1 | /*
|
---|
2 | * GLIEntityResolver.java
|
---|
3 | * Copyright (C) 2008 New Zealand Digital Library, http://www.nzdl.org
|
---|
4 | *
|
---|
5 | * This program is free software; you can redistribute it and/or modify
|
---|
6 | * it under the terms of the GNU General Public License as published by
|
---|
7 | * the Free Software Foundation; either version 2 of the License, or
|
---|
8 | * (at your option) any later version.
|
---|
9 | *
|
---|
10 | * This program is distributed in the hope that it will be useful,
|
---|
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
13 | * GNU General Public License for more details.
|
---|
14 | *
|
---|
15 | * You should have received a copy of the GNU General Public License
|
---|
16 | * along with this program; if not, write to the Free Software
|
---|
17 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
18 | */
|
---|
19 |
|
---|
20 | package org.greenstone.gatherer.util;
|
---|
21 |
|
---|
22 | import org.greenstone.gatherer.Configuration;
|
---|
23 | import org.greenstone.gatherer.DebugStream;
|
---|
24 | import org.greenstone.gatherer.Gatherer;
|
---|
25 |
|
---|
26 | import org.xml.sax.InputSource;
|
---|
27 | import org.xml.sax.EntityResolver;
|
---|
28 | import java.io.File;
|
---|
29 | import java.net.URL;
|
---|
30 | import java.util.ArrayList;
|
---|
31 | import java.util.Iterator;
|
---|
32 |
|
---|
33 |
|
---|
34 | /**
|
---|
35 | * Uses a list of default search paths, or the file's own path if provided, to resolve entities referenced
|
---|
36 | * in the file. The search paths, including the file's own, have to be within GS.
|
---|
37 | * By default, the GS3HOME/WEB-INF and toplevel GLI user directory are added to the list of search paths,
|
---|
38 | * since this EntityResolver is used for web.xml to resolve its inclusion of servlets.xml, which can be
|
---|
39 | * located in GS3HOME/WEB-INF for a local GLI or is extracted into the gli user directory for client-GLI.
|
---|
40 | * If everything fails, it will try to resolve entities using the classloader, not otherwise used by GLI,
|
---|
41 | * but this part has been copied over from GS3's GSEntityResolver.java.
|
---|
42 | * If ever needed, maybe to make things faster, can maintain a static list of default search paths.
|
---|
43 | */
|
---|
44 | public class GLIEntityResolver implements EntityResolver {
|
---|
45 |
|
---|
46 | protected ArrayList<File> list_of_local_search_paths = new ArrayList<File>();
|
---|
47 | ClassLoader class_loader = null;
|
---|
48 |
|
---|
49 |
|
---|
50 | public GLIEntityResolver() {
|
---|
51 | // Add the basic search paths:
|
---|
52 |
|
---|
53 | // add the GS3 web/WEB-INF folder as search location for servlets.xml, which web.xml
|
---|
54 | // includes and which is the reason for adding a GLIEntityResolver class into GLI
|
---|
55 |
|
---|
56 | if(Gatherer.GS3) {
|
---|
57 | //list_of_local_search_paths.add(new File(Configuration.gsdl3_path));
|
---|
58 | list_of_local_search_paths.add(new File(Configuration.gsdl3_path, "WEB-INF"));
|
---|
59 | }
|
---|
60 |
|
---|
61 | // Add gli user dir for remote GS, since that is where the
|
---|
62 | // web.xml and server.xml downloaded from the remote server will be unpacked to
|
---|
63 | if(Gatherer.isGsdlRemote) {
|
---|
64 | list_of_local_search_paths.add(new File(Configuration.gli_user_directory_path));
|
---|
65 | }
|
---|
66 | }
|
---|
67 |
|
---|
68 | public GLIEntityResolver(File file) {
|
---|
69 | this(); // adds default search paths
|
---|
70 |
|
---|
71 | // add file's own path as first in search list
|
---|
72 | addSafeSearchPath(file, true);
|
---|
73 | }
|
---|
74 |
|
---|
75 | /* Methods with the ClassLoader parameter are unused at present */
|
---|
76 | public GLIEntityResolver(ClassLoader loader) {
|
---|
77 | this.class_loader = loader;
|
---|
78 | }
|
---|
79 |
|
---|
80 | public GLIEntityResolver(File f, ClassLoader loader) {
|
---|
81 | this(f);
|
---|
82 | this.class_loader = loader;
|
---|
83 | }
|
---|
84 |
|
---|
85 | public void setClassLoader(ClassLoader loader) {
|
---|
86 | this.class_loader = loader;
|
---|
87 | }
|
---|
88 |
|
---|
89 | public void addSafeSearchPath(File file, boolean prepend) {
|
---|
90 |
|
---|
91 | // add the file's directory to list of search paths,
|
---|
92 | // if it is within the greenstone installation and not already in the list
|
---|
93 | if((!Gatherer.GS3 && file.getAbsolutePath().startsWith(Configuration.gsdl_path))
|
---|
94 | || (Gatherer.GS3 && file.getAbsolutePath().startsWith(Configuration.gsdl3_src_path))) {
|
---|
95 |
|
---|
96 | File path = file.getParentFile();
|
---|
97 | if(!list_of_local_search_paths.contains(path)) {
|
---|
98 | if(prepend) {
|
---|
99 | list_of_local_search_paths.add(0, path);
|
---|
100 | } else {
|
---|
101 | list_of_local_search_paths.add(path);
|
---|
102 | }
|
---|
103 | }
|
---|
104 | } else {
|
---|
105 | DebugStream.println("### Location of file " + file + " not within GS. Not adding to list of search paths.");
|
---|
106 | }
|
---|
107 | }
|
---|
108 |
|
---|
109 | /**
|
---|
110 | * resolveEntity() is not called for every file:
|
---|
111 | * http://www.postseek.com/meta/37735b65e6a459a6aa631f048cc5a0b6
|
---|
112 | *
|
---|
113 | * "I think the parser will call [resolveEntity() on an EntityResolver that has been set], if
|
---|
114 | * [the parser] is unable to find the DTD that your XML file refers to. So if you try to parse
|
---|
115 | * an XML file that doesn't refer to a DTD, it won't be called. And if you try to parse an XML
|
---|
116 | * file where the parser can find the DTD, it won't be called either."
|
---|
117 | *
|
---|
118 | */
|
---|
119 | public InputSource resolveEntity (String public_id, String system_id) {
|
---|
120 |
|
---|
121 | DebugStream.println("### resolveEntity() called for " + system_id);
|
---|
122 |
|
---|
123 | String temp_id = system_id;
|
---|
124 | if (temp_id.startsWith("file://")) {
|
---|
125 | File f = new File(system_id);
|
---|
126 | if (f.exists()) {
|
---|
127 |
|
---|
128 | // check if inside GS2 or GS3 installation
|
---|
129 | if(f.getAbsolutePath().startsWith(Configuration.gsdl_path)
|
---|
130 | || f.getAbsolutePath().startsWith(Configuration.gsdl3_src_path)) {
|
---|
131 |
|
---|
132 | DebugStream.println("### file denoted by systemID is inside GS: " + f.getAbsolutePath());
|
---|
133 |
|
---|
134 | return new InputSource(system_id); // problem solved
|
---|
135 | } else {
|
---|
136 | DebugStream.println("\t### file denoted by systemID exists, but not located inside GS: "
|
---|
137 | + f.getAbsolutePath());
|
---|
138 | }
|
---|
139 | } else {
|
---|
140 |
|
---|
141 | temp_id = f.getName();
|
---|
142 |
|
---|
143 | //check in list of search paths
|
---|
144 | Iterator<File> i = list_of_local_search_paths.iterator();
|
---|
145 | while(i.hasNext()) {
|
---|
146 | File searchPath = i.next();
|
---|
147 | DebugStream.println("### searching for entity '" + temp_id + "' in: "
|
---|
148 | + searchPath.getAbsolutePath());
|
---|
149 |
|
---|
150 | File searchFile = new File(searchPath, temp_id);
|
---|
151 |
|
---|
152 | if(searchFile.exists()) {
|
---|
153 | DebugStream.println("\t Found " + searchFile.getAbsolutePath());
|
---|
154 |
|
---|
155 | String newpath = searchFile.getAbsolutePath();
|
---|
156 | if(Utility.isWindows()) {
|
---|
157 | newpath = "file:///" + newpath.replace("\\", "/");
|
---|
158 | } else { // linux version, file protocol starts with file:// and slashes are already URL-style
|
---|
159 | newpath = "file://" + newpath;
|
---|
160 | }
|
---|
161 | return new InputSource(newpath);
|
---|
162 |
|
---|
163 | }
|
---|
164 | }
|
---|
165 |
|
---|
166 | // else, external entity/file denoted by systemid is not in list of search paths, try classloader
|
---|
167 |
|
---|
168 | }
|
---|
169 | } else {
|
---|
170 |
|
---|
171 | DebugStream.println("### Entity is not a file: " + system_id);
|
---|
172 |
|
---|
173 | if (temp_id.indexOf("/")!= -1) {
|
---|
174 | temp_id = temp_id.substring(temp_id.lastIndexOf("/")+1);
|
---|
175 | }
|
---|
176 | }
|
---|
177 |
|
---|
178 | DebugStream.println("### Using classloader to attempt to resolve entity: " + temp_id);
|
---|
179 |
|
---|
180 | // try using a class loader. If none provided, use current class loader
|
---|
181 | if (this.class_loader==null) {
|
---|
182 | this.class_loader = this.getClass().getClassLoader();
|
---|
183 | }
|
---|
184 | URL url = class_loader.getResource(temp_id);
|
---|
185 | if (url == null) {
|
---|
186 | return null;
|
---|
187 | }
|
---|
188 |
|
---|
189 | return new InputSource("file://"+url.getFile());
|
---|
190 | }
|
---|
191 | }
|
---|