UNIX for Advanced & Expert Users

hi

I have a ubuntu machine and want to know the endianness of the system....
How would i get to know.....

The information of my machine is:

Code:

uname -a

42-Ubuntu SMP Mon Apr 11 03:31:50 UTC 201
1 i686 i686 i386 GNU/Linux

and

Code:

lsb_release -a

No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 11.04
Release: 11.04
Codename: natty


I have done it using C Program......but dont want to know using C Program.......
Want some linux command to know endianness of a machine

Plzz Help...
Thanks in advance....

You are on an x86 machine, which is little-endian by definition.

If you have perl:

Code:

perl -MConfig -e 'print "$Config{byteorder}\n";'

Unfortunately I don't have a big-endian machine to test it with anymore, one of the last holdouts got replaced by a more modern mac...

Google a lil' bit and found this thread.
Modified the suggestion a little bit:

Code:

echo I | tr -d [:space:] | od -to2 | awk 'FNR==1{ print substr($2,6,1)}'

'0' for Big.
'1' for Little.

Interesting suggestion, it appears to work here.

I suggest putting single-quotes around '[:space:]' in case you have a weird-named file or folder in your directory(in my case, c) -- your shell might glob it like mine did!

I agree. Interesting. However, the technique could be further improved and simplified:

Code:

printf '\1' | od -dAn

Little Endian: 1
Big Endian: 256

A simple, [ 1 -eq ... ] integer comparison can be used to test.

printf can portably avoid printing newlines, making tr unnecessary. It can also directly specify a byte value, removing the dependency between a character and its encoding.

To simplify the output, the input offset is suppressed with -An.


The linked thread mentions that the approach fails on some systems, but the behavior on which it depends has been in the POSIX od manual for at least 15 years:

In my opinion, it's reasonable to expect it to work with modern systems (even when the definition of "modern" is loosely applied)

Regards,
Alister

We still some some Sparc systems, so I gave it a try...

First on Linux:

Code:

$ uname -is
Linux x86_64
$ perl -MConfig -e 'print "$Config{byteorder}\n";'
12345678
$ echo I | tr -d [:space:] | od -to2 | awk 'FNR==1{ print substr($2,6,1)}'
1
$ printf '\1' | od -dAn
     1
$

Now on Solaris:

Code:

$
$ uname -is
SunOS SUNW,Sun-Blade-1000
$ perl -MConfig -e 'print "$Config{byteorder}\n";'
87654321
$
$ echo I | tr -d [:space:] | od -to2 | awk 'FNR==1{ print substr($2,6,1)}'
$
$
$ echo I | tr -d [:space:] | od -to2 | nawk 'FNR==1{ print substr($2,6,1)}'
0
$ printf '\1' | od -dAn
         00256

$

I was surprised at first that the second one gave me trouble. Then I remembered that famous Solaris awk vs nawk thing and got it to work. My favorite is the perl line.