xmloutputter.java

一个用于搜索本地文件内容的小型搜索引擎

        out.write(";");
    }

    /**
     * This will handle printing of <code>{@link CDATA}</code> text.
     *
     * @param cdata <code>CDATA</code> to output.
     * @param out <code>Writer</code> to use.
     */
    protected void printCDATA(Writer out, CDATA cdata) throws IOException {
        String str = (currentFormat.mode == Format.TextMode.NORMALIZE)
                     ? cdata.getTextNormalize()
                     : ((currentFormat.mode == Format.TextMode.TRIM) ?
                             cdata.getText().trim() : cdata.getText());
        out.write("<![CDATA[");
        out.write(str);
        out.write("]]>");
    }

    /**
     * This will handle printing of <code>{@link Text}</code> strings.
     *
     * @param text <code>Text</code> to write.
     * @param out <code>Writer</code> to use.
     */
    protected void printText(Writer out, Text text) throws IOException {
        String str = (currentFormat.mode == Format.TextMode.NORMALIZE)
                     ? text.getTextNormalize()
                     : ((currentFormat.mode == Format.TextMode.TRIM) ?
                          text.getText().trim() : text.getText());
        out.write(escapeElementEntities(str));
    }

    /**
     * This will handle printing a string.  Escapes the element entities,
     * trims interior whitespace, etc. if necessary.
     */
    private void printString(Writer out, String str) throws IOException {
        if (currentFormat.mode == Format.TextMode.NORMALIZE) {
            str = Text.normalizeString(str);
        }
        else if (currentFormat.mode == Format.TextMode.TRIM) {
            str = str.trim();
        }
        out.write(escapeElementEntities(str));
    }

    /**
     * This will handle printing of a <code>{@link Element}</code>,
     * its <code>{@link Attribute}</code>s, and all contained (child)
     * elements, etc.
     *
     * @param element <code>Element</code> to output.
     * @param out <code>Writer</code> to use.
     * @param level <code>int</code> level of indention.
     * @param namespaces <code>List</code> stack of Namespaces in scope.
     */
    protected void printElement(Writer out, Element element,
                                int level, NamespaceStack namespaces)
                       throws IOException {

        List attributes = element.getAttributes();
        List content = element.getContent();

        // Check for xml:space and adjust format settings
        String space = null;
        if (attributes != null) {
            space = element.getAttributeValue("space",
                                               Namespace.XML_NAMESPACE);
        }

        Format previousFormat = currentFormat;

        if ("default".equals(space)) {
            currentFormat = userFormat;
        }
        else if ("preserve".equals(space)) {
            currentFormat = preserveFormat;
        }

        // Print the beginning of the tag plus attributes and any
        // necessary namespace declarations
        out.write("<");
        printQualifiedName(out, element);

        // Mark our namespace starting point
        int previouslyDeclaredNamespaces = namespaces.size();

        // Print the element's namespace, if appropriate
        printElementNamespace(out, element, namespaces);

        // Print out additional namespace declarations
        printAdditionalNamespaces(out, element, namespaces);

        // Print out attributes
        if (attributes != null)
            printAttributes(out, attributes, element, namespaces);

        // Depending on the settings (newlines, textNormalize, etc), we may
        // or may not want to print all of the content, so determine the
        // index of the start of the content we're interested
        // in based on the current settings.

        int start = skipLeadingWhite(content, 0);
        int size = content.size();
        if (start >= size) {
            // Case content is empty or all insignificant whitespace
            if (currentFormat.expandEmptyElements) {
                out.write("></");
                printQualifiedName(out, element);
                out.write(">");
            }
            else {
                out.write(" />");
            }
        }
        else {
            out.write(">");

            // For a special case where the content is only CDATA
            // or Text we don't want to indent after the start or
            // before the end tag.

            if (nextNonText(content, start) < size) {
                // Case Mixed Content - normal indentation
                newline(out);
                printContentRange(out, content, start, size,
                                  level + 1, namespaces);
                newline(out);
                indent(out, level);
            }
            else {
                // Case all CDATA or Text - no indentation
                printTextRange(out, content, start, size);
            }
            out.write("</");
            printQualifiedName(out, element);
            out.write(">");
        }

        // remove declared namespaces from stack
        while (namespaces.size() > previouslyDeclaredNamespaces) {
            namespaces.pop();
        }

        // Restore our format settings
        currentFormat = previousFormat;
    }

    /**
     * This will handle printing of content within a given range.
     * The range to print is specified in typical Java fashion; the
     * starting index is inclusive, while the ending index is
     * exclusive.
     *
     * @param content <code>List</code> of content to output
     * @param start index of first content node (inclusive.
     * @param end index of last content node (exclusive).
     * @param out <code>Writer</code> to use.
     * @param level <code>int</code> level of indentation.
     * @param namespaces <code>List</code> stack of Namespaces in scope.
     */
    private void printContentRange(Writer out, List content,
                                     int start, int end, int level,
                                     NamespaceStack namespaces)
                       throws IOException {
        boolean firstNode; // Flag for 1st node in content
        Object next;       // Node we're about to print
        int first, index;  // Indexes into the list of content

        index = start;
        while (index < end) {
            firstNode = (index == start) ? true : false;
            next = content.get(index);

            //
            // Handle consecutive CDATA, Text, and EntityRef nodes all at once
            //
            if ((next instanceof Text) || (next instanceof EntityRef)) {
                first = skipLeadingWhite(content, index);
                // Set index to next node for loop
                index = nextNonText(content, first);

                // If it's not all whitespace - print it!
                if (first < index) {
                    if (!firstNode)
                        newline(out);
                    indent(out, level);
                    printTextRange(out, content, first, index);
                }
                continue;
            }

            //
            // Handle other nodes
            //
            if (!firstNode) {
                newline(out);
            }

            indent(out, level);

            if (next instanceof Comment) {
                printComment(out, (Comment)next);
            }
            else if (next instanceof Element) {
                printElement(out, (Element)next, level, namespaces);
            }
            else if (next instanceof ProcessingInstruction) {
                printProcessingInstruction(out, (ProcessingInstruction)next);
            }
            else {
                // XXX if we get here then we have a illegal content, for
                //     now we'll just ignore it (probably should throw
                //     a exception)
            }

            index++;
        } /* while */
    }

    /**
     * This will handle printing of a sequence of <code>{@link CDATA}</code>
     * or <code>{@link Text}</code> nodes.  It is an error to have any other
     * pass this method any other type of node.
     *
     * @param content <code>List</code> of content to output
     * @param start index of first content node (inclusive).
     * @param end index of last content node (exclusive).
     * @param out <code>Writer</code> to use.
     */
    private void printTextRange(Writer out, List content, int start, int end
                                  ) throws IOException {
        String previous; // Previous text printed
        Object node;     // Next node to print
        String next;     // Next text to print

        previous = null;

        // Remove leading whitespace-only nodes
        start = skipLeadingWhite(content, start);

        int size = content.size();
        if (start < size) {
            // And remove trialing whitespace-only nodes
            end = skipTrailingWhite(content, end);

            for (int i = start; i < end; i++) {
                node = content.get(i);

                // Get the unmangled version of the text
                // we are about to print
                if (node instanceof Text) {
                    next = ((Text) node).getText();
                }
                else if (node instanceof EntityRef) {
                    next = "&" + ((EntityRef) node).getValue() + ";";
                }
                else {
                    throw new IllegalStateException("Should see only " +
                                                   "CDATA, Text, or EntityRef");
                }

                // This may save a little time
                if (next == null || "".equals(next)) {
                    continue;
                }

                // Determine if we need to pad the output (padding is
                // only need in trim or normalizing mode)
                if (previous != null) { // Not 1st node
                    if (currentFormat.mode == Format.TextMode.NORMALIZE ||
                        currentFormat.mode == Format.TextMode.TRIM) {
                            if ((endsWithWhite(previous)) ||
                                (startsWithWhite(next))) {
                                    out.write(" ");
                            }
                    }
                }

                // Print the node
                if (node instanceof CDATA) {
                    printCDATA(out, (CDATA) node);
                }
                else if (node instanceof EntityRef) {
                    printEntityRef(out, (EntityRef) node);
                }
                else {
                    printString(out, next);
                }

                previous = next;
            }
        }
    }

    /**
     * This will handle printing of any needed <code>{@link Namespace}</code>
     * declarations.
     *
     * @param ns <code>Namespace</code> to print definition of
     * @param out <code>Writer</code> to use.
     */
    private void printNamespace(Writer out, Namespace ns,
                                NamespaceStack namespaces)
                     throws IOException {
        String prefix = ns.getPrefix();
        String uri = ns.getURI();

        // Already printed namespace decl?
        if (uri.equals(namespaces.getURI(prefix))) {
            return;
        }

        out.write(" xmlns");
        if (!prefix.equals("")) {
            out.write(":");
            out.write(prefix);
        }
        out.write("=\"");
        out.write(escapeAttributeEntities(uri));
        out.write("\"");
        namespaces.push(ns);
    }

    /**
     * This will handle printing of a <code>{@link Attribute}</code> list.
     *
     * @param attributes <code>List</code> of Attribute objcts
     * @param out <code>Writer</code> to use
     */
    protected void printAttributes(Writer out, List attributes, Element parent,
                                   NamespaceStack namespaces)
                       throws IOException {

        // I do not yet handle the case where the same prefix maps to
        // two different URIs. For attributes on the same element
        // this is illegal; but as yet we don't throw an exception
        // if someone tries to do this
        // Set prefixes = new HashSet();
        for (int i = 0; i < attributes.size(); i++) {
            Attribute attribute = (Attribute) attributes.get(i);
            Namespace ns = attribute.getNamespace();
            if ((ns != Namespace.NO_NAMESPACE) &&
                (ns != Namespace.XML_NAMESPACE)) {
                    printNamespace(out, ns, namespaces);
            }

            out.write(" ");
            printQualifiedName(out, attribute);
            out.write("=");

            out.write("\"");
            out.write(escapeAttributeEntities(attribute.getValue()));
            out.write("\"");
        }
    }

    private void printElementNamespace(Writer out, Element element,
                                       NamespaceStack namespaces)
                             throws IOException {
        // Add namespace decl only if it's not the XML namespace and it's
        // not the NO_NAMESPACE with the prefix "" not yet mapped
        // (we do output xmlns="" if the "" prefix was already used and we
        // need to reclaim it for the NO_NAMESPACE)
        Namespace ns = element.getNamespace();
        if (ns == Namespace.XML_NAMESPACE) {
            return;
        }
        if ( !((ns == Namespace.NO_NAMESPACE) &&
               (namespaces.getURI("") == null))) {
            printNamespace(out, ns, namespaces);
        }
    }

    private void printAdditionalNamespaces(Writer out, Element element,
                                           NamespaceStack namespaces)
                                throws IOException {
        List list = element.getAdditionalNamespaces();
        if (list != null) {
            for (int i = 0; i < list.size(); i++) {
                Namespace additional = (Namespace)list.get(i);
                printNamespace(out, additional, namespaces);
            }
        }
    }

    // * * * * * * * * * * Support methods * * * * * * * * * *
    // * * * * * * * * * * Support methods * * * * * * * * * *

    /**
     * This will print a newline only if indent is not null.
     *
     * @param out <code>Writer</code> to use
     */
    private void newline(Writer out) throws IOException {
        if (currentFormat.indent != null) {
            out.write(currentFormat.lineSeparator);
        }
    }