Md5 hash diffing to skip generating unchanged files

[ninjamuffin99]

Simple logic to generate a md5 hash when parsing the .xml files. If the hash is different, it will generate as normal. Otherwise it will skip to the next .xml file to parse and update.

This saves a bit of time, running it on the test files (npx haxe --run Make pages) went from 1.5s to generate the pages, to about 0.9s, so about 150ms saved per .xml

[Simn]

I can see the appeal, but I don't know if it's really acceptable behavior to generate such auxiliary output files into the input directory...

Plus we create a problem if someone wipes their output directory in order to generate, and then nothing happens because the input didn't change.

How about generating this somewhere under the output directory instead?

[ninjamuffin99]

I've changed it to spit it out into the cfg.outputPath, so the dox test site generates the md5 hashes into bin/pages/hashes. Could be somewhere else though or add more specific configuration for paths if required.