The Problem
This situation occurs when using the .net 2.0 classes to transform XML into HTML with a style sheet the transform (System.Xml.Xsl.XslCompiledTransform). the output will include whitespace inside of certain HTML tags that may occasionally display within form fields when output. This can be a pain especially when displaying TextArea elements. In IE the behaviour appeared to change unexpectedly when some other element appeared before the TextArea – such as a dropdown list (select).
The Solution
A solution is to remove the whitespaces from these tags after it has been transformed.
private string CleanUpWhitespaceInTags(string xmlString)
{
return cleanUpWhitespaceInTagsRegEx.Replace(xmlString, cleanUpWhitespaceInTagsReplacement);
}
private static readonly Regex cleanUpWhitespaceInTagsRegEx = new Regex(@"<(?<tagName>[\w-]+)\b(?<attributes>[^>]*)>\s*(?<content>[^<]*?)\s*</\1>");
private const string cleanUpWhitespaceInTagsReplacement = "<${tagName}${attributes}>${content}</${tagName}>";
The above code uses a regular expression to match tags that have no inner xml/html and gets named captures of the tag name, it’s attributes and content (without any surrounding whitespace).
To break it down:
| < | match the opening tag’s starting brace |
| (?<tagName>[\w-]+) | capture any valid tag name (containing one or more word characters or dashes) |
| \b | check that we are at a word boundary |
| (?<attributes>[^>]*) | capture any characters after the tag name and before an end brace as attributes |
| > | match the opening tag’s end brace |
| \s* | match any whitespace on the left of the content |
| (?<content>[^<]*?) | capture any character that isn’t a starting brace but don’t be greedy |
| \s* | match any whitespace on the right of the content |
| </\1> | match the closing tag |
Disclaimer: Whether this regex covers all scenarios has not been tested but it works for the situation we encountered.
The replacement string then reconstitutes the tag, attributes and content without the leading or trailing whitespace.
