exclude duplicates with different names

[madsurgeon]

Hi, for my it would be very useful to have a switch to exclude all matches with different filenames, so that files would only be considered duplicates if they have both the same contents and the same filename.

[arafesthain]

hi, same requirement here :)

[jbruchon]

I'll see what I can do.

[justbrowsing]

Quick and dirty method:

Assume previous jdupes run like

time sudo jdupes \
    --one-file-system \
    --recurse \
    --size \
    --permissions \
    --order=time \
    -X size+=:5M \
    -M \
    /path/to/scan/ | tee jdupes_log.txt

and log output

11574824 bytes each:
/path/one/filename.txt
/path/backup/two/filename.txt
/path/somedir/backup/three/filename.txt

Define a function to parse log file for exact same filenames in each set

strict_link() {
IFS=$'\n'; for line in $(cat "$1"); do 
    basename=$(basename "$line" 2>/dev/null); 
    if [[ "$line" =~ "bytes each:" ]]; then
        unset link oldline;
    elif [[ -n "$link" ]]; then
        if [[ "$basename" == "$link" ]]; then
            echo "$oldline -> $line";
            [[ "$2" == "list" ]] && ls -l "$oldline" "$line";
            [[ "$2" == "size" ]] && du -csh "$oldline" "$line";
            [[ "$2" =~ sum ]] && $2 "$oldline" "$line";
            [[ "$2" == "hard" ]] && ln -f "$oldline" "$line";
            [[ "$2" == "soft" ]] && ln -sf "$oldline" "$line";
        else
            unset link oldline;
        fi;
    else
        link="$basename";
        oldline="$line";
    fi;
done
}

USAGE: strict_link jdupes_log.txt [list|size|md5sum|hard|soft]