Regular expressions.

Regular expressions.

Regular expressions are used to describe character patterns in strings. So you can look for all paragraphs in a tagged text no matter that style the paragraph uses:

<ParaStyle:Style1>...
<ParaStyle:Stil1>...
<ParaStyle:>
<ParaStyle:2\>3>

Search    <ParaStyle:(<0x[0-9A-F]{4,4}>|\\\\>|\\\\<|[^<>])*>

You can use regular expression in the following situations:

• strreplace
• strstr
• strstrpos
• Text layout condition Contains
• Text layout rule Search and replace

• GNU conform regular expressions
• PCRE, (Perl compatible regular expressions). since v3.3.1 R4000 and at least CS5

Since v3.4 R9000 and CS5 regulare expresssions are parsed by PCRE only. Only CS4 is using the old implementation for GNU compatible REs.

GNU-conform regular expressions supporting the base functionality of regular expressions ((Characters and substrings, Counter, Word boundaries und Sub expressions).

• Lookarounds
• Conditions
• Unicode characters ([äöü]{0,7} or \x{00E6}) The \u00E6 syntax is not supported by PCRE.
• Unicode Character Properties (\p{Ll} or \p{Thai})
• Modifiers (z.B. (?i))
• Comments

If you not familar with regular expressions - you can find a lot of descriptions and examples in the net. If you want learn something about the theorie of regular expression, search for Formlar languages, regular expressions are a subfamily of the formal languages.

An excellent description of PCRE you can find here..
Extremely helpful is the Online regex tester..

To give strreplace and strstrpos a hint, that the search string is a regular expression, use the prefixes

    regexp: or pcre:

Since v3.4 R9000 and CS5 the prefix regexp: points to PCRE too.

Characters and substrings in regular expressions are expressed by itself.

Search all 1234's in a given string

456 1234 6748 441234567 64641329 4321 4321 999

Search   1234

Of course, a simple search can solve this problem too. Things getting harder if you allow any order of the digits 1, 2, 3, 4. (You may search for the 24 combinations ...)

Look for all occurrances of 1234 in any order of the digits:

456 1234 6748 441234567 64641329 4321 999

Search    [1234][1234][1234][1234]

Every [] expression matches exactly one of the characters inside the brackets. The brackets may contain any number of Ascii characters or Ascii character ranges. So [a-zA-Z_] will mach any letter and the Underscore. Use [^...] if the character shall not match. [^ \t\r\n] will match any character except the blank, the tab and the line delimiters.

Regular expression may find any UTF8 character. But the expression itself must not contain Ascii characters greater 127 like ä, ö, ü.

If you want look for letters used to describe a regular expressions, use the backslash to escape it.
And take care to escpae this backslash in cScript string itself!

Using character ranges in the above expression, will make thing not really better:

Look for all occurrances of 1234 in any order of the digits:

456 1234 6748 441234567 64641329 4321 999

Search    [1-4][1-4][1-4][1-4]

Even counters will lead to better expressions :

Using counters the above expression looks well.

Look for all occurrances of 1234 in any order of the digits:

456 1234 6748 441234567 64641329 4321 999

Search    [1-4]{4}

With \b you can describe word boundaries. If you want find word onl in the above expression, simple add \b at the beginning and the end of the expression.

Look for all words of 1234 in any order of the digits:

456 1234 6748 441234567 64641329 4321 999

\b[1-4]{4}\b

Any part of a regular expression can set in parenthesis. You can refer to this sub parts of the expression using the following descriptors:

Find all 3-digit substrings enclosed by letters and swap this letters:

000v5664w358x00l345m50v523w1f789g6040h928i01
000v5664x358w00m345l50w523v1g789f6040i928h01

Search    ([a-z])([1-9]{3})([a-z])
Replace   \3\2\1

Find a words containing a r and replace it by the same word but in upper cases:

Search   \b([^[:space:][:punct:]]*r[^[:space:][:punct:]]*)\b
Replace  \u0

Replace all paragraph tags of a tagged text by a HTML comment including the upper cased style name.

Serach   <ParaStyle:((([^\>])|(\\>))*)>
Replace  <--\u1-->

Look for all tags of a tagged text :

Search    (<[a-zA-Z][a-zA-Z0-9_]*:(<)?)((([^><])|(\\>)|(\\<))*)(>)*

PCRE handles the search for regular expressions in three steps:

1. Compile (pcre_compile) : This step compiles a regular expression into an internal form.
2. Study (pcre_study) : This step studies a compiled pattern, to see if additional information can be extracted that might speed up matching.
3. Execute (pcre_exec) : This step matches a compiled regular expression against a given subject string, using a matching algorithm that is similar to Perl's.

Each of the three steps, that automatically executed for every search for regular expressions using PCRE, can receive additional options. In the following you can see the available options. For more information about the options, please search the web for the option names (e.g., PCRE_BSR_ANYCRLF).

The option names are not defined in cScript, please use the corresponding numbers. To use several options, you can add the numbers by | (logical or). The options are used exclusively for the processing regular expressions with PCRE.