parsercontroller.java

业界著名的tomcat服务器的最新6.0的源代码。

            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;
            }

            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;
                }
            }
        }
        
    }

    /*
     * Determines page source encoding for page or tag file in JSP syntax,
     * by reading (in this order) the value of the 'pageEncoding' page
     * directive attribute, or the charset value of the 'contentType' page
     * directive attribute.
     *
     * @return The page encoding, or null if not found
     */
    private String getPageEncodingForJspSyntax(JspReader jspReader,
            Mark startMark)
    throws JasperException {

        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;
    }

    /*
     * Scans the given attributes for the attribute with the given name,
     * which is either 'pageEncoding' or 'contentType', and returns the
     * specified page encoding.
     *
     * In the case of 'contentType', the page encoding is taken from the
     * content type's 'charset' component.
     *
     * @param attrs The page directive attributes
     * @param attrName The name of the attribute to search for (either
     * 'pageEncoding' or 'contentType')
     *
     * @return The page encoding, or null
     */
    private String getPageEncodingFromDirective(Attributes attrs,
            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;
    }

    /*
     * Resolve the name of the file and update baseDirStack() to keep track of
     * the current base directory for each included file.
     * The 'root' file is always an 'absolute' path, so no need to put an
     * initial value in the baseDirStack.
     */
    private String resolveFileName(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;
    }

    /*
     * Checks to see if the given page contains, as its first element, a <root>
     * element whose prefix is bound to the JSP namespace, as in:
     *
     * <wombat:root xmlns:wombat="http://java.sun.com/JSP/Page" version="1.2">
     *   ...
     * </wombat:root>
     *
     * @param reader The reader for this page
     *
     * @return true if this page contains a root element whose prefix is bound
     * to the JSP namespace, and false otherwise
     */
    private boolean hasJspRoot(JspReader reader) throws JasperException {

        // <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;
    }

    private JarFile getJarFile(URL jarFileUrl) throws IOException {
        JarFile jarFile = null;

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

        return jarFile;
    }

}