Closed GoogleCodeExporter closed 9 years ago
Oops... forgot to fix UART.c
Original comment by stephan....@gmail.com
on 14 Nov 2010 at 2:08
Attachments:
applied and commited. Thanks! Please also consider joining us in the forum,
it's a lot more active there :)
How was the Linux compile? Do you like MPLABx?
Original comment by ianles...@gmail.com
on 14 Nov 2010 at 2:33
Thanks for the quick reply.
There was a problem with your commit: eetext.s should have been renamed to
eetext.S (capital S).
Apparently there's a bug in SVN because renamed files don't show up in the
patch.
Also compiling the BPv4 variant does not currently work. I'll have to figure
out where to get the USB header files...
Original comment by stephan....@gmail.com
on 14 Nov 2010 at 4:11
Hi Stephan,
I've fixed eetext.s into eetext.S
Your makefile works against the current SVN.
But there is also a new Makefile made automatically from my MPLABX conversion,
in busPirate.X/busPirate which works too.
What I don't understand is the difference in size between your .hex and mine:
main.hex: 154804
busPirate.production.hex: 205080
Original comment by yob...@gmail.com
on 16 Nov 2010 at 4:51
[deleted comment]
Hi,
Thanks for fixing my Makefile! As for the file size, I'm using "-Os" to
optimize for size. It's strange that the could should be 25% smaller, but I
checked that all modules are included.
For BP_MAIN I get a main.hex of 176792 bytes
For BP_ADDONS it's 154804 bytes
If I use "-Os" in MPLABX, I get a hex size of 147992 and 169980, respectively
Original comment by stephan....@gmail.com
on 16 Nov 2010 at 6:19
Hum there is definitively something to explore here...
I could make a firmware containing both MAIN & ADDONS module (except I2C_HW for
which compil fails ATM) and the device seems working fine!! (I tried self-test
& UART sniffing)
Original comment by yob...@gmail.com
on 16 Nov 2010 at 7:37
I tried to compare your Makefile with MPLABX one and I see the following diffs:
Yours:
space optimization: -Os
ee.c & eetext.S
MPLABX's:
via elf format: -omf=elf
def intern symbols, probably not needed:
-Wa,--defsym=__MPLAB_BUILD=1
-Wl,--defsym=__MPLAB_BUILD=1
ee?? is it a kind of easter egg?? And people say the firmware is full, pfff :-)
So I think all diffs you saw come from the optimization flag and this ee
stuff...
What do you suggest for both Makefiles?
Can we add properly the -Os flag in the IDE so we can keep the sole MPLABX
Makefile?
It's a bit less neat than your but, well, it's there anyway and managed by the
IDE...
Original comment by yob...@gmail.com
on 16 Nov 2010 at 10:24
Would it be possible to move this to the development forum:
http://dangerousprototypes.com/forum/index.php?board=28.0
Can I also give you SVN access so you can commit these patches as needed?
Original comment by ianles...@gmail.com
on 17 Nov 2010 at 10:25
Done, see
http://dangerousprototypes.com/forum/index.php?topic=1450.msg13354#msg13354 for
Makefile discussion
Original comment by yob...@gmail.com
on 17 Nov 2010 at 11:03
Original issue reported on code.google.com by
stephan....@gmail.com
on 14 Nov 2010 at 2:00Attachments: