GFM meta parser ignores blank lines; this is non-standard

The GFM reader skips blank lines when parsing meta-data.
This is in contrast with the standard Pelican parsers which treat a blank line as the end of the meta-data block.

This matters because:
- it is non-standard
- the parser can treat the start of body content as a meta-data item, if the body starts with something that looks like a keyword. This is unlikely, but not impossible.

It looks as though this behaviour has been relied on in some ASF website sources, in particular the main website.

Some options are:
- just document the behaviour
- add an option to warn if such blank lines are found
- add an option to disable the behaviour