GFM reader may fail to be registered for markdown files

[sebbASF]

Pelican looks for readers by scanning BaseReader.subclasses(). The last one in the list wins.
There is no warning if an extension has multiple readers.

It looks like the gfm reader is currently located after the built-in MarkdownReader, but that may not always be the case.

If the wrong reader is used, this may cause hard to find issues (especially if the ordering depends on the context!).

It should be easy enough to use the readers_init signal to check which reader is being used for Markdown files, and log an error if it is not GFM.

[sebbASF]

Turns out that the code already updates the class for the 'md' extension, but fails to do so for 'mkd' etc.