regexpr.pas

delphi 正则表达例程。。。 正则表达式在实际应用中非常广泛

unit RegExpr;

{
     TRegExpr class library
     Delphi Regular Expressions

 Copyright (c) 1999-2004 Andrey V. Sorokin, St.Petersburg, Russia

 You may use this software in any kind of development,
 including comercial, redistribute, and modify it freely,
 under the following restrictions :
 1. This software is provided as it is, without any kind of
    warranty given. Use it at Your own risk.The author is not
    responsible for any consequences of use of this software.
 2. The origin of this software may not be mispresented, You
    must not claim that You wrote the original software. If
    You use this software in any kind of product, it would be
    appreciated that there in a information box, or in the
    documentation would be an acknowledgement like

     Partial Copyright (c) 2004 Andrey V. Sorokin
                                http://RegExpStudio.com
                                mailto:[e-mail]

 3. You may not have any income from distributing this source
    (or altered version of it) to other developers. When You
    use this product in a comercial package, the source may
    not be charged seperatly.
 4. Altered versions must be plainly marked as such, and must
    not be misrepresented as being the original software.
 5. RegExp Studio application and all the visual components as 
    well as documentation is not part of the TRegExpr library 
    and is not free for usage.

                                    mailto:[e-mail]
                                    http://RegExpStudio.com
                                    http://anso.da.ru/
}

interface

// ======== Determine compiler
{$IFDEF VER80} Sorry, TRegExpr is for 32-bits Delphi only. Delphi 1 is not supported (and whos really care today?!). {$ENDIF}
{$IFDEF VER90} {$DEFINE D2} {$ENDIF} // D2
{$IFDEF VER93} {$DEFINE D2} {$ENDIF} // CPPB 1
{$IFDEF VER100} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // D3
{$IFDEF VER110} {$DEFINE D4} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // CPPB 3
{$IFDEF VER120} {$DEFINE D4} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // D4
{$IFDEF VER130} {$DEFINE D5} {$DEFINE D4} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // D5
{$IFDEF VER140} {$DEFINE D6} {$DEFINE D5} {$DEFINE D4} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // D6
{$IFDEF VER150} {$DEFINE D7} {$DEFINE D6} {$DEFINE D5} {$DEFINE D4} {$DEFINE D3} {$DEFINE D2} {$ENDIF} // D7

// ======== Define base compiler options
{$BOOLEVAL OFF}
{$EXTENDEDSYNTAX ON}
{$LONGSTRINGS ON}
{$OPTIMIZATION ON}
{$IFDEF D6}
  {$WARN SYMBOL_PLATFORM OFF} // Suppress .Net warnings
{$ENDIF}
{$IFDEF D7}
  {$WARN UNSAFE_CAST OFF} // Suppress .Net warnings
  {$WARN UNSAFE_TYPE OFF} // Suppress .Net warnings
  {$WARN UNSAFE_CODE OFF} // Suppress .Net warnings
{$ENDIF}
{$IFDEF FPC}
 {$MODE DELPHI} // Delphi-compatible mode in FreePascal
{$ENDIF}

// ======== Define options for TRegExpr engine
{.$DEFINE UniCode} // Unicode support
{$DEFINE RegExpPCodeDump} // p-code dumping (see Dump method)
{$IFNDEF FPC} // the option is not supported in FreePascal
 {$DEFINE reRealExceptionAddr} // exceptions will point to appropriate source line, not to Error procedure
{$ENDIF}
{$DEFINE ComplexBraces} // support braces in complex cases
{$IFNDEF UniCode} // the option applicable only for non-UniCode mode
 {$DEFINE UseSetOfChar} // Significant optimization by using set of char
{$ENDIF}
{$IFDEF UseSetOfChar}
 {$DEFINE UseFirstCharSet} // Fast skip between matches for r.e. that starts with determined set of chars
{$ENDIF}

// ======== Define Pascal-language options
// Define 'UseAsserts' option (do not edit this definitions).
// Asserts used to catch 'strange bugs' in TRegExpr implementation (when something goes
// completely wrong). You can swith asserts on/off with help of {$C+}/{$C-} compiler options.
{$IFDEF D3} {$DEFINE UseAsserts} {$ENDIF}
{$IFDEF FPC} {$DEFINE UseAsserts} {$ENDIF}

// Define 'use subroutine parameters default values' option (do not edit this definition).
{$IFDEF D4} {$DEFINE DefParam} {$ENDIF}

// Define 'OverMeth' options, to use method overloading (do not edit this definitions).
{$IFDEF D5} {$DEFINE OverMeth} {$ENDIF}
{$IFDEF FPC} {$DEFINE OverMeth} {$ENDIF}

uses
 Classes,  // TStrings in Split method
 SysUtils; // Exception

type
 {$IFDEF UniCode}
 PRegExprChar = PWideChar;
 RegExprString = WideString;
 REChar = WideChar;
 {$ELSE}
 PRegExprChar = PChar;
 RegExprString = AnsiString; //###0.952 was string
 REChar = Char;
 {$ENDIF}
 TREOp = REChar; // internal p-code type //###0.933
 PREOp = ^TREOp;
 TRENextOff = integer; // internal Next "pointer" (offset to current p-code) //###0.933
 PRENextOff = ^TRENextOff; // used for extracting Next "pointers" from compiled r.e. //###0.933
 TREBracesArg = integer; // type of {m,n} arguments
 PREBracesArg = ^TREBracesArg;

const
 REOpSz = SizeOf (TREOp) div SizeOf (REChar); // size of p-code in RegExprString units
 RENextOffSz = SizeOf (TRENextOff) div SizeOf (REChar); // size of Next 'pointer' -"-
 REBracesArgSz = SizeOf (TREBracesArg) div SizeOf (REChar); // size of BRACES arguments -"-

type
 TRegExprInvertCaseFunction = function (const Ch : REChar) : REChar
                               of object;

const
  EscChar = '\'; // 'Escape'-char ('\' in common r.e.) used for escaping metachars (\w, \d etc).
  RegExprModifierI : boolean = False;    // default value for ModifierI
  RegExprModifierR : boolean = True;     // default value for ModifierR
  RegExprModifierS : boolean = True;     // default value for ModifierS
  RegExprModifierG : boolean = True;     // default value for ModifierG
  RegExprModifierM : boolean = False;    // default value for ModifierM
  RegExprModifierX : boolean = False;    // default value for ModifierX
  RegExprSpaceChars : RegExprString =    // default value for SpaceChars
  ' '#$9#$A#$D#$C;
  RegExprWordChars : RegExprString =     // default value for WordChars
    '0123456789' //###0.940
  + 'abcdefghijklmnopqrstuvwxyz'
  + 'ABCDEFGHIJKLMNOPQRSTUVWXYZ_';
  RegExprLineSeparators : RegExprString =// default value for LineSeparators
   #$d#$a{$IFDEF UniCode}+#$b#$c#$2028#$2029#$85{$ENDIF}; //###0.947
  RegExprLinePairedSeparator : RegExprString =// default value for LinePairedSeparator
   #$d#$a;
  { if You need Unix-styled line separators (only \n), then use:
  RegExprLineSeparators = #$a;
  RegExprLinePairedSeparator = '';
  }


const
 NSUBEXP = 15; // max number of subexpression //###0.929
 // Cannot be more than NSUBEXPMAX
 // Be carefull - don't use values which overflow CLOSE opcode
 // (in this case you'll get compiler erorr).
 // Big NSUBEXP will cause more slow work and more stack required
 NSUBEXPMAX = 255; // Max possible value for NSUBEXP. //###0.945
 // Don't change it! It's defined by internal TRegExpr design.

 MaxBracesArg = $7FFFFFFF - 1; // max value for {n,m} arguments //###0.933

 {$IFDEF ComplexBraces}
 LoopStackMax = 10; // max depth of loops stack //###0.925
 {$ENDIF}

 TinySetLen = 3;
 // if range includes more then TinySetLen chars, //###0.934
 // then use full (32 bytes) ANYOFFULL instead of ANYOF[BUT]TINYSET
 // !!! Attension ! If you change TinySetLen, you must
 // change code marked as "//!!!TinySet"


type

{$IFDEF UseSetOfChar}
 PSetOfREChar = ^TSetOfREChar;
 TSetOfREChar = set of REChar;
{$ENDIF}

 TRegExpr = class;

 TRegExprReplaceFunction = function (ARegExpr : TRegExpr): string
                               of object;

 TRegExpr = class
   private
    startp : array [0 .. NSUBEXP - 1] of PRegExprChar; // founded expr starting points
    endp : array [0 .. NSUBEXP - 1] of PRegExprChar; // founded expr end points

    {$IFDEF ComplexBraces}
    LoopStack : array [1 .. LoopStackMax] of integer; // state before entering loop
    LoopStackIdx : integer; // 0 - out of all loops
    {$ENDIF}

    // The "internal use only" fields to pass info from compile
    // to execute that permits the execute phase to run lots faster on
    // simple cases.