Re: Bug in od

[Bob Proulx]

Keith P. Boruff wrote:
> I discovered an 'endian' bug when executing od in the following manner: 
> od -Ax -t x2 baseball.txt on my Linux O/S (with an Intel processor).

Thanks for the report.  But I looked at the files but I could not see
a bug there.  The data looked fine on each machine, big and little,
endian.  What output were you expecting?

> The dump for this command is in the file 'little_endian.txt'.
>
> Seeing that the data layout of this execution was corrupted due to the 
> endian problem, I also ran the same command (same args) on a Sun 
> workstation. The results are on big_endian.txt

Corrupted how?  Please be specific.

> Also included is the file 'baseball.txt' for which od processed.

Let's try something simple.

  echo abcd | od -tx1
  0000000 61 62 63 64 0a

But you were using -tx2 which means to use two bytes.  Which means
that it will be different on different machines.  The bytes might be
in "backwords" storage order.

On an ascii little endian machine:

  echo abcd | od -tx1
  0000000 61 62 63 64 0a

  echo abcd | od -tx2
  0000000 6261 6463 000a

  echo abcd | od -tx4
  0000000 64636261 0000000a

On an ascii big endian machine:

  echo abcd | od -tx1
  0000000 61 62 63 64 0a

  echo abcd | od -tx2
  0000000 6162 6364 0a00

  echo abcd | od -tx4
  0000000 61626364 0a000000

Please report back to the mailing list how this resolves for you.

Thanks
Bob