|
1 /**************************************************************************** |
|
2 ** |
|
3 ** Copyright (C) 2010 Nokia Corporation and/or its subsidiary(-ies). |
|
4 ** All rights reserved. |
|
5 ** Contact: Nokia Corporation (qt-info@nokia.com) |
|
6 ** |
|
7 ** This file is part of the Qt Mobility Components. |
|
8 ** |
|
9 ** $QT_BEGIN_LICENSE:LGPL$ |
|
10 ** No Commercial Usage |
|
11 ** This file contains pre-release code and may not be distributed. |
|
12 ** You may use this file in accordance with the terms and conditions |
|
13 ** contained in the Technology Preview License Agreement accompanying |
|
14 ** this package. |
|
15 ** |
|
16 ** GNU Lesser General Public License Usage |
|
17 ** Alternatively, this file may be used under the terms of the GNU Lesser |
|
18 ** General Public License version 2.1 as published by the Free Software |
|
19 ** Foundation and appearing in the file LICENSE.LGPL included in the |
|
20 ** packaging of this file. Please review the following information to |
|
21 ** ensure the GNU Lesser General Public License version 2.1 requirements |
|
22 ** will be met: http://www.gnu.org/licenses/old-licenses/lgpl-2.1.html. |
|
23 ** |
|
24 ** In addition, as a special exception, Nokia gives you certain additional |
|
25 ** rights. These rights are described in the Nokia Qt LGPL Exception |
|
26 ** version 1.1, included in the file LGPL_EXCEPTION.txt in this package. |
|
27 ** |
|
28 ** If you have questions regarding the use of this file, please contact |
|
29 ** Nokia at qt-info@nokia.com. |
|
30 ** |
|
31 ** |
|
32 ** |
|
33 ** |
|
34 ** |
|
35 ** |
|
36 ** |
|
37 ** |
|
38 ** $QT_END_LICENSE$ |
|
39 ** |
|
40 ****************************************************************************/ |
|
41 |
|
42 #include "qversitreader_p.h" |
|
43 #include "qversitdocument.h" |
|
44 #include "versitutils_p.h" |
|
45 #include "qmobilityglobal.h" |
|
46 #include <QTextCodec> |
|
47 #include <QMutexLocker> |
|
48 #include <QVariant> |
|
49 #include <QBuffer> |
|
50 |
|
51 QTM_USE_NAMESPACE |
|
52 |
|
53 // Some big enough value for nested versit documents to prevent infinite recursion |
|
54 #define MAX_VERSIT_DOCUMENT_NESTING_DEPTH 20 |
|
55 |
|
56 /*! |
|
57 \class LineReader |
|
58 \brief The LineReader class is a wrapper around a QIODevice that allows line-by-line reading. |
|
59 \internal |
|
60 |
|
61 This class keeps an internal buffer which it uses to temporarily store data which it has read from |
|
62 the device but not returned to the user. |
|
63 */ |
|
64 |
|
65 /*! |
|
66 Constructs a LineReader that reads from the given \a device using the given \a codec. |
|
67 \a chunkSize is the number of bytes to read at a time (it is useful for testing but shouldn't |
|
68 otherwise be set). |
|
69 */ |
|
70 LineReader::LineReader(QIODevice* device, QTextCodec *codec, int chunkSize) |
|
71 : mDevice(device), |
|
72 mCodec(codec), |
|
73 mChunkSize(chunkSize), |
|
74 mCrlfList(*VersitUtils::newlineList(mCodec)), |
|
75 mBuffer(VersitCursor(QByteArray())), |
|
76 mOdometer(0) |
|
77 { |
|
78 } |
|
79 |
|
80 /*! |
|
81 Attempts to read a line and returns a VersitCursor describing the line. The cursor returned |
|
82 includes the data, as well as the position and selection index bounds. Data within those bounds |
|
83 represents the line. Data outside those bounds should not be used. |
|
84 */ |
|
85 VersitCursor LineReader::readLine() |
|
86 { |
|
87 mBuffer.position = mBuffer.selection; |
|
88 mSearchFrom = mBuffer.position; |
|
89 |
|
90 // First, look for a newline in the already-existing buffer. If found, return the line. |
|
91 if (tryReadLine(mBuffer, false)) { |
|
92 mBuffer.dropOldData(); |
|
93 mOdometer += mBuffer.selection - mBuffer.position; |
|
94 return mBuffer; |
|
95 } |
|
96 |
|
97 // Otherwise, keep reading more data until either a CRLF is found, or there's no more to read. |
|
98 while (!mDevice->atEnd()) { |
|
99 QByteArray temp = mDevice->read(mChunkSize); |
|
100 if (!temp.isEmpty()) { |
|
101 mBuffer.data.append(temp); |
|
102 if (tryReadLine(mBuffer, false)) { |
|
103 mBuffer.dropOldData(); |
|
104 mOdometer += mBuffer.selection - mBuffer.position; |
|
105 return mBuffer; |
|
106 } |
|
107 } else { |
|
108 mDevice->waitForReadyRead(500); |
|
109 } |
|
110 } |
|
111 |
|
112 // We've reached the end of the stream. Find a newline from the buffer (or return what's left). |
|
113 tryReadLine(mBuffer, true); |
|
114 mBuffer.dropOldData(); |
|
115 mOdometer += mBuffer.selection - mBuffer.position; |
|
116 return mBuffer; |
|
117 } |
|
118 |
|
119 /*! |
|
120 How many bytes have been returned in the VersitCursor in the lifetime of the LineReader. |
|
121 */ |
|
122 int LineReader::odometer() |
|
123 { |
|
124 return mOdometer; |
|
125 } |
|
126 |
|
127 /*! |
|
128 Returns true if there are no more lines left for readLine() to return. It is possible for atEnd() |
|
129 to return false and for there to be no more data left (eg. if there are trailing newlines at the |
|
130 end of the input. In this case, readLine() will return an empty line (ie. position == selection). |
|
131 */ |
|
132 bool LineReader::atEnd() |
|
133 { |
|
134 return mDevice->atEnd() && mBuffer.selection == mBuffer.data.size(); |
|
135 } |
|
136 |
|
137 /*! |
|
138 Returns the codec that the LineReader reads with. |
|
139 */ |
|
140 QTextCodec* LineReader::codec() |
|
141 { |
|
142 return mCodec; |
|
143 } |
|
144 |
|
145 /*! |
|
146 * Get the next line of input from \a device to parse. Also performs unfolding by removing |
|
147 * sequences of newline-space from the retrieved line. Skips over any newlines at the start of the |
|
148 * input. |
|
149 * |
|
150 * Returns a VersitCursor containing and selecting the line. |
|
151 */ |
|
152 bool LineReader::tryReadLine(VersitCursor &cursor, bool atEnd) |
|
153 { |
|
154 int crlfPos = -1; |
|
155 |
|
156 QByteArray space = VersitUtils::encode(' ', mCodec); |
|
157 QByteArray tab = VersitUtils::encode('\t', mCodec); |
|
158 int spaceLength = space.length(); |
|
159 |
|
160 forever { |
|
161 foreach(const QByteArrayMatcher& crlf, mCrlfList) { |
|
162 int crlfLength = crlf.pattern().length(); |
|
163 crlfPos = crlf.indexIn(cursor.data, mSearchFrom); |
|
164 if (crlfPos == cursor.position) { |
|
165 // Newline at start of line. Set position to directly after it. |
|
166 cursor.position += crlfLength; |
|
167 mSearchFrom = cursor.position; |
|
168 break; |
|
169 } else if (crlfPos > cursor.position) { |
|
170 // Found the CRLF. |
|
171 if (QVersitReaderPrivate::containsAt(cursor.data, space, crlfPos + crlfLength) |
|
172 || QVersitReaderPrivate::containsAt(cursor.data, tab, crlfPos + crlfLength)) { |
|
173 // If it's followed by whitespace, collapse it. |
|
174 cursor.data.remove(crlfPos, crlfLength + spaceLength); |
|
175 mSearchFrom = crlfPos; |
|
176 break; |
|
177 } else if (!atEnd && crlfPos + crlfLength + spaceLength >= cursor.data.size()) { |
|
178 // If our CRLF is at the end of the current buffer but there's more to read, |
|
179 // it's possible that a space could be hiding on the next read from the device. |
|
180 // Just pretend we didn't see the CRLF and pick it up the next time round. |
|
181 mSearchFrom = crlfPos; |
|
182 return false; |
|
183 } else { |
|
184 // Found the CRLF. |
|
185 cursor.selection = crlfPos; |
|
186 return true; |
|
187 } |
|
188 } |
|
189 } |
|
190 if (crlfPos == -1) { |
|
191 // No CRLF found. |
|
192 cursor.selection = cursor.data.size(); |
|
193 return false; |
|
194 } |
|
195 } |
|
196 } |
|
197 |
|
198 /*! Links the signals from this to the signals of \a reader. */ |
|
199 void QVersitReaderPrivate::init(QVersitReader* reader) |
|
200 { |
|
201 qRegisterMetaType<QVersitReader::State>("QVersitReader::State"); |
|
202 connect(this, SIGNAL(stateChanged(QVersitReader::State)), |
|
203 reader, SIGNAL(stateChanged(QVersitReader::State)),Qt::DirectConnection); |
|
204 connect(this, SIGNAL(resultsAvailable()), |
|
205 reader, SIGNAL(resultsAvailable()), Qt::DirectConnection); |
|
206 } |
|
207 |
|
208 /*! Construct a reader. */ |
|
209 QVersitReaderPrivate::QVersitReaderPrivate() |
|
210 : mIoDevice(0), |
|
211 mDocumentNestingLevel(0), |
|
212 mDefaultCodec(QTextCodec::codecForName("UTF-8")), |
|
213 mState(QVersitReader::InactiveState), |
|
214 mError(QVersitReader::NoError), |
|
215 mIsCanceling(false) |
|
216 { |
|
217 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("AGENT")), |
|
218 QVersitProperty::VersitDocumentType); |
|
219 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("AGENT")), |
|
220 QVersitProperty::VersitDocumentType); |
|
221 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("N")), |
|
222 QVersitProperty::CompoundType); |
|
223 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("N")), |
|
224 QVersitProperty::CompoundType); |
|
225 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("ADR")), |
|
226 QVersitProperty::CompoundType); |
|
227 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("ADR")), |
|
228 QVersitProperty::CompoundType); |
|
229 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("GEO")), |
|
230 QVersitProperty::CompoundType); |
|
231 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("GEO")), |
|
232 QVersitProperty::CompoundType); |
|
233 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("ORG")), |
|
234 QVersitProperty::CompoundType); |
|
235 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("ORG")), |
|
236 QVersitProperty::CompoundType); |
|
237 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("NICKNAME")), |
|
238 QVersitProperty::ListType); |
|
239 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("NICKNAME")), |
|
240 QVersitProperty::ListType); |
|
241 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("CATEGORIES")), |
|
242 QVersitProperty::ListType); |
|
243 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("CATEGORIES")), |
|
244 QVersitProperty::ListType); |
|
245 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("X-CHILDREN")), |
|
246 QVersitProperty::ListType); |
|
247 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("X-CHILDREN")), |
|
248 QVersitProperty::ListType); |
|
249 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard21Type, QString::fromAscii("X-NICKNAME")), |
|
250 QVersitProperty::ListType); |
|
251 mValueTypeMap.insert(qMakePair(QVersitDocument::VCard30Type, QString::fromAscii("X-NICKNAME")), |
|
252 QVersitProperty::ListType); |
|
253 } |
|
254 |
|
255 /*! Destroy a reader. */ |
|
256 QVersitReaderPrivate::~QVersitReaderPrivate() |
|
257 { |
|
258 } |
|
259 |
|
260 /*! |
|
261 * Inherited from QThread, called by QThread when the thread has been started. |
|
262 */ |
|
263 void QVersitReaderPrivate::run() |
|
264 { |
|
265 read(); |
|
266 } |
|
267 |
|
268 /*! |
|
269 * Does the actual reading and sets the error and state as appropriate. |
|
270 * If \a async, then stateChanged() signals are emitted as the reading happens. |
|
271 */ |
|
272 void QVersitReaderPrivate::read() |
|
273 { |
|
274 mMutex.lock(); |
|
275 mVersitDocuments.clear(); |
|
276 mMutex.unlock(); |
|
277 bool canceled = false; |
|
278 |
|
279 LineReader lineReader(mIoDevice, mDefaultCodec); |
|
280 while(!lineReader.atEnd()) { |
|
281 if (isCanceling()) { |
|
282 canceled = true; |
|
283 break; |
|
284 } |
|
285 QVersitDocument document; |
|
286 int oldPos = lineReader.odometer(); |
|
287 bool ok = parseVersitDocument(lineReader, document); |
|
288 |
|
289 if (ok) { |
|
290 if (document.isEmpty()) |
|
291 break; |
|
292 else { |
|
293 QMutexLocker locker(&mMutex); |
|
294 mVersitDocuments.append(document); |
|
295 emit resultsAvailable(); |
|
296 } |
|
297 } else { |
|
298 setError(QVersitReader::ParseError); |
|
299 if (lineReader.odometer() == oldPos) |
|
300 break; |
|
301 } |
|
302 }; |
|
303 if (canceled) |
|
304 setState(QVersitReader::CanceledState); |
|
305 else |
|
306 setState(QVersitReader::FinishedState); |
|
307 } |
|
308 |
|
309 void QVersitReaderPrivate::setState(QVersitReader::State state) |
|
310 { |
|
311 mMutex.lock(); |
|
312 mState = state; |
|
313 mMutex.unlock(); |
|
314 emit stateChanged(state); |
|
315 } |
|
316 |
|
317 QVersitReader::State QVersitReaderPrivate::state() const |
|
318 { |
|
319 QMutexLocker locker(&mMutex); |
|
320 return mState; |
|
321 } |
|
322 |
|
323 void QVersitReaderPrivate::setError(QVersitReader::Error error) |
|
324 { |
|
325 QMutexLocker locker(&mMutex); |
|
326 mError = error; |
|
327 } |
|
328 |
|
329 QVersitReader::Error QVersitReaderPrivate::error() const |
|
330 { |
|
331 QMutexLocker locker(&mMutex); |
|
332 return mError; |
|
333 } |
|
334 |
|
335 void QVersitReaderPrivate::setCanceling(bool canceling) |
|
336 { |
|
337 QMutexLocker locker(&mMutex); |
|
338 mIsCanceling = canceling; |
|
339 } |
|
340 |
|
341 bool QVersitReaderPrivate::isCanceling() |
|
342 { |
|
343 QMutexLocker locker(&mMutex); |
|
344 return mIsCanceling; |
|
345 } |
|
346 |
|
347 /*! |
|
348 * Parses a versit document. Returns true if the parsing was successful. |
|
349 */ |
|
350 bool QVersitReaderPrivate::parseVersitDocument(LineReader& lineReader, QVersitDocument& document, |
|
351 bool foundBegin) |
|
352 { |
|
353 if (mDocumentNestingLevel >= MAX_VERSIT_DOCUMENT_NESTING_DEPTH) |
|
354 return false; // To prevent infinite recursion |
|
355 |
|
356 bool parsingOk = true; |
|
357 mDocumentNestingLevel++; |
|
358 |
|
359 // TODO: Various readers should be made subclasses and eliminate assumptions like this. |
|
360 // We don't know what type it is: just assume it's a vCard 3.0 |
|
361 document.setType(QVersitDocument::VCard30Type); |
|
362 |
|
363 QVersitProperty property; |
|
364 |
|
365 if (!foundBegin) { |
|
366 property = parseNextVersitProperty(document.type(), lineReader); |
|
367 if (property.name() == QLatin1String("BEGIN") |
|
368 && property.value().trimmed().toUpper() == QLatin1String("VCARD")) { |
|
369 foundBegin = true; |
|
370 } else if (property.isEmpty()) { |
|
371 // A blank document (or end of file) was found. |
|
372 document = QVersitDocument(); |
|
373 } else { |
|
374 // Some property other than BEGIN was found. |
|
375 parsingOk = false; |
|
376 } |
|
377 } |
|
378 |
|
379 if (foundBegin) { |
|
380 do { |
|
381 /* Grab it */ |
|
382 property = parseNextVersitProperty(document.type(), lineReader); |
|
383 |
|
384 /* Discard embedded vcard documents - not supported yet. Discard the entire vCard */ |
|
385 if (property.name() == QLatin1String("BEGIN") && |
|
386 QString::compare(property.value().trimmed(), |
|
387 QLatin1String("VCARD"), Qt::CaseInsensitive) == 0) { |
|
388 parsingOk = false; |
|
389 QVersitDocument nestedDocument; |
|
390 if (!parseVersitDocument(lineReader, nestedDocument, true)) |
|
391 break; |
|
392 } |
|
393 |
|
394 // See if this is a version property and continue parsing under that version |
|
395 if (!setVersionFromProperty(document, property)) { |
|
396 parsingOk = false; |
|
397 break; |
|
398 } |
|
399 |
|
400 /* Nope, something else.. just add it */ |
|
401 if (property.name() != QLatin1String("VERSION") && |
|
402 property.name() != QLatin1String("END")) |
|
403 document.addProperty(property); |
|
404 } while (property.name().length() > 0 && property.name() != QLatin1String("END")); |
|
405 if (property.name() != QLatin1String("END")) |
|
406 parsingOk = false; |
|
407 } |
|
408 mDocumentNestingLevel--; |
|
409 if (!parsingOk) |
|
410 document = QVersitDocument(); |
|
411 |
|
412 return parsingOk; |
|
413 } |
|
414 |
|
415 /*! |
|
416 * Parses a versit document and returns whether parsing succeeded. |
|
417 */ |
|
418 QVersitProperty QVersitReaderPrivate::parseNextVersitProperty( |
|
419 QVersitDocument::VersitType versitType, |
|
420 LineReader& lineReader) |
|
421 { |
|
422 VersitCursor cursor = lineReader.readLine(); |
|
423 if (cursor.position >= cursor.selection) |
|
424 return QVersitProperty(); |
|
425 |
|
426 // Otherwise, do stuff. |
|
427 QPair<QStringList,QString> groupsAndName = |
|
428 extractPropertyGroupsAndName(cursor, lineReader.codec()); |
|
429 |
|
430 QVersitProperty property; |
|
431 property.setGroups(groupsAndName.first); |
|
432 property.setName(groupsAndName.second); |
|
433 // set the propertyValueType |
|
434 QPair<QVersitDocument::VersitType, QString> key = |
|
435 qMakePair(versitType, property.name()); |
|
436 if (mValueTypeMap.contains(key)) |
|
437 property.setValueType(mValueTypeMap.value(key)); |
|
438 |
|
439 if (versitType == QVersitDocument::VCard21Type) |
|
440 parseVCard21Property(cursor, property, lineReader); |
|
441 else if (versitType == QVersitDocument::VCard30Type) |
|
442 parseVCard30Property(cursor, property, lineReader); |
|
443 |
|
444 return property; |
|
445 } |
|
446 |
|
447 /*! |
|
448 * Parses the property according to vCard 2.1 syntax. |
|
449 */ |
|
450 void QVersitReaderPrivate::parseVCard21Property(VersitCursor& cursor, QVersitProperty& property, |
|
451 LineReader& lineReader) |
|
452 { |
|
453 property.setParameters(extractVCard21PropertyParams(cursor, lineReader.codec())); |
|
454 |
|
455 QByteArray value = extractPropertyValue(cursor); |
|
456 if (property.valueType() == QVersitProperty::VersitDocumentType) { |
|
457 // Hack to handle cases where start of document is on the same or next line as "AGENT:" |
|
458 bool foundBegin = false; |
|
459 if (value == "BEGIN:VCARD") { |
|
460 foundBegin = true; |
|
461 } else if (value.isEmpty()) { |
|
462 } else { |
|
463 property = QVersitProperty(); |
|
464 return; |
|
465 } |
|
466 QVersitDocument subDocument; |
|
467 if (!parseVersitDocument(lineReader, subDocument, foundBegin)) { |
|
468 property = QVersitProperty(); |
|
469 } else { |
|
470 property.setValue(QVariant::fromValue(subDocument)); |
|
471 } |
|
472 } else { |
|
473 QTextCodec* codec; |
|
474 bool isBinary = unencode(value, cursor, property, lineReader); |
|
475 if (isBinary) { |
|
476 property.setValue(value); |
|
477 property.setValueType(QVersitProperty::BinaryType); |
|
478 } |
|
479 else { |
|
480 property.setValue(decodeCharset(value, property, lineReader.codec(), &codec)); |
|
481 splitStructuredValue(property, false); |
|
482 } |
|
483 } |
|
484 } |
|
485 |
|
486 /*! |
|
487 * Parses the property according to vCard 3.0 syntax. |
|
488 */ |
|
489 void QVersitReaderPrivate::parseVCard30Property(VersitCursor& cursor, QVersitProperty& property, |
|
490 LineReader& lineReader) |
|
491 { |
|
492 property.setParameters(extractVCard30PropertyParams(cursor, lineReader.codec())); |
|
493 |
|
494 QByteArray value = extractPropertyValue(cursor); |
|
495 |
|
496 QTextCodec* codec; |
|
497 |
|
498 if (property.valueType() == QVersitProperty::VersitDocumentType) { |
|
499 QString valueString(decodeCharset(value, property, lineReader.codec(), &codec)); |
|
500 removeBackSlashEscaping(valueString); |
|
501 // Make a line reader from the value of the property. |
|
502 QByteArray subDocumentValue(codec->fromUnicode(valueString)); |
|
503 QBuffer subDocumentData(&subDocumentValue); |
|
504 subDocumentData.open(QIODevice::ReadOnly); |
|
505 subDocumentData.seek(0); |
|
506 LineReader subDocumentLineReader(&subDocumentData, codec); |
|
507 |
|
508 QVersitDocument subDocument; |
|
509 if (!parseVersitDocument(subDocumentLineReader, subDocument)) { |
|
510 property = QVersitProperty(); |
|
511 } else { |
|
512 property.setValue(QVariant::fromValue(subDocument)); |
|
513 } |
|
514 } else { |
|
515 bool isBinary = unencode(value, cursor, property, lineReader); |
|
516 if (isBinary) { |
|
517 property.setValue(value); |
|
518 property.setValueType(QVersitProperty::BinaryType); |
|
519 } else { |
|
520 property.setValue(decodeCharset(value, property, lineReader.codec(), &codec)); |
|
521 bool isList = splitStructuredValue(property, true); |
|
522 // Do backslash unescaping |
|
523 if (isList) { |
|
524 QStringList list = property.value<QStringList>(); |
|
525 for (int i = 0; i < list.length(); i++) { |
|
526 removeBackSlashEscaping(list[i]); |
|
527 } |
|
528 property.setValue(list); |
|
529 } else { |
|
530 QString value = property.value(); |
|
531 removeBackSlashEscaping(value); |
|
532 property.setValue(value); |
|
533 } |
|
534 } |
|
535 } |
|
536 } |
|
537 |
|
538 /*! |
|
539 * Sets version to \a document if \a property contains a supported version. |
|
540 */ |
|
541 bool QVersitReaderPrivate::setVersionFromProperty(QVersitDocument& document, const QVersitProperty& property) const |
|
542 { |
|
543 bool valid = true; |
|
544 if (property.name() == QLatin1String("VERSION")) { |
|
545 QString value = property.value().trimmed(); |
|
546 QStringList encodingParameters = property.parameters().values(QLatin1String("ENCODING")); |
|
547 QStringList typeParameters = property.parameters().values(QLatin1String("TYPE")); |
|
548 if (encodingParameters.contains(QLatin1String("BASE64"), Qt::CaseInsensitive) |
|
549 || typeParameters.contains(QLatin1String("BASE64"), Qt::CaseInsensitive)) |
|
550 value = QLatin1String(QByteArray::fromBase64(value.toAscii())); |
|
551 if (value == QLatin1String("2.1")) { |
|
552 document.setType(QVersitDocument::VCard21Type); |
|
553 } else if (value == QLatin1String("3.0")) { |
|
554 document.setType(QVersitDocument::VCard30Type); |
|
555 } else { |
|
556 valid = false; |
|
557 } |
|
558 } |
|
559 return valid; |
|
560 } |
|
561 |
|
562 /*! |
|
563 * On entry, \a value should be the byte array to unencode. It is modified to be the unencoded |
|
564 * version. Returns true if and only if the value was base-64 encoded. \a cursor and |
|
565 * \a lineReader are supplied in case more lines need to be read (for quoted-printable). The |
|
566 * \a property is supplied so we know what kind of encoding was used. |
|
567 */ |
|
568 bool QVersitReaderPrivate::unencode(QByteArray& value, VersitCursor& cursor, |
|
569 QVersitProperty& property, |
|
570 LineReader& lineReader) const |
|
571 { |
|
572 QStringList encodingParameters = property.parameters().values(QLatin1String("ENCODING")); |
|
573 QStringList typeParameters = property.parameters().values(QLatin1String("TYPE")); |
|
574 if (encodingParameters.contains(QLatin1String("QUOTED-PRINTABLE"), Qt::CaseInsensitive)) { |
|
575 // At this point, we need to accumulate bytes until we hit a real line break (no = before |
|
576 // it) value already contains everything up to the character before the newline |
|
577 while (value.endsWith('=')) { |
|
578 value.chop(1); // Get rid of '=' |
|
579 // We add each line (minus the escaped = and newline chars) |
|
580 cursor = lineReader.readLine(); |
|
581 QByteArray line = cursor.data.mid(cursor.position, cursor.selection-cursor.position); |
|
582 value.append(line); |
|
583 } |
|
584 decodeQuotedPrintable(value); |
|
585 // Remove the encoding parameter as the value is now decoded |
|
586 property.removeParameters(QLatin1String("ENCODING")); |
|
587 return false; |
|
588 } else if (encodingParameters.contains(QLatin1String("BASE64"), Qt::CaseInsensitive) |
|
589 || encodingParameters.contains(QLatin1String("B"), Qt::CaseInsensitive) |
|
590 || typeParameters.contains(QLatin1String("BASE64"), Qt::CaseInsensitive) |
|
591 || typeParameters.contains(QLatin1String("B"), Qt::CaseInsensitive)) { |
|
592 value = QByteArray::fromBase64(value); |
|
593 // Remove the encoding parameter as the value is now decoded |
|
594 property.removeParameters(QLatin1String("ENCODING")); |
|
595 return true; |
|
596 } |
|
597 return false; |
|
598 } |
|
599 |
|
600 /*! |
|
601 * Decodes \a value, after working out what charset it is in using the context of \a property and |
|
602 * returns it. The codec used to decode is returned in \a codec. |
|
603 */ |
|
604 QString QVersitReaderPrivate::decodeCharset(const QByteArray& value, |
|
605 QVersitProperty& property, |
|
606 QTextCodec* defaultCodec, |
|
607 QTextCodec** codec) const |
|
608 { |
|
609 const QString charset(QLatin1String("CHARSET")); |
|
610 if (property.parameters().contains(charset)) { |
|
611 QString charsetValue = *property.parameters().find(charset); |
|
612 property.removeParameters(charset); |
|
613 *codec = QTextCodec::codecForName(charsetValue.toAscii()); |
|
614 if (*codec != NULL) { |
|
615 return (*codec)->toUnicode(value); |
|
616 } else { |
|
617 *codec = defaultCodec; |
|
618 return defaultCodec->toUnicode(value); |
|
619 } |
|
620 } |
|
621 *codec = defaultCodec; |
|
622 return defaultCodec->toUnicode(value); |
|
623 } |
|
624 |
|
625 /*! |
|
626 * Decodes Quoted-Printable encoded (RFC 1521) characters in /a text. |
|
627 */ |
|
628 void QVersitReaderPrivate::decodeQuotedPrintable(QByteArray& text) const |
|
629 { |
|
630 for (int i=0; i < text.length(); i++) { |
|
631 char current = text.at(i); |
|
632 if (current == '=' && i+2 < text.length()) { |
|
633 char next = text.at(i+1); |
|
634 char nextAfterNext = text.at(i+2); |
|
635 if (((next >= 'a' && next <= 'f') || |
|
636 (next >= 'A' && next <= 'F') || |
|
637 (next >= '0' && next <= '9')) && |
|
638 ((nextAfterNext >= 'a' && nextAfterNext <= 'f') || |
|
639 (nextAfterNext >= 'A' && nextAfterNext <= 'F') || |
|
640 (nextAfterNext >= '0' && nextAfterNext <= '9'))) { |
|
641 bool ok; |
|
642 char decodedChar(text.mid(i+1, 2).toInt(&ok,16)); |
|
643 if (ok) { |
|
644 text[i] = decodedChar; |
|
645 text.remove(i+1, 2); |
|
646 } |
|
647 } else if (next == '\r' && nextAfterNext == '\n') { |
|
648 // Newlines can still be found here if they are encoded in a non-default charset. |
|
649 text.remove(i, 3); |
|
650 } |
|
651 } |
|
652 } |
|
653 } |
|
654 |
|
655 /*! |
|
656 * Extracts the groups and the name of the property using \a codec to determine the delimiters |
|
657 * |
|
658 * On entry, \a line should select a whole line. |
|
659 * On exit, \a line will be updated to point after the groups and name. |
|
660 */ |
|
661 QPair<QStringList,QString>QVersitReaderPrivate::extractPropertyGroupsAndName( |
|
662 VersitCursor& line, QTextCodec *codec) const |
|
663 { |
|
664 const QByteArray semicolon = VersitUtils::encode(';', codec); |
|
665 const QByteArray colon = VersitUtils::encode(':', codec); |
|
666 const QByteArray backslash = VersitUtils::encode('\\', codec); |
|
667 QPair<QStringList,QString> groupsAndName; |
|
668 int length = 0; |
|
669 Q_ASSERT(line.data.size() >= line.position); |
|
670 |
|
671 int separatorLength = semicolon.length(); |
|
672 for (int i = line.position; i < line.selection - separatorLength + 1; i++) { |
|
673 if ((containsAt(line.data, semicolon, i) |
|
674 && !containsAt(line.data, backslash, i-separatorLength)) |
|
675 || containsAt(line.data, colon, i)) { |
|
676 length = i - line.position; |
|
677 break; |
|
678 } |
|
679 } |
|
680 if (length > 0) { |
|
681 QString trimmedGroupsAndName = |
|
682 codec->toUnicode(line.data.mid(line.position, length)).trimmed(); |
|
683 QStringList parts = trimmedGroupsAndName.split(QLatin1Char('.')); |
|
684 if (parts.count() > 1) { |
|
685 groupsAndName.second = parts.takeLast(); |
|
686 groupsAndName.first = parts; |
|
687 } else { |
|
688 groupsAndName.second = trimmedGroupsAndName; |
|
689 } |
|
690 line.setPosition(length + line.position); |
|
691 } |
|
692 |
|
693 return groupsAndName; |
|
694 } |
|
695 |
|
696 /*! |
|
697 * Extracts the value of the property. |
|
698 * Returns an empty string if the value was not found. |
|
699 * |
|
700 * On entry \a line should point to the value anyway. |
|
701 * On exit \a line should point to newline after the value |
|
702 */ |
|
703 QByteArray QVersitReaderPrivate::extractPropertyValue(VersitCursor& line) const |
|
704 { |
|
705 QByteArray value = line.data.mid(line.position, line.selection - line.position); |
|
706 |
|
707 /* Now advance the cursor in all cases. */ |
|
708 line.position = line.selection; |
|
709 return value; |
|
710 } |
|
711 |
|
712 /*! |
|
713 * Extracts the property parameters as a QMultiHash using \a codec to determine the delimiters. |
|
714 * The parameters without names are added as "TYPE" parameters. |
|
715 * |
|
716 * On entry \a line should contain the entire line. |
|
717 * On exit, line will be updated to point to the start of the value. |
|
718 */ |
|
719 QMultiHash<QString,QString> QVersitReaderPrivate::extractVCard21PropertyParams( |
|
720 VersitCursor& line, QTextCodec *codec) const |
|
721 { |
|
722 QMultiHash<QString,QString> result; |
|
723 QList<QByteArray> paramList = extractParams(line, codec); |
|
724 while (!paramList.isEmpty()) { |
|
725 QByteArray param = paramList.takeLast(); |
|
726 QString name = paramName(param, codec); |
|
727 QString value = paramValue(param, codec); |
|
728 result.insert(name,value); |
|
729 } |
|
730 |
|
731 return result; |
|
732 } |
|
733 |
|
734 /*! |
|
735 * Extracts the property parameters as a QMultiHash using \a codec to determine the delimiters. |
|
736 * The parameters without names are added as "TYPE" parameters. |
|
737 */ |
|
738 QMultiHash<QString,QString> QVersitReaderPrivate::extractVCard30PropertyParams( |
|
739 VersitCursor& line, QTextCodec *codec) const |
|
740 { |
|
741 QMultiHash<QString,QString> result; |
|
742 QList<QByteArray> paramList = extractParams(line, codec); |
|
743 while (!paramList.isEmpty()) { |
|
744 QByteArray param = paramList.takeLast(); |
|
745 QString name(paramName(param, codec)); |
|
746 removeBackSlashEscaping(name); |
|
747 QString values = paramValue(param, codec); |
|
748 QStringList valueList = splitValue(values, QLatin1Char(','), QString::SkipEmptyParts, true); |
|
749 foreach (QString value, valueList) { |
|
750 removeBackSlashEscaping(value); |
|
751 result.insert(name, value); |
|
752 } |
|
753 } |
|
754 return result; |
|
755 } |
|
756 |
|
757 |
|
758 /*! |
|
759 * Extracts the parameters as delimited by semicolons using \a codec to determine the delimiters. |
|
760 * |
|
761 * On entry \a line should point to the start of the parameter section (past the name). |
|
762 * On exit, \a line will be updated to point to the start of the value. |
|
763 */ |
|
764 QList<QByteArray> QVersitReaderPrivate::extractParams(VersitCursor& line, QTextCodec *codec) const |
|
765 { |
|
766 const QByteArray colon = VersitUtils::encode(':', codec); |
|
767 QList<QByteArray> params; |
|
768 |
|
769 /* find the end of the name¶ms */ |
|
770 int colonIndex = line.data.indexOf(colon, line.position); |
|
771 if (colonIndex > line.position && colonIndex < line.selection) { |
|
772 QByteArray nameAndParamsString = line.data.mid(line.position, colonIndex - line.position); |
|
773 params = extractParts(nameAndParamsString, VersitUtils::encode(';', codec), codec); |
|
774 |
|
775 /* Update line */ |
|
776 line.setPosition(colonIndex + colon.length()); |
|
777 } else if (colonIndex == line.position) { |
|
778 // No parameters.. advance past it |
|
779 line.setPosition(line.position + colon.length()); |
|
780 } |
|
781 |
|
782 return params; |
|
783 } |
|
784 |
|
785 /*! |
|
786 * Extracts the parts separated by separator discarding the separators escaped with a backslash |
|
787 * encoded with \a codec |
|
788 */ |
|
789 QList<QByteArray> QVersitReaderPrivate::extractParts( |
|
790 const QByteArray& text, const QByteArray& separator, QTextCodec* codec) const |
|
791 { |
|
792 QList<QByteArray> parts; |
|
793 int partStartIndex = 0; |
|
794 int textLength = text.length(); |
|
795 int separatorLength = separator.length(); |
|
796 QByteArray backslash = VersitUtils::encode('\\', codec); |
|
797 int backslashLength = backslash.length(); |
|
798 |
|
799 for (int i=0; i < textLength-separatorLength+1; i++) { |
|
800 if (containsAt(text, separator, i) |
|
801 && (i < backslashLength |
|
802 || !containsAt(text, backslash, i-backslashLength))) { |
|
803 int length = i-partStartIndex; |
|
804 QByteArray part = extractPart(text,partStartIndex,length); |
|
805 if (part.length() > 0) |
|
806 parts.append(part); |
|
807 partStartIndex = i+separatorLength; |
|
808 } |
|
809 } |
|
810 |
|
811 // Add the last or only part |
|
812 QByteArray part = extractPart(text,partStartIndex); |
|
813 if (part.length() > 0) |
|
814 parts.append(part); |
|
815 return parts; |
|
816 } |
|
817 |
|
818 /*! |
|
819 * Extracts a substring limited by /a startPosition and /a length. |
|
820 */ |
|
821 QByteArray QVersitReaderPrivate::extractPart( |
|
822 const QByteArray& text, int startPosition, int length) const |
|
823 { |
|
824 QByteArray part; |
|
825 if (startPosition >= 0) |
|
826 part = text.mid(startPosition,length).trimmed(); |
|
827 return part; |
|
828 } |
|
829 |
|
830 /*! |
|
831 * Extracts the name of the parameter using \a codec to determine the delimiters. |
|
832 * No name is interpreted as an implicit "TYPE". |
|
833 */ |
|
834 QString QVersitReaderPrivate::paramName(const QByteArray& parameter, QTextCodec* codec) const |
|
835 { |
|
836 if (parameter.trimmed().length() == 0) |
|
837 return QString(); |
|
838 QByteArray equals = VersitUtils::encode('=', codec); |
|
839 int equalsIndex = parameter.indexOf(equals); |
|
840 if (equalsIndex > 0) { |
|
841 return codec->toUnicode(parameter.left(equalsIndex)).trimmed(); |
|
842 } |
|
843 |
|
844 return QLatin1String("TYPE"); |
|
845 } |
|
846 |
|
847 /*! |
|
848 * Extracts the value of the parameter using \a codec to determine the delimiters |
|
849 */ |
|
850 QString QVersitReaderPrivate::paramValue(const QByteArray& parameter, QTextCodec* codec) const |
|
851 { |
|
852 QByteArray value(parameter); |
|
853 QByteArray equals = VersitUtils::encode('=', codec); |
|
854 int equalsIndex = parameter.indexOf(equals); |
|
855 if (equalsIndex > 0) { |
|
856 int valueLength = parameter.length() - (equalsIndex + equals.length()); |
|
857 value = parameter.right(valueLength).trimmed(); |
|
858 } |
|
859 |
|
860 return codec->toUnicode(value); |
|
861 } |
|
862 |
|
863 /*! |
|
864 * Returns true if and only if \a text contains \a ba at \a index |
|
865 * |
|
866 * On entry, index must be >= 0 |
|
867 */ |
|
868 bool QVersitReaderPrivate::containsAt(const QByteArray& text, const QByteArray& match, int index) |
|
869 { |
|
870 int n = match.length(); |
|
871 if (text.length() - index < n) |
|
872 return false; |
|
873 const char* textData = text.constData(); |
|
874 const char* matchData = match.constData(); |
|
875 return memcmp(textData+index, matchData, n) == 0; |
|
876 } |
|
877 |
|
878 /*! |
|
879 * If the \a type and the \a property's name is known to contain a structured value, \a property's |
|
880 * value is split according to the type of structuring (compound vs. list) it is known to have. |
|
881 * Returns true if and only if such a split happened (ie. the property value holds a QStringList on |
|
882 * exit). |
|
883 */ |
|
884 bool QVersitReaderPrivate::splitStructuredValue( |
|
885 QVersitProperty& property, |
|
886 bool hasEscapedBackslashes) const |
|
887 { |
|
888 QVariant variant = property.variantValue(); |
|
889 if (property.valueType() == QVersitProperty::CompoundType) { |
|
890 variant.setValue(splitValue(variant.toString(), QLatin1Char(';'), |
|
891 QString::KeepEmptyParts, hasEscapedBackslashes)); |
|
892 property.setValue(variant); |
|
893 return true; |
|
894 } else if (property.valueType() == QVersitProperty::ListType) { |
|
895 variant.setValue(splitValue(variant.toString(), QLatin1Char(','), |
|
896 QString::SkipEmptyParts, hasEscapedBackslashes)); |
|
897 property.setValue(variant); |
|
898 return true; |
|
899 } |
|
900 return false; |
|
901 } |
|
902 |
|
903 /*! |
|
904 * Splits the \a string into substrings wherever \a sep occurs. |
|
905 * If \a hasEscapedBackslashes is false, then a \a sep preceded by a backslash is not considered |
|
906 * a split point (but the backslash is removed). |
|
907 * If \a hasEscapedBackslashes is true, then a \a sep preceded by an odd number of backslashes is |
|
908 * not considered a split point (but one backslash is removed). |
|
909 */ |
|
910 QStringList QVersitReaderPrivate::splitValue(const QString& string, |
|
911 const QChar& sep, |
|
912 QString::SplitBehavior behaviour, |
|
913 bool hasEscapedBackslashes) |
|
914 { |
|
915 QStringList list; |
|
916 bool isEscaped = false; // is the current character escaped |
|
917 int segmentStartIndex = 0; |
|
918 QString segment; |
|
919 for (int i = 0; i < string.length(); i++) { |
|
920 if (string.at(i) == QLatin1Char('\\')) { |
|
921 if (hasEscapedBackslashes) |
|
922 isEscaped = !isEscaped; // two consecutive backslashes make isEscaped false |
|
923 else |
|
924 isEscaped = true; |
|
925 } else if (string.at(i) == sep) { |
|
926 if (isEscaped) { |
|
927 // we see an escaped separator - remove the backslash |
|
928 segment += string.midRef(segmentStartIndex, i-segmentStartIndex-1); |
|
929 segment += sep; |
|
930 } else { |
|
931 // we see a separator |
|
932 segment += string.midRef(segmentStartIndex, i - segmentStartIndex); |
|
933 if (behaviour == QString::KeepEmptyParts || !segment.isEmpty()) |
|
934 list.append(segment); |
|
935 segment.clear(); |
|
936 } |
|
937 segmentStartIndex = i+1; |
|
938 isEscaped = false; |
|
939 } else { // normal character - keep going |
|
940 isEscaped = false; |
|
941 } |
|
942 } |
|
943 // The rest of the string after the last sep. |
|
944 segment += string.midRef(segmentStartIndex); |
|
945 if (behaviour == QString::KeepEmptyParts || !segment.isEmpty()) |
|
946 list.append(segment); |
|
947 return list; |
|
948 } |
|
949 |
|
950 /*! |
|
951 * Removes backslash escaping for line breaks (CRLFs), colons, semicolons, backslashes and commas |
|
952 * according to RFC 2426. This is called on parameter names and values and property values. |
|
953 * Colons ARE unescaped because the text of RFC2426 suggests that they should be. |
|
954 */ |
|
955 void QVersitReaderPrivate::removeBackSlashEscaping(QString& text) |
|
956 { |
|
957 if (!(text.startsWith(QLatin1Char('"')) && text.endsWith(QLatin1Char('"')))) { |
|
958 /* replaces \; with ; |
|
959 \, with , |
|
960 \: with : |
|
961 \\ with \ |
|
962 */ |
|
963 text.replace(QRegExp(QLatin1String("\\\\([;,:\\\\])")), QLatin1String("\\1")); |
|
964 // replaces \n with a CRLF |
|
965 text.replace(QLatin1String("\\n"), QLatin1String("\r\n"), Qt::CaseInsensitive); |
|
966 } |
|
967 } |
|
968 |
|
969 |
|
970 #include "moc_qversitreader_p.cpp" |