Looping through find output in Bash where file name contains white spaces

I try to search for files that may contain white spaces I try to use -print0 and set IFS here is my script

IFS=$'\0';find people -name '*.svg' -print0 | while read file; do
    grep '<image' $file > /dev/null && echo $file | tee -a embeded_images.txt;
done

I try to fine all svg file containing embeded images, it work without -print0 but fail one one file so I stop the script. Here is simpler example that don't work too

IFS=$'\0';find . -print0 | while read file; do echo $file; done

it don't display anything.


Solution 1:

Though Dennis Williamson's answer is absolutely correct, it creates a subshell, which will prevent you from setting any variables inside the loop. You may consider using process substitution, as so:

while IFS= read -d '' -r file; do
    grep '<image' "$file" > /dev/null && echo "$file" | tee -a embeded_images.txt
done < <(find people -name '*.svg' -print0)

The first < indicates that you're reading from a file, and the <(find...) is replaced by a filename (usually a handle to a pipe) that returns the output from find directly. Because while reads from a file instead of a pipe, your loop can set variables that are accessible from outside the scope.

Solution 2:

Use read -d '' -r file and set IFS only for the context of read:

find people -name '*.svg' -print0 | while IFS= read -d '' -r file; do
    grep '<image' "$file" > /dev/null && echo "$file" | tee -a embeded_images.txt;
done

And quote your variables.