- nfc_device is now an opaque type;
- PN53x specific errors are not public anymore;
- nfc_device_name() renamed to nfc_device_get_name() for the sake of consistency;
- examples/*, utils/* uses the new nfc_device_get_name() function instead of access directly to struct's content;
- new error defined: NFC_ERFTRANS for notifying about RF transmission error, its used by mifare.c to detect permissions error on mifare;
- drivers initiator_transceive_bytes() function now returns libnfc's error code on failure (<0), and received bytes count on success (>=0);
- remove some unused errors.
- Add timeout on InJumpForDEP (pn53x) and initiator_select_dep (libnfc API)
- test_dep now works again (except 424Kbps)
- Fix ndi.ndm feeling when select a DEP target
Important: test_dep does not work on PN53x_USB devices after running previous tests
- Provide a nfc_get_default_device() that allow to grab the connstring stored
in LIBNFC_DEFAULT_DEVICE environnement variable or returns the first
available device if not set;
- nfc_connect(NULL) now takes the default device (see
nfc_get_default_device());
- Removes nfc_driver_desc_t from public types
- Defines nfc_connstring as char[1024]
- examples/*: use nfc_connstring
- examples/nfc-poll: now uses only the default device (instead of all devices
availables)
- Removes parse_args() from nfc-utils.[hc]
* Removes internal DBG/WARN/ERR macros (but keep them in examples);
* Removes almost all DEBUG define references;
* Uses the flexible log4c library to log all messages.
The probing functions assume that at least one device should be found and
passed to the caller, and only after finding such a device they will check that
they should eventually stop looking for more devices. While we could make each
probing function capable of handling the special-case of finding 0 devices, it
is much more consistent to perform the check from the calling function.
Initialisation of RC-S360 in mode 0
Better generic nfc-list, does not rely only on DeSelect to avoid duplicates
Restore SetParameters for RC-S360
Handle Diagnose communication test properly
Tested:
- nfc-list => ok except for B' & 3B
- nfc-anticol => fails to send raw frames before select
- nfc-mfultralight => ok
- nfc-mfclassic => ok
- pn53x-tamashell => ok
- pn53x-diagnose => ok
- nfc-dep-initiator => ok
- nfc-dep-target => fails in InitAsTarget
- nfc-emulate-* => fails in InitAsTarget
Commands momentarily disabled for RC-S330:
- InDeselect
- InRelease
- select_passive_target for B'& 3B
NDO_AUTO_ISO14443_4 is now set to _true_ by default with nfc_target_init();
nfc_emulate_target() now fails (return _false_) when emulator->state_machine->io() failed;
Initialize nfc_device_t variables to _false_ in order to send correctly (see notes in nfc-device.c);
nfc_emulate_forum_tag4 prints error when nfc_emulate_target() failed.
nfc_disconnect() now switch NFC device into idle before disconnecting;
pn53x_data now have a operating_mode enum to know the current running mode (initiator, target or idle);
new pn53x_PowerDown wrapper for PowerDown (PN532) command;
Note: ARYGON idle mode now to be implemented and ACR122 does not support this mode.
iAbortFd file descriptor array have been removed from nfc_device_t;
nfc_abort_command() can now failed (return false);
nfc_abort_command() now call abort_command pointer from drivers;
pn532_uart and arygon drivers use a pipe-based mecanism (similar from previous one);
pn53x_usb driver use a boolean flag-based mecanism (the previous one does not work as expected);
pn53x_usb now print smarter messages on error at usb connection;
pn53x_usb now handle a strange case: sometimes, the first sent command is not ACKed by PN53x USB device, a dummy command is now sent.
- Move CMake modules from cmake_modules/ to cmake/modules/
- CMake now use cmake/config_windows.h.cmake to create config.h on Windows platform
- contrib/windows.h header is automagically included by config.h
- Put missing NFC_EXPORT macro on front of emulation API
- nfc-mfclassic and nfc-mfcultralight examples are now compiled under Windows
Many thanks to Glenn Ergeerts which provide the initial patch.
Suppress any PN53x references in nfc.c
Improve nfc_driver_t struct to embedded HAL API
Merge macros from nfc-messages.h into nfc-internal.h
Remove useless files: nfc-messages.h, buses.h and chips.h
Implement HAL for ARYGON driver
Move send/receive callbacks from nfc_driver_t to internal chip io callbacks (since there are dedicated to pn53x framing)
use a new way to handle drivers
use absolute include path instead of relative ones
move some nfc_device_t members in a better place
nfc_device_t now embeddeds driver data and chip data pointers (useful to be more generic)
use more readable variables instead of strange coding convention
move PRINT_HEX macro into nfc-internal.h
silent warnings with more strict CFLAGS
chips/pn53x: use the powerful C99 writing to construct PN53x commands
chips/pn53x: remove almost all memcpy()
chips/pn53x: WriteRegister, ReadRegister and SetParameters command wrappers are correctly named
chips/pn53x: introduce chip state (SLEEP, NORMAL or EXECUTE)
chips/pn53x: add SAMConfiguration command wrapper (need to be improved)
chips/pn53x: remove almost all const arrays
chips/pn53x: use human readable defines for commands instead of hex values
chips/pn53x: in debug mode, the PN53x command is shown in human-readable string, awesome isn't it? ;-)
drivers: split transceive() into send() and receive() to be able to handle more cases (differed replies, abort commands, etc) later
drivers: use a const structure of functions instead of -dirty- callbacks array
drivers/pn532_uart: major improvement of UART handling
drivers/pn532_uart: check PN53x frames when received
buses/uart: receive() is now based on expected bytes instead of calculated timeouts..
buses/uart: use a smart way to determine available ports on POSIX systems (tested on Linux and FreeBSD)
Always use nfc_initiator_transceive_bytes(). If you where using advanced
features and already relying on nfc_initiator_transceive_bytes(), then your
code has to be updated to unset the NDO_EASY_FRAMING option. See an example of
such a change in the libfreefare's repository:
http://code.google.com/p/nfc-tools/source/detail?r=566
Updates issue 106
Status: Feedback
Romuald: I am not sure about the option enum values. I took 0x02 thinking it
would not hurt but am not really sure about that because I can see many 'holes'
in the sequence.
