UNIX - formatting another shell script - using awk in sed

Question

Hi I would like to write some program to format random another script-code. You have a testdata.sh code in a form like:

#!/bin/sh
# usage: fsplit file1 file2
total=0; lost=0
while
    read next
do
total=`expr $total + 1`
for i in *; do
if test -d $dir/$i
then
cd $dir/$i
while echo "$i:" ; read x ; do eval $x ; done
cd ..
fi
done
case "$next" in
*[A-Za-z]*)  echo "$next" >> $1 ;;
*[0-9]*)     echo "$next" >> $2 ;; *)           lost=`expr $lost + 1`
esac
done ; echo "$total lines read, $lost thrown away"

You call your script "format_it.sh" ...

 sh format_it.sh -t < testdata.sh

which mean to use tabulators for formatting. Otherwise you can use

 sh format_it.sh -s5 < testdata.sh

which means f.e. to use 5spaces for formatting.

What I want to do, is to format it in this way. Always after while/for/case/if.

#!/bin/sh
# usage: fsplit file1 file2
total=0; lost=0
     while
          read next
          do 
          total=`expr $total + 1`
               for i in *; do
                    if test -d $dir/$i
                         then
                         cd $dir/$i
                              while echo "$i:" ; read x ; do eval $x ; done
                         cd ..
                    fi
               done
           case "$next" in 
                *[A-Za-z]*)  echo "$next" >> $1 ;;
                *[0-9]*)     echo "$next" >> $2 ;; 
                *)           lost=`expr $lost + 1`
           esac
     done ;
 echo "$total lines read, $lost thrown away"

Till now my script looks like that..

#!/bin/sh

function use_t(){
layer=0

while read a; do
     if [ -n "$(echo $a | egrep "if|case|while|for")" ]; then
          layer=$((layer+1))
          sed -r  "s#(if|case|while|for)#\n awk "BEGIN \{ ORS=\"\"; for (i=0;i<=$layer;i++)    
          print \"\t\" \}" \1 #g"

          elif [ -n "$(echo $a | egrep "fi|esac|done")" ]; then layer=$(layer-1)) 
      fi
 done
}


if [ "$1" = "-t" ] ; then
use_t
elif [ -n "$(echo "$1" | grep  "^\-s[1-9][0-9]*$")" ] ; then
n_spaces=$(echo $1| sed 's/'-s'//' )
#use_s 
else
   echo "ERROR -  wrong input of parameters"
fi

The problem line is :

sed -r  "s#(if|case|while|for)#\n awk "BEGIN \{ ORS=\"\"; for (i=0;i<=$layer;i++)    
print \"\t\" \}" \1 #g"

What I want this line do, is to for every if/case/while/for in a text replace it with a /n {to get it to new line) and then use "layer times" tab/space to have it nice formatted. I dont know how to make this line with sed/awk work. Advices?

Maybe you see another way to do it as well, I would be glad for some tips.

petrbel · Accepted Answer · 2014-06-30 17:40:06Z

2

The only information you need to store is depth of current block of code.

Choose some character and add it to sed 'memory' when you enter new block (e.g. begin|do|...) and remove one when you exit block (e.g. end|done|...). At the beginning of each line add spaces according to size of the memory - you can achieve it by moving characters one by one to the second memory.

But I recommend using either awk or bash as they are much more appropriate.

answered Jun 30, 2014 at 17:40

petrbel

2,5585 gold badges31 silver badges52 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ed Morton · Accepted Answer · 2014-05-21 20:45:54Z

1

You need a parser that understands the language to do this, you can't do it reliably with a simple script. Google "shell script beautifier" and you'll find a few tools that claim to do it. MAYBE the C beautifiers "indent" or "cb" would even work for a shell script but I doubt it.

answered May 21, 2014 at 20:45

Ed Morton

209k18 gold badges90 silver badges212 bronze badges

Collectives™ on Stack Overflow

UNIX - formatting another shell script - using awk in sed

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related