FileDocCategorySizeDatePackage
ParserController.javaAPI DocApache Tomcat 6.0.1422047Fri Jul 20 04:20:34 BST 2007org.apache.jasper.compiler

ParserController

public class ParserController extends Object implements TagConstants
Controller for the parsing of a JSP page.

The same ParserController instance is used for a JSP page and any JSP segments included by it (via an include directive), where each segment may be provided in standard or XML syntax. This class selects and invokes the appropriate parser for the JSP page and its included segments.

author
Pierre Delisle
author
Jan Luehe

Fields Summary
private static final String
CHARSET
private org.apache.jasper.JspCompilationContext
ctxt
private Compiler
compiler
private ErrorDispatcher
err
private boolean
isXml
private Stack
baseDirStack
private boolean
isEncodingSpecifiedInProlog
private boolean
isBomPresent
private int
skip
private String
sourceEnc
private boolean
isDefaultPageEncoding
private boolean
isTagFile
private boolean
directiveOnly
Constructors Summary
public ParserController(org.apache.jasper.JspCompilationContext ctxt, Compiler compiler)


    /*
     * Constructor
     */
         
        this.ctxt = ctxt; 
        this.compiler = compiler;
        this.err = compiler.getErrorDispatcher();
Methods Summary
private voiddetermineSyntaxAndEncoding(java.lang.String absFileName, java.util.jar.JarFile jarFile, java.lang.String jspConfigPageEnc)
Determines the syntax (standard or XML) and page encoding properties for the given file, and stores them in the 'isXml' and 'sourceEnc' instance variables, respectively.


        isXml = false;

        /*
         * 'true' if the syntax (XML or standard) of the file is given
         * from external information: either via a JSP configuration element,
         * the ".jspx" suffix, or the enclosing file (for included resources)
         */
        boolean isExternal = false;

        /*
         * Indicates whether we need to revert from temporary usage of
         * "ISO-8859-1" back to "UTF-8"
         */
        boolean revert = false;

        JspConfig jspConfig = ctxt.getOptions().getJspConfig();
        JspConfig.JspProperty jspProperty = jspConfig.findJspProperty(
                absFileName);
        if (jspProperty.isXml() != null) {
            // If <is-xml> is specified in a <jsp-property-group>, it is used.
            isXml = JspUtil.booleanValue(jspProperty.isXml());
            isExternal = true;
        } else if (absFileName.endsWith(".jspx")
                || absFileName.endsWith(".tagx")) {
            isXml = true;
            isExternal = true;
        }

        if (isExternal && !isXml) {
            // JSP (standard) syntax. Use encoding specified in jsp-config
            // if provided.
            sourceEnc = jspConfigPageEnc;
            if (sourceEnc != null) {
                return;
            }
            // We don't know the encoding, so use BOM to determine it
            sourceEnc = "ISO-8859-1";
        } else {
            // XML syntax or unknown, (auto)detect encoding ...
            Object[] ret = XMLEncodingDetector.getEncoding(absFileName,
                    jarFile, ctxt, err);
            sourceEnc = (String) ret[0];
            if (((Boolean) ret[1]).booleanValue()) {
                isEncodingSpecifiedInProlog = true;
            }
            if (((Boolean) ret[2]).booleanValue()) {
                isBomPresent = true;
            }
            skip = ((Integer) ret[3]).intValue();

            if (!isXml && sourceEnc.equals("UTF-8")) {
                /*
                 * We don't know if we're dealing with XML or standard syntax.
                 * Therefore, we need to check to see if the page contains
                 * a <jsp:root> element.
                 *
                 * We need to be careful, because the page may be encoded in
                 * ISO-8859-1 (or something entirely different), and may
                 * contain byte sequences that will cause a UTF-8 converter to
                 * throw exceptions. 
                 *
                 * It is safe to use a source encoding of ISO-8859-1 in this
                 * case, as there are no invalid byte sequences in ISO-8859-1,
                 * and the byte/character sequences we're looking for (i.e.,
                 * <jsp:root>) are identical in either encoding (both UTF-8
                 * and ISO-8859-1 are extensions of ASCII).
                 */
                sourceEnc = "ISO-8859-1";
                revert = true;
            }
        }

        if (isXml) {
            // (This implies 'isExternal' is TRUE.)
            // We know we're dealing with a JSP document (via JSP config or
            // ".jspx" suffix), so we're done.
            return;
        }

        /*
         * At this point, 'isExternal' or 'isXml' is FALSE.
         * Search for jsp:root action, in order to determine if we're dealing 
         * with XML or standard syntax (unless we already know what we're 
         * dealing with, i.e., when 'isExternal' is TRUE and 'isXml' is FALSE).
         * No check for XML prolog, since nothing prevents a page from
         * outputting XML and still using JSP syntax (in this case, the 
         * XML prolog is treated as template text).
         */
        JspReader jspReader = null;
        try {
            jspReader = new JspReader(ctxt, absFileName, sourceEnc, jarFile,
                    err);
        } catch (FileNotFoundException ex) {
            throw new JasperException(ex);
        }
        jspReader.setSingleFile(true);
        Mark startMark = jspReader.mark();
        if (!isExternal) {
            jspReader.reset(startMark);
            if (hasJspRoot(jspReader)) {
                if (revert) {
                    sourceEnc = "UTF-8";
                }
                isXml = true;
                return;
            } else {
                if (revert && isBomPresent) {
                    sourceEnc = "UTF-8";
                }
                isXml = false;
            }
        }

        /*
         * At this point, we know we're dealing with JSP syntax.
         * If an XML prolog is provided, it's treated as template text.
         * Determine the page encoding from the page directive, unless it's
         * specified via JSP config.
         */
        if (!isBomPresent) {
            sourceEnc = jspConfigPageEnc;
            if (sourceEnc == null) {
                sourceEnc = getPageEncodingForJspSyntax(jspReader, startMark);
                if (sourceEnc == null) {
                    // Default to "ISO-8859-1" per JSP spec
                    sourceEnc = "ISO-8859-1";
                    isDefaultPageEncoding = true;
                }
            }
        }
private Node.NodesdoParse(java.lang.String inFileName, Node parent, java.net.URL jarFileUrl)
Parses the JSP page or tag file with the given path name.

param
inFileName The name of the JSP page or tag file to be parsed.
param
parent The parent node (non-null when processing an include directive)
param
isTagFile true if file to be parsed is tag file, and false if it is a regular JSP page
param
directivesOnly true if the file to be parsed is a tag file and we are only interested in the directives needed for constructing a TagFileInfo.
param
jarFile The JAR file from which to read the JSP page or tag file, or null if the JSP page or tag file is to be read from the filesystem


        Node.Nodes parsedPage = null;
        isEncodingSpecifiedInProlog = false;
        isBomPresent = false;
        isDefaultPageEncoding = false;

        JarFile jarFile = getJarFile(jarFileUrl);
        String absFileName = resolveFileName(inFileName);
        String jspConfigPageEnc = getJspConfigPageEncoding(absFileName);

        // Figure out what type of JSP document and encoding type we are
        // dealing with
        determineSyntaxAndEncoding(absFileName, jarFile, jspConfigPageEnc);

        if (parent != null) {
            // Included resource, add to dependent list
            compiler.getPageInfo().addDependant(absFileName);
        }

        if ((isXml && isEncodingSpecifiedInProlog) || isBomPresent) {
            /*
             * Make sure the encoding explicitly specified in the XML
             * prolog (if any) matches that in the JSP config element
             * (if any), treating "UTF-16", "UTF-16BE", and "UTF-16LE" as
             * identical.
             */
            if (jspConfigPageEnc != null && !jspConfigPageEnc.equals(sourceEnc)
                    && (!jspConfigPageEnc.startsWith("UTF-16")
                            || !sourceEnc.startsWith("UTF-16"))) {
                err.jspError("jsp.error.prolog_config_encoding_mismatch",
                        sourceEnc, jspConfigPageEnc);
            }
        }

        // Dispatch to the appropriate parser
        if (isXml) {
            // JSP document (XML syntax)
            // InputStream for jspx page is created and properly closed in
            // JspDocumentParser.
            parsedPage = JspDocumentParser.parse(this, absFileName,
                    jarFile, parent,
                    isTagFile, directiveOnly,
                    sourceEnc,
                    jspConfigPageEnc,
                    isEncodingSpecifiedInProlog,
                    isBomPresent);
        } else {
            // Standard syntax
            InputStreamReader inStreamReader = null;
            try {
                inStreamReader = JspUtil.getReader(absFileName, sourceEnc,
                        jarFile, ctxt, err, skip);
                JspReader jspReader = new JspReader(ctxt, absFileName,
                        sourceEnc, inStreamReader,
                        err);
                parsedPage = Parser.parse(this, jspReader, parent, isTagFile,
                        directiveOnly, jarFileUrl,
                        sourceEnc, jspConfigPageEnc,
                        isDefaultPageEncoding, isBomPresent);
            } finally {
                if (inStreamReader != null) {
                    try {
                        inStreamReader.close();
                    } catch (Exception any) {
                    }
                }
            }
        }

        if (jarFile != null) {
            try {
                jarFile.close();
            } catch (Throwable t) {}
        }

        baseDirStack.pop();

        return parsedPage;
public CompilergetCompiler()

        return compiler;
private java.util.jar.JarFilegetJarFile(java.net.URL jarFileUrl)

        JarFile jarFile = null;

        if (jarFileUrl != null) {
            JarURLConnection conn = (JarURLConnection) jarFileUrl.openConnection();
            conn.setUseCaches(false);
            conn.connect();
            jarFile = conn.getJarFile();
        }

        return jarFile;
public org.apache.jasper.JspCompilationContextgetJspCompilationContext()

        return ctxt;
private java.lang.StringgetJspConfigPageEncoding(java.lang.String absFileName)


        JspConfig jspConfig = ctxt.getOptions().getJspConfig();
        JspConfig.JspProperty jspProperty
            = jspConfig.findJspProperty(absFileName);
        return jspProperty.getPageEncoding();
private java.lang.StringgetPageEncodingForJspSyntax(JspReader jspReader, Mark startMark)


        String encoding = null;
        String saveEncoding = null;

        jspReader.reset(startMark);

        /*
         * Determine page encoding from directive of the form <%@ page %>,
         * <%@ tag %>, <jsp:directive.page > or <jsp:directive.tag >.
         */
        while (true) {
            if (jspReader.skipUntil("<") == null) {
                break;
            }
            // If this is a comment, skip until its end
            if (jspReader.matches("%--")) {
                if (jspReader.skipUntil("--%>") == null) {
                    // error will be caught in Parser
                    break;
                }
                continue;
            }
            boolean isDirective = jspReader.matches("%@");
            if (isDirective) {
                jspReader.skipSpaces();
            }
            else {
                isDirective = jspReader.matches("jsp:directive.");
            }
            if (!isDirective) {
                continue;
            }

            // compare for "tag ", so we don't match "taglib"
            if (jspReader.matches("tag ") || jspReader.matches("page")) {

                jspReader.skipSpaces();
                Attributes attrs = Parser.parseAttributes(this, jspReader);
                encoding = getPageEncodingFromDirective(attrs, "pageEncoding");
                if (encoding != null) {
                    break;
                }
                encoding = getPageEncodingFromDirective(attrs, "contentType");
                if (encoding != null) {
                    saveEncoding = encoding;
                }
            }
        }

        if (encoding == null) {
            encoding = saveEncoding;
        }

        return encoding;
private java.lang.StringgetPageEncodingFromDirective(org.xml.sax.Attributes attrs, java.lang.String attrName)

        String value = attrs.getValue(attrName);
        if (attrName.equals("pageEncoding")) {
            return value;
        }

        // attrName = contentType
        String contentType = value;
        String encoding = null;
        if (contentType != null) {
            int loc = contentType.indexOf(CHARSET);
            if (loc != -1) {
                encoding = contentType.substring(loc + CHARSET.length());
            }
        }

        return encoding;
private booleanhasJspRoot(JspReader reader)


        // <prefix>:root must be the first element
        Mark start = null;
        while ((start = reader.skipUntil("<")) != null) {
            int c = reader.nextChar();
            if (c != '!" && c != '?") break;
        }
        if (start == null) {
            return false;
        }
        Mark stop = reader.skipUntil(":root");
        if (stop == null) {
            return false;
        }
        // call substring to get rid of leading '<'
        String prefix = reader.getText(start, stop).substring(1);

        start = stop;
        stop = reader.skipUntil(">");
        if (stop == null) {
            return false;
        }

        // Determine namespace associated with <root> element's prefix
        String root = reader.getText(start, stop);
        String xmlnsDecl = "xmlns:" + prefix;
        int index = root.indexOf(xmlnsDecl);
        if (index == -1) {
            return false;
        }
        index += xmlnsDecl.length();
        while (index < root.length()
                && Character.isWhitespace(root.charAt(index))) {
            index++;
        }
        if (index < root.length() && root.charAt(index) == '=") {
            index++;
            while (index < root.length()
                    && Character.isWhitespace(root.charAt(index))) {
                index++;
            }
            if (index < root.length() && root.charAt(index++) == '""
                && root.regionMatches(index, JSP_URI, 0,
                        JSP_URI.length())) {
                return true;
            }
        }

        return false;
public Node.Nodesparse(java.lang.String inFileName)
Parses a JSP page or tag file. This is invoked by the compiler.

param
inFileName The path to the JSP page or tag file to be parsed.

        // If we're parsing a packaged tag file or a resource included by it
        // (using an include directive), ctxt.getTagFileJar() returns the 
        // JAR file from which to read the tag file or included resource,
        // respectively.
        isTagFile = ctxt.isTagFile();
        directiveOnly = false;
        return doParse(inFileName, null, ctxt.getTagFileJarUrl());
public Node.Nodesparse(java.lang.String inFileName, Node parent, java.net.URL jarFileUrl)
Processes an include directive with the given path.

param
inFileName The path to the resource to be included.
param
parent The parent node of the include directive.
param
jarFile The JAR file from which to read the included resource, or null of the included resource is to be read from the filesystem

        // For files that are statically included, isTagfile and directiveOnly
        // remain unchanged.
        return doParse(inFileName, parent, jarFileUrl);
public Node.NodesparseTagFileDirectives(java.lang.String inFileName)
Extracts tag file directive information from the tag file with the given name. This is invoked by the compiler

param
inFileName The name of the tag file to be parsed.

        boolean isTagFileSave = isTagFile;
        boolean directiveOnlySave = directiveOnly;
        isTagFile = true;
        directiveOnly = true;
        Node.Nodes page = doParse(inFileName, null,
                ctxt.getTagFileJarUrl(inFileName));
        directiveOnly = directiveOnlySave;
        isTagFile = isTagFileSave;
        return page;
private java.lang.StringresolveFileName(java.lang.String inFileName)

        String fileName = inFileName.replace('\\", '/");
        boolean isAbsolute = fileName.startsWith("/");
        fileName = isAbsolute ? fileName 
                : (String) baseDirStack.peek() + fileName;
        String baseDir = 
            fileName.substring(0, fileName.lastIndexOf("/") + 1);
        baseDirStack.push(baseDir);
        return fileName;