Sign in Join
Wiki copytemporary

RegEx packages

Bahasa Indonesia · Deutsch · English · Español · Polski

Free Pascal includes the RegExpr package, which includes several "engines" for regular expressions. The main of these engines is TRegExpr, others are deprecated and not developed anymore.

Read the Wikipedia page about regular expressions to get introduced to the subject: POSIX Basic Regular Expressions.

TRegExpr by Sorokin

This is the most complete implementation, it is available in FPC in the file "packages/regexpr/src/regexpr.pas". This package implements a subset of Perl's regular expressions and the syntax provided by the package is documented here.

Official repository: GitHub. The development is going on there, FPC package is following this repo with some delay. Version from this repo works OK in FPC.

Improvements in 2019/2020

In 2019 year, after about 15 years of missed patches from anybody, work on TRegExpr was continued by Alexey Torgashin, User:Alextp. These changes were merged to FPC Git 'main' branch.

Changes are:

Improvements in 2023

In 2023, Martin Friebe (SynEdit maintainer) made lot of good changes. These changes were also merged to FPC TRegExpr fork.

And from Alexey Torgashin:

Change case on replaces

You can change case of found fragments, use modifiers in replace-with field (case change works after position of modifier):

E.g. if found a word, use replace-with field "\L$0" to change word to lowercase (here $0 is group 0, found text).

In addition \n will be replaced with the line-break character(s) (LF or CR LF, this depends on internal constant).

Example

Using TRegExpr to check if an expression is present in a string is very easy, just create an instance of TRegExpr, then place your regular expression in the property TRegExpr.Expression and use the method Exec to verify if there are any matches for this regular expression. Exec will return true if the expression matches the string passed to it.

pascal
var
  RegexObj: TRegExpr;
begin
  RegexObj := TRegExpr.Create;
  RegexObj.Expression := '.*login.*';
  if RegexObj.Exec('Please try to login here') then WriteLn('The login was found!');
  RegexObj.Free;
end;

Subexpression match:

pascal
program Project1;

uses
  RegExpr;

var
  re: TRegExpr;
begin
  re := TRegExpr.Create('hello (.*?)!');
  if re.Exec('hello world! hello pascal!') then
  begin
    WriteLn(re.Match[1]);
    while re.ExecNext do
    begin
      WriteLn(re.Match[1]);
    end;
  end;
  re.Free;
end.

Output:

pascal
world
pascal

FLRE - Fast Light Regular Expressions

FLRE (Fast Light Regular Expressions) is a fast, safe and efficient regular expression library, which is implemented in Object Pascal (Delphi and Free Pascal) but which is even usable from other languages like C/C++ and so on. It can handle Unicode and UTF-8 strings.

It implements the many of the most common Perl and POSIX features, except irregular expression features like forward references and nested back references and so on, which aren't supported at FLRE, only real "back" references are supported, hence also the word "Light" at the FLRE name. It also finds the leftmost-first match, the same match that Perl and PCRE would, and can return submatch information. But it also features a flag for a yet experimental POSIX-style leftmost-longest match behaviour mode.

FLRE is licensed under the LGPL v2.1 with static-linking-exception.

GitHub repository

Example

Implementing PHP parse_url function (based on php.js parse_url implementation)

pascal

uses Classes, SysUtils, flre;

type

  RUrlParser = record
  private
    Fcomponents: array[1..13] of string;
  private
    function getComponent( aIndex : integer ) : string;
    procedure setComponent( aIndex : integer; const aValue : string );
  public
    function parse( const aUrl : UTF8String ) : boolean;
  public
    property scheme : string index 1 read getComponent write setComponent; // e.g. http
    property authority : string index 2 read getComponent;
    property userInfo : string index 3 read getComponent;
    property user : string index 4 read getComponent;
    property pass : string index 5 read getComponent;
    property host : string index 6 read getComponent;
    property port : string index 7 read getComponent;
    property path : string index 9 read getComponent;
    property directory : string index 10 read getComponent;
    property fileName : string index 11 read getComponent;
    property query : string  index 12 read getComponent; // after the question mark ?
    property fragment : string  index 13 read getComponent; // after the hashmark #
  end;

implementation

function RUrlParser.getComponent( aIndex : integer ) : string;
begin
    Result := Fcomponents[ aIndex ];
end;

procedure RUrlParser.setComponent( aIndex : integer; const aValue : string );
begin
    Fcomponents[ aIndex ] := aValue;
end;

function RUrlParser.parse( const aUrl : UTF8String ) : boolean; overload;
var
    i : integer;
    re : TFLRE;
    parts : TFLREMultiStrings;
begin
    re := TFLRE.Create( '(?:([^:\\/\?#]+):)?'
        + '(?:\/\/()(?:(?:()(?:([^:@\/]*):?([^:@\/]*))?@)?([^:\/?#]*)(?::(\d*))?))?'
        + '()'
        + '(?:(()(?:(?:[^?#\/]*\/)*)()(?:[^?#]*))(?:\?([^#]*))?(?:#(.*))?)'
        , [ rfUTF8 ]
    ); 

    parts := nil;
    Result := re.UTF8ExtractAll( aUrl, parts );
    if ( Result ) then
    begin
        for i := 1 to Length( parts[0] ) - 1 do
        begin
            setComponent( i, string( parts[0][i] ) );
        end;
    end;

    // Free regexp memory
    for i := 0 to Length( parts ) - 1 do
    begin
        SetLength( parts[i], 0 );
    end;
    parts := nil;

    re.Free();
end;

Regexpr by Joost

Regexpr by Joost (units oldregexpr.pp and regex.pp) is a very basic regex (Regular Expression) unit, it handles most regular expressions as GNU regexpr.

The current unit is far from complete, and still misses very simple syntax support of POSIX or more complex syntax such as Perl regex, Java Regex, Ruby Regex etc...

The unit contains 4 functions for now:

There is also one test:

Regexpr by Florian

This is the oldest implementation. It is present in packages/regexpr/src/old and is not currently compiled by the makefiles, so it is not available precompiled by default in released FPC/Lazarus versions.

See also

Go to back Packages List

From the Free Pascal wiki, page "RegEx packages", as it stood on 2023-11-24 after a change by Alextpp, out of the wiki's dump of 2024-05-26. Copyright stays with the author who wrote it, according to the old wiki's Licensing page. If you have concerns about this material, message the admin.

Revision 1, last changed by WikiArchive .

Talk about this page

Nobody has opened a bench for this page yet. Sign in to open one.