blob: f73407cd5e2bdfac27499e9157dc96de9a355b50 [file] [log] [blame]
/*
* Copyright (c) 1998, 2015, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
package jdk.javadoc.internal.doclets.formats.html;
import java.io.*;
import java.util.ArrayList;
import java.util.Collection;
import java.util.List;
import java.util.ListIterator;
import java.util.Set;
import java.util.TreeSet;
import jdk.javadoc.internal.doclets.formats.html.markup.HtmlStyle;
import jdk.javadoc.internal.doclets.formats.html.markup.HtmlTag;
import jdk.javadoc.internal.doclets.formats.html.markup.HtmlTree;
import jdk.javadoc.internal.doclets.formats.html.markup.StringContent;
import jdk.javadoc.internal.doclets.toolkit.Content;
import jdk.javadoc.internal.doclets.toolkit.util.DocPath;
import jdk.javadoc.internal.doclets.toolkit.util.DocPaths;
import jdk.javadoc.internal.doclets.toolkit.util.DocletAbortException;
import jdk.javadoc.internal.doclets.toolkit.util.IndexBuilder;
/**
* Generate Separate Index Files for all the member names with Indexing in
* Unicode Order. This will create "index-files" directory in the current or
* destination directory and will generate separate file for each unicode index.
*
* <p><b>This is NOT part of any supported API.
* If you write code that depends on this, you do so at your own risk.
* This code and its internal interfaces are subject to change or
* deletion without notice.</b>
*
* @see java.lang.Character
* @author Atul M Dambalkar
* @author Bhavesh Patel (Modified)
*/
public class SplitIndexWriter extends AbstractIndexWriter {
/**
* Previous unicode character index in the built index.
*/
protected int prev;
/**
* Next unicode character in the built index.
*/
protected int next;
private List<Character> indexElements;
/**
* Construct the SplitIndexWriter. Uses path to this file and relative path
* from this file.
*
* @param path Path to the file which is getting generated.
* @param indexbuilder Unicode based Index from {@link IndexBuilder}
*/
public SplitIndexWriter(ConfigurationImpl configuration,
DocPath path,
IndexBuilder indexbuilder,
Collection<Character> elements,
int prev, int next) throws IOException {
super(configuration, path, indexbuilder);
this.indexElements = new ArrayList<>(elements);
this.prev = prev;
this.next = next;
}
/**
* Generate separate index files, for each Unicode character, listing all
* the members starting with the particular unicode character.
*
* @param indexbuilder IndexBuilder built by {@link IndexBuilder}
* @throws DocletAbortException
*/
public static void generate(ConfigurationImpl configuration,
IndexBuilder indexbuilder) {
SplitIndexWriter indexgen;
DocPath filename = DocPath.empty;
DocPath path = DocPaths.INDEX_FILES;
try {
Set<Character> keys = new TreeSet<>(indexbuilder.getIndexMap().keySet());
keys.addAll(configuration.tagSearchIndexKeys);
ListIterator<Character> li = new ArrayList<>(keys).listIterator();
while (li.hasNext()) {
Object ch = li.next();
filename = DocPaths.indexN(li.nextIndex());
indexgen = new SplitIndexWriter(configuration,
path.resolve(filename),
indexbuilder, keys, li.previousIndex(), li.nextIndex());
indexgen.generateIndexFile((Character) ch);
if (!li.hasNext()) {
indexgen.createSearchIndexFiles();
}
indexgen.close();
}
} catch (IOException exc) {
configuration.standardmessage.error(
"doclet.exception_encountered",
exc.toString(), filename.getPath());
throw new DocletAbortException(exc);
}
}
/**
* Generate the contents of each index file, with Header, Footer,
* Member Field, Method and Constructor Description.
*
* @param unicode Unicode character referring to the character for the
* index.
*/
protected void generateIndexFile(Character unicode) throws IOException {
String title = configuration.getText("doclet.Window_Split_Index",
unicode.toString());
HtmlTree body = getBody(true, getWindowTitle(title));
HtmlTree htmlTree = (configuration.allowTag(HtmlTag.HEADER))
? HtmlTree.HEADER()
: body;
addTop(htmlTree);
addNavLinks(true, htmlTree);
if (configuration.allowTag(HtmlTag.HEADER)) {
body.addContent(htmlTree);
}
HtmlTree divTree = new HtmlTree(HtmlTag.DIV);
divTree.addStyle(HtmlStyle.contentContainer);
addLinksForIndexes(divTree);
if (configuration.tagSearchIndexMap.get(unicode) == null) {
addContents(unicode, indexbuilder.getMemberList(unicode), divTree);
} else if (indexbuilder.getMemberList(unicode) == null) {
addSearchContents(unicode, configuration.tagSearchIndexMap.get(unicode), divTree);
} else {
addContents(unicode, indexbuilder.getMemberList(unicode),
configuration.tagSearchIndexMap.get(unicode), divTree);
}
addLinksForIndexes(divTree);
body.addContent((configuration.allowTag(HtmlTag.MAIN)) ? HtmlTree.MAIN(divTree) : divTree);
if (configuration.allowTag(HtmlTag.FOOTER)) {
htmlTree = HtmlTree.FOOTER();
}
addNavLinks(false, htmlTree);
addBottom(htmlTree);
if (configuration.allowTag(HtmlTag.FOOTER)) {
body.addContent(htmlTree);
}
printHtmlDocument(null, true, body);
}
/**
* Add links for all the Index Files per unicode character.
*
* @param contentTree the content tree to which the links for indexes will be added
*/
protected void addLinksForIndexes(Content contentTree) {
for (int i = 0; i < indexElements.size(); i++) {
int j = i + 1;
contentTree.addContent(getHyperLink(DocPaths.indexN(j),
new StringContent(indexElements.get(i).toString())));
contentTree.addContent(getSpace());
}
}
/**
* Get link to the previous unicode character.
*
* @return a content tree for the link
*/
public Content getNavLinkPrevious() {
Content prevletterLabel = getResource("doclet.Prev_Letter");
if (prev == -1) {
return HtmlTree.LI(prevletterLabel);
}
else {
Content prevLink = getHyperLink(DocPaths.indexN(prev),
prevletterLabel);
return HtmlTree.LI(prevLink);
}
}
/**
* Get link to the next unicode character.
*
* @return a content tree for the link
*/
public Content getNavLinkNext() {
Content nextletterLabel = getResource("doclet.Next_Letter");
if (next == -1) {
return HtmlTree.LI(nextletterLabel);
}
else {
Content nextLink = getHyperLink(DocPaths.indexN(next),
nextletterLabel);
return HtmlTree.LI(nextLink);
}
}
}