Quote:
Originally Posted by Liberated
maybe it can check for like a % match? I mean if even 50% matches that's already quite alot.
|
It's not quite that simple. The way it works is that it looks for heads with the exact same file size (down to the byte) from a list of corresponding head paths and file sizes. If it matches, it's most likely a duplicate. However, there are rare cases where there are two heads with the same file size, so it also compares the contents with a string check.
Changing this to actually load every file, parse it as a PNG/GIF/MNG, then compare each pixel would be nearly impossible and definitely not worth the effort (except maybe for experience in programming, I guess).