Open GoogleCodeExporter opened 9 years ago
Problem: The heap is not aligned to 4-byte boundary. heap_init() detects this
and exits with an error code.
Try: The following fix did not work, the heap was still only 2-byte aligned.
int main(void)
{
uint8_t heap[HEAP_SIZE] __attribute__((aligned(4)));
...
Original comment by dwhall...@gmail.com
on 16 Mar 2011 at 11:33
- "Fixed" heap alignment by easing the alignment based on architecture bit
size. Now, arch's with a 16-bit pointer width only need to align to 2-byte
boundaries.
- Also had to reduce the heap size for the Arduino Mega platform; I guess
avr-gcc 4.3.3 produces less-compact code.
- Added "VARIANT=2560" option to Makefile to allow build for Arduino Mega 2560
Original comment by dwhall...@gmail.com
on 18 Mar 2011 at 8:47
This issue was closed by revision 46c0e3df818b.
Original comment by dwhall...@gmail.com
on 18 Mar 2011 at 8:57
It still is not working. I just cloned from svn and still having the same issue
with Mega 2560
I can compile and upload the program but I can't interact with it using ipm. No
error, no exceptions returned but only a delay then the prompt back.
Original comment by pen...@gmail.com
on 12 Aug 2011 at 9:37
I have heard the same thing from others, but forgot to do this paperwork.
Thank you for the reminder; I've reopened this issue. I still don't have an
Arduino Mega 2560, so I can't fix this one myself. My 'Mega is the 1280.
Original comment by dwhall...@gmail.com
on 12 Aug 2011 at 11:57
I just tried to compile with avr-gcc 3.4.6 without the VARIANT=2560 and upload
the program using make program with VARIANT=2560 now ipm works for list, dict
and print but doesn't work for anything involved with integers, so if I do a =
[1] print a will not print stuff in the list but empty
if I do a= ['1'] print a however it will print the list ['1'] as it is. This is
weird consider that the program wasn't compiled for 2560 but it works on 2560
and ipm got the response I guess your makefile should be corrected for 2560.
You should definitely buy one to test it out. Comeon, it doesn't cost too much
for one right?
Original comment by pen...@gmail.com
on 12 Aug 2011 at 1:48
It is definitely an issue with the new compiler, right now if I compile with
avr-gcc 4.3.3 without VARIANT=2560, ipm would not return anything as before on
the 2560
Original comment by pen...@gmail.com
on 13 Aug 2011 at 8:19
Hi guys, I have one of these and am getting the same response (or no response
as the case may be) from ipm. If anybody would like me to test a patch for them
just let me know.
Original comment by carl.nob...@gmail.com
on 14 Aug 2011 at 1:07
Original issue reported on code.google.com by
dwhall...@gmail.com
on 16 Mar 2011 at 8:57