Convert 32 bit hex value into fields in decimal

UNIX for Advanced & Expert Users
1

I have 32 bit value in hex that I want to separate into fields and then convert the fields into decimal values.

Input file has 2 words of 32 bit hex values:

000001ac
ca85210e

Output both words separated into individual bit fields:

ca85210e: f1(31:9), f2(8:0)
f7c392ac:  f1(31:14), f2(13:6), f3(5:3), f4(2), f5(1), f6(0)

Output file should look like this with original hex value and fields in decimal:

ca85210e: 0, 428
f7c392ac: 253710, 74, 5, 1, 0, 0

Thank You,
morrbie

2

I think you have a cut and paste issue with your sample data. Going on what I believe you meant, this will read two lines from stdin and output the 'words' from each line using the indicated bit masking:

#!/usr/bin/env ksh


l=0
while read word
do
    if (( l < 1 ))
    then
        printf "%08x %d %d\n" 0x$word $(( 0x$word >> 9 )) $(( 0x$word & 0x1ffff ))
    else
        printf "%08x %d %d %d %d %d %d\n"  0x$word $(( 0x$word >> 14 )) $(( (0x$word >> 6)  & 0xff)) $(( (0x$word >> 3)  & 0x07)) $(( (0x$word >> 2)  & 0x01)) $(( (0x$word >> 1)  & 0x01)) $(( 0x$word & 0x01 ))
        break
    fi

    l=1
done

---------- Post updated at 21:07 ---------- Previous update was at 21:03 ----------

From your original post:

Shouldn't the two values in the file have been

000001ac
f7c392ac

given the output you listed, or the output should have been:

000001ac 0 428
ca85210e 207380 132 1 1 1 0

---------- Post updated at 21:16 ---------- Previous update was at 21:07 ----------

And one more note about the solution I posted. I originally read into word and assigned 0x to the variable so that I wouldn't need to prefix $word in each expression. This resulted in interesting behaviour in ksh -- bash seemed to get it right both ways. I need to test with a current build of ksh -- the most recent build I have access to here is 2011/02.

Code that has issues in Kshell:

#!/usr/bin/env ksh
l=0
while read word
do
    word="0x$word"
    if (( l < 1 ))
    then
        printf "%08x %d %d\n" $word $(( $word >> 9 )) $(( $word & 0x1ffff ))
    else
        printf "%08x %d %d %d %d %d %d\n"  $word $(( $word >> 14 )) $(( ($word >> 6)  & 0xff)) $(( ($word >> 3)  & 0x07)) $(( ($word >> 2)  & 0x01)) $(( ($word >> 1)  & 0x01)) $(( $word & 0x01 ))
        break
    fi

    l=1
done

It prints the first output line incorrectly, but the second is correct. Don't know if anybody wants to verify or offer an explanation, but it would be appreciated.

3

Yes, the output was posted incorrectly. Sorry about that. I must have been in a rush. Forgot to cut and paste the correct hex words in the output.
It should be:

000001ac 0 428
ca85210e 207380 132 1 1 1 0

Thank You for the reply. I will try it out.