Shell script recursion doesnt work - bash

Question

I have a shell script to count all my files and directories sizes using recursive function
Here's my code:
#!/bin/bash
count() {
local file
total=$2
files=(`ls $1`)
for file in "$files"
do
if [ -d "$1/$file" ]
then
count "$1/$file" $total
else
#size=`du $file | grep -o [0-9]*`
#total=$(($2 + $size))
echo "$1/$file"
fi
done
}
total=0
count . $total
echo "$total"
I have error somewhere it just goes into the first directory prints the file and stops. Where's my error? :)

Answer 1

This line is wrong:
for file in "$files"
It should be:
for file in "${files[#]}"
$files just expands to the first element of the array.

Answer 2

Note: The recursive shell-function approach is suboptimal in real life (given that there are specific utilities such as du that do the job), but needed to satisfy the OP's specific requirements.
Update: The original answer mistakenly simply counted files instead of determining the combined file size - this has been corrected.
A revised version of your code that demonstrates several advanced bash techniques; note that the function has been renamed to sumFileSizes to better reflect its purpose:
Declares local variables, including one with -i to type it as an integer
Uses a string composed of quoted and unquoted elements (wildcards) for safe globbing (pathname expansion) - "$1/"*
Uses stdout output to "return" the desired result and captures it with command substitution ($(...)) rather than trying to pass a variable "by reference" (which bash doesn't directly support).
Use of process substitution via stdin (< <(...)) to provide a command's output as input to another command.
Shows relevant shell options (set with shopt) that govern globbing (pathname expansion) behavior.
#!/bin/bash
# Recursive function to report the *combined size of all files*
# in the specified directory's *subtree*.
sumFileSizes() {
# Declare the variable as an integer (-i) to ensure
# that += assignments performs *arithmetic*.
local -i size=0
local file
# Loop over all files/subdirectories
for file in "$1/"*; do
if [[ -d $file ]]; then # item is a *directory*
# Recurse, adding to the size so far.
size+=$($FUNCNAME "$file")
else # a *file*
# Add this file's size to the size so far.
# Note: `du` reports `{size} {file}`, so we need to
# extract the 1st token, which we do with `read` and
# process substitution, followed by printing the 1st token
# and capturing the output via command substitution.
size+=$(read thisSize unused < <(du -- "$file"); printf $thisSize)
fi
done
# Output combined size.
printf $size
}
# Ensure that:
# - globs expand to *nothing* in case there are *no* matching files/directories:
# option `nullglob`
# - hidden files/directories (those whose names start with '.') are included:
# option `dotglob`
shopt -s nullglob dotglob
# Make `du` report sizes in KB (this is the default on Linux).
export BLOCKSIZE=1024
# Invoke the recursive function and capture
# its outoput.
totalSize=$(sumFileSizes .)
# Output combined size of all files
# in multiples of 1KB.
echo "$totalSize"

Answer 3

If you want to get the size of a directory recursively, try this:
du -sh
-h is for human readable and -s is for summarize