org.apache.royale.compiler.internal.mxml.MXMLTextData Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of compiler Show documentation
Show all versions of compiler Show documentation
The Apache Royale Compiler
/*
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*/
package org.apache.royale.compiler.internal.mxml;
import static org.apache.royale.compiler.constants.IMXMLCoreConstants.*;
import java.util.Collection;
import org.apache.royale.compiler.common.ISourceLocation;
import org.apache.royale.compiler.internal.parsing.ISourceFragment;
import org.apache.royale.compiler.internal.parsing.SourceFragment;
import org.apache.royale.compiler.internal.parsing.mxml.MXMLToken;
import org.apache.royale.compiler.mxml.IMXMLTextData;
import org.apache.royale.compiler.parsing.MXMLTokenTypes;
import org.apache.royale.compiler.problems.ICompilerProblem;
/**
* Encapsulation of a block of text in MXML
*/
public class MXMLTextData extends MXMLUnitData implements IMXMLTextData
{
/**
* Constructor.
*/
public MXMLTextData(MXMLToken textToken)
{
text = textToken.getText();
setOffsets(textToken.getStart(), textToken.getEnd());
setLine(textToken.getLine());
setColumn(textToken.getColumn());
setEndLine(textToken.getEndLine());
setEndColumn(textToken.getEndColumn());
switch (textToken.getType())
{
case MXMLTokenTypes.TOKEN_TEXT:
{
type = TextType.TEXT;
break;
}
case MXMLTokenTypes.TOKEN_WHITESPACE:
{
type = TextType.WHITESPACE;
break;
}
case MXMLTokenTypes.TOKEN_CDATA:
{
type = TextType.CDATA;
break;
}
case MXMLTokenTypes.TOKEN_COMMENT:
{
type = TextType.COMMENT;
break;
}
case MXMLTokenTypes.TOKEN_ASDOC_COMMENT:
{
type = TextType.ASDOC;
break;
}
}
}
/**
* The represented text
*/
private String text;
private TextType type;
//
// Object overrides
//
// For debugging only. This format is nice in the Eclipse debugger.
@Override
public String toString()
{
StringBuilder sb = new StringBuilder();
// Display TEXT, WHITESPACE, etc.
sb.append(getTextType());
sb.append(' ');
// Display the text characters.
sb.append('"');
sb.append(getEscapedContent());
sb.append('"');
sb.append(' ');
// Display line, column, start, and end as "17:5 160-188".
sb.append(super.toString());
return sb.toString();
}
//
// MXMLUnitData overrides
//
@Override
public boolean isText()
{
return true;
}
// For debugging only.
@Override
public String getTypeString()
{
StringBuilder sb = new StringBuilder();
sb.append(super.getTypeString());
sb.append(":");
sb.append(getTextType());
return sb.toString();
}
// For debugging only. This format is nice in a text file.
@Override
public String toDumpString()
{
StringBuilder sb = new StringBuilder();
sb.append(super.toDumpString());
sb.append('\t');
sb.append('|');
sb.append(getEscapedContent());
sb.append('|');
return sb.toString();
}
//
// IMXMLTextData implementations
//
@Override
public String getContent()
{
return text;
}
@Override
public TextType getTextType()
{
return type;
}
//
// Other methods
//
// For debugging only.
private String getEscapedContent()
{
String s = getContent();
s = s.replaceAll("\n", "\\\\n");
s = s.replaceAll("\r", "\\\\r");
s = s.replaceAll("\t", "\\\\t");
return s;
}
@Override
public String getCompilableText()
{
switch (type)
{
case TEXT:
case CDATA:
case WHITESPACE:
{
return getContents();
}
case ASDOC:
case COMMENT:
{
return "";
}
}
assert false;
return null;
}
@Override
public int getCompilableTextStart()
{
return getContentsStart();
}
@Override
public int getCompilableTextEnd()
{
return getContentsEnd();
}
@Override
public int getCompilableTextLine()
{
return getLine();
}
@Override
public int getCompilableTextColumn()
{
return (getContentsStart() - getAbsoluteStart()) + getColumn();
}
/**
* Return the text contained within this unit, without any opening or
* closing delimiters
*
* @return a normalized string
*/
public String getContents()
{
String tokenString = text;
switch (type)
{
case CDATA:
{
if (tokenString.endsWith(cDataEnd))
{
tokenString = tokenString.substring(cDataStart.length(), tokenString.length() - cDataEnd.length());
}
else
{
tokenString = tokenString.substring(cDataStart.length());
}
return tokenString;
}
case ASDOC:
{
if (tokenString.endsWith(asDocEnd))
{
tokenString = tokenString.substring(asDocStart.length(), tokenString.length() - asDocEnd.length());
}
else
{
tokenString = tokenString.substring(asDocStart.length());
}
return tokenString;
}
case COMMENT:
{
if (tokenString.endsWith(commentEnd))
{
tokenString = tokenString.substring(commentStart.length(), tokenString.length() - commentEnd.length());
}
else
{
tokenString = tokenString.substring(commentStart.length());
}
return tokenString;
}
default:
{
break;
}
}
return text;
}
/**
* @return The start offset of actual content
*/
@SuppressWarnings("incomplete-switch")
public int getContentsStart()
{
switch (type)
{
case CDATA:
return getAbsoluteStart() + cDataStart.length();
case ASDOC:
return getAbsoluteStart() + asDocStart.length();
case COMMENT:
return getAbsoluteStart() + commentStart.length();
}
return getAbsoluteStart();
}
/**
* @return The end offset of content
*/
@SuppressWarnings("incomplete-switch")
public int getContentsEnd()
{
switch (type)
{
case CDATA:
return text.endsWith(cDataEnd) ? getAbsoluteEnd() - cDataEnd.length() : getAbsoluteEnd();
case ASDOC:
return text.endsWith(asDocEnd) ? getAbsoluteEnd() - asDocEnd.length() : getAbsoluteEnd();
case COMMENT:
return text.endsWith(commentEnd) ? getAbsoluteEnd() - commentEnd.length() : getAbsoluteEnd();
}
return getAbsoluteEnd();
}
public int getContentsLine()
{
return getLine();
}
@SuppressWarnings("incomplete-switch")
public int getContentsColumn()
{
switch (type)
{
case CDATA:
return getColumn() + cDataStart.length();
case ASDOC:
return getColumn() + asDocStart.length();
case COMMENT:
return getColumn() + commentStart.length();
}
return getColumn();
}
@Override
public ISourceFragment[] getFragments(Collection problems)
{
ISourceLocation location = this;
switch (type)
{
case TEXT:
{
// TEXT might contain one or more entities,
// in which case we will return multiple fragments.
MXMLDialect mxmlDialect = getMXMLDialect();
return EntityProcessor.parse(text, location, mxmlDialect, problems);
}
case WHITESPACE:
{
// WHITESPACE creates only a single fragment.
ISourceFragment fragment = new SourceFragment(text, location);
return new ISourceFragment[] {fragment};
}
case CDATA:
{
// CDATA creates only a single fragment.
ISourceFragment fragment = new SourceFragment(text, getContents(), getContentsStart(), getContentsLine(), getContentsColumn());
return new ISourceFragment[] {fragment};
}
default:
{
assert false : "Unexpected type of MXMLTextData";
}
}
return null;
}
}