Memos From the Cube

I’ve got a fearsome cold at the moment, which is threatening to turn into a chest infection, so I’m pretty much out of action at the moment. Add to that my huge pile of University work and their looming deadlines, and I’m resigned to the fact that I simply can’t finish the Bluetooth Host code for the next release. On the upside, that gives me a good 1-2 month window to get it all spit-and-polished for the mid-year release.

Today I’m releasing the official 100505 LUFA BETA, for public testing. This beta release will hopefully undergo a full two week testing by the community before the full release, to iron out any bugs. I’m sticking to the schedule this time, because LUFA is being used more and more widely as it becomes more well known, and with an increased user base comes more support required from me if I don’t get the official releases right — plus it also means a 1-2 month wait for an official patched version with any discovered bugs squashed.

So please download the new beta, try it out, and report back your findings (good or bad). This is very nearly 100% drop-in replacement for the previous version – the exception is the TWI driver – so LUFA users should have no excuse not to give the new code a whirl.

Download the beta release here:

http://lufa-lib.googlecode.com/files/LUFA-100505-BETA.zip

Download the prebuilt Doxygen documentation here:

http://fourwalledcubicle.com/files/MyUSB/Doc/LUFA 0100505 BETA Documentation.zip

As always, the full Changelog is appended to the end of this post. As always, a big thanks to everyone who contributed and donated over the past two months!

---

Changelog for 100505 BETA:

New:

• Added incomplete MIDIToneGenerator project
• Added new Relay Controller Board project (thanks to OBinou)
• Added board hardware driver support for the Teensy, USBTINY MKII, Benito and JM-DB-U2 lines of third party USB AVR boards
• Added new ATTR_NO_INIT variable attribute

Changed:

• AVRISP programmer project now has a more robust timeout system, allowing for an increase of the software USART speed for PDI and TPI programming
• Added a timeout value to the TWI_StartTransmission() function, within which the addressed device must respond
• Webserver project now uses the board LEDs to indicate the current IP configuration state
• Added ENABLE_TELNET_SERVER compile time option to the Webserver project to disable the TELNET server if desired
• Increased throughput of the USBtoSerial demo on systems that send multiple bytes per packet (thanks to Opendous Inc.)
• Double bank CDC endpoints in the XPLAIN Bridge project, re-enable JTAG once the mode selection pin has been sampled.
• Standardized the naming scheme given to configuration descriptor sub-elements in the Device mode demos, bootloaders and projects
• All Class Driver Host mode demos now correctly set the board LEDs to READY once the enumeration process has completed
• Added LIBUSB_FILTERDRV_COMPAT compile time option to the AVRISP programmer project to make the code compatible with Windows builds of avrdude at the expense of AVRStudio compatibility
• Removed two-step endpoint/pipe bank clear and switch sequence for smaller, faster endpoint/pipe code

Fixed:

• Fixed software PDI/TPI programming mode in the AVRISP project not correctly toggling just the clock pin
• Fixed TWI_StartTransmission() corrupting the contents of the GPIOR0 register
• Fixed TWI driver not aborting when faced with no response after attempting to address a device on the bus
• Fixed ADC routines not correctly returning the last result when multiple channels were read
• Fixed ADC routines failing to read the extended channels (Channels 8 to 13, Internal Temperature Sensor) on the U4 series USB AVR parts
• Fixed LowLevel MassStorage demo broken on the U2 series USB AVRs due to unsupported double-banked endpoint modes used
• Fixed compilation error in the AudioInput demos when MICROPHONE_BIASED_TO_HALF_RAIL is defined (thanks to C. Scott Ananian)
• Fixed incorrect definition of HID_ALIGN_DATA() causing incorrect HID report item data alignment
• Fixed Still Image Host class driver not resetting the transaction ID when a new session is opened, fixed driver not sending a valid session ID to the device
• Removed invalid dfu and flip related targets from the bootloaders – bootloaders can only be replaced with an external programmer
• Fixed Set/Clear Feature requests directed to a non-configured endpoint not returning a stall to the host
• Fixed HID Device Class Driver not allocating a temporary buffer when the host requests a report via the control endpoint and the user has set the PrevReportINBuffer driver configuration element to NULL (thanks to Lars Noschinski)