|
1 // Copyright (c) 2001-2009 Nokia Corporation and/or its subsidiary(-ies). |
|
2 // All rights reserved. |
|
3 // This component and the accompanying materials are made available |
|
4 // under the terms of "Eclipse Public License v1.0" |
|
5 // which accompanies this distribution, and is available |
|
6 // at the URL "http://www.eclipse.org/legal/epl-v10.html". |
|
7 // |
|
8 // Initial Contributors: |
|
9 // Nokia Corporation - initial contribution. |
|
10 // |
|
11 // Contributors: |
|
12 // |
|
13 // Description: |
|
14 // MDXMLEntityConverter.cpp |
|
15 // @file |
|
16 // This class represents a generic entity converter for an XML parser or composer. |
|
17 // |
|
18 // |
|
19 |
|
20 #include <f32file.h> |
|
21 #include <stringpool.h> |
|
22 |
|
23 #include <gmxmlconstants.h> |
|
24 #include <gmxmlcomposer.h> |
|
25 #include <gmxmlentityconverter.h> |
|
26 #include "xmlentityreferences.h" |
|
27 |
|
28 const TInt KMaxBuiltinEntityLength = 20; |
|
29 |
|
30 // |
|
31 // CMDXMLEntityConverter // |
|
32 // |
|
33 |
|
34 EXPORT_C CMDXMLEntityConverter::CMDXMLEntityConverter() |
|
35 // |
|
36 // Constructor |
|
37 // |
|
38 { |
|
39 } |
|
40 |
|
41 EXPORT_C CMDXMLEntityConverter::~CMDXMLEntityConverter() |
|
42 { |
|
43 } |
|
44 |
|
45 //================================================================================== |
|
46 |
|
47 TInt CMDXMLEntityConverter::EntityToTextL(TDes& aTextToConvert) |
|
48 // |
|
49 // Takes a block of text and converts any entity references found to the |
|
50 // appropriate text. Because built-in and character entity references are |
|
51 // longer than the replacement text, this takes place in-situ. |
|
52 // @param aTextToConvert Text to be converted - replacement text goes |
|
53 // out in the same |
|
54 // |
|
55 { |
|
56 TInt error = KErrNone; |
|
57 TInt beginEntityOffset = 0; |
|
58 TInt beginSearchOffset = 0; |
|
59 TPtrC searchPtr = TPtrC(aTextToConvert); |
|
60 |
|
61 // Need to convert all entity refs in this bit of data. Don't want to convert |
|
62 // things more than once so we keep a marker and move it on so each time round |
|
63 // the loop we start looking for entity references after the end of the previous |
|
64 // one. |
|
65 while((beginEntityOffset = searchPtr.Locate('&')) != KErrNotFound) |
|
66 { |
|
67 searchPtr.Set(aTextToConvert.Mid(beginSearchOffset + beginEntityOffset)); |
|
68 TInt endEntityOffset = searchPtr.Locate(';'); |
|
69 if(endEntityOffset == KErrNotFound) |
|
70 { |
|
71 // No point continuing past here - even if we find more & they can't be |
|
72 // terminated either. |
|
73 error = KErrXMLBadEntity; |
|
74 break; |
|
75 } |
|
76 endEntityOffset += beginEntityOffset; |
|
77 // Can't construct a modifiable descriptor that's just part of another one. Looks |
|
78 // like we need to copy the section containing the reference and twiddle that, then |
|
79 // copy the result back if it's valid. |
|
80 HBufC* entityRef = searchPtr.Mid(0, (endEntityOffset - beginEntityOffset)+1).AllocLC(); |
|
81 TInt convertErr = KErrNone; |
|
82 TPtr entityRefDes = entityRef->Des(); |
|
83 if((convertErr = ConvertEntityRefL(entityRefDes)) == KErrNotSupported) |
|
84 { |
|
85 convertErr = DTDEntityToText(entityRefDes); |
|
86 } |
|
87 |
|
88 if(!convertErr) |
|
89 { |
|
90 // We need to replace this section of aTextToConvert with entityRefDes |
|
91 // we don't know whether DTDEntityToText has actually changed anything - |
|
92 // because of constraints on the published behaviour of DTDEntityToText we |
|
93 // can't compel DTDEntityToText to return KErrUnsupported - it thinks it |
|
94 // could be dealing with multiple entity refs and just cos it doesn't support |
|
95 // one of them doesn't mean it doesn't support the rest. Probably true that |
|
96 // the entity ref will be a builtin, so won't need to call DTDEntityToText |
|
97 // anyway. ConvertEntityRefL will return KErrNone only if it converted so |
|
98 // we probably won't be copying things gratuitously in the common case. |
|
99 aTextToConvert.Replace(beginEntityOffset + beginSearchOffset, (endEntityOffset - beginEntityOffset)+1, entityRefDes); |
|
100 } |
|
101 else |
|
102 { |
|
103 error = KErrXMLBadEntity; |
|
104 } |
|
105 |
|
106 beginSearchOffset += beginEntityOffset + entityRefDes.Length(); |
|
107 searchPtr.Set(aTextToConvert.Mid(beginSearchOffset)); |
|
108 CleanupStack::PopAndDestroy(entityRef); |
|
109 } |
|
110 |
|
111 return error; |
|
112 } |
|
113 |
|
114 //================================================================================== |
|
115 |
|
116 EXPORT_C TInt CMDXMLEntityConverter::OutputComposedTextL( CMDXMLComposer* aComposer, const TDesC& aTextToConvert) |
|
117 // |
|
118 // Outputs a block of text to a composer with offending characters replaced by |
|
119 // entity references. |
|
120 // @param aComposer the composer to be used for output |
|
121 // @param aTextToConvert The text to be converted. |
|
122 // @return Returns KErrNone if successful or a file write error |
|
123 // @leave can Leave due to OOM |
|
124 // |
|
125 { |
|
126 TBuf<1> oneChar; |
|
127 TInt offset = 0; |
|
128 TInt error = KErrNone; |
|
129 TInt textLen = aTextToConvert.Length(); |
|
130 |
|
131 while(error == KErrNone && (offset < textLen)) |
|
132 { |
|
133 oneChar = aTextToConvert.Mid(offset, 1); |
|
134 offset++; |
|
135 |
|
136 // Built in entities |
|
137 RStringPool pool; |
|
138 CleanupClosePushL(pool); |
|
139 pool.OpenL(XMLEntityReferences::Table); |
|
140 |
|
141 |
|
142 TBool found = EFalse; |
|
143 for (TUint i=0; i<XMLEntityReferences::Table.iCount; i++) |
|
144 { |
|
145 RStringF entity = pool.StringF(i, XMLEntityReferences::Table); |
|
146 TBuf<KMaxBuiltinEntityLength> buf; |
|
147 buf.Copy(entity.DesC()); |
|
148 TLex string(buf); |
|
149 |
|
150 TPtrC entityRef = string.NextToken(); |
|
151 if(oneChar.Compare(string.NextToken()) == 0) |
|
152 { |
|
153 error = aComposer->OutputDataL(entityRef); |
|
154 found = ETrue; |
|
155 break; |
|
156 } |
|
157 |
|
158 } |
|
159 |
|
160 CleanupStack::PopAndDestroy(); // close pool |
|
161 |
|
162 if(!found) |
|
163 error = aComposer->OutputDataL(oneChar); |
|
164 else |
|
165 found = EFalse; |
|
166 |
|
167 } |
|
168 |
|
169 return error; |
|
170 } |
|
171 |
|
172 //================================================================================== |
|
173 |
|
174 TInt CMDXMLEntityConverter::DTDEntityToText(TDes& /*aTextToConvert*/) |
|
175 // |
|
176 // DTD Specific entity to text converter |
|
177 // Takes a block of text and converts any entity references found to the |
|
178 // appropriate text. We hope that this can happen in-situ. |
|
179 // @param aTextToConvert Text to be converted - replacement text goes |
|
180 // out in the same |
|
181 // |
|
182 { // Nothing in the generic case. |
|
183 return KErrNone; |
|
184 } |
|
185 |
|
186 //================================================================================== |
|
187 |
|
188 TInt CMDXMLEntityConverter::ConvertEntityRefL(TDes& aTextToConvert) |
|
189 { |
|
190 #ifdef _DEBUG |
|
191 _LIT(KEntityConverter, "GMXML Entity Converter"); |
|
192 __ASSERT_ALWAYS(aTextToConvert.Length() > 1, User::Panic(KEntityConverter, KErrArgument)); |
|
193 #endif |
|
194 |
|
195 TInt err = KErrNotSupported; |
|
196 TBool charRef = EFalse; |
|
197 TInt idLen = 0; |
|
198 TRadix elementRadix = EDecimal; |
|
199 |
|
200 if(aTextToConvert.Find(KXMLHexidecimalId) != KErrNotFound) |
|
201 { |
|
202 charRef = ETrue; |
|
203 idLen = TPtrC(KXMLHexidecimalId).Length(); |
|
204 elementRadix = EHex; |
|
205 } |
|
206 else if(aTextToConvert.Find(KXMLDecimalId) != KErrNotFound) |
|
207 { |
|
208 charRef = ETrue; |
|
209 idLen = TPtrC(KXMLDecimalId).Length(); |
|
210 elementRadix = EDecimal; |
|
211 } |
|
212 |
|
213 if(charRef) |
|
214 { |
|
215 // Character entities |
|
216 /* |
|
217 algorithm explanation: |
|
218 The entity reference in the descriptor holding the string is known by it's positions. |
|
219 offset = offset to beginning of '&#x' from start of supplied descriptor. (0 indicates start of descriptor) |
|
220 entitylen = offset to ';' character from the start of the entity reference. |
|
221 |
|
222 the length of the entity preamble (&#x in this case) is known. |
|
223 |
|
224 The algorithm assigns the required portion of the descriptor to a TLex16 object so that the |
|
225 value of the code can be read out. |
|
226 |
|
227 The math for the assign is as follows: |
|
228 start of number = offset + length of preamble |
|
229 num of characters to take = entityLen - length of preamble |
|
230 |
|
231 The entity is then deleted from the string, and the new code is inserted in the |
|
232 correct location. |
|
233 The deleted characters are 'entityLen' charcters from 'offset'. |
|
234 |
|
235 |
|
236 e.g. for the string <a href="http://www.symbian.com"> |
|
237 |
|
238 offset = 16 |
|
239 entitylen = 5 |
|
240 string assigned to lexer = '77' |
|
241 code = 0x77 (extracted from lexer) |
|
242 |
|
243 descriptor after delete operation: |
|
244 |
|
245 <a href="http://&ww.symbian.com"> |
|
246 |
|
247 array dereference fills in the descriptor such that the ampersand is replaced. |
|
248 <a href="http://www.symbian.com"> |
|
249 */ |
|
250 TInt entityLen = aTextToConvert.Length(); |
|
251 |
|
252 TLex16 element; |
|
253 element.Assign(aTextToConvert.Mid(idLen, (entityLen - idLen))); |
|
254 |
|
255 TUint16 code; |
|
256 err = element.Val(code, elementRadix); |
|
257 |
|
258 if(!err) |
|
259 { |
|
260 aTextToConvert.Delete(1, entityLen); |
|
261 aTextToConvert[0] = code; |
|
262 } |
|
263 } |
|
264 else |
|
265 { |
|
266 // Built in entities |
|
267 RStringPool pool; |
|
268 CleanupClosePushL(pool); |
|
269 pool.OpenL(XMLEntityReferences::Table); |
|
270 |
|
271 |
|
272 for (TUint i=0; (i<XMLEntityReferences::Table.iCount) && (err != KErrNone); i++) |
|
273 { |
|
274 RStringF entity = pool.StringF(i, XMLEntityReferences::Table); |
|
275 TBuf<KMaxBuiltinEntityLength> buf; |
|
276 buf.Copy(entity.DesC()); |
|
277 TLex string(buf); |
|
278 |
|
279 TPtrC token = string.NextToken(); |
|
280 TInt offset; |
|
281 if((offset = aTextToConvert.Find(token)) != KErrNotFound) |
|
282 { |
|
283 aTextToConvert.Replace(offset, token.Length(), string.NextToken()); |
|
284 err = KErrNone; |
|
285 } |
|
286 } |
|
287 CleanupStack::PopAndDestroy(); //close pool |
|
288 } |
|
289 |
|
290 return err; |
|
291 } |
|
292 |
|
293 // End Of File |