/*
 * Copyright (c) 2005, 2006, 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 com.sun.xml.internal.dtdparser;

import org.xml.sax.EntityResolver;
import org.xml.sax.InputSource;

import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.net.URL;
import java.net.URLConnection;
import java.util.Hashtable;

/**
 * This entity resolver class provides a number of utilities which can help
 * managment of external parsed entities in XML.  These are commonly used
 * to hold markup declarations that are to be used as part of a Document
 * Type Declaration (DTD), or to hold text marked up with XML.
 * <p/>
 * <P> Features include: <UL>
 * <p/>
 * <LI> Static factory methods are provided for constructing SAX InputSource
 * objects from Files, URLs, or MIME objects.  This eliminates a class of
 * error-prone coding in applications.
 * <p/>
 * <LI> Character encodings for XML documents are correctly supported: <UL>
 * <p/>
 * <LI> The encodings defined in the RFCs for MIME content types
 * (2046 for general MIME, and 2376 for XML in particular), are
 * supported, handling <em>charset=...</em> attributes and accepting
 * content types which are known to be safe for use with XML;
 * <p/>
 * <LI> The character encoding autodetection algorithm identified
 * in the XML specification is used, and leverages all of
 * the JDK 1.1 (and later) character encoding support.
 * <p/>
 * <LI> The use of MIME typing may optionally be disabled, forcing the
 * use of autodetection, to support web servers which don't correctly
 * report MIME types for XML.  For example, they may report text that
 * is encoded in EUC-JP as being US-ASCII text, leading to fatal
 * errors during parsing.
 * <p/>
 * <LI> The InputSource objects returned by this class always
 * have a <code>java.io.Reader</code> available as the "character
 * stream" property.
 * <p/>
 * </UL>
 * <p/>
 * <LI> Catalog entries can map public identifiers to Java resources or
 * to local URLs.  These are used to reduce network dependencies and loads,
 * and will often be used for external DTD components.  For example, packages
 * shipping DTD files as resources in JAR files can eliminate network traffic
 * when accessing them, and sites may provide local caches of common DTDs.
 * Note that no particular catalog syntax is supported by this class, only
 * the notion of a set of entries.
 * <p/>
 * </UL>
 * <p/>
 * <P> Subclasses can perform tasks such as supporting new URI schemes for
 * URIs which are not URLs, such as URNs (see RFC 2396) or for accessing
 * MIME entities which are part of a <em>multipart/related</em> group
 * (see RFC 2387).  They may also be used to support particular catalog
 * syntaxes, such as the <a href="http://www.oasis-open.org/html/a401.htm">
 * SGML/Open Catalog (SOCAT)</a> which supports the SGML notion of "Formal
 * Public Identifiers (FPIs).
 *
 * @author David Brownell
 * @author Janet Koenig
 * @version 1.3 00/02/24
 */
public class Resolver implements EntityResolver {
    private boolean ignoringMIME;

    // table mapping public IDs to (local) URIs
    private Hashtable id2uri;

    // tables mapping public IDs to resources and classloaders
    private Hashtable id2resource;
    private Hashtable id2loader;

    //
    // table of MIME content types (less attributes!) known
    // to be mostly "OK" to use with XML MIME entities.  the
    // idea is to rule out obvious braindamage ("image/jpg")
    // not the subtle stuff ("text/html") that might actually
    // be (or become) safe.
    //
    private static final String types [] = {
        "application/xml",
        "text/xml",
        "text/plain",
        "text/html", // commonly mis-inferred
        "application/x-netcdf", // this is often illegal XML
        "content/unknown"
    };

    /**
     * Constructs a resolver.
     */
    public Resolver() {
    }

    /**
     * Returns an input source, using the MIME type information and URL
     * scheme to statically determine the correct character encoding if
     * possible and otherwise autodetecting it.  MIME carefully specifies
     * the character encoding defaults, and how attributes of the content
     * type can change it.  XML further specifies two mandatory encodings
     * (UTF-8 and UTF-16), and includes an XML declaration which can be
     * used to internally label most documents encoded using US-ASCII
     * supersets (such as Shift_JIS, EUC-JP, ISO-2022-*, ISO-8859-*, and
     * more).
     * <p/>
     * <P> This method can be used to access XML documents which do not
     * have URIs (such as servlet input streams, or most JavaMail message
     * entities) and to support access methods such as HTTP POST or PUT.
     * (URLs normally return content using the GET method.)
     * <p/>
     * <P> <em> The caller should set the system ID in order for relative URIs
     * found in this document to be interpreted correctly.</em> In some cases,
     * a custom resolver will need to be used; for example, documents
     * may be grouped in a single MIME "multipart/related" bundle, and
     * relative URLs would refer to other documents in that bundle.
     *
     * @param contentType The MIME content type for the source for which
     *                    an InputSource is desired, such as <em>text/xml;charset=utf-8</em>.
     * @param stream      The input byte stream for the input source.
     * @param checkType   If true, this verifies that the content type is known
     *                    to support XML documents, such as <em>application/xml</em>.
     * @param scheme      Unless this is "file", unspecified MIME types
     *                    default to US-ASCII.  Files are always autodetected since most
     *                    file systems discard character encoding information.
     */
    public static InputSource createInputSource(String contentType,
                                                InputStream stream,
                                                boolean checkType,
                                                String scheme) throws IOException {
        InputSource retval;
        String charset = null;

        if (contentType != null) {
            int index;

            contentType = contentType.toLowerCase();
            index = contentType.indexOf(';');
            if (index != -1) {
                String attributes;

                attributes = contentType.substring(index + 1);
                contentType = contentType.substring(0, index);

                // use "charset=..." if it's available
                index = attributes.indexOf("charset");
                if (index != -1) {
                    attributes = attributes.substring(index + 7);
                    // strip out subsequent attributes
                    if ((index = attributes.indexOf(';')) != -1)
                        attributes = attributes.substring(0, index);
                    // find start of value
                    if ((index = attributes.indexOf('=')) != -1) {
                        attributes = attributes.substring(index + 1);
                        // strip out rfc822 comments
                        if ((index = attributes.indexOf('(')) != -1)
                            attributes = attributes.substring(0, index);
                        // double quotes are optional
                        if ((index = attributes.indexOf('"')) != -1) {
                            attributes = attributes.substring(index + 1);
                            attributes = attributes.substring(0,
                                    attributes.indexOf('"'));
                        }
                        charset = attributes.trim();
                        // XXX "\;", "\)" etc were mishandled above
                    }
                }
            }

            //
            // Check MIME type.
            //
            if (checkType) {
                boolean isOK = false;
                for (int i = 0; i < types.length; i++)
                    if (types[i].equals(contentType)) {
                        isOK = true;
                        break;
                    }
                if (!isOK)
                    throw new IOException("Not XML: " + contentType);
            }

            //
            // "text/*" MIME types have hard-wired character set
            // defaults, as specified in the RFCs.  For XML, we
            // ignore the system "file.encoding" property since
            // autodetection is more correct.
            //
            if (charset == null) {
                contentType = contentType.trim();
                if (contentType.startsWith("text/")) {
                    if (!"file".equalsIgnoreCase(scheme))
                        charset = "US-ASCII";
                }
                // "application/*" has no default
            }
        }

        retval = new InputSource(XmlReader.createReader(stream, charset));
        retval.setByteStream(stream);
        retval.setEncoding(charset);
        return retval;
    }


    /**
     * Creates an input source from a given URI.
     *
     * @param uri       the URI (system ID) for the entity
     * @param checkType if true, the MIME content type for the entity
     *                  is checked for document type and character set encoding.
     */
    static public InputSource createInputSource(URL uri, boolean checkType)
            throws IOException {

        URLConnection conn = uri.openConnection();
        InputSource retval;

        if (checkType) {
            String contentType = conn.getContentType();
            retval = createInputSource(contentType, conn.getInputStream(),
                    false, uri.getProtocol());
        } else {
            retval = new InputSource(XmlReader.createReader(conn.getInputStream()));
        }
        retval.setSystemId(conn.getURL().toString());
        return retval;
    }


    /**
     * Creates an input source from a given file, autodetecting
     * the character encoding.
     */
    static public InputSource createInputSource(File file)
            throws IOException {
        InputSource retval;
        String path;

        retval = new InputSource(XmlReader.createReader(new FileInputStream(file)));

        // On JDK 1.2 and later, simplify this:
        //    "path = file.toURL ().toString ()".
        path = file.getAbsolutePath();
        if (File.separatorChar != '/')
            path = path.replace(File.separatorChar, '/');
        if (!path.startsWith("/"))
            path = "/" + path;
        if (!path.endsWith("/") && file.isDirectory())
            path = path + "/";

        retval.setSystemId("file:" + path);
        return retval;
    }


    /**
     * <b>SAX:</b>
     * Resolve the given entity into an input source.  If the name can't
     * be mapped to a preferred form of the entity, the URI is used.  To
     * resolve the entity, first a local catalog mapping names to URIs is
     * consulted.  If no mapping is found there, a catalog mapping names
     * to java resources is consulted.  Finally, if neither mapping found
     * a copy of the entity, the specified URI is used.
     * <p/>
     * <P> When a URI is used, <a href="#createInputSource">
     * createInputSource</a> is used to correctly deduce the character
     * encoding used by this entity.  No MIME type checking is done.
     *
     * @param name Used to find alternate copies of the entity, when
     *             this value is non-null; this is the XML "public ID".
     * @param uri  Used when no alternate copy of the entity is found;
     *             this is the XML "system ID", normally a URI.
     */
    public InputSource resolveEntity(String name, String uri)
            throws IOException {
        InputSource retval;
        String mappedURI = name2uri(name);
        InputStream stream;

        // prefer explicit URI mappings, then bundled resources...
        if (mappedURI == null && (stream = mapResource(name)) != null) {
            uri = "java:resource:" + (String) id2resource.get(name);
            retval = new InputSource(XmlReader.createReader(stream));

            // ...and treat all URIs the same (as URLs for now).
        } else {
            URL url;
            URLConnection conn;

            if (mappedURI != null)
                uri = mappedURI;
            else if (uri == null)
                return null;

            url = new URL(uri);
            conn = url.openConnection();
            uri = conn.getURL().toString();
            // System.out.println ("++ URI: " + url);
            if (ignoringMIME)
                retval = new InputSource(XmlReader.createReader(conn.getInputStream()));
            else {
                String contentType = conn.getContentType();
                retval = createInputSource(contentType,
                        conn.getInputStream(),
                        false, url.getProtocol());
            }
        }
        retval.setSystemId(uri);
        retval.setPublicId(name);
        return retval;
    }


    /**
     * Returns true if this resolver is ignoring MIME types in the documents
     * it returns, to work around bugs in how servers have reported the
     * documents' MIME types.
     */
    public boolean isIgnoringMIME() {
        return ignoringMIME;
    }

    /**
     * Tells the resolver whether to ignore MIME types in the documents it
     * retrieves.  Many web servers incorrectly assign text documents a
     * default character encoding, even when that is incorrect.  For example,
     * all HTTP text documents default to use ISO-8859-1 (used for Western
     * European languages), and other MIME sources default text documents
     * to use US-ASCII (a seven bit encoding).  For XML documents which
     * include text encoding declarations (as most should do), these server
     * bugs can be worked around by ignoring the MIME type entirely.
     */
    public void setIgnoringMIME(boolean value) {
        ignoringMIME = value;
    }


    // maps the public ID to an alternate URI, if one is registered
    private String name2uri(String publicId) {
        if (publicId == null || id2uri == null)
            return null;
        return (String) id2uri.get(publicId);
    }


    /**
     * Registers the given public ID as corresponding to a particular
     * URI, typically a local copy.  This URI will be used in preference
     * to ones provided as system IDs in XML entity declarations.  This
     * mechanism would most typically be used for Document Type Definitions
     * (DTDs), where the public IDs are formally managed and versioned.
     *
     * @param publicId The managed public ID being mapped
     * @param uri      The URI of the preferred copy of that entity
     */
    public void registerCatalogEntry(String publicId,
                                     String uri) {
        if (id2uri == null)
            id2uri = new Hashtable(17);
        id2uri.put(publicId, uri);
    }


    // return the resource as a stream
    private InputStream mapResource(String publicId) {
        // System.out.println ("++ PUBLIC: " + publicId);
        if (publicId == null || id2resource == null)
            return null;

        String resourceName = (String) id2resource.get(publicId);
        ClassLoader loader = null;

        if (resourceName == null)
            return null;
        // System.out.println ("++ Resource: " + resourceName);

        if (id2loader != null)
            loader = (ClassLoader) id2loader.get(publicId);
        // System.out.println ("++ Loader: " + loader);
        if (loader == null)
            return ClassLoader.getSystemResourceAsStream(resourceName);
        return loader.getResourceAsStream(resourceName);
    }

    /**
     * Registers a given public ID as corresponding to a particular Java
     * resource in a given class loader, typically distributed with a
     * software package.  This resource will be preferred over system IDs
     * included in XML documents.  This mechanism should most typically be
     * used for Document Type Definitions (DTDs), where the public IDs are
     * formally managed and versioned.
     * <p/>
     * <P> If a mapping to a URI has been provided, that mapping takes
     * precedence over this one.
     *
     * @param publicId     The managed public ID being mapped
     * @param resourceName The name of the Java resource
     * @param loader       The class loader holding the resource, or null if
     *                     it is a system resource.
     */
    public void registerCatalogEntry(String publicId,
                                     String resourceName,
                                     ClassLoader loader) {
        if (id2resource == null)
            id2resource = new Hashtable(17);
        id2resource.put(publicId, resourceName);

        if (loader != null) {
            if (id2loader == null)
                id2loader = new Hashtable(17);
            id2loader.put(publicId, loader);
        }
    }
}