Todays topic is speed!
Commiting two very effective optimizations to the gui loading code.

One that stops doing split, join all the time when searching for duplicates, which halves(!!) the loading time for opening qt's enormous tarball.

Another one uses a shortcut that assumes the next file to be read is probably from the same directory as the same one (correct about 90% of the time).

Loading time is now almost the same as just listing the file contents with tar.