1 Release: version 0.8 - 2007/05/??
2 ------------------------------------------------------------------------------
4 - added doc/TroubleShooting.txt
7 - fixed bug in vcard.cc that was using the FN field
8 for the broken-down name instead of N
9 - added dbId and dbName to DatabaseSyncState
10 - added Reconnect() functionality to BarryEnvironment class
11 and moved connect logic out of barry_sync.cc
12 - changed the batch commit functionality in barry_sync.cc
13 to single-change-commit. Why? Because opensync
14 0.22 groups all batch changes into one group
15 and calls one batch commit function, even if you
16 registered two. Moving to single change commit
17 mode removes a chunk of unneccessary code
18 from the plugin as well, and can in theory
19 ease memory pressure in the library.
20 - added FinishSync() to perform actions needed on successful
22 - disconnect on successful sync completion, and reconnect,
23 since the device's dirty flags don't seem to
24 be updated until a disconnect occurs.
26 - added ReadDataArray() to data.h, using an istream instead of
28 - added raw data version of BuildField() to record-internal.h
29 This is so that btool doesn't add extra null terminators
30 to unknown data, when restoring with option -s
31 - fixed bugs in r_contact.cc:
32 - fixed bug that caused a LastName/FirstName swap
33 if FirstName was empty and you wrote a contact
34 record into the device.... First/Last Name
35 fields are position sensitive, and the code
36 was missing the empty FirstName case on writing
37 - fixed bug in BuildFields() and Dump() where they
38 were not updated to handle the new field
39 table added for the postal address changes
42 - changed the plugin config file, so it is possible to
43 turn debug output on/off without recompiling
44 - moved the standalone Map2Uid() function into
45 the DatabaseSyncState class, and changed the
46 code to produce "contact-#" and "calendar-#"
47 UID strings, instead of just numbers
48 - changed default PIN to -1, so that if the user doesn't
49 configure properly, it will fail... this is a
50 safety measure to avoid syncing with different
51 devices and corrupting the cache/idmap...
52 this should be fixed someday, to handle this
54 On second thought... is this right? Perhaps
55 it is better to fail if the PIN is not what
56 is expected, so a user doesn't accidentally
57 sync the wrong device and lose data.
58 Actions to take on new PIN:
61 - incorporated the new VCardConverter code into the
62 main barry_sync.cc plugin code
63 - fixed bug that did not clear the devices dirty flags
64 in the case where a sync only read data from it,
66 - changed PIN error message number to hex
68 - minor wording change in backup GUI prompt: "Backup working..." to
69 "Backup in progress..."
70 - renamed Contact class's Title field to JobTitle for clarity
71 - added record function ParseFieldString(), which properly handles
72 strings from the device that have multiple null terminators,
73 and changed the record parsing code to use this new function
74 - moved FullName() function from ldif.cc to r_contact.cc since it
75 is generally useful in areas outside of LDIF processing
76 - refactored postal addresses in the Contact record class into
77 a new class called PostalAddress
78 - renamed Address class in record.h to EmailAddress
79 - moved Contact::GetPostalAddress() to record.cc as part of the
80 new PostalAddress class and renamed it
81 PostalAddress::GetLabel()
83 - refactored vformat helpers into separate base class
85 - added vCard converter class
86 - commented out the code that skipped over the adding
87 of empty attribute values and parameters,
88 since VCARD depends on some of this behaviour
89 - movified vformat.c:vformat_find_attribute() to take
90 an extra nth argument, in order to search
91 for multiple attributes with the same name...
92 the way this is coded is a bit inefficient...
93 but we're moving away from vformat when we move
94 to OpenSync 0.3x / 0.40, so this is just temporary
97 - added SUBSYSTEM=="usb_device" to example udev rules, in order
98 to avoid running bcharge for each endpoint that is
99 added by udev, when first plugged in. Running bcharge
100 back-to-back multiple times can crash some
104 - fixed ID mapping bug in the case of a non-numeric UID
105 from syncing external -> Barry, and then the
106 new Barry ID would look like a new change
107 when syncing back from Barry -> external.
108 - added support for slow sync
109 - added clear() to idmap class
110 - added vformat attr parameter extraction class
111 - added more trace logging during vformat/Barry conversions
112 - added Notification time support
113 - handled "all day" vformat special case, when DTEND
114 does not exist in a vformat data block
115 - changed the min1900_t time conversion functions to handle
116 the 0xffffffff special case, when notifications
117 are turned off on a calendar event (src/time.cc)
118 - also made the calendar data dump output clearer for
119 disabled timestamps, and added debug output
120 for the raw min1900 value, in case there are more
121 special cases in the future
123 - fixed null pointer access in opensync's HasMultipleVEvents()
125 - upper management directive: revert all OpenSync 0.3x changes
126 and proceed with OpenSync 0.22. Reverting the
127 opensync-plugin/src tree to Barry 0.7 status.
129 - reorganized exception and trace code in opensync plugin
130 - added XmlToCalendar() function, based on unfinished opensync
131 library API that is not yet finished... will be
132 committed to opensync svn tree when complete
133 - added OSyncXMLField logger to plugin trace class
135 - applied email MessageRecord patch from Brian Edginton (thanks!)
136 - updated Message::Dump() to output more valid mbox data...
137 the goal is to be able to do:
138 btool -d "Messages" > email.mbox
140 - changed some TODO's to FIXME's to keep it grep-consistent
141 - applied Folder support patch from Brian Edginton
142 - fixed pedantic compile warnings, and changed order in src/Makefile.am
143 - removed unneeded cast in src/r_folder.cc
145 - minor adjustment to ctags building in buildgen.sh
146 - applied patch from Brian Edginton for better boost detection
148 - added boost usage message to btool -h output
150 - more porting of OpenSync module to 0.30
151 - added ctags build for opensync source tree to buildgen.sh
153 - changed PKG_CONFIG_PATH setting in rpm script, so that systems
154 like SuSE that have an extensive PKG_CONFIG_PATH
155 in general use will not break on build
156 - added PKG_CONFIG_PATH to debian script too, just in case :-)
157 - added debian/changelog to doc/ReleaseChecklist.txt
158 - added comment to rpm/barry.spec regarding SuSE RPM builds
160 - applied patch from Niels de Vos fixing debian package build
162 - applied Brian Edginton's Saved Email Messages patch
163 - updated r_saved_messages.cc comment to match header
164 - added GetRecType()/GetUniqueId() to r_saved_messages.h, since
165 SetIds() is already there
166 - removed Address version of operator<<() from message related headers
167 and moved into record.h
168 - added std::hex to message related dumps, and changed output to
170 - updated doc/CodingStyle.txt to cover switch() statements
171 - opensync compiles again, with 0.30! (not functional yet)
172 - moving closer to more modular and reusable plugin functions
173 - updated release checklist for opensync version number
175 - added doc/CodingStyle.txt... preliminary version
177 - applied Brian Edginton's PIN message patch
178 - changes to PIN message patch:
179 - fixed missing backslash in src/Makefile.am
180 - fixed tab alignment throughout
181 - removed typedef from Address struct
182 - added clear() to Address... might as well, if it it is
183 going to be a standalone type
184 - fixed missing comma in PINMessageFieldLinks table
185 - changed MessageRecordId in PINMessage to load from a new
186 uint32_t field in CommonField, instead of piggy
187 backing on a field for other use
188 - did some testing and found that PINMessage's MessageRecordId
189 is the same as the RecordId given through the SetIds()
190 API... but these Id's are duplicated in the protocol...
191 RecordId is stored in the protocol header, while
192 MessageRecordId is stored in the record it self as
193 a common field... added RecType and RecordId to
194 PIN Message so that we store both of these
197 - added item to gui/TODO
198 - applied Brian Edginton's s11n-boost.h patch
200 - renamed template values to more consistent naming scheme
201 - applied memos and tasks patch from Brian Edginton, slightly
202 modified to split into the new separate files system
203 - changes to memos and tasks classes:
204 - reformatted the code to match coding style
206 - added copyright notice
207 - changed class to singular (Memo, not Memos)
208 - placed in Barry namespace
209 - added #ifdef header protectors
210 - added new files to Makefile.am
211 - added new r_*.h headers to record.h
212 - added support for Memos and Tasks to tools/btool.cc, based on
213 Brian Edginton's patch
214 (currently commented out, for Boost library reasons):
215 - added doc/CommitPolicy.txt
218 - added ctags generation to buildgen.sh
219 - applied record.{h,cc} fix patch from Brian Edginton
223 - fixed broken formatting in above patch
224 - refactored record parser classes into separate files for each
225 - changed the ConvertHtoB<> template to inline, as it caused
226 linker errors with the new file layout
228 - fixed bug in Debian packaging that put util commands in
229 /bin and /sbin instead of /usr/bin and /usr/sbin
230 - commited byteswap.h replacement code from gm2net
232 - removed unneeded debug messages on stdout
233 - changed the default for debug output to "off"
234 and added a command line option to turn
235 it on if necessary (--debug-output)
236 - command line option --help now also works
237 - fixed bug in Calendar class where it didn't initialize
238 RecType properly... and cleaned up other RecType
239 code to use the GetDefaultRecType() functions,
240 localizing the constants in one place
242 - applied patches from Brian Edginton, parsing more
243 fields in Contact and Calendar records
244 - added Brian Edginton to AUTHORS
245 - reformatted record.{h,cc} to match the rest of the code,
246 removed commented out code, and removed mention of
247 Outlook from the comments.
248 - updated udev/10-blackberry.rules.Debian for Debian Etch
249 - updated debian/rules to use new udev rules file
250 - removed debian/postinst, as stable versions of Debian now
251 have a recent udev, and don't need the special handling
252 - updated Debian build scripts to build separate packages:
257 - added menu support to Debian barrybackup-gui package
258 - made the buttons bigger in the backup GUI, as they were too
261 - more progress in porting to opensync 0.30, heavily
262 based on the example plugin in the opensync
263 svn tree... this is a work in progress, and so far
267 - partial porting to opensync 0.30 (in progress!)
268 - fixed bug in sync_done where it reported context error
269 and then success afterward
270 - checking in new INSTALL files based on Etch's autoconf run
271 - added item to gui's todo list
272 - bumped version numbers to 0.8 in:
276 - gui/src/BackupWindow.cc (about dialog)
280 Release: version 0.7 - 2007/05/02
281 ------------------------------------------------------------------------------
283 - added rpm building to maintainer/release scripts
284 - added "HEAD" feature to tarball creation script, for easier testing
285 - updated maintainer/README
286 - updated maintainer/release.sh to use new HEAD feature too
288 - added pppob to rpm barry.spec
289 - bumped release number in barry.spec
292 - fixed size check bugs (strlen() == 0 when empty)
293 - finished refactoring calendar conversion routines
294 - started timezone parsing, but this may go away
295 if we use opensync 0.30's time functions and
297 - added development note to README
299 - updated README and opensync-plugin/README with more current
302 - refactored some calendar conversion routines
303 - added Barry->vCalendar conversion for recurring appointments
306 - removed some unneeded trace logs
307 - fixed VEventConverter constructor bug, with uninitialized
309 - eased up the requirement that every calendar event
310 must have a summary... defaults to "<blank subject>"
312 - added more VCALENDAR fields to match Barry::Calendar record
313 - fixed case sensitive compares in vformat.c
315 - opensync module work:
316 - added id map class to map string uid's from OpenSync to
317 numeric ID's from the Blackberry
318 - fixed bug in CommitChange() where it was still using
319 a hard coded calendar record state table
320 - added lots of logging throughout
321 - fixed bug where newly added records didn't use the
322 change object's uid, but created a new ID,
323 causing another sync on the next run
324 i.e. (CommitData_t was missing an argument)
325 - fixed bug where uninitialized notification time was
326 written to the Blackberry
327 - more opensync module work:
328 - refactored the environment class into another subclass
329 called DatabaseSyncState, and moved some stand-
330 alone functions from barry_sync.cc into
331 the environment classes
332 - added a tagged constructor to the trace class
333 - fixed idmap::Load() to clear map before loading
335 - renamed barry-config to barry-sync, the default configuration
336 file for the opensync module... and filled it with
337 a sample configuration
338 - added code to update the calendar cache, and write it to disk
339 on success (opensync module)
340 - added code to clear dirty flags in the device (opensync module)
341 - allowed comments in the config file
342 - added vformat.c from libopensync 0.22, since the vformat
343 routines are not available from the plugin side
346 - added -d switch to bcharge to switch back from 0001 to 0004 mode,
347 changed the auto-detection logic to ignore iProduct and
348 only use idProduct, and updated man page
349 - updated breset to recognize the Pearl in 00004 mode as well
350 - changed src/probe.cc logic to assume that the second endpoint
351 pair is the one needed for database communication, instead
352 of doing a reverse search from the end... newer Pearls
353 in Dual mode seem to have at least 3 or 4 endpoint pairs,
354 and the old search didn't always work
355 - applied sourceforge patch #1696884 from Peter Silva, updating
356 the Debian build for Etch, and including all files in one deb
357 - added Peter Silva to AUTHORS file
358 - uncommented the "make distclean" in debian/rules, and removed
359 the reliance on /tmp when installing 99-barry-perms.rules
360 - applied bb_task_format.txt patch from Peter Silva
361 - fixed each misspelling of "recurrance" to "recurrence"
362 - tweaked src/probe.cc again for paranoia's sake... start at
363 offset 1 if more than 1 endpoint pair, otherwise start
365 - split up opensync module into multiple source files, and started
366 using the opensync vformat routines for parsing and building
367 the vevent20 data (experimental)
369 - more incremental work on the opensync plugin, adding ADDED
370 and MODIFIED functionality... (experimental!)
372 - added code to dump extra descriptors during device discovery,
374 - bumped opensync version number in configure.ac, and made
375 the configure scripts more consistent with the rest
377 - fixed printf-format size bug in opensync module
379 - applied patch from Niels de Vos, fixing deb udev rules for Pearl
381 - fixed extra space bug in LDIF output (added space when creating
382 FullName, even if there was no last name)
383 - fixed Address Book record parsing in Contact class... the
384 Blackberry database uses field id 0x20 for both the
385 first and last names, and if only a last name exists,
386 the first name will be empty. Fixed parser to handle
387 this special case, without relying on string length.
388 Thanks to Troy Engel for reporting above two LDIF bugs.
389 - added a few more constants to bcharge.cc
390 - renamed opensync's configure.in to configure.ac
392 - fixed bcharge to avoid endless reset loop... oops
394 - reorganized bcharge.cc to make it possible to switch between
395 Pearl modes 0004 and 0001
396 - added flag to BadPassword exception class, to signal whether
397 the library considers the device "out of retries" or not
398 - updated socket.cc for new BadPassword exception
399 - added special case to Controller class destructor to reset the
400 device if being destroyed in a half-open state... this is
401 a temporary fix until we understand how to handle
402 the "already open" message we get when opening the Desktop
404 - added password prompt dialog to the backup GUI
405 - fixed possible memory leak in GUI's DeviceInterface::Connect()
407 - committed Troy Engel's bbrules_mass_symlinks.patch, fixing
408 udev rules for the Pearl
409 - updated 10-blackberry.rules.Debian to match 10-blackberry.rules's
411 - applied Troy Engel's barry.spec patch to generate all required
412 RPM's... Thank you! Notes in the README taken from
413 the sourceforge tracker entry containing the patch.
414 - added libusb lib flags to tools/ and examples/ makefiles,
415 and tweaked configure.ac with a better default,
416 to build cleanly on RHEL 4
417 - modified rpm/barry.spec:
418 - does not create doc-only barry base package anymore
419 - fixed "barry-backup" typo
420 - removed the patch step, as version 0.7 shouldn't need it
421 - added license file to each package
422 - commented out console perms scripts, pending mailing list
425 - applied Troy Engel's barry.spec patch to fix RHEL4 compile issues
426 - fixed 80 column wrapping error in barry.spec, thanks to Troy Engel
428 - applied Troy Engel's udev permissions patch for Redhat/Fedora
429 systems, which automatically set the ownership of a
430 newly plugged in device to the currently logged in console
432 - added udev/README to document what all those files are for...
433 part of the text comes from the sourceforge comment
434 on the above udev permissions patch from Troy Engel
435 - uncommented the console perms scripts for rpm/barry.spec
436 - added libglademm-2.4 to gui/README's dependency list
437 - made usbwrap.* calls more threadsafe
438 - added Usb::Timeout exception for timeout errors
439 - added preliminary UsbSerData support to library for GPRS modem
440 functionality, using Rick Scott's XmBlackBerry's serdata.c
442 - added overloaded Socket::Send() for sending without receiving
443 - added preliminary pppob command line tool
449 - gui/src/BackupWindow.cc (about dialog)
452 - build system tweaks for opensync-plugin, and added a buildgen.sh
454 - added better debug messages in probe.cc and usbwrap.cc's
455 device discovery code
456 - changed error handling for null interface->altsetting pointers
457 when doing device discovery in usbwrap.cc. Seems that
458 some devices report a higher number in bNumInterfaces
459 than are really available... in this case, let's
460 just ignore the missing data and assume success
461 - applied udev script patch #1663986 from Troy Engel
462 - applied barry-bcharge.spec patch #1672178 from Troy Engel
463 - ReleaseChecklist update
464 - added rpm/README based on Troy Engel's patch notes
465 - updated new barry-bcharge.spec for version 0.7, removing patch
468 Release: version 0.6 - 2007/02/28
469 ------------------------------------------------------------------------------
471 - fixed size checks for device database records larger than 64K
472 Thanks to Michael Brown for the bug report.
473 - added initial release maintainer scripts
474 - fixed odd backup success message in GUI when doing a backup with
475 no databases selected in configuration
476 Thanks to Michael Brown for the bug report.
477 - added udev permissions file patch from Michael Brown
478 - added Michael Brown to AUTHORS
479 - updated README, gui/README, and doc/USB-capture.txt
480 - added -o command line switch to bcharge.cc, and updated man page
482 - changed library behaviour so it only turns on libusb debug output
483 if in data dump mode. In btool, -v controls this output.
484 - minor stream formatting fix in btool
485 - added OpenSSL dependency to build for password hash support
486 - made debug dout() output conditional on Init() flag... In btool,
487 -v now controls this as well
488 - reverse engineered more of the initial probe packets, and
489 implemented probing in struct based code... changes include:
490 - added support for "attribute fetch" packets, used at
491 the beginning of USB conversations
492 - defined object/attribute id codes for the PIN number
493 - made Protocol::SocketCommand more robust, and implemented
494 its sequence number behaviour
495 - moved ModeSelectCommand into SocketCommand, as it is a
496 socket 0 level command
497 - added initial protocol struct: PasswordChallenge
498 - renamed Barry::Packet to DBPacket
499 - added ZeroPacket for socket 0 packet creation and analysis
500 - Barry::Packet is now the base class for the above
501 - implemented socket 0 behaviour in socket.cc
502 - changed return values for Send/Receive/Packet/NextRecord
503 members in the Socket class from bool to void, as they
504 don't generate errors themselves, but only can
505 have an error if Usb::Device throws an exception.
506 Cleaned up all code that called these members as well.
507 - removed "last status" from Socket class as it was unused
508 - updated probe.cc to use new socket 0 implementation
509 - removed two of the hard coded packet handshakes in favour
511 - updated controller.cc:
512 - uses new socket 0 implementation
513 - removed "flag" hack when opening sockets and
514 selecting modes, as it should use the
515 socket 0 sequence number
516 - added initial, incomplete prototype header for the C API
517 - updated some .cvsignore files
519 - updated ReleaseChecklist.txt, adding rpm spec file
520 - added the following to ZeroPacket:
525 - moved Command() to base Packet class
526 - added comments to protostructs.h and size constants for the
527 new PasswordChallenge struct
528 - added password support to Socket and Controller classes
529 Thanks to Rick Scott's XmBlackBerry for the openssl
530 password hashing logic.
531 - updated btool to use new password support, and updated its man page
533 - updated Doxygen input files and version, and added to ReleaseChecklist
534 - fixed btool PIN output for LDIF mode, now commented in output
535 - major LDIF overhaul:
536 - split LDIF specific code out of the record class and
538 - fixed "full name" behaviour that added an extra space
539 in dn attributes, as reported by Troy Engel
540 - added support for mapping of LDIF attributes to Barry
542 - fixed base64 handling, so attributes are only so encoded
544 - fixed base64 attribute reading inconsistency... some
545 notes fields were missed in old version if not
547 - added support for specifying alternate attribute for
548 constructing a FQDN, instead of just "cn"
549 - updated btool and upldif to use new features
550 - large update to btool manpage
551 - fixed initialization bug in Contact constructor (missed RecType)
553 - added ktrans.cc to make reading of kernel usbfs_snoop logs easier
554 - removed commented code from translate.cc
555 - added Pearl handshake support to bcharge.cc, which resets the
556 Pearl so that the vendor specific 0xFF class exists.
557 This should allow use of btool on the Pearl.
558 - added Data::AppendHexString to data.cc
559 - added better probing support for the Pearl, so it detects the
560 Product ID 0x0004 mode, and does a search for class 0xFF
561 instead of hardcoding it.
563 - applied README patch from Ian Darwin
564 - minor tweaks to README
565 - added missing errno.h header in src/usbwrap.cc for AMD64
566 Thanks to Jonathan Hudson for reporting the bug.
567 - tightened up size checks in Data::ReleaseBuffer()
568 - stable libusb *does* return the actual length of the USB packet...
569 fixed this in usbwrap.cc to set the actual size of read data,
570 and removed the size hacks elsewhere in the code:
571 probe.cc, protocol.cc, socket.cc, usbwrap.cc
573 - added proper return codes for error conditions in btool.cc
574 - clarified some stream output code in the library, for hex/dec numbers
575 - removed old, non-working connect.cc test program
576 - added iomanip to debug.h
577 - changed usbwrap to call libusb again if EINTR and EAGAIN
578 This should fix one of the OpenBSD issues
581 - fixed bug in barrybackup GUI that didn't create the target path
582 if the user didn't do Edit | Config first.
583 Thanks to Ian Darwin for reporting the bug.
584 - applied gui/src/ConfigFile.cc header fix patch for OpenBSD
587 - added initial btool.1 man page from Ian Darwin
588 - updated man/Makefile.am for new btool.1
590 - applied btool -h help clarification patch from Ian Darwin
592 - reworked exception hierarchy, so applications can handle all
593 Barry related exceptions, including Usb errors, with
594 one base class: Barry::Error
595 - fixed all the ripple changes this caused throughout the codebase
596 - added BadPassword exception
602 - gui/src/BackupWindow.cc (about dialog)
603 - updated doc/ReleaseChecklist.txt
604 - added usb-level breset.cc command line tool
605 - fixed bug in gui/src/tarfile.* for systems where
606 sizeof(int) != sizeof(void*)
607 - TarFile class now uses a plugin style for the compression
608 operations, to allow for threadsafe versions if needed
610 - also fixed leaked file handle bug in open_compressed()
611 - Thanks to Jonathan Hudson for reporting this for 64bit systems
612 - minor signed comparison fixes in the library
613 - added better protocol data dumping code in probe.cc
614 - added Usb::Device::BulkDrain(), to prevent Barry from hanging
617 Release: version 0.5 - 2007/01/26
618 ------------------------------------------------------------------------------
620 - added missing errno.h header to gui/src/util.cc
621 - lots more endian fixes after testing on iMac
622 - iostream output size fixes
623 - added option to btool to reset the device via software
624 - beefed up exception handling in barrybackup
625 - checked in opensync work-in-progress: trace logging, pointer bugfix
626 this is still experimental
627 - added aboutdialog.* files from gtkmm24 2.10.6, so we have a dialog
628 class for systems running older versions of gtkmm, like
630 - uncommented the about box menu handler
631 - added "cleanall" option to root buildgen.sh script
632 - update ReleaseChecklist.txt
633 - removed CVS auto-history strings from aboutdialog.*
635 - found a link between the unknown field in tagged protocol headers
636 and the unknown field in RecordStateTable entries. This
637 field matters when uploading certain databases, such as
638 Browser Options. It is unknown what this field really
639 means, but we're calling it RecType. Updated the API
640 to use this in all parser objects and builder objects
641 and record objects that use it.
642 - changed some uses of uint64_t to uint32_t, in Contact and
643 Calendar record objects... unique IDs only seem to be
645 - added new static function to record classes: GetDefaultRecType()
646 This returns the default rectype that should work
647 for that record. Ideally this should be retrieved from
648 the device, but in the case where a database has no
649 records, this info would be good to know.
650 - updated convenience template function Controller::AddRecordByType()
651 to use GetDefaultRecType()
652 - documented the recurrence data and how to use it in Calendar objects
653 - added data debug output for incoming USB packets in the
654 socket class... when we ported to the stable libusb
655 we lost some of the verbose output... this should get it
656 back. (In the official release, you need to uncomment
657 ddout() in debug.h for the full blast of debug output).
658 - backup GUI updates:
659 - changed error_done signal into a pure error signal,
660 and changed the logic so the restore continues
661 even if there was a protocol error on one database
662 - changed status bar messages to show what's happening
663 i.e. Backup or Restore
664 - default to not restore the "Handheld Agent" database as
665 it appears to be read-only on the device
666 - added temporary Restore & Backup mode for debugging
667 - added extra hyphen to backup files to show date/time better
668 - updated GUI to save and restore the new RecType field
669 supported now by the library
671 - added future considerations to doc/ReleaseChecklist.txt
672 - applied big endian patch from Jonathan Hudson
674 - removed endian.h include from barry.h, since applications
675 using the library should not care about endianness
676 - hard coded the path to config.h in endian.h
677 so that endian.h never makes it into the final
679 - removed DEFAULT_INCLUDES setting from src/Makefile.am
681 - removed config.h dependency in btool.cc and used library
683 - added big/little endian text to the Barry::Version call
685 - fixed install location of bcharge when building with ./configure;
686 now goes to sbin/ for both source compile and binary packages
687 Thanks to Jonathan Hudson for reporting this bug.
689 - added smarter timeout values to controller, and allowed
690 timeout overrides in the packet, socket, and usbwrap code...
691 extra timeouts are needed for erasing databases for
693 - added an explicit DataDumpParser class to btool.cc, as the
694 library doesn't always have extreme verbose debugging on...
695 this makes sure anyone can always get raw database packets
696 through btool, for development
697 - added thread callback signals for error finishes and erase_db
698 operations, so the GUI is updated properly
699 - more GUI cleanups (status bar updates, etc)
700 - finished implementation of GUI restore
701 - fixed gtkmm/glibmm exception handling when the exceptions
702 occur in sigc signal handlers... in that case,
703 exceptions need to be handled with
704 Glib::add_exception_handler()
705 - added filename-only reading support to TarFile class
706 - added pending gui/TODO items
708 - backup GUI now properly shows progress
709 - the buttons become insensitive when backup is busy
710 - refactored directory check code, and added mkdir check
711 when to the backup setup (so the directory exists
712 when we try to write to it)
713 - updated code to use Barry::Data in new namespace
715 - moved Data class into Barry namespace
717 - updated copyright dates for the new year
718 - clarified exception message in controller.cc
719 - fixed pointer initialization bug in controller.cc
720 - added documentation comments to parser.h and builder.h
721 - updated ReleaseChecklist.txt
723 - changed Barry library configure scripts to make better
724 use of pkg-config for autodetecting dependencies
725 - added initial Barry Backup GUI (gui/ directory), capable of making
726 backups only at this point
728 - added doc/ReleaseChecklist.txt
729 - added version API, returning version number and string
730 - added Troy Engel to AUTHORS
731 - added special case to bcharge.cc for the Blackberry Pearl,
732 which doesn't reset itself after the charge handshake
733 Thanks to Troy Engel for testing.
735 - added pkg-config support for the library (libbarry-0.pc.in)
736 - fixed compile error on g++ 3.3 systems (missing stdint.h
738 - fixed protocol hang at end of Controller operation... in
739 controller destructor, it was deleting the USB interface
740 before closing the socket at the BlackBerry level, thereby
741 leaving the device in Desktop mode
743 - support the newer udev packages on Ubuntu, which need slightly
745 - modified bcharge.cc to search for Pearl devices too.
747 - added Pearl product ID to probe code, but likely not functional
748 yet until we find someone able to get a USB capture for us
750 - version bump in configure.ac and src/Makefile.am
752 Release: version 0.4 - 2006/12/29
753 ------------------------------------------------------------------------------
755 - added bcharge.cc to the tools/ directory (whoohoo!)
756 - added initial release version of rpm/barry.spec (aka bcharge.spec)
757 This begins the barry binary release, currently only
758 including the bcharge utility, but possibly more in
759 the future. Future binary packages will likely include
760 libbarry, libbarry-devel, etc, which barry will then
762 - added udev rule script for BlackBerry devices
763 - added man/ directory and bcharge.1 manpage
764 - added debian/ directory for building deb packages, with initial
766 - added rpm/make-rpm-tarball.sh for building RPMs
767 - split out technical docs into separate libbarry.docs file for
768 a future Debian install
769 - added more cleaning to buildgen.sh
770 - added new udev rule file for Debian stable
771 - hardcoded g++ to 3.3 in debian/rules as I have multiple versions
772 on my system and need to compile releases with stable
774 - ported Barry to use the stable branch of libusb
775 NOTE: no longer depends on the devel libusb tarball!
776 Just use your distro's libusb packages.
777 - fixed more endian issues
778 - minor update to TODO
780 - moved Exceptions, Hacking, and VersionNotes to doc/
781 - added doc/USB-capture.txt
783 - added Controller::AddRecord() and template helper
784 - added beginning version of generic RecordFetch<> template
785 - set default of Barry::Init() to false (no data dump mode)
786 - added examples/ directory, with first addcontact.cc example
788 - added autoconf support based on Peter McAlpine's patch
789 - renamed Makefiles to Makefile.orig and added script to turn
790 on old build system if wanted
791 - fixed boost-specific error path in btool
792 - moved platform specific code out of time.h and into time.cc
793 - moved btool.cc, translate.cc, and upldif.cc out of src/ and
794 into their own tools/ subdirectory
795 - added NEWS file to make autotools happy
796 - added a "clean" option to the buildgen.sh script... maintainer's
798 - reworked autoconf and automake files, to support
799 barry as a library, to move the programs into their
800 own tools directory, and support a proper installation
801 - fixed headers in parser.h, btool.cc, and upldif.cc to work
802 properly when installed standalone
803 - added tools/Makefile.plain for testing compiling against
804 installed barry library
805 - added commentary on version number system
806 - removed Makefile.conf and Makefile.orig, as new autoconf
808 - added support for boost serialization library in autoconf build
809 - added library version number to autoconf build
810 - changed boost flag to __BARRY_BOOST_MODE__ to avoid collisions
811 - removed debug.h dependency from btool.cc
812 - removed libusb check in favour of specific --with-libusb
813 option... may need to revisit this later
814 - quick installation update in README
816 - fixed some missed endian conversions in record.cc and added
818 - fixed endian issues in time.h
819 - added 2 more functions to the time zone API, and adjusted
820 the time zone table for better defaults in North America
821 - added btool and upldif to the make install sequence
822 - added support for Calendar recurrence data
823 - small doc/TimeZones.txt update
825 - added BlackBerry time zone code list, reverse engineered from a 7750
826 - added GetTimeZone() to convert device time zone codes to useful data
828 Release: version 0.0.3 - 2006/11/24
829 ------------------------------------------------------------------------------
831 - added return code check to Controller::SaveDatabase()
832 - fixed Data() class constructor, so it doesn't auto-convert
834 - fixed size checking bug in Packet::ReturnCode() that caused
835 uploading to device to fail
837 - added Exceptions doc file, documenting the C++ exceptions used,
839 - fixed small bug in 'make install' which didn't build first
840 - minor documentation and cleanup in upldif.cc
841 - added utility functions to the RecordStateTable class,
842 GetIndex() and MakeNewRecordId()
843 - changed btool so that the -d command modifiers can be
844 specified multiple times to work with multiple records
846 - added Calendar sync code to the opensync module (not complete)
847 - cleaned up exception handling in the opensync module, making
848 more use of the base exception to make sure no
849 exceptions leak into the C library code
851 Release: version 0.0.2 - 2006/10/12
852 ------------------------------------------------------------------------------
854 - fixed casting error in opensync module
855 - minor updates to documentation to prepare for release
857 - added stdint.h to parser.h for uint32_t. Thanks Ron Gage for
860 - large refactoring of the packet parsing and building code,
861 splitting the code into 3 groups:
862 - record: handles record specific parsing and building...
863 this is mostly subfields with CommonField structs
864 as their headers, but sometimes has a specific
865 header of its own, in the case of email
866 Note: email's header is as yet undecoded.
867 Record does not know the details of the packet
868 formats, only the formats of its own record data.
869 - packet: handles building of complete command packets
870 and handles parsing of commonly used header values,
871 providing an API for it... packet does not know
872 the record format details, but does know the
873 details of each Database Operation (protocol.h)
875 So far, it seems that there is a real separation
876 of operation formats and record formats, with the
877 record formats being the same even with different
878 database operation codes (GET_RECORDS /
879 OLD_GET_RECORDS), which was not confirmed before.
880 See the header size of the email record after
882 - controller: handles device state, and the management of
883 sockets, commands, and checking for response values.
884 The Controller object does not know the format
885 of either the record or the packet it is sending,
886 but does know the protocol handshake logic used
887 to talk to the device.
888 - implemented SetRecord (SET_RECORD_BY_INDEX)... syncing, here
890 - spelling correction in opensync-plugin/AUTHORS
892 - added sample hotplug scripts
893 - added .cvsignore files for doxygen and opensync directories
894 - added src/endian.h... still need to add configure support to
895 handle non-little-endian machines
896 - added exception handler in opensync plugin commit_change()
897 - removed the device reset code in probe.cc, as resetting a device
898 can renumber the devices in /proc on kernel 2.4.x,
899 and cause a failure to reach the device...
900 FIXME - this should be changed someday to reset if probing
901 fails, and restart the probe on reset, as sometimes
902 the blackberry devices respond differently when not reset
903 Ideally, the probe protocol should be properly reverse
905 - added support for retrieving record state table, which is needed
907 - added support for retrieving, deleting, and clearing the dirty
908 flags of individual blackberry device records
909 - added endian macros throughout the code
911 - added opensync-plugin to root Makefile's clean
912 - updated time conversion calls to match opensync's latest SVN
913 - added opensync_change_set_changetype() (only add for now,
915 - removed slow sync functions for now
917 - added quick "install" makefile target, to src/install by default
918 - added static library target... this is temporary until autoconf
919 stuff is added properly... needed for the opensync plugin
920 - documented in comments that Barry::BError is intended to be
921 the base classes for all exceptions
922 - added first round of implementation of an opensync plugin, based
923 on the example plugin code from the opensync sources
925 - added to Contact record class:
927 - ReadLdif() to load and parse LDAP data
928 - added fragmented send support to Socket class
929 - added Probe::FindActive() to make client programs more streamlined
930 - fixed hex/decimal printing of record counts in btool
931 - added upldif.cc, a utility to take ldapsearch output on stdin
932 and upload it to a blackberry
933 - updated src/.cvsignore
935 - added parsing of service book fields:
937 - old and new unique ID (UID)
939 - old and new description
941 - bes domain / domain
942 - date change in legal.txt
943 - added initial parsing of ServiceBook field 0x09, and calling it
944 ServiceBookConfig for now, for lack of a better name
945 - added parsing for ServiceBook field 0x02 on 72xx Blackberries...
946 seems to be a name or description, but doesn't appear
947 in the device's own GUI... called "Hidden Name" for now
949 - added ServiceBook record class and parser code, based on some
950 of Ron Gage's reverse engineering... code is by Chris Frey
952 - fixed the libusb set_configuration/claim_interface order
953 bug in controller.{h,cc}... already fixed in probe,
954 but not controller. Thanks Ron Gage for finding this.
956 - renamed Changelog to ChangeLog in preparation for autoconf
957 - renamed Todo to TODO in preparation for autoconf
959 - make clean cleans up tests properly now
961 - turned source code browsing on in Doxyfile
962 - added LoadDatabaseByName() and LoadDatabaseByType() template members
963 to the Controller class
964 - new file: controllertmpl.h
965 - documentation fixes in builder.h
966 - added static GetDBName() functions to all record classes, in
967 support of the template Controller functions
968 - fleshed out DatabaseDatabase::GetDBNumber() and GetDBName(),
969 returning errors properly
970 - documented sample btool.cc code better, and added example
971 of Controller template member usage
973 - copyright dates updated for 2006
975 - backed out update to latest libusb DEVEL tree, since it is buggy.
976 Stick with libusb 2005/11/26
978 - added clean target to convenience Makefile
979 - added Boost to list of dependencies in README
980 - fixed minor compile warnings in base64.cc
981 - fixed new USB discovery code, and changed endpoint usage to use
982 it, instead of the READ_ENDPOINT and WRITE_ENDPOINT
983 constants... constants are now removed
984 common.h, controller.cc, probe.cc, probe.h, usbwrap.cc
985 - added ClearHalt() to Usb::Device (usbwrap.cc)
986 - updated usbwrap.h to match latest libusb CVS devel tree
987 - updated README to refer to 2005/12/30 libusb CVS devel tree
988 - added roadmap to Todo list
990 - added convenience Makefile in root directory
991 - added Hacking document
992 - minor touchups to README
993 - added caution warning to README, since this release will support
994 delete / overwrite / upload functionality
995 - minor compile order change in Makefile (speed reasons)
996 - added dynamic endpoint discovery to probe.cc
997 - changed claim_interface/set_configuration() function call order in
998 probe.cc based on feedback on the libusb mailing list
999 - added USB discovery code to usbwrap.{h,cc}
1001 - fixed packing bug in new upload header struct
1002 - renamed CommonField union from data to u to be consistent
1003 - added return_code field to DBAccess struct, for error checking
1004 - added response checking during upload: if response is non-zero
1005 throw exception in controller.cc
1006 - added sorting operators to record parser classes... when writing
1007 contact data to the device, it may be important to write
1008 group link items last, since they reference other contacts...
1009 This needs to be tested, and if not required, removed.
1010 - added GROUP_FLAG field code to Contact parser class
1011 - fixed offset bug in Contact and Calendar Build() functions
1012 - fixed name bug in Contact::Build(), now output only with data
1013 - added appointment type flag support to Calendar record parser
1014 class, and "all day event" flag support
1016 - added autoconf to Todo list
1017 - refactored the protocol structures so that record structs can
1018 be reused for upload and download
1019 - with new protostructs, more header control is handled by
1020 controller.cc instead of record.cc parsers and builders
1021 - parser code now accepts an offset, to govern start of record data
1022 - changed Data sizes from int to size_t and removed casts
1023 - removed unneeded Parser::GetHeaderSize() (new protostructs make
1025 - added Calendar::Build() for calendar uploads
1027 - added s11n-boost.h serialization templates for boost::serialization
1028 - added boost::serialization to build
1029 - minor change in usbwrap.cc to display error codes in decimal
1030 - added SizePacket special case code to Socket::Send()... this is
1031 required when uploaded packet size is a multiple of 0x40,
1032 a special 3 byte size packet is sent
1033 - moved some record class variables to public: so they can be
1035 - fixed variable name but in RecordBuilder template
1036 - added -f switch to btool, for saving and loading of record data
1037 - added -s switch to btool, for uploading saved data back to handheld
1038 - added Controller::SaveDatabase()
1040 - fixed potential casting / pointer bug in controller.cc
1041 - added builder.h, to support uploading
1042 - minor comment fixes in parser.h
1043 - added preliminary Contact protocol record building support
1044 - fixed field order bug in ContactRecord protocol struct
1045 - Contact record.h class now saves unknown Group Link field,
1048 - added -fno-strict-aliasing to compiler flags so gcc-4.0.x won't
1049 optimize away protocol struct casts
1051 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25235
1052 http://mail-index.netbsd.org/tech-kern/2003/08/11/0001.html
1053 http://groups.google.ca/group/comp.lang.c++.moderated/ \
1054 browse_thread/thread/e7bf096832526f8e/5714701b \
1055 02a2a3cc?hl=en#5714701b02a2a3cc
1056 - changed headers so that any low level protocol-specific sizes and
1057 structs are completely hidden in the library...
1058 this adds library-only header: protostructs.h
1060 Release: version 0.0.1 - 2005/11/25
1061 ------------------------------------------------------------------------------
1063 - added barry.h for documentation purposes and application usage
1064 - updated README for release
1065 - renamed SBError exception class to BError
1066 - updated Doxyfile to version 1.4.5
1067 - added more doxygen comments
1069 - added conversion routines (and tests) between time_t and min1900_t
1070 - added -v option to btool to dynamically control protocol dumping
1071 - removed duplicated database database table display in controller.cc
1072 - minor spelling corrections
1073 - added copyright to usbwrap.{h,cc} as per earlier meeting
1074 - added support for Old protocol Calendar entries
1076 - added legal.txt, to save typing when adding new source files
1077 - ignore test-base64 file (cvsignore)
1078 - added contact-to-ldif generation, and updated supporting code
1079 - added argument to btool to support contact-to-ldif operation
1080 - added Contact::GetPostalAddress()
1081 - added README documentation in preparation for future release
1083 - added parser.{h,cc} virtual wrapper to generalize LoadDatabase()
1084 - added code to socket.cc's Packet() command to re-receive on
1085 reception of empty packets... this seems to be a signal
1086 from the device that it needs more time for long transfers
1087 Current limit: 10 blank packets before exception is thrown
1088 Observed maximum blank count: 1
1089 - added MessageRecord and OldMessageRecord to protocol.h
1090 - changed common record field structs to one CommonField struct
1091 - increased default USB timeout to 10 seconds (some loaded devices
1092 seem to need more time...)
1093 - reorganized record.{h,cc} to try to reduce the code duplication
1094 - record.h classes now "know" the sizes of their corresponding
1095 protocol structs... this may cause trouble down the line
1096 with Python, if used(?)
1097 - added Message class for Email records
1098 - removed the "Get*" database functions and replaced with
1099 general LoadDatabase() call, which loads any database
1100 available, and parses it with a Parser object
1101 - added GetDBID() for searching for database numbers via name
1102 - added test code to retrieve and display all email
1103 - cleaned up comments
1104 - renamed blackberry.{h,cc} to controller.{h,cc}
1105 - renamed bbtool.cc to btool.cc
1106 - added Makefile.conf
1107 - added copyright notices to source files
1108 - added COPYING file with GPL
1109 - put parser in Barry namespace, and added auto_ptr-like mode for
1110 store object pointers
1111 - added doxygen docs
1112 - added header comments to convo.awk script
1113 - removed Test() function in Controller object, and made class
1114 better suited to its general API purpose
1115 - changed OpenMode() so it loads command table and DBDB... it is now
1116 a public API function
1117 - moved Controller test code to btool.cc and made tool more
1118 generically useful... can now download any DB by name on
1120 - added mode checks to Controller
1121 - added -t and -d command line options to btool
1122 - added base64 routines from John Walker / Fourmilab and Citadel/UX
1123 - converted base64 routines into an API, instead of hardcoded
1126 - changed project name to Barry, including namespaces
1127 - renamed sbcommon.* to common.*
1128 - increased USB default timeout to 2 seconds
1129 - fixed formatting bug in Data operator<< output
1130 - cleaned up the protocol packet structs, getting rid of the
1131 confusing param/simple split... param is now the tableCmd
1132 from the Command Table
1133 - moved protocol structs from record.h to protocol.h
1134 - added support for older version of the USB protocol, which matches
1135 the Cassis spec more closely
1136 - fixed the hardcoded offsets in blackberry.cc's protocol parsing
1137 - fixed Socket::Close() so on error, socket flags are reset.
1138 This stops Close() from being called twice on exception
1141 - added Doxyfile to the project
1143 - added mode selection support to Blackberry class (RIM Desktop,
1144 RIM Bypass, and RIM_JavaLoader)
1145 Mode selection governs which socket is used in subsequent
1147 - added Command Table parsing and protocol support
1148 Blackberry class, and CommandTable record class
1149 - added Database database parsing and protocol support
1150 Blackberry class and DatabaseDatabase record class
1151 - added Address book retrieval support - dumps data to stdout for now
1152 - fixed bug in CopyOnWrite mode in Data class, where could allocate
1153 less buffer than data copied
1154 - added "easy exception" debug output macro
1155 - fixed minor spelling error in protocol.cc
1156 - fixed ModeSelectCommand protocol structure - it is not a parameter
1157 command, but uses the unknown slots for socket and flag
1159 - removed hardcoded socket numbers from protocol.h
1160 - fixed bug in socket.cc:Open() where packet size was not properly
1162 - implemented socket::Close()
1163 - fixed sequence number checks, so it is reset properly on socket 0
1164 - added Socket::NextRecord() for multiple record data transfers
1165 - fixed input handling bug in translate.cc
1167 - added .cvsignore files to project
1168 - added README notes to document the source architecture
1169 - added dbsetup data script, which shows how to get into database
1170 mode, using the RIM Desktop mode, and open a socket
1171 - added blackberry.{h,cc} class
1172 - added command line parameters to bbtool.cc
1174 -p = specify pin of blackberry device
1175 - added blackberry test code in bbtool.cc - debugging the initial
1176 opening protocol: sockets, modes, etc... see dbsetup for
1178 - added copy-on-write functionality to Data class, so it can be used
1179 with plain data buffers with little performance impact
1180 - added eout() and ddout() debugging - to separate:
1181 - regular debug output
1182 - exception debug output (should probably be logged to
1183 syslog someday, and never turned off)
1184 - raw data dump output
1185 - changed errno SBError exception handler to try to decode the error
1186 numbers into english (limited success)
1187 - fixed probe and socket USB Bulk message handling, so they don't
1188 use PollCompletions(), but use IO::Wait() (libusb_io_wait())
1189 - added common packet size checker in protocol.cc
1190 - fixed unpacked protocol packet structs... using gcc's
1191 __attribute__ ((packed)) ... see protocol.h
1192 - added ModeSelectCommand structure (protocol.h)
1193 - finished Socket::Open() call (Close() not yet implemented)
1194 - moved Socket::Send/Receive() out of private, as they can be used
1195 for socket 0 operations.
1196 - cleaned up pointer management with macros:
1200 - added IO::Wait() and IO::Cancel() to correspond with libusb API
1201 - added datadumping in usbwrap for easy protocol capture
1202 - added preliminary record.{h,cc} classes
1203 - added Contact record parsing, and test code
1205 - syncberry tree started