Bash script to delete oldest file causes error - but used to work before

The best way to handle special characters in file paths is to use null terminated lists; a feature supported by many standard Linux tools

find /home/lmc/tmp/abc/ -type f -printf '%T+ %p\0' | \
sort -z | head -z -n5 | cut -z -c 32- | xargs -r0 rm -v

find ... -printf '%T+ %p\0'
sort -z
cut -z
xargs -r0

Result

removed '/home/lmc/tmp/abc/with space/10574705_780745301964694_1921917122_n.png'
removed '/home/lmc/tmp/abc/with space/10689463_10204933967012054_5293239505955413653_n.jpg'
removed '/home/lmc/tmp/abc/with space/10940622_852091444848970_5646746288209785678_n_002.jpg'
removed '/home/lmc/tmp/abc/with space/10940622_852091444848970_5646746288209785678_n.jpg'
removed '/home/lmc/tmp/abc/with space/10956906_449528018528925_457296511_n.png'

As mentioned in comments, your code is not able to correctly handle directory (and file) names that contain spaces. And since awk's default field separator includes spaces, your awk script sees your full paths as multiple fields ($2, $3, ...).

Since you have the ability to designate the field delimiter (via find/printf) one simple fix would be to pick a delimiter you know won't show up in your directory (and file) names.

For example, assume | does not show up in directory (or file) names then we can make 3 small changes:

OLD: find /media/2C1F-444F/usb_musik/newMusik -type f -printf '%T+ %p\n' | sort | head -n 5 |  awk        '{print $2}'  | xargs       sudo rm -v
NEW: find /media/2C1F-444F/usb_musik/newMusik -type f -printf '%T+|%p\n' | sort | head -n 5 |  awk -F '|' '{print $2}'  | xargs -I {} sudo rm -v {}
                                                                  ^                                ^^^^^^                       ^^^^^            ^^

Where:

• printf '%T+|%p\n' - use | as delimiter between timestamp and path to file
• awk -F '|' - tell awk to treat | as the field delimiter
• xargs -I {} sudo rm -v {} - tell xargs to (effectively) not split the input on spaces (akin to wrapping the path/file combo in quotes)

Taking for a test drive:

$ cd /home/username/tmp
$ mkdir -p 'this dir/that dir'
$ touch 'this dir'/file1.txt 'this dir/that dir/file2.txt'

$ find "$tmpdir/testd" -type f -printf '%T+|%p\n' | awk -F'|' '{print $2}'
/home/username/tmp/testd/this dir/file1.txt
/home/username/tmp/testd/this dir/that dir/file2.txt

$ find "$tmpdir/testd" -type f -printf '%T+|%p\n' | awk -F'|' '{print $2}' | xargs -I {} sudo rm -v {}
removed '/home/username/tmp/testd/this dir/file1.txt'
removed '/home/username/tmp/testd/this dir/that dir/file2.txt'

$ find "$tmpdir/testd" -type f -printf '%T+|%p\n' | awk -F'|' '{print $2}'
    -- no output

NOTE: if the only issue causing problems is spaces in the directory (and/or file) names then this solution should be sufficient; if you need to worry about other special characters (eg, embedded linefeeds) then you'll need to expand this solution to include a custom record separator (eg, \0) and/or look at other approaches (eg, see the other answers you've received for this question)

Your approach looks pretty fragile about special characters like spaces.

What I would do instead:

#!/bin/bash

cd "$1" || exit 1

array_list_temp=( * )
readarray -td '' array_list_temp_files < <(
    perl -l0e '
        print join "\0",
        sort { -M $b <=> -M $a }
        grep -f, @ARGV
    '  -- "${array_list_temp[@]}"
)
printf -- '%s\n' "${array_list_temp_files[@]:0:5}"

Usage:

cd /media/2C1F-444F/usb_musik/newMusik
./script.sh .

If the output looks good, you can do:

cd /media/2C1F-444F/usb_musik/newMusik
./script.sh . | xargs rm -f