(significant size impact: 14 bytes!)
Initialized "address" to eliminate compiler warning (4 bytes!)
Add "atmega168" as a more accurate target name than "diecimila"
(keep diecimila as well for backward compatibility)
Reduce the .hex and .lst targets that are stored in source control
to the three basics: atmega8, atmega168, atmega328. The other
targets remain in the makefile and makeall, but will need to be
built from source if wanted. Which should be less of a problem
now that the source is buildable without installing crosspack.
Optiboot does not support ArduinoasISP programmer.
When avrdude runs and talks to an arduino running ArduinoISP,
it needs the optiboot (entered due to auto-reset) to abort and
start the ArduinoISP "application" when it sees communications
at the wrong serial speed. Unfortunately, optiboot treats all
unrecognized command characters as "no-ops" and responds/loops
for more commands, leading to a nice loop that never gets to
the sketch. This patch causes characters received with Framing
errors (the most likely error for speed mis-matches) to NOT
reset the watchdog timer (normally done in getch()), which will
cause the application to start if it continues for "a while."
(tested. Works! Running ArduinoISP at speeds as high as 57600
still causes the bootloader to start the sketch (although it fails
later on for other reasons.))
http://code.google.com/p/arduino/issues/detail?id=554
end of flash memory where they can be read (at least in theory) by
device programmers, hex-file examination, or application programs.
This is done by putting the version number in a separate section
(".version"), and using linker/objcopy magic to locate that section as
appropriate for the target chip. (See
http://lists.gnu.org/archive/html/avr-gcc-list/2011-02/msg00016.html
for some discussion on the details.)
Start the version at 4.1 (the last "packaged" version of optiboot was
called version 3, so the "top of source" would be 4.0, and adding the
version number makes 4.1)
Refactor LDSECTION in the Makefile to LDSECTIONS so that multiple
section start addresses can be defined.
Change the _isp makefile definitions to make the bootloader section
readable (but not writable) by the application section. (This would
need to be done elsewhere as well to handle all bootloader programming
techniques. Notably Arduino's boards.txt
Note that this change does not change the "code" portion of optiboot
at all. The only diffs in the .hex files are the added version word
at the end of flash memory.
Allows building within the Arduino Source tree, and within the Arduino
IDE tree, as well as using CrossPack on Mac.
Adds README.TXT to track arduino-specific changes (and documents the
new build options.)
This addresses Arduino issue:
http://code.google.com/p/arduino/issues/detail?id=487
And optiboot issue
http://code.google.com/p/optiboot/issues/detail?id=1
(which can be thought of as a subset of the Arduno issue.)
Note that the binaries produced after these Makefile changes (using any
of the compile environments) are identical to those produced by the
crosspack-20100115 environment on a Mac.
Because UBBR is only 12 bits, we were overflowing it at 300 baud because of the use of the U2X bit. Now we turn off U2X if it would yield a UBBR value that would overflow.
Note that this breaks 300 baud communication with the computer on the Uno and Mega 2560 because the 8U2 USB-serial firmware has this same bug (and previously they cancelled each other out). Since, however, it seems more likely that people will need to use 300 baud to communicate with other (legacy) hardware than with the computer, I'm making this change. Issue for 8U2 firmware bug: http://code.google.com/p/arduino/issues/detail?id=542http://code.google.com/p/arduino/issues/detail?id=522
Because UBBR is only 12 bits, we were overflowing it at 300 baud because of the use of the U2X bit. Now we turn off U2X if it would yield a UBBR value that would overflow.
Note that this breaks 300 baud communication with the computer on the Uno and Mega 2560 because the 8U2 USB-serial firmware has this same bug (and previously they cancelled each other out). Since, however, it seems more likely that people will need to use 300 baud to communicate with other (legacy) hardware than with the computer, I'm making this change. Issue for 8U2 firmware bug: http://code.google.com/p/arduino/issues/detail?id=542http://code.google.com/p/arduino/issues/detail?id=522
This makes explicit the String constructors that take numeric types and chars and removes the versions of concat() and operator=() and operator+() that accept numberic types.
It also replaces the operator bool() with a operator that converts to a function pointer. This allows for uses like "if (s)" but not "s + 123". See: http://www.artima.com/cppsource/safebool.html. This allowed removing the disambiguating operator+() functions and relying solely on StringSumHelper and anonymous temporaries once again.
Also, now treating unsigned char's like int when constructing Strings from them, i.e. String(byte(65)) is now "65" not "A". This is consistent with the new behavior of Serial.print(byte).
The operator bool() means that you could implicitly convert a String to a bool and then add it to it an int, for example. Which means our operator+ has to match exactly or it will be ambiguous.
Changing toLowerCase(), toUpperCase(), trim() and replace() to return void instead of a reference to the string that's just been changed. That way, it's clear that the functions modify the string they've been called on.
Now Serial.write() places characters in the transmit buffer, and the data register empty interrupt reads and transmit them. Based loosely on the implementation here: ftp://wookey.org.uk/arduino.
http://code.google.com/p/arduino/issues/detail?id=262
That is, there's now a pins/ directory in a platform, which includes multiple directories, each of which has its own pins_arduino.h. The boards.txt gets a new preferences, <BOARD>.build.pins, whose values is a sub-directory of the pins/ directory (possibly with a "platform:" prefix). That sub-directory is then placed in the include path during compilation.
Also, removing the inline version of digitalPinToTimer() (since we're not optimizing the functions that use it anyway). The noAnalogWrite() function is in wiring_analog.c, deriving from the previous turnOffPWM() which has moved from wiring_digital.c.
http://code.google.com/p/arduino/issues/detail?id=476
Factoring out the implementation of digitalWrite(), digitalRead(), and pinMode() into macros that can either be inlined (for constant pin numbers) or executed within a function (non-constant pins). Removing testing for timers on pins in digitalWrite(), digitalRead(), and pinMode(). Moving pin to port macros from pins_arduino.h to wiring.h.
Enabling INTERRUPT_CONTROL_ENDPOINT in the makefile so we don't miss control messages from Linux. Also deleted two unused USB events to save a few bytes of space (the Mega 2560 .hex is now exactly 4096 bytes).
- moving back to multple cores per platform
- using target instead of platform
- moving per-board and per-programmer preferences out of Preferences.java and into a new Target class
- adding a new "target" preference
- support for platform:value values in board preferences for bootloader path and core
- XXX: need to support platform:value syntax for board upload.using preferences.
