More flexibility in choosing the files to process

[johann-petrak]

Currently all the extensions defined for any of the file types are processed (and in addition the extensions added for a type).

As we add more types, it becomes more likely that this is not wanted or that extensions can be used for several different types.

So it would be good to limit what is processed by type, by extension, or even exclude whole file path patterns. On the other hand it may be useful to define file path patterns which should be processed in a particular way (so maybe because of the base name, not just the extension).
One example is something like "robots.txt" which is a specific type of file, but not just a general text (.txt) file.

For this it may also be useful to change the way how file types are defined to not just use the last extension but arbitrary path patterns, of which extensions are a special case.