2.11Regular Expression

Regular expressions are a powerful mean to search for patterns in strings and to extract substrings matching a certain criterion.

Falcon uses PCRE, or "PERL Compatible Regular Expression", version 7.6, as codebase for the Regex module. Scanning and matching international and wide strings is fully supported. Version 7.6 of PCRE offers also extended functionalities for some Unicode category, in example it can recognize also Unicode "wide" whitespaces. Discussion of usage of regular expression, and specifically, of PERL compatible regular expressions is beyond the scope of this guide. Further material on the topic can be found at:

The original codebase, with complete documentation, is at http:www.pcre.org.

Basic usage of Falcon Regular Expression module is quite simple. After importing the Regex module, one regular expression can be compiled using the Regex() class constructor, and then the instance can be used to match repeatedly strings.

The Falcon Regex module is based on concepts and pattern usages that are meant to minimize wastes and inefficiencies that are usually bound with scripting language usages. In example, it is quite common to extract substrings from the pattern using "parenthesized" expressions, that are actually subpatterns in parenthesis. But as parenthesis could be used to simply build recognition patterns, the substring they match is not always needed by the calling application. Other times, the calling application is more interested in knowing where a substring starts and ends rather than in its content. While other scripting languages build a whole set of matched substrings as the match happens, Falcon Regex module provides the user with set of range items that can be applied on the original string either to know where the matches took place or to extract a substring in one single VM step.

A minimal example would look like the following:


   load regex

   re = Regex( "find:\\s*(.*);" )  // Use "\\" to get "\", see below
   string = "this is something to find: the text or 言葉; (and this will be ignored)"

   if re.match( string )
      range = re.captured(1)
      printl( "Match was found at: ", re.captured(0) )
      printl( "Parenthesized capture: >", string[ range ], "<" )
   else
      printl( "No match" )
   end

The above code will create the re instance, and then will use it to search any text (.*) between "find:", eventually followed by any number of whitespaces (s*), and the first ";". The capture method will return the nth closed range that has been matched; the "0" range is the complete match, that is the match including "find" and ";". As any Falcon range, the match result can be cached in any variable (range in our example) and applied to the string later.

Please, notice the double "\" in the above Regex() constructor; Falcon string parsing would trasform s into a "silent whitespace", and would parse "n" as a physical new line, t as a tab and so on. Regular expression pattern may include many of those characters, giving them different meanings. For this reason, it is necessary to use the "\" escape where in regular expression documentation (and in user interfaces) a single "" is required. If you receive the string from an external source (i.e. a file or a user input), there is no need for this escape; this is just for Falcon compiler to produce a correct regular expression.

The above program will output:


   Match was found at: [21:42]
   Parenthesized capture: >the text or 言葉<

The parser provides single quoted literal strings, where the backslash has no special meaning.

In example:


   re = Regex( '^\s*\w' )

This would match any word near the beginning of a line, and it is equivalent to:


   re = Regex( "^\\s*\\w" )

Of course, the former is more readable and natural when dealing with regular expressions.

Contents of this module

Made with http://www.falconpl.org