[SOLVED] Integer limit??

[jesushero]

Ok, this is a pretty stupid question... But why is the maximum integer value that bash can handle 9223372036854775807? Is that signed or not?

What happens in cases of floating point numbers?

Is there a way around this in bash?

When it wraps, in 9223372036854775807 + 1, it goes to -9223372036854775808.. How is this possible and 9223372036854775808 is not?


user@machine:~$ a=9223372036854775807
user@machine:~$ echo $a
9223372036854775807
user@machine:~$ let a++
user@machine:~$ echo $a
-9223372036854775808
user@machine:~$ let a++
user@machine:~$ echo $a
-9223372036854775807

[ghostdog74]

it depends on your machine. bash performs maths in the largest integer size your machine supports (intmax_t).

[unutbu]

According to http://en.wikipedia.org/wiki/Integer_(computer_science)
a 64-bit signed long int has the range you describe.
(See also /usr/include/limits.h). It seems likely based on this that bash implements integers as 64-bit signed long ints.

You can see by looking at the Two's complement chart (http://en.wikipedia.org/wiki/Signed_number_representations) that the Two's complement interpretation of

0111...111 is 9223372036854775807, and
1000...000 is -9223372036854775808

[mssever]

Is there a way around this in bash?
Store numbers as strings and use bc for math:
~:$ a=9223372036854775807
~:$ ((b = $a + 1))
~:$ echo $b
-9223372036854775808
~:$ echo $a + 1 | bc
9223372036854775808(Bash can't handle floating point numbers at all, so you have to use something like bc for that.

[slavik]

I would say that if you are using a two digit number in a bash script, you need to rethink your design (unless this is for educational purpose, then there are exceptions).

[mssever]

I would say that if you are using a two digit number in a bash script, you need to rethink your design (unless this is for educational purpose, then there are exceptions).
Wait. A two digit number is a sign of bad design? Huh?

[jesushero]

I would say that if you are using a two digit number in a bash script, you need to rethink your design (unless this is for educational purpose, then there are exceptions).

Would you recommend Python or Perl for scripting then? Or something else? I use bash if I need to use a lot of system commands and little else.

[ghostdog74]

what other things do you want to do , beside the calculation?


awk 'BEGIN{printf "%.2f", 9223372036854775807 + 1}'

[mssever]

Would you recommend Python or Perl for scripting then? Or something else? I use bash if I need to use a lot of system commands and little else.
See this thread: http://ubuntuforums.org/showthread.php?t=910779

[slavik]

Wait. A two digit number is a sign of bad design? Huh?
You're changing my words and misunderstanding the point. Please re-read my post.

[pmasiar]

You're changing my words and misunderstanding the point. Please re-read my post.

I did read it again and I am no wiser. How are two digits related to anything? Some background info maybe?

BTW, OP: see wiki in my sig for links to learn Python (excellent as replacement of bash, with unlimited size integers when you need them), including training tasks.

[slavik]

I did read it again and I am no wiser. How are two digits related to anything? Some background info maybe?

BTW, OP: see wiki in my sig for links to learn Python (excellent as replacement of bash, with unlimited size integers when you need them), including training tasks.
There's a reason the basic datatype in a shell is a string. Shells aren't intended to perform mathematical calculations even if they can.

[ghostdog74]

There's a reason the basic datatype in a shell is a string. Shells aren't intended to perform mathematical calculations even if they can.

some shells, like zsh, or ksh99, can perform floating point, eg zsh


# echo $(( 1.0 / 2.0 ))
0.5

otherwise, standard utility for the shell is to use bc, or dc.