Remove validation for image type and format

[lubomir]

This is supposed to catch invalid values when creating metadata, but doesn't work particularly well.

The validation is applied at parsing time too, causing breakage when new type/format is added but users haven't yet upgraded.

This is backwards compatible in that any file that was parseable with the old version will still be parseable. Some previously invalid files will now become valid.

[lubomir]

@ktdreyer , you run into this problem last time. What do you think of this change?

[ktdreyer]

Yeah, I don't know the original purpose of having these values hard-coded into productmd. It's probably better to do this validation in the applications that use productmd (if at all).

[AdamWill]

I quite like the validation, because it means we notice unexpected changes when we e.g. try to land new tools for building images (as we're doing at the moment in Fedora). I think the general idea is to try and keep control of the range of 'expected' images we ought to be building, and that there aren't unexpected changes to that.

I guess I agree that in theory maybe we should do that in pungi or something, but...we don't, because productmd does it, so if we drop this from productmd, it isn't being done anywhere.