|
1
5
6 package com.hp.hpl.jena.rdf.arp;
7
8 import com.hp.hpl.jena.util.CharEncoding;
9
10 import java.io.IOException  ;
11 import java.io.InputStream ;
12 import java.io.InputStreamReader ;
13 import java.io.Reader ;
14 import java.io.UTFDataFormatException ;
15 import java.nio.charset.Charset ;
16
17 import org.apache.xerces.parsers.SAXParser;
18 import org.apache.xerces.parsers.StandardParserConfiguration;
19 import org.apache.xerces.util.EncodingMap;
20 import org.apache.xerces.xni.Augmentations;
21 import org.apache.xerces.xni.parser.XMLInputSource;
22 import org.apache.xerces.xni.parser.XMLPullParserConfiguration;
23 import org.xml.sax.*;
24
25
34 class SingleThreadedParser extends XMLHandler {
35
36 private XMLPullParserConfiguration pullParser;
37
38 private SAXParser saxParser;
39
40 private String readerXMLEncoding = null;
41
42 private String xmlEncoding = null;
43
44 private SingleThreadedParser(SAXParser rdr,
45 XMLPullParserConfiguration config) {
46 super();
47 pullParser = config;
48 saxParser = rdr;
49 try {
50 SAX2RDF.installHandlers(rdr, this);
51 } catch (SAXException e) {
52 throw new RuntimeException ("Supposedly impossible:", e);
53 }
54 }
56
57 SAXParser getSAXParser() {
58 return saxParser;
59 }
60
61 static private class MySAXParser extends SAXParser {
62 MySAXParser(StandardParserConfiguration c) {
63 super(c);
64 try {
65 setFeature("http://xml.org/sax/features/string-interning",
66 false);
67 } catch (SAXException e) {
68 }
72 }
73
74 SingleThreadedParser a;
75
76 public void xmlDecl(String version, String encoding, String standalone,
77 Augmentations augs) {
78 a.setEncoding(encoding == null ? "UTF" : encoding);
79 super.xmlDecl(version, encoding, standalone, augs);
80 }
81
88 }
89
90 static SingleThreadedParser create() {
91 StandardParserConfiguration c = new StandardParserConfiguration();
92 MySAXParser msp = new MySAXParser(c);
93 SingleThreadedParser a = new SingleThreadedParser(msp, c);
94 msp.a = a;
95 return a;
96 }
97
98 boolean parseSome() {
99 try {
100 return pullParser.parse(false);
101 } catch (UTFDataFormatException e) {
102 try {
103 generalError(ERR_UTF_ENCODING, e);
104 } catch (SAXParseException e1) {
105 }
107 return false;
108 } catch (IOException e) {
109 try {
110 generalError(ERR_GENERIC_IO, e);
111 } catch (SAXParseException e1) {
112 }
114 return false;
115 } catch (FatalParsingErrorException e) {
116 return false;
117 }
118 }
119
120 RDFParser rdfParser;
121
122 synchronized public void parse(InputSource input) throws IOException ,
123 SAXException {
124 parse(input, input.getSystemId());
125 }
126
127 synchronized public void parse(InputSource input, String base)
128 throws IOException , SAXException {
129 initParse(base);
132 pipe = new PullingTokenPipe(this);
134 pullParser.setInputSource(convert(input));
135
136 SAX2RDF.installHandlers(saxParser, this);
137 saxParser.reset();
138
139 try {
141 try {
142 rdfParser = new RDFParser(pipe, SingleThreadedParser.this);
143 if (getOptions().getEmbedding())
144 rdfParser.embeddedFile(documentContext);
145 else
146 rdfParser.rdfFile(documentContext);
147 } catch (WrappedException wrapped) {
148 wrapped.throwMe();
149 } catch (ParseException parse) {
150
153 userError(parse);
154
157 }
158 } finally {
159 endBnodeScope();
160 }
161
162 }
163
164 XMLInputSource convert(InputSource in) {
165 Reader rdr = in.getCharacterStream();
166 InputStream str = in.getByteStream();
167 String publicID = in.getPublicId();
168 String systemID = in.getSystemId();
169 readerXMLEncoding = null;
170 encodingProblems = false;
171 if (rdr == null && str == null) {
172 return new XMLInputSource(publicID, systemID, systemID);
173 } else if (rdr == null) {
174 return new XMLInputSource(publicID, systemID, systemID, str, null);
175 } else if (str == null) {
176 if (rdr instanceof InputStreamReader ) {
177 String javaEnc = ((InputStreamReader ) rdr).getEncoding();
178 readerXMLEncoding = CharEncoding.create(javaEnc).name();
179 }
180 return new XMLInputSource(publicID, systemID, systemID, rdr, null);
181 }
182 return null;
183 }
184
185 void setEncoding(String original) {
186
187 CharEncoding encodingInfo = CharEncoding.create(original);
188 String e = encodingInfo.name();
189 if (xmlEncoding == null) {
191 if (e.equals("UTF") && readerXMLEncoding != null
193 && readerXMLEncoding.startsWith("UTF")) {
194 xmlEncoding = readerXMLEncoding;
195 return;
196 }
197 xmlEncoding = e;
198 try {
199
200 if (readerXMLEncoding != null
201 && !readerXMLEncoding.equalsIgnoreCase(e)) {
202 putWarning(
203 WARN_ENCODING_MISMATCH,
204 new Location(locator),
205 "Encoding on InputStreamReader or FileReader does not match that of XML document. Use FileInputStream. ["
206 + readerXMLEncoding + " != " + e + "]");
207 encodingProblems = true;
208 }
209
210 if (e.equals("UTF"))
211 return;
212
213 if (!encodingInfo.isIANA()) {
214 putWarning(
215 encodingInfo.isInNIO() ? WARN_NON_IANA_ENCODING
216 : WARN_UNSUPPORTED_ENCODING, new Location(
217 locator), encodingInfo.warningMessage());
218 } else if (!original.equalsIgnoreCase(e)) {
219 putWarning(
220 WARN_NONCANONICAL_IANA_NAME,
221 new Location(locator),
222 "The encoding \""
223 + original
224 + "\" is not the canonical name at IANA, suggest \""
225 + e
226 + "\" would give more interoperability.");
227
228 }
229 } catch (SAXParseException e1) {
230 }
232
233 }
234 }
235
236 }
237
238
263
264 |
Popular Tags |
Java API By Example, From Geeks To Geeks.