shell scripting: "formatting" issue with text files
 

Hello!

I'm trying to replace a value in a file, and to do this, I'm reading in line-by-line. If the line doesn't match I copy it over to a temporary file, and if it does match, I make my change and copy that over to the temp file. This works great, but I'm losing white spaces in the line (tabs and indents only). So it's basically pushing everything along the left edge, which is horrid. How can I go about fixing this? I tried using "sed" to change the value, but it doesn't work well when you include variables. Here's the current code that I'm using just in case someone is interested.


while read line
do
if echo $line | grep -q "^default.*" ; then
echo "default=${myval}" >> /tmp/lilo-copy.conf
else
echo $line >> /tmp/lilo-copy.conf
fi
done < /etc/lilo.conf
mv /tmp/lilo-copy.conf /etc/lilo.conf


original sed command I used:
sed -e "/^default/ s/^default=.*/default=$myval/" /etc/lilo.conf

Hi,

You don't need the first part of the sed command:

sed "s/^default=.*/default=$myval/" /etc/lilo.conf

Should replace all instances of default=.*, at the beginning of the line with default= and the content of $myval.

If you have sed 4.x (sed --version) you can use the -i flag to instantly edit the infile (man sed for details).

Hope this helps.

You probably loose that whitespace because of
Replace it with
in both instances, as echo will interpret $line as a collection of separate arguments instead of a single argument that contains space characters.
But then, the sed command should work fine unless $myval contains "/". Note that you may use any other character for that purpose in sed, as in

Thanks for the help guys! I ended up using your sed command (so much prettier).