By avoiding the use of regular expressions, the Uize.Str.Whitespace module can achieve improved performance in performance critital applications such as parser implementations.

In addition to avoiding regular expressions, the methods of the Uize.Str.Whitespace module also achieve improved performnce by implementing an optimized handling for the special case of single character source strings that avoids looping.

The various index type methods of the Uize.Str.Whitespace module provide a more convenient and semantically elegant way of finding the index of whitespace or non-whitespace characters in a string.

To illustrate this, consider the following example of how an index could be obtained using a whitespace matcher regular expression versus using the Uize.Str.Whitespace module...

BEFORE

var regExp = /\s/g;
regExp.exec (sourceStr);
var whitespacePos = regExp.lastIndex - 1;

Using a regular expression, we have to create the regular expression and assign it to a local variable. Then, we call the exec method on the regular expression instance. Finally, we compute the index of the matched whitespace character by using the regular expression instance's lastIndex property. In order for this property to have a meaningful value, the regular expression instance must be created with the "g" flag.

All of this is not so intuitive. In contrast, using the Uize.Str.Whitespace.indexOfWhitespace static method produces a statement that is easy to read and make sense of...

AFTER

var whitespacePos = Uize.Str.Whitespace.indexOfWhitespace (sourceStr);

The index type methods of the Uize.Str.Whitespace module provide any easy and understandable way to specify a start position for a search for whitespace or non-whitespace characters.

Consider the following example of how a start position for a search can be achieved using regular expression versus using the Uize.Str.Whitespace module...

BEFORE

var regExp = /\s/g;
regExp.lastIndex = startPos;
regExp.exec (sourceStr);
var whitespacePos = regExp.lastIndex - 1;

Using a regular expression, we have to set the start position as the value for its lastIndex property before we call its exec method. Combining this with the other steps we need to perform, we end up with something that is far less elegant than just using the Uize.Str.Whitespace.indexOfWhitespace static and specifying the start position using the optional second argument...

AFTER

var whitespacePos = Uize.Str.Whitespace.indexOfWhitespace (sourceStr,startPos);

The Uize.Str.Whitespace.lastIndexOfWhitespace and Uize.Str.Whitespace.lastIndexOfNonWhitespace methods support backwards scanning to find the last whitespace or non-whitespace character in a source string.

This can be achieved with regular expressions by applying a bit of trickery, but there can be a performance cost. Consider the following example of how a last index of whitespace could be obtained using a regular expression versus using the Uize.Str.Whitespace module...

BEFORE

var regExp = /\s\S*$/g;
var match = regExp.exec (sourceStr);
var whitespacePos = match ? regExp.lastIndex - match [0].length : -1;

In order to achieve a backwards scan for the last whitespace character using a regular expression, we have to create a regular expression than matches a whitespace character, followed by any number of non-whitespace characters, and that is anchored to the end of the source string. Now, because our match could contain more than one character, we need to use the length of the first element in the match array to adjust the value of the lastIndex property.

We don't need to deal with this kind of trickery if we just use the dedicated Uize.Str.Whitespace.lastIndexOfWhitespace static method...

AFTER

var whitespacePos = Uize.Str.Whitespace.lastIndexOfWhitespace (sourceStr);

3.4.1. Backwards Scanning and Start Position

var regExp = /\s\S*$/g;
var match = regExp.exec (sourceStr.slice (0,startPos + 1));
var whitespacePos = match ? regExp.lastIndex - match [0].length : -1;

var whitespacePos = Uize.Str.Whitespace.lastIndexOfWhitespace (sourceStr,startPos);