[SOLVED] Bash script to count lines of code?

Rogue45 · 11-26-2012, 11:26 AM

I want to write a bash script to count the lines of code vs comments in a secondary bash script i have. Comments start with a # and can have spaces or tabs preceeding it. A simple grep -c returns the number of comments but doesn't give an accurate count of comment lines. How can i do this?

shivaa · 11-26-2012, 11:31 AM

Sounds some home work... ain't it?
Commands such as wc will tell you number count, plus vi editor will also tell you total no. of lines using :se nu option in it's colon mode.
However mention that what you've tried so far.

Rogue45 · 11-26-2012, 01:48 PM

Not homework.

I've successfully gotten my line counts but its pretty awful.

I used a sed to identify comment lines easily
sed -i 's/#.*/removed/' root/Desktop/test.txt

then i removed blank lines
sed -i '/^$/d' /root/Desktop/test.txt

then i grep -c removed ,to get comment line count

then total lines - comment lines = actual loc

Anyways that method sucks so please show me a prettier method.

steelneck · 11-26-2012, 04:16 PM

I do not think it qualifies as pretty, but i think this script will work for you. Just feed it with the file you want to count lines in as argument.

Code:

#!/bin/sh
noblank=$(sed '/^$/d' $1)
echo "Lines with code: $(echo "$noblank" |  grep -cv "#")"
echo "Lines with comments: $(echo "$noblank" |  grep -c "#")"

fakie_flip · 11-27-2012, 06:35 AM

Counting the lines that begin with #.
grep -c '^#' script.sh

Just using grep '#' will also match these lines. That line is not a comment is it?

Code:

[root@beastlinux ~]# grep '#' script 
echo "product #5"
[root@beastlinux ~]# grep '^#' script 
[root@beastlinux ~]#

I have the script I helped another user with on the forum.

Code:

[root@beastlinux ~]# cat recycle 
#!/bin/bash

recycleBin=/tmp/recycle_bin

function recycle_files {
  mv -v "$1" $recycleBin
}

case "$1" in
  -r) recycle_files $2 ;;
   *) echo "-r wasn't used" ;;
esac
[root@beastlinux ~]#

This greps only the code. Any lines that are blank, contain only white space, or begin with a comment are removed from the results leaving only the lines of code.

Code:

[root@beastlinux ~]# egrep -v '^$|^#|^[  ]$' recycle 
recycleBin=/tmp/recycle_bin
function recycle_files {
  mv -v "$1" $recycleBin
}
case "$1" in
  -r) recycle_files $2 ;;
   *) echo "-r wasn't used" ;;
esac
[root@beastlinux ~]#

The [ ] contains a space and a tab character. To insert the tab, push control + v, then tab.

So if you want to count those different things, just add -c to grep.

grail · 11-27-2012, 07:59 AM

I must say I am a little concerned by all the code that is changing files when all we are looking for is information. I refer to both the use of moving files around or using sed to directly
changes files. Hopefully not to be used in a live environment I guess.

Anyhoo, how about something simple like:

Code:

awk '{if($1 ~ /^#/)comment++; else no_comment++}END{print "line of code =",no_comment,"and number of comments =",comment}' file

fakie_flip · 11-27-2012, 04:59 PM

Quote:

Originally Posted by grail

I must say I am a little concerned by all the code that is changing files when all we are looking for is information.

I am not suggesting this code be used. I was just using it as an example for counting lines with comments and counting lines of actual code excluding comment lines, white space lines, and blank lines. The code was something I fixed for another user on the forum. See:

https://www.linuxquestions.org/quest...ts-4175438882/