I now have a post-build step which formats all the code using
astyle -n -z2 -Z -A14 *.hpp *.cpp
The options are to preserve the dates of each file, force the
Linux line endings, and format the files directly omitting the
creation of the .orig file. The -A14 is "Google" style which
happens to be the style I've always personally used, which is
handy.
The makebin.sh script takes all the .wav files in a directory,
resamples them to 8kHz, 8-bit unsigned integer RAW files. It then
concatenates these all together, while creating a text listing of
how long each clip is and the start position for that clip. It also
spits out some useful C code that does require some modification
before being pasted into the code, but makes things much easier
for fast prototyping.
The programFlash.py script connects to the chip over UART, sends it
64 bytes at a time to program into the flash chip, and repeats this
until all the data has been transferred. This mechanism does seem
to work, the only part that needs to be tested still is the SPI
connection to the flash chip.
Interrupts finally solved, bloody confusing literature.
All the pins *should* be set up correctly now. Next step is to
test SPI (and possibly split SPI out into its own class). The DAC
update function may eventually be put in a DAC class if there are
other related functions that need to be associated.
Haven't tested on-chip yet, but I think things look correct,
so I will wire it up and begin testing soon.
Eventually this will be merged back into master.
A.M. Rowsell