/android/import-contacts

To get this branch, use:
bzr branch http://bzr.ed.am/android/import-contacts
6 by edam
- added GPL header comments to all files
1
/*
2
 * VCFImporter.java
3
 *
50 by edam
updated all URLs, email addresses and package names to ed.am
4
 * Copyright (C) 2009 to 2011 Tim Marston <tim@ed.am>
6 by edam
- added GPL header comments to all files
5
 *
6
 * This file is part of the Import Contacts program (hereafter referred
7
 * to as "this program"). For more information, see
50 by edam
updated all URLs, email addresses and package names to ed.am
8
 * http://ed.am/dev/android/import-contacts
6 by edam
- added GPL header comments to all files
9
 *
10
 * This program is free software: you can redistribute it and/or modify
11
 * it under the terms of the GNU General Public License as published by
12
 * the Free Software Foundation, either version 3 of the License, or
13
 * (at your option) any later version.
14
 *
15
 * This program is distributed in the hope that it will be useful,
16
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
18
 * GNU General Public License for more details.
19
 *
20
 * You should have received a copy of the GNU General Public License
21
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
22
 */
23
50 by edam
updated all URLs, email addresses and package names to ed.am
24
package am.ed.importcontacts;
1 by edam
Initial import
25
26
import java.io.BufferedReader;
27
import java.io.File;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
28
import java.io.FileInputStream;
1 by edam
Initial import
29
import java.io.FileNotFoundException;
30
import java.io.FileReader;
31
import java.io.FilenameFilter;
32
import java.io.IOException;
33
import java.io.UnsupportedEncodingException;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
34
import java.nio.ByteBuffer;
37 by edam
- updated TODO and NEWS
35
import java.util.ArrayList;
1 by edam
Initial import
36
import java.util.Arrays;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
37
import java.util.HashMap;
1 by edam
Initial import
38
import java.util.HashSet;
36 by edam
- formatting: removed some double-indents on overrunning lines
39
import java.util.Iterator;
1 by edam
Initial import
40
import java.util.List;
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
41
import java.util.Locale;
37 by edam
- updated TODO and NEWS
42
import java.util.NoSuchElementException;
1 by edam
Initial import
43
import java.util.Set;
44
import java.util.Vector;
45
import java.util.regex.Matcher;
46
import java.util.regex.Pattern;
47
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
48
import android.annotation.SuppressLint;
1 by edam
Initial import
49
import android.content.SharedPreferences;
50
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
51
public class VcardImporter extends Importer
1 by edam
Initial import
52
{
41 by edam
- updated TODO
53
	private int _vcard_count = 0;
1 by edam
Initial import
54
	private int _progress = 0;
55
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
56
	public VcardImporter( Doit doit )
1 by edam
Initial import
57
	{
58
		super( doit );
59
	}
60
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
61
	@SuppressLint( "SdCardPath" )
1 by edam
Initial import
62
	@Override
63
	protected void onImport() throws AbortImportException
64
	{
65
		SharedPreferences prefs = getSharedPreferences();
66
67
		// update UI
68
		setProgressMessage( R.string.doit_scanning );
69
70
		// get a list of vcf files
71
		File[] files = null;
72
		try
73
		{
74
			// open directory
19 by edam
- added file chooser
75
			String path = "/sdcard" + prefs.getString( "location", "/" );
76
			File file = new File( path );
77
			if( !file.exists() )
1 by edam
Initial import
78
				showError( R.string.error_locationnotfound );
79
15 by edam
- added facility to enter a filename (instead of a directory to scan) and just use that
80
			// directory, or file?
19 by edam
- added file chooser
81
			if( file.isDirectory() )
15 by edam
- added facility to enter a filename (instead of a directory to scan) and just use that
82
			{
83
				// get files
84
				class VCardFilter implements FilenameFilter {
85
					public boolean accept( File dir, String name ) {
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
86
						return name.toLowerCase( Locale.US ).endsWith( ".vcf" );
15 by edam
- added facility to enter a filename (instead of a directory to scan) and just use that
87
					}
13 by edam
- converted project to use Android 1.5 SDK
88
				}
19 by edam
- added file chooser
89
				files = file.listFiles( new VCardFilter() );
15 by edam
- added facility to enter a filename (instead of a directory to scan) and just use that
90
			}
91
			else
92
			{
93
				// use just this file
94
				files = new File[ 1 ];
19 by edam
- added file chooser
95
				files[ 0 ] = file;
15 by edam
- added facility to enter a filename (instead of a directory to scan) and just use that
96
			}
1 by edam
Initial import
97
		}
98
		catch( SecurityException e ) {
99
			showError( R.string.error_locationpermissions );
100
		}
101
102
		// check num files and set progress max
103
		if( files != null && files.length > 0 )
104
			setProgressMax( files.length );
105
		else
106
			showError( R.string.error_locationnofiles );
107
108
		// scan through the files
109
		setTmpProgress( 0 );
110
		for( int i = 0; i < files.length; i++ ) {
111
			countVCardFile( files[ i ] );
112
			setTmpProgress( i );
113
		}
41 by edam
- updated TODO
114
		setProgressMax( _vcard_count );	// will also update tmp progress
1 by edam
Initial import
115
116
		// import them
117
		setProgress( 0 );
118
		for( int i = 0; i < files.length; i++ )
119
			importVCardFile( files[ i ] );
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
120
		setProgress( _vcard_count );
1 by edam
Initial import
121
	}
122
123
	private void countVCardFile( File file ) throws AbortImportException
124
	{
125
		try
126
		{
127
			// open file
128
			BufferedReader reader = new BufferedReader(
36 by edam
- formatting: removed some double-indents on overrunning lines
129
				new FileReader( file ) );
1 by edam
Initial import
130
131
			// read
132
			String line;
41 by edam
- updated TODO
133
			boolean in_vcard = false;
1 by edam
Initial import
134
			while( ( line = reader.readLine() ) != null )
135
			{
41 by edam
- updated TODO
136
				if( !in_vcard ) {
1 by edam
Initial import
137
					// look for vcard beginning
36 by edam
- formatting: removed some double-indents on overrunning lines
138
					if( line.matches( "^BEGIN:VCARD" ) ) {
41 by edam
- updated TODO
139
						in_vcard = true;
140
						_vcard_count++;
1 by edam
Initial import
141
					}
142
				}
36 by edam
- formatting: removed some double-indents on overrunning lines
143
				else if( line.matches( "^END:VCARD" ) )
41 by edam
- updated TODO
144
					in_vcard = false;
1 by edam
Initial import
145
			}
146
147
		}
148
		catch( FileNotFoundException e ) {
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
149
			showError( getText( R.string.error_filenotfound ) +
150
				file.getName() );
1 by edam
Initial import
151
		}
152
		catch( IOException e ) {
153
			showError( getText( R.string.error_ioerror ) + file.getName() );
154
		}
155
	}
156
157
	private void importVCardFile( File file ) throws AbortImportException
158
	{
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
159
		// check file is good
160
		if( !file.exists() )
161
			showError( getText( R.string.error_filenotfound ) +
162
				file.getName() );
163
		if( file.length() == 0 )
164
			showError( getText( R.string.error_fileisempty ) +
165
				file.getName() );
166
1 by edam
Initial import
167
		try
168
		{
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
169
			// open/read file
170
			FileInputStream istream = new FileInputStream( file );
171
			byte[] content = new byte[ (int)file.length() ];
172
			istream.read( content );
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
173
			istream = null;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
174
175
			// import
176
			importVCardFileContent( content, file.getName() );
1 by edam
Initial import
177
		}
178
		catch( FileNotFoundException e ) {
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
179
			showError( getText( R.string.error_filenotfound ) +
180
				file.getName() );
1 by edam
Initial import
181
		}
182
		catch( IOException e ) {
183
			showError( getText( R.string.error_ioerror ) + file.getName() );
184
		}
185
	}
186
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
187
	private void importVCardFileContent( byte[] content, String fileName )
36 by edam
- formatting: removed some double-indents on overrunning lines
188
		throws AbortImportException
1 by edam
Initial import
189
	{
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
190
		// go through lines
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
191
		Vcard vcard = null;
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
192
		int vcard_start_line = 0;
36 by edam
- formatting: removed some double-indents on overrunning lines
193
		ContentLineIterator cli = new ContentLineIterator( content );
194
		while( cli.hasNext() )
1 by edam
Initial import
195
		{
36 by edam
- formatting: removed some double-indents on overrunning lines
196
			ByteBuffer buffer = cli.next();
197
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
198
			// get a US-ASCII version of the line for processing
199
			String line;
200
			try {
36 by edam
- formatting: removed some double-indents on overrunning lines
201
				line = new String( buffer.array(), buffer.position(),
202
					buffer.limit() - buffer.position(), "US-ASCII" );
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
203
			}
204
			catch( UnsupportedEncodingException e ) {
205
				// we know US-ASCII is supported, so appease the compiler...
206
				line = "";
207
			}
1 by edam
Initial import
208
41 by edam
- updated TODO
209
			if( vcard == null ) {
1 by edam
Initial import
210
				// look for vcard beginning
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
211
				if( line.matches( "^BEGIN[ \t]*:[ \t]*VCARD" ) ) {
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
212
					setProgress( _progress++ );
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
213
					vcard = new Vcard();
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
214
					vcard_start_line = cli.getLineNumber();
1 by edam
Initial import
215
				}
216
			}
217
			else {
218
				// look for vcard content or ending
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
219
				if( line.matches( "^END[ \t]*:[ \t]*VCARD" ) )
1 by edam
Initial import
220
				{
43 by edam
- refactored some code to do with how contacts are imported
221
					// finalise the vcard/contact
1 by edam
Initial import
222
					try {
43 by edam
- refactored some code to do with how contacts are imported
223
						vcard.finaliseVcard();
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
224
225
						// pass the finalised contact to the importer
226
						importContact( vcard );
1 by edam
Initial import
227
					}
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
228
					catch( Vcard.ParseException e ) {
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
229
						if( !showContinue(
230
							getText( R.string.error_vcf_parse ).toString()
231
							+ fileName +
232
							getText( R.string.error_vcf_parse_line ).toString()
233
							+ cli.getLineNumber() + ":\n" + e.getMessage() ) )
234
						{
235
							finish( ACTION_ABORT );
236
						}
237
						else
238
							skipContact();
239
					}
240
					catch( ContactData.ContactNotIdentifiableException e ) {
241
						if( !showContinue(
242
							getText( R.string.error_vcf_parse ).toString()
243
							+ fileName +
244
							getText( R.string.error_vcf_parse_line ).toString()
245
							+ vcard_start_line + ":\n" + getText(
246
								R.string.error_vcf_notenoughinfo ).toString()
247
						) )
248
						{
249
							finish( ACTION_ABORT );
250
						}
251
						else
252
							skipContact();
253
					}
254
255
					// discard this vcard
41 by edam
- updated TODO
256
					vcard = null;
1 by edam
Initial import
257
				}
258
				else
259
				{
260
					// try giving the line to the vcard
261
					try {
41 by edam
- updated TODO
262
						vcard.parseLine( buffer, line,
36 by edam
- formatting: removed some double-indents on overrunning lines
263
							cli.doesNextLineLookFolded() );
1 by edam
Initial import
264
					}
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
265
					catch( Vcard.ParseException e ) {
1 by edam
Initial import
266
						skipContact();
267
						if( !showContinue(
36 by edam
- formatting: removed some double-indents on overrunning lines
268
							getText( R.string.error_vcf_parse ).toString()
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
269
							+ fileName +
270
							getText( R.string.error_vcf_parse_line ).toString()
271
							+ cli.getLineNumber() + "\n" + e.getMessage() ) )
36 by edam
- formatting: removed some double-indents on overrunning lines
272
						{
3 by edam
- added "all done" message
273
							finish( ACTION_ABORT );
36 by edam
- formatting: removed some double-indents on overrunning lines
274
						}
1 by edam
Initial import
275
276
						// although we're continuing, we still need to abort
277
						// this vCard. Further lines will be ignored until we
278
						// get to another BEGIN:VCARD line.
41 by edam
- updated TODO
279
						vcard = null;
1 by edam
Initial import
280
					}
43 by edam
- refactored some code to do with how contacts are imported
281
					catch( Vcard.SkipImportException e ) {
1 by edam
Initial import
282
						skipContact();
283
						// abort this vCard. Further lines will be ignored until
284
						// we get to another BEGIN:VCARD line.
41 by edam
- updated TODO
285
						vcard = null;
1 by edam
Initial import
286
					}
287
				}
288
			}
289
		}
290
	}
291
36 by edam
- formatting: removed some double-indents on overrunning lines
292
	class ContentLineIterator implements Iterator< ByteBuffer >
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
293
	{
36 by edam
- formatting: removed some double-indents on overrunning lines
294
		protected byte[] _content = null;
295
		protected int _pos = 0;
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
296
		protected int _line = 0;
36 by edam
- formatting: removed some double-indents on overrunning lines
297
298
		public ContentLineIterator( byte[] content )
299
		{
300
			_content = content;
301
		}
302
303
		@Override
304
		public boolean hasNext()
305
		{
306
			return _pos < _content.length;
307
		}
308
309
		@Override
310
		public ByteBuffer next()
311
		{
312
			int initial_pos = _pos;
313
314
			// find newline
315
			for( ; _pos < _content.length; _pos++ )
316
				if( _content[ _pos ] == '\n' )
317
				{
318
					// adjust for a \r preceding the \n
319
					int to = ( _pos > 0 && _content[ _pos - 1 ] == '\r' &&
320
						_pos > initial_pos )? _pos - 1 : _pos;
321
					_pos++;
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
322
					_line++;
36 by edam
- formatting: removed some double-indents on overrunning lines
323
					return ByteBuffer.wrap( _content, initial_pos,
324
						to - initial_pos );
325
				}
326
327
			// we didn't find one, but were there bytes left?
328
			if( _pos != initial_pos ) {
329
				int to = _pos;
330
				_pos++;
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
331
				_line++;
36 by edam
- formatting: removed some double-indents on overrunning lines
332
				return ByteBuffer.wrap( _content, initial_pos,
333
					to - initial_pos );
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
334
			}
36 by edam
- formatting: removed some double-indents on overrunning lines
335
336
			// no bytes left
337
			throw new NoSuchElementException();
338
		}
339
340
		@Override
341
		public void remove()
342
		{
343
			throw new UnsupportedOperationException();
344
		}
345
346
		/**
347
		 * Does the next line, if there is one, look like it should be folded
348
		 * onto the end of this one?
349
		 * @return
350
		 */
351
		public boolean doesNextLineLookFolded()
352
		{
353
			return _pos > 0 && _pos < _content.length &&
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
354
				_content[ _pos - 1 ] == '\n' &&
355
				( _content[ _pos ] == ' ' || _content[ _pos ] == '\t' );
36 by edam
- formatting: removed some double-indents on overrunning lines
356
		}
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
357
358
		public int getLineNumber()
359
		{
360
			return _line;
361
		}
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
362
	}
363
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
364
	private class Vcard extends ContactData
1 by edam
Initial import
365
	{
366
		private final static int NAMELEVEL_NONE = 0;
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
367
		private final static int NAMELEVEL_N = 1;
368
		private final static int NAMELEVEL_FN = 2;
1 by edam
Initial import
369
37 by edam
- updated TODO and NEWS
370
		private final static int MULTILINE_NONE = 0;
371
		private final static int MULTILINE_ENCODED = 1;	// v2.1 quoted-printable
372
		private final static int MULTILINE_ESCAPED = 2;	// v2.1 \\CRLF
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
373
		private final static int MULTILINE_FOLDED = 3;	// MIME-DIR folding
37 by edam
- updated TODO and NEWS
374
1 by edam
Initial import
375
		private String _version = null;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
376
		private Vector< ByteBuffer > _buffers = null;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
377
		private int _name_level = NAMELEVEL_NONE;
37 by edam
- updated TODO and NEWS
378
		private int _parser_multiline_state = MULTILINE_NONE;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
379
		private String _parser_current_name_and_params = null;
380
		private String _parser_buffered_value_so_far = "";
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
381
		private String _cached_organisation = null;
382
		private String _cached_title = null;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
383
384
		protected class UnencodeResult
385
		{
386
			private boolean _another_line_required;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
387
			private ByteBuffer _buffer;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
388
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
389
			public UnencodeResult( boolean another_line_required,
390
				ByteBuffer buffer )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
391
			{
392
				_another_line_required = another_line_required;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
393
				_buffer = buffer;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
394
			}
395
396
			public boolean isAnotherLineRequired()
397
			{
398
				return _another_line_required;
399
			}
400
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
401
			public ByteBuffer getBuffer()
402
			{
403
				return _buffer;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
404
			}
405
		}
1 by edam
Initial import
406
14 by edam
- got rid of the pretend ImportContacts activity alltogether (and made the Intro activity the startup one)
407
		@SuppressWarnings("serial")
1 by edam
Initial import
408
		protected class ParseException extends Exception
409
		{
14 by edam
- got rid of the pretend ImportContacts activity alltogether (and made the Intro activity the startup one)
410
			@SuppressWarnings("unused")
1 by edam
Initial import
411
			public ParseException( String error )
412
			{
413
				super( error );
414
			}
415
416
			public ParseException( int res )
417
			{
42 by edam
- renamed VCFImporter to VcardImporter and VCard to Vcard
418
				super( VcardImporter.this.getText( res ).toString() );
1 by edam
Initial import
419
			}
420
		}
421
14 by edam
- got rid of the pretend ImportContacts activity alltogether (and made the Intro activity the startup one)
422
		@SuppressWarnings("serial")
43 by edam
- refactored some code to do with how contacts are imported
423
		protected class SkipImportException extends Exception { }
1 by edam
Initial import
424
36 by edam
- formatting: removed some double-indents on overrunning lines
425
		private String extractCollonPartFromLine( ByteBuffer buffer,
426
			String line, boolean former )
427
		{
428
			String ret = null;
429
430
			// get a US-ASCII version of the line for processing, unless we were
431
			// supplied with one
432
			if( line == null ) {
433
				try {
434
					line = new String( buffer.array(), buffer.position(),
435
						buffer.limit() - buffer.position(), "US-ASCII" );
436
				}
437
				catch( UnsupportedEncodingException e ) {
438
					// we know US-ASCII is supported, so appease the compiler...
439
					line = "";
440
				}
441
			}
442
443
			// split line into name and value parts and check to make sure we
444
			// only got 2 parts and that the first part is not zero in length
445
			String[] parts = line.split( ":", 2 );
446
			if( parts.length == 2 && parts[ 0 ].length() > 0 )
447
				ret = parts[ former? 0 : 1 ];
448
449
			return ret;
450
		}
451
452
		private String extractNameAndParamsFromLine( ByteBuffer buffer,
453
			String line )
454
		{
455
			return extractCollonPartFromLine( buffer, line, true );
456
		}
457
458
		private String extractValueFromLine( ByteBuffer buffer, String line )
459
		{
460
			return extractCollonPartFromLine( buffer, line, false );
461
		}
462
463
		public void parseLine( ByteBuffer buffer, String line,
464
			boolean next_line_looks_folded )
43 by edam
- refactored some code to do with how contacts are imported
465
			throws ParseException, SkipImportException,
36 by edam
- formatting: removed some double-indents on overrunning lines
466
			AbortImportException
467
		{
468
			// do we have a version yet?
1 by edam
Initial import
469
			if( _version == null )
470
			{
36 by edam
- formatting: removed some double-indents on overrunning lines
471
				// tentatively get name and params from line
472
				String name_and_params =
473
					extractNameAndParamsFromLine( buffer, line );
474
475
				// is it a version line?
476
				if( name_and_params != null &&
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
477
					name_and_params.equalsIgnoreCase( "VERSION" ) )
1 by edam
Initial import
478
				{
36 by edam
- formatting: removed some double-indents on overrunning lines
479
					// yes, get it!
480
					String value = extractValueFromLine( buffer, line );
481
					if( !value.equals( "2.1" ) && !value.equals( "3.0" ) )
1 by edam
Initial import
482
						throw new ParseException( R.string.error_vcf_version );
36 by edam
- formatting: removed some double-indents on overrunning lines
483
					_version = value;
1 by edam
Initial import
484
36 by edam
- formatting: removed some double-indents on overrunning lines
485
					// parse any buffers we've been accumulating while we waited
486
					// for a version
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
487
					if( _buffers != null )
488
						for( int i = 0; i < _buffers.size(); i++ )
36 by edam
- formatting: removed some double-indents on overrunning lines
489
							parseLine( _buffers.get( i ), null,
490
								i + 1 < _buffers.size() &&
491
								_buffers.get( i + 1 ).hasRemaining() &&
492
								_buffers.get( i + 1 ).get(
493
									_buffers.get( i + 1 ).position() ) == ' ' );
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
494
					_buffers = null;
1 by edam
Initial import
495
				}
496
				else
497
				{
36 by edam
- formatting: removed some double-indents on overrunning lines
498
					// no, so stash this line till we get a version
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
499
					if( _buffers == null )
500
						_buffers = new Vector< ByteBuffer >();
501
					_buffers.add( buffer );
1 by edam
Initial import
502
				}
503
			}
504
			else
505
			{
36 by edam
- formatting: removed some double-indents on overrunning lines
506
				// name and params and the position in the buffer where the
507
				// "value" part of the line start
508
				String name_and_params;
509
				int pos;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
510
37 by edam
- updated TODO and NEWS
511
				if( _parser_multiline_state != MULTILINE_NONE )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
512
				{
513
					// if we're currently in a multi-line value, use the stored
514
					// property name and parameters
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
515
					name_and_params = _parser_current_name_and_params;
516
37 by edam
- updated TODO and NEWS
517
					// skip some initial line characters, depending on the type
518
					// of multi-line we're handling
36 by edam
- formatting: removed some double-indents on overrunning lines
519
					pos = buffer.position();
37 by edam
- updated TODO and NEWS
520
					switch( _parser_multiline_state )
521
					{
522
					case MULTILINE_FOLDED:
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
523
						pos++;
37 by edam
- updated TODO and NEWS
524
						break;
525
					case MULTILINE_ENCODED:
36 by edam
- formatting: removed some double-indents on overrunning lines
526
						while( pos < buffer.limit() && (
527
							buffer.get( pos ) == ' ' ||
528
							buffer.get( pos ) == '\t' ) )
529
						{
530
							pos++;
531
						}
37 by edam
- updated TODO and NEWS
532
						break;
533
					default:
534
						// do nothing
535
					}
536
537
					// take us out of multi-line so that we can re-detect that
538
					// this line is a multi-line or not
539
					_parser_multiline_state = MULTILINE_NONE;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
540
				}
541
				else
542
				{
36 by edam
- formatting: removed some double-indents on overrunning lines
543
					// get name and params from line, and since we're not
544
					// parsing a subsequent line in a multi-line, this should
545
					// not fail, or it's an error
546
					name_and_params =
547
						extractNameAndParamsFromLine( buffer, line );
548
					if( name_and_params == null )
549
						throw new ParseException(
550
							R.string.error_vcf_malformed );
551
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
552
					// calculate how many chars to skip from beginning of line
553
					// so we skip the property "name:" part
36 by edam
- formatting: removed some double-indents on overrunning lines
554
					pos = buffer.position() + name_and_params.length() + 1;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
555
556
					// reset the saved multi-line state
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
557
					_parser_current_name_and_params = name_and_params;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
558
					_parser_buffered_value_so_far = "";
559
				}
560
36 by edam
- formatting: removed some double-indents on overrunning lines
561
				// get value from buffer, as raw bytes
562
				ByteBuffer value;
563
				value = ByteBuffer.wrap( buffer.array(), pos,
564
					buffer.limit() - pos );
565
1 by edam
Initial import
566
				// get parameter parts
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
567
				String[] name_param_parts = name_and_params.split( ";", -1 );
568
				for( int i = 0; i < name_param_parts.length; i++ )
569
					name_param_parts[ i ] = name_param_parts[ i ].trim();
1 by edam
Initial import
570
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
571
				// determine whether we care about this entry
572
				final HashSet< String > interesting_fields =
47 by edam
- bump version no. to 1.2
573
					new HashSet< String >( Arrays.asList( new String[] { "N",
574
						"FN", "ORG", "TITLE", "TEL", "EMAIL", "ADR", "LABEL" }
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
575
				) );
576
				boolean is_interesting_field =
577
					interesting_fields.contains( name_param_parts[ 0 ] );
578
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
579
				// parse encoding parameter
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
580
				String encoding = checkParam( name_param_parts, "ENCODING" );
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
581
				if( encoding != null )
582
					encoding = encoding.toUpperCase( Locale.US );
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
583
				if( is_interesting_field && encoding != null &&
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
584
					!encoding.equalsIgnoreCase( "8BIT" ) &&
585
					!encoding.equalsIgnoreCase( "QUOTED-PRINTABLE" ) )
586
					//&& !encoding.equalsIgnoreCase( "BASE64" ) )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
587
				{
588
					throw new ParseException( R.string.error_vcf_encoding );
589
				}
590
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
591
				// parse charset parameter
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
592
				String charset = checkParam( name_param_parts, "CHARSET" );
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
593
				if( charset != null )
594
					charset = charset.toUpperCase( Locale.US );
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
595
				if( charset != null &&
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
596
					!charset.equalsIgnoreCase( "US-ASCII" ) &&
597
					!charset.equalsIgnoreCase( "ASCII" ) &&
598
					!charset.equalsIgnoreCase( "UTF-8" ) )
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
599
				{
600
					throw new ParseException( R.string.error_vcf_charset );
601
				}
602
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
603
				// do unencoding (or default to a fake unencoding result with
604
				// the raw string)
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
605
				UnencodeResult unencoding_result = null;
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
606
				if( encoding != null &&
607
					encoding.equalsIgnoreCase( "QUOTED-PRINTABLE" ) )
608
				{
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
609
					unencoding_result = unencodeQuotedPrintable( value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
610
				}
611
//				else if( encoding != null &&
612
//					encoding.equalsIgnoreCase( "BASE64" ) )
613
//				{
34 by edam
- check for empty data "values" after parsing line parameters, so that we catch parameter errors (such as unknown encoding types).
614
//					unencoding_result = unencodeBase64( props[ 1 ], charset );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
615
//				}
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
616
				if( unencoding_result != null ) {
617
					value = unencoding_result.getBuffer();
37 by edam
- updated TODO and NEWS
618
					if( unencoding_result.isAnotherLineRequired() )
619
						_parser_multiline_state = MULTILINE_ENCODED;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
620
				}
621
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
622
				// convert 8-bit US-ASCII charset to UTF-8 (where no charset is
623
				// specified for a v2.1 vcard entry, we assume it's US-ASCII)
624
				if( ( charset == null && _version.equals( "2.1" ) ) ||
625
					( charset != null && (
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
626
						charset.equalsIgnoreCase( "ASCII" ) ||
627
						charset.equalsIgnoreCase( "US-ASCII" ) ) ) )
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
628
				{
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
629
					value = transcodeAsciiToUtf8( value );
630
				}
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
631
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
632
				// process charset (value is now in UTF-8)
36 by edam
- formatting: removed some double-indents on overrunning lines
633
				String string_value;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
634
				try {
36 by edam
- formatting: removed some double-indents on overrunning lines
635
					string_value = new String( value.array(), value.position(),
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
636
						value.limit() - value.position(), "UTF-8" );
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
637
				} catch( UnsupportedEncodingException e ) {
638
					throw new ParseException( R.string.error_vcf_charset );
639
				}
640
37 by edam
- updated TODO and NEWS
641
				// for some entries that have semicolon-separated value parts,
642
				// check to see if the value ends in an escape character, which
643
				// indicates that we have a multi-line value
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
644
				if( ( name_param_parts[ 0 ].equalsIgnoreCase( "N" ) ||
645
					name_param_parts[ 0 ].equalsIgnoreCase( "ORG" ) ||
646
					name_param_parts[ 0 ].equalsIgnoreCase( "ADR" ) ) &&
37 by edam
- updated TODO and NEWS
647
					doesStringEndInAnEscapeChar( string_value ) )
648
				{
649
					_parser_multiline_state = MULTILINE_ESCAPED;
650
					string_value = string_value.substring( 0,
651
						string_value.length() - 1 );
652
				}
653
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
654
				// if we know we're not in an encoding-based multi-line, check
655
				// to see if we're in a folded multi-line
37 by edam
- updated TODO and NEWS
656
				if( _parser_multiline_state == MULTILINE_NONE &&
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
657
					next_line_looks_folded )
37 by edam
- updated TODO and NEWS
658
				{
659
					_parser_multiline_state = MULTILINE_FOLDED;
660
				}
36 by edam
- formatting: removed some double-indents on overrunning lines
661
37 by edam
- updated TODO and NEWS
662
				// handle multi-lines by buffering them and parsing them when we
663
				// are processing the last line in a multi-line sequence
664
				if( _parser_multiline_state != MULTILINE_NONE ) {
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
665
					_parser_buffered_value_so_far += string_value;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
666
					return;
667
				}
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
668
				String complete_value =
37 by edam
- updated TODO and NEWS
669
					( _parser_buffered_value_so_far + string_value ).trim();
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
670
34 by edam
- check for empty data "values" after parsing line parameters, so that we catch parameter errors (such as unknown encoding types).
671
				// ignore empty values
672
				if( complete_value.length() < 1 ) return;
673
1 by edam
Initial import
674
				// parse some properties
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
675
				if( name_param_parts[ 0 ].equalsIgnoreCase( "N" ) )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
676
					parseN( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
677
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "FN" ) )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
678
					parseFN( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
679
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "ORG" ) )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
680
					parseORG( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
681
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "TITLE" ) )
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
682
					parseTITLE( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
683
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "TEL" ) )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
684
					parseTEL( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
685
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "EMAIL" ) )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
686
					parseEMAIL( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
687
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "ADR" ) )
37 by edam
- updated TODO and NEWS
688
					parseADR( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
689
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "LABEL" ) )
47 by edam
- bump version no. to 1.2
690
					parseLABEL( name_param_parts, complete_value );
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
691
				else if( name_param_parts[ 0 ].equalsIgnoreCase( "NOTE" ) )
65 by edam
added support for notes; rewrote backends so that all normalising of data is now done within the contacts cache; made the vCard unescape routine slightly more acceptant of non-standard escaped characters
692
					parseNOTE( name_param_parts, complete_value );
37 by edam
- updated TODO and NEWS
693
			}
694
		}
695
696
		private boolean doesStringEndInAnEscapeChar( String string )
697
		{
698
			// count the number of backslashes at the end of the string
699
			int count = 0;
700
			for( int a = string.length() - 1; a >= 0; a-- )
701
				if( string.charAt( a ) == '\\' )
702
					count++;
703
				else
704
					break;
705
706
			// if there are an even number of backslashes then the final one
707
			// doesn't count
708
			return ( count & 1 ) == 1;
709
		}
710
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
711
		private String[] splitValueByCharacter( String value, char character )
37 by edam
- updated TODO and NEWS
712
		{
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
713
			// split string in to parts by specified character
37 by edam
- updated TODO and NEWS
714
			ArrayList< String > parts = new ArrayList< String >(
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
715
				Arrays.asList( value.split( "" + character ) ) );
37 by edam
- updated TODO and NEWS
716
717
			// go through parts
718
			for( int a = 0; a < parts.size(); a++ )
719
			{
720
				String str = parts.get( a );
721
722
				// look for parts that end in an escape character, but ignore
723
				// the final part. We've already detected escape chars at the
724
				// end of the final part in parseLine() and handled multi-lines
725
				// accordingly.
726
				if( a < parts.size() - 1 &&
727
					doesStringEndInAnEscapeChar( str ) )
728
				{
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
729
					// append the escaped character, join the next part to this
730
					// part and remove the next part
37 by edam
- updated TODO and NEWS
731
					parts.set( a, str.substring( 0, str.length() - 1 ) +
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
732
						character + parts.get( a + 1 ) );
37 by edam
- updated TODO and NEWS
733
					parts.remove( a + 1 );
734
735
					// re-visit this part
736
					a--;
737
					continue;
738
				}
739
740
				// trim and replace string
741
				str = str.trim();
742
				parts.set( a, str );
743
			}
744
745
			String[] ret = new String[ parts.size() ];
746
			return parts.toArray( ret );
1 by edam
Initial import
747
		}
748
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
749
		private String unescapeValue( String value )
750
		{
751
			StringBuilder ret = new StringBuilder( value.length() );
752
			boolean in_escape = false;
753
			for( int a = 0; a < value.length(); a++ )
754
			{
755
				int c = value.codePointAt( a );
756
757
				// process a normal character
758
				if( !in_escape ) {
759
					if( c == '\\' )
760
						in_escape = true;
761
					else
762
						ret.append( Character.toChars( c ) );
763
					continue;
764
				}
765
766
				// process an escape sequence
767
				in_escape = false;
768
				switch( c )
769
				{
65 by edam
added support for notes; rewrote backends so that all normalising of data is now done within the contacts cache; made the vCard unescape routine slightly more acceptant of non-standard escaped characters
770
				case 'T':
771
				case 't':
772
					// add tab (invalid/non-standard, but accepted)
773
					ret.append( '\t' );
774
					break;
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
775
				case 'N':
776
				case 'n':
777
					// add newline
778
					ret.append( '\n' );
779
					break;
780
				case '\\':
781
				case ',':
782
				case ';':
783
					// add escaped character
784
					ret.append( Character.toChars( c ) );
785
					break;
786
				default:
787
					// unknown escape sequence, so add it unescaped
65 by edam
added support for notes; rewrote backends so that all normalising of data is now done within the contacts cache; made the vCard unescape routine slightly more acceptant of non-standard escaped characters
788
					// (invalid/non-standard, but accepted)
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
789
					ret.append( "\\" );
790
					ret.append( Character.toChars( c ) );
791
					break;
792
				}
793
			}
794
795
			return ret.toString();
796
		}
797
1 by edam
Initial import
798
		private void parseN( String[] params, String value )
799
		{
800
			// already got a better name?
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
801
			if( _name_level >= NAMELEVEL_N ) return;
1 by edam
Initial import
802
803
			// get name parts
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
804
			String[] name_parts = splitValueByCharacter( value, ';' );
1 by edam
Initial import
805
806
			// build name
807
			value = "";
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
808
			final int[] part_order = { 3, 1, 2, 0, 4 };
809
			for( int a = 0; a < part_order.length; a++ )
810
				if( name_parts.length > part_order[ a ] &&
811
					name_parts[ part_order[ a ] ].length() > 0 )
812
				{
813
					// split this part in to it's comma-separated bits
814
					String[] name_part_parts = splitValueByCharacter(
815
						name_parts[ part_order[ a ] ], ',' );
816
					for( int b = 0; b < name_part_parts.length; b++ )
817
						if( name_part_parts[ b ].length() > 0 )
818
						{
819
							if( value.length() == 0 ) value += " ";
820
							value += name_part_parts[ b ];
821
						}
822
				}
1 by edam
Initial import
823
824
			// set name
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
825
			setName( unescapeValue( value ) );
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
826
			_name_level = NAMELEVEL_N;
1 by edam
Initial import
827
		}
828
829
		private void parseFN( String[] params, String value )
830
		{
831
			// already got a better name?
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
832
			if( _name_level >= NAMELEVEL_FN ) return;
1 by edam
Initial import
833
834
			// set name
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
835
			setName( unescapeValue( value ) );
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
836
			_name_level = NAMELEVEL_FN;
1 by edam
Initial import
837
		}
838
839
		private void parseORG( String[] params, String value )
840
		{
841
			// get org parts
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
842
			String[] org_parts = splitValueByCharacter( value, ';' );
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
843
			if( org_parts == null || org_parts.length < 1 ) return;
844
845
			// build organisation name
846
			StringBuilder builder = new StringBuilder(
847
				String.valueOf( org_parts[ 0 ] ) );
848
			for( int a = 1; a < org_parts.length; a++ )
849
				builder.append( ", " ).append( org_parts[ a ] );
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
850
			String organisation = unescapeValue( builder.toString() );
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
851
852
			// set organisation name (using a title we've previously found)
853
			addOrganisation( organisation, _cached_title, true );
854
855
			// if we've not previously found a title, store this organisation
856
			// name (we'll need it when we find a title to update the
857
			// organisation, by name), else if we *have* previously found a
858
			// title, clear it (since we just used it)
859
			if( _cached_title == null )
860
				_cached_organisation = organisation;
861
			else
862
				_cached_title = null;
863
		}
864
865
		private void parseTITLE( String[] params, String value )
866
		{
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
867
			value = unescapeValue( value );
868
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
869
			// if we previously had an organisation, look it up and append this
870
			// title to it
871
			if( _cached_organisation != null && hasOrganisations() ) {
872
				HashMap< String, ExtraDetail > datas = getOrganisations();
873
				ExtraDetail detail = datas.get( _cached_organisation );
874
				if( detail != null )
875
					detail.setExtra( value );
876
			}
877
878
			// same as when handling organisation, if we've not previously found
879
			// an organisation we store this title, else we clear it (since we
880
			// just appended this title to it)
881
			if( _cached_organisation == null )
882
				_cached_title = value;
883
			else
884
				_cached_organisation = null;
1 by edam
Initial import
885
		}
886
887
		private void parseTEL( String[] params, String value )
888
		{
889
			if( value.length() == 0 ) return;
890
891
			Set< String > types = extractTypes( params, Arrays.asList(
36 by edam
- formatting: removed some double-indents on overrunning lines
892
				"PREF", "HOME", "WORK", "VOICE", "FAX", "MSG", "CELL",
893
				"PAGER", "BBS", "MODEM", "CAR", "ISDN", "VIDEO" ) );
1 by edam
Initial import
894
895
			// here's the logic...
41 by edam
- updated TODO
896
			boolean is_preferred = types.contains( "PREF" );
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
897
			int type;
1 by edam
Initial import
898
			if( types.contains( "FAX" ) )
899
				if( types.contains( "HOME" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
900
					type = TYPE_FAX_HOME;
1 by edam
Initial import
901
				else
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
902
					type = TYPE_FAX_WORK;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
903
			else if( types.contains( "CELL" ) || types.contains( "VIDEO" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
904
				type = TYPE_MOBILE;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
905
			else if( types.contains( "PAGER" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
906
				type = TYPE_PAGER;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
907
			else if( types.contains( "WORK" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
908
				type = TYPE_WORK;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
909
			else
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
910
				type = TYPE_HOME;
24 by edam
- import phone numbers even when they have no specified type (default to mobile)
911
912
			// add phone number
41 by edam
- updated TODO
913
			addNumber( value, type, is_preferred );
1 by edam
Initial import
914
		}
915
916
		public void parseEMAIL( String[] params, String value )
917
		{
918
			if( value.length() == 0 ) return;
919
920
			Set< String > types = extractTypes( params, Arrays.asList(
36 by edam
- formatting: removed some double-indents on overrunning lines
921
				"PREF", "WORK", "HOME", "INTERNET" ) );
1 by edam
Initial import
922
37 by edam
- updated TODO and NEWS
923
			// add email address
41 by edam
- updated TODO
924
			boolean is_preferred = types.contains( "PREF" );
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
925
			int type;
1 by edam
Initial import
926
			if( types.contains( "WORK" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
927
				type = TYPE_WORK;
1 by edam
Initial import
928
			else
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
929
				type = TYPE_HOME;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
930
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
931
			addEmail( unescapeValue( value ), type, is_preferred );
1 by edam
Initial import
932
		}
933
37 by edam
- updated TODO and NEWS
934
		private void parseADR( String[] params, String value )
935
		{
936
			// get address parts
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
937
			String[] adr_parts = splitValueByCharacter( value, ';' );
37 by edam
- updated TODO and NEWS
938
939
			// build address
940
			value = "";
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
941
			for( int a = 0; a < adr_parts.length; a++ )
942
				if( adr_parts[ a ].length() > 0 )
943
				{
69 by edam
compatibility and flexibility fixes (accept spaces in BEGIN/END lines; accept tabs as folding character; ignore case of item names, encodings and charsets; only parse comma-separated subparts of address parts for v3.0 vcards; accept new-line + white-space folding in v2.1 vCards)
944
					// version 3.0 vCards allow further splitting by comma
945
					if( _version.equals( "3.0" ) )
946
					{
947
						// split this part in to it's comma-separated bits and
948
						// add them on individual lines
949
						String[] adr_part_parts =
950
							splitValueByCharacter( adr_parts[ a ], ',' );
951
						for( int b = 0; b < adr_part_parts.length; b++ )
952
							if( adr_part_parts[ b ].length() > 0 )
953
							{
954
								if( value.length() > 0 ) value += "\n";
955
								value += adr_part_parts[ b ];
956
							}
957
					}
958
					else
959
					{
960
						// add this part on an individual line
961
						if( value.length() > 0 ) value += "\n";
962
						value += adr_parts[ a ];
963
					}
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
964
				}
37 by edam
- updated TODO and NEWS
965
966
			Set< String > types = extractTypes( params, Arrays.asList(
47 by edam
- bump version no. to 1.2
967
				"PREF", "WORK", "HOME" ) );
968
969
			// add address
970
			int type;
971
			if( types.contains( "WORK" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
972
				type = TYPE_WORK;
47 by edam
- bump version no. to 1.2
973
			else
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
974
				type = TYPE_HOME;
47 by edam
- bump version no. to 1.2
975
976
			addAddress( unescapeValue( value ), type );
977
		}
978
979
		private void parseLABEL( String[] params, String value )
980
		{
981
			Set< String > types = extractTypes( params, Arrays.asList(
982
				"PREF", "WORK", "HOME" ) );
37 by edam
- updated TODO and NEWS
983
984
			// add address
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
985
			int type;
37 by edam
- updated TODO and NEWS
986
			if( types.contains( "WORK" ) )
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
987
				type = TYPE_WORK;
37 by edam
- updated TODO and NEWS
988
			else
61 by edam
made contacts backend throw exceptions when it can't create contacts on the device; specified Locale in lower/upper case conversions; stripped old Contacts types from Importer (and replaced with our own types, which are now converted in the backend; switched using Integer() constructors to Integer.valueOf(); rewrote the main importer routine a bit
989
				type = TYPE_HOME;
40 by edam
- fixed logic for vcard field types (home, work, cell, etc) so it works
990
45 by edam
- fix UTF-8 unencoding by default on v3.0 vCards
991
			addAddress( unescapeValue( value ), type );
37 by edam
- updated TODO and NEWS
992
		}
993
65 by edam
added support for notes; rewrote backends so that all normalising of data is now done within the contacts cache; made the vCard unescape routine slightly more acceptant of non-standard escaped characters
994
		private void parseNOTE( String[] params, String value )
995
		{
996
			addNote( unescapeValue( value ) );
997
		}
998
43 by edam
- refactored some code to do with how contacts are imported
999
		public void finaliseVcard()
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
1000
			throws ParseException, ContactNotIdentifiableException
1 by edam
Initial import
1001
		{
1002
			// missing version (and data is present)
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1003
			if( _version == null && _buffers != null )
1 by edam
Initial import
1004
				throw new ParseException( R.string.error_vcf_malformed );
1005
43 by edam
- refactored some code to do with how contacts are imported
1006
			// finalise the parent class
44 by edam
- added checks for Doit.this == null when handling dialog buttons (I managed to abort an import as a duplicate contacts dialog was shown, but can't reproduce it now)
1007
			finalise();
1 by edam
Initial import
1008
		}
1009
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1010
		/**
1011
		 * Amongst the params, find the value of the first, only, of any with
1012
		 * the specified name
1013
		 * @param params
1014
		 * @param name
1015
		 * @return a value, or null
1016
		 */
1 by edam
Initial import
1017
		private String checkParam( String[] params, String name )
1018
		{
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1019
			String[] res = checkParams( params, name );
1020
			return res.length > 0? res[ 0 ] : null;
1021
		}
1022
1023
		/**
1024
		 * Amongst the params, find the values of any with the specified name
1025
		 * @param params
1026
		 * @param name
1027
		 * @return an array of values, or null
1028
		 */
1029
		private String[] checkParams( String[] params, String name )
1030
		{
1031
			HashSet< String > ret = new HashSet< String >();
1032
35 by edam
- accept parameters that are quoted (this doesn't appear to be part of the standards AFAICT, but Evolution apparently quotes parameter values)
1033
			Pattern p = Pattern.compile(
68 by edam
be less sensitive to case when parsing vCard params
1034
				"^" + name + "[ \\t]*=[ \\t]*(\"?)(.*)\\1$",
1035
				Pattern.CASE_INSENSITIVE );
1 by edam
Initial import
1036
			for( int i = 0; i < params.length; i++ ) {
1037
				Matcher m = p.matcher( params[ i ] );
1038
				if( m.matches() )
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1039
					ret.add( m.group( 2 ) );
1 by edam
Initial import
1040
			}
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1041
1042
			return (String[]) ret.toArray( new String[ ret.size() ] );
1 by edam
Initial import
1043
		}
1044
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1045
		/**
1046
		 * Amongst the params, return any type values present. For v2.1 vCards,
1047
		 * those types are just parameters. For v3.0, they are prefixed with
1048
		 * "TYPE=". There may also be multiple type parameters.
68 by edam
be less sensitive to case when parsing vCard params
1049
		 * @param params an array of params to look for types in
1050
		 * @param valid_types an list of upper-case type values to look for
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1051
		 * @return a set of present type values
1052
		 */
1 by edam
Initial import
1053
		private Set< String > extractTypes( String[] params,
36 by edam
- formatting: removed some double-indents on overrunning lines
1054
			List< String > valid_types )
1 by edam
Initial import
1055
		{
1056
			HashSet< String > types = new HashSet< String >();
1057
1058
			// get 3.0-style TYPE= param
46 by edam
- properly handle multiple TYPE= params in one entry in a v3.0 vCard
1059
			String type_params[] = checkParams( params, "TYPE" );
1060
			for( int a = 0; a < type_params.length; a++ )
1061
			{
1062
				// check for a comma-separated list of types (why? this isn't in
1063
				// the specs!)
1064
				String[] parts = type_params[ a ].split( "," );
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
1065
				for( int i = 0; i < parts.length; i++ )
68 by edam
be less sensitive to case when parsing vCard params
1066
					parts[ i ] = parts[ i ].toUpperCase( Locale.US );
1067
				for( int i = 0; i < parts.length; i++ )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
1068
					if( valid_types.contains( parts[ i ] ) )
1069
						types.add( parts[ i ] );
1 by edam
Initial import
1070
			}
1071
1072
			// get 2.1-style type param
1073
			if( _version.equals( "2.1" ) ) {
1074
				for( int i = 1; i < params.length; i++ )
25 by edam
- fixed bug where parts[0] was assumed to exists after calling split()
1075
					if( valid_types.contains( params[ i ] ) )
1 by edam
Initial import
1076
						types.add( params[ i ] );
1077
			}
1078
1079
			return types;
1080
		}
1081
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1082
		private UnencodeResult unencodeQuotedPrintable( ByteBuffer in )
1 by edam
Initial import
1083
		{
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1084
			boolean another = false;
1085
36 by edam
- formatting: removed some double-indents on overrunning lines
1086
			// unencode quoted-printable encoding, as per RFC1521 section 5.1
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1087
			byte[] out = new byte[ in.limit() - in.position() ];
1 by edam
Initial import
1088
			int j = 0;
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1089
			for( int i = in.position(); i < in.limit(); i++ )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1090
			{
1091
				// get next char and process...
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1092
				byte ch = in.array()[ i ];
1093
				if( ch == '=' && i < in.limit() - 2 )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1094
				{
1095
					// we found a =XX format byte, add it
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1096
					out[ j ] = (byte)(
36 by edam
- formatting: removed some double-indents on overrunning lines
1097
							Character.digit( in.array()[ i + 1 ], 16 ) * 16 +
1098
							Character.digit( in.array()[ i + 2 ], 16 ) );
1 by edam
Initial import
1099
					i += 2;
1100
				}
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1101
				else if( ch == '=' && i == in.limit() - 1 )
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1102
				{
1103
					// we found a '=' at the end of a line signifying a multi-
1104
					// line string, so we don't add it.
1105
					another = true;
1106
					continue;
1107
				}
1 by edam
Initial import
1108
				else
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1109
					// just a normal char...
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1110
					out[ j ] = (byte)ch;
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1111
				j++;
1 by edam
Initial import
1112
			}
18 by edam
- changed case on charset and encoding warning strings (it looked bad)
1113
22 by edam
- read vCard files in as raw bytes now and convert to string only tentatively to check for version no.s and property names and params
1114
			return new UnencodeResult( another, ByteBuffer.wrap( out, 0, j ) );
1115
		}
1116
1117
		private ByteBuffer transcodeAsciiToUtf8( ByteBuffer in )
1118
		{
1119
			// transcode
1120
			byte[] out = new byte[ ( in.limit() - in.position() ) * 2 ];
1121
			int j = 0;
1122
			for( int a = in.position(); a < in.limit(); a++ )
1123
			{
1124
				// if char is < 127, keep it as-is
1125
				if( in.array()[ a ] >= 0 )
1126
					out[ j++ ] = in.array()[ a ];
1127
1128
				// else, convert it to UTF-8
1129
				else {
1130
					int b = 0xff & (int)in.array()[ a ];
1131
					out[ j++ ] = (byte)( 0xc0 | ( b >> 6 ) );
1132
					out[ j++ ] = (byte)( 0x80 | ( b & 0x3f ) );
1133
				}
1134
			}
1135
1136
			return ByteBuffer.wrap( out, 0, j );
1 by edam
Initial import
1137
		}
1138
	}
1139
}