[libdfm] Support lexer weeding configuration

[Dadie]

Currently the weeder for the tokens while having a configuration it's not available for users of the library. Each possible likely configuration should be tested to make sure a valid parse tree can be generated from the more or less weeded tokens.

It could be as simple as adding another argument to dfm::lexer::weed().

Argument was added to the function dfm::lexer::weed() but tests are still missing.