Description Chander Prakash 2008-06-06 05:53:12 EDT

Build ID: 3.4.0

Steps To Reproduce:

Code:

public final ITextRegion getNextToken() throws IOException {
		
//			text = new StringBuffer(yytext());
			start = yychar;
			textLength = length = yylength();
		}

		if (context != null) {
			if (context == CSS_JSP_SCRIPTLET || context == CSS_JSP_EL){
				nextTokenType = primGetNextToken();
				while (nextTokenType != CSS_JSP_END && nextTokenType != CSS_EL_END && nextTokenType != CSS_JSP_COMMENT) {
//					text.append(yytext());
					textLength += yylength();
					length = textLength;
					



Assume there are 3 lines in the editor that uses JSPedCSSSourceParser [styles.jsp]

(Line1) <%@ page session="false" contentType="text/css;charset=ISO-8859-1" %>
(Line2) <% pageContext.getRequest().setCharacterEncoding("UTF-8"); %>
(Line3) <%@ include file="./styles_fragment.jspf" %>

If you edit line2, then the below hierarchy is called


Hierarchy

CSSStructuredDocumentReParser(StructuredDocumentReParser).reparse(IStructuredDocumentRegion, IStructuredDocumentRegion)->CSSStructuredDocumentReParser(StructuredDocumentReParser).core_reparse(int, int, CoreNodeList, boolean) >CSSStructuredDocumentReParser(StructuredDocumentReParser)._core_reparse_text(int, int) ->JSPedCSSSourceParser(CSSSourceParser).getDocumentRegions()->JSPedCSSSourceParser(CSSSourceParser).parseNodes()->JSPedCSSSourceParser(CSSSourceParser).getNextRegion() ->JSPedCSSTokenizer.getNextToken() 	



Why node/region calculation fails

CSSStructuredDocumentReParser(StructuredDocumentReParser).core_reparse(int, int, CoreNodeList, boolean) in newNodesHead it is having <% pageContext.getRequest().setCharacterEncoding("UTF-8"); %>
<%@ include file="./styles_fragment.jspf" %> both the lines as a single node and its nextnode is null. From above hierarchy you can see that finally it calls JSPedCSSTokenizer.getNextToken() method 
In this case value of  nextToeknType it does not change its value to CSS_JSP_END when it encounter with “%>” character, Instead of this it start counting the character it goes to 3 lines, when it get “%>” character in 3rd line and it value of nextToeknType become CSS_JSP_END and it consider both the lines as a single node. So in this case value of newNodesHead is <% pageContext.getRequest().setCharacterEncoding("UTF8"); %>
<%@ include file="./styles_fragment.jspf" %>
 As a single node instead of considering it two different node.

 

Why node/region calculation passes if line2 is modified to have @ or =

 But if we modify 2nd line and make it either <%@ pageContext.getRequest().setCharacterEncoding("UTF-8"); %> or <% int i=0 %> then newNodesHead is consider it as a single node and its next node is <%@ include file="./styles_fragment.jspf" %>. This problem came from fStructuredDocument.getParser().getDocumentRegions From above hierarchy you can see that finally it calls JSPedCSSTokenizer.getNextToken() method and value of nextTokenType start with CSS_JSP_Directive and it start counting the characters and when it reaches to “%>” [end of line] character, value of nextTokenType become CSS_JSP_End  and it consider it as a different node. This is because of presence of either ‘=’ or ‘@’ . Again it start with CSS_JSP_DIRECTIVE for 3rd line and end with CSS_JSP_END when it reaches “%>”character and also it store it as a different node so in this case it is having 2 nodes in newNodesHead.




What is our expectation.

When we include 2nd line it is not able to recognize the Styles those are applied on styles_fragment.jspf (3rd line) file on current file styles.jsp.
Our expectation is when we include second line it should able to recognize the styles those are applies in 3rd lines. Or It should consider second line and third line as a two separate nodes so that it can apply styles of third line on current file (in which we have included these lines). This will happen if just ‘%>’ is considered to be causing the end of CSS_JSP_END.