
var startState = new Id <DfaState <LetterId> >(startStateId);
sm(ref startState, ' ');
var lapJit = sw.ElapsedTicks;
sm(ref startState, ' ');
sw.Stop();
this.output.WriteLine("Build ticks: " + lapBuild);
this.output.WriteLine("Compile ticks: " + (lapCompile - lapBuild));
this.output.WriteLine("JIT ticks: " + (lapJit - lapCompile));
this.output.WriteLine("Exec ticks: " + (sw.ElapsedTicks - lapJit));
this.output.WriteLine("Total ms: " + sw.ElapsedMilliseconds);
this.output.WriteLine(stateMachine.ToReadableString());
}
public void CharacterInSet(string ch, bool inSet, string unicodeName)
{
Codepoint parsedChar;
if (ch.StartsWith("\\"))
{
Assert.True(((RangeSetHandle.Static)RegexMatchSet.ParseEscape(ch)).TryGetSingle(out parsedChar));
}
else
{
parsedChar = ch.Single();
}
Assert.Equal(inSet, UnicodeRanges.FromUnicodeName(unicodeName).Contains(parsedChar));
}
public static RangeSetHandle ParseEscape(string escape)
{
var match = rxEscape.Match(escape);
if (!match.Success)
{
throw new ArgumentException("Escape is invalid", "escape");
}
if (match.Groups["name"].Success)
{
return(new RangeSetHandle.Static(UnicodeRanges.FromUnicodeName(match.Groups["name"].Value), match.Groups["c"].Value == "P"));
}
if (match.Groups["hex"].Success)
{
return(new RangeSetHandle.Static(Codepoint.Parse(match.Groups["hex"].Value)));
}
var c = match.Groups["c"].Value[0];
switch (c)
{
case '0':
return(new RangeSetHandle.Static('\0'));
case 'r':
return(new RangeSetHandle.Static('\r'));
case 'n':
return(new RangeSetHandle.Static('\n'));
case 't':
return(new RangeSetHandle.Static('\t'));
case 'a':
return(new RangeSetHandle.Static('\x07'));
case 'e':
return(new RangeSetHandle.Static('\x1B'));
case 'f':
return(new RangeSetHandle.Static('\x0C'));
case 'v':
return(new RangeSetHandle.Static('\x0B'));
case 'd':
return(new RangeSetHandle.Class(CharSetClass.Digit, false));
case 'D':
return(new RangeSetHandle.Class(CharSetClass.Digit, true));
case 'w':
return(new RangeSetHandle.Class(CharSetClass.Word, false));
case 'W':
return(new RangeSetHandle.Class(CharSetClass.Word, true));
case 's':
return(new RangeSetHandle.Class(CharSetClass.Space, false));
case 'S':
return(new RangeSetHandle.Class(CharSetClass.Space, true));
default:
if (char.IsLetterOrDigit(c))
{
throw new ArgumentOutOfRangeException(nameof(escape), "Invalid escape character " + c);
}
return(new RangeSetHandle.Static(c));
}
}
public static RegexMatchSet FromUnicode(string name, bool negate = false)
{
return(new RegexMatchSet($@"\{(negate ? 'P' : 'p')}{{{name}}}", new RangeSetHandle.Static(UnicodeRanges.FromUnicodeName(name), negate)));
}
