Re: [AVaRICE-user] Debug trace the next

[Joerg W. <j...@ur...>]

As Andreas Løhre wrote:

> The full output is:

An interesting list, thanks!

> [jtag2rw.cc:180]: (error) Memory leak: response
> [jtag3rw.cc:182]: (error) Memory leak: response

Both these are actually the same (with the jtag3 code being cloned
from the jtag2 implementation).  All this looks confusing, and I have
to think about it again (there's a usage of "resp" versus "response"
inside these functions).

> [jtag2usb.cc:422]: (error) Mismatching allocation and deallocation:
> devnamecopy

Created as an array, but deleted as a non-array.  Fixed.

> [jtag2usb.cc:302]: (error) Memory leak: devnamecopy

Yes, in cas of an error; fixed.

> [jtag3io.cc:495]: (error) Array 'cmd[4]' accessed at index 4, which is out
> of bounds

As you've guessed, 0, 1, 2, 3 is the correct sequence.  Besides
accessing the array out of bounds, the bug lead to the request for
applying an external reset (throgh nSRST) not being obeyed at all.
Fixed.

I'm afraid the next step would be to strace both invocations, and see
what is done differently in the version as compiled by GCC 4.7.
-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)