I have started to play with and arduino UNO quite recently (without any prior experience with micro-controllers). I would like to use emacs instead of the IDE, and I'd also like to know what the IDE does under the hood, in order to be able to write my own makefile. The tutorials I've found are either outdated, or are presented as a series of steps without any explanation. I'd appreciate it if someone could explain to me how the whole compliation/linking/upload process works using gcc-avr and avr-dude, and how it is used by the IDE.
If you want an exact duplication of what the IDE does but want it driven from the command line, that's what Ino is for. The full Arduino build process involves copying a lot of files from a lot of places, and is generally not trivial to duplicate.
If you're ready to let go of .ino files and the Arduino libraries, you get a much simpler toolset. avr-gcc compiles, avrdude uploads, and you're done. Here's one of my makefiles from a simple project:
CC=avr-gcc CXX=avr-c++ CXXFLAGS=-Wall -Wextra -mmcu=atmega1284p -Os CFLAGS=$(CXXFLAGS) BINARY=ledmatrix OBJECTS= all: $(BINARY) ↹@avr-size $< $(BINARY): $(OBJECTS) clean: ↹@rm -f $(BINARY) $(BINARY).hex $(OBJECTS) upload: $(BINARY).hex ↹@avrdude -c usbasp -p m1284p -U flash:w:$<:i %.hex: % ↹@avr-objcopy -j .text -j .data -O ihex $< $@ .PHONY: all clean upload
If copying-and-pasting, be sure to replace all "↹" with tab characters.
I have created a repository with my buildsystem on Github.