Since #+pandoc-emphasis-pre and #+pandoc-emphasis-post have somewhat
different implementations, it is wise to test them
separately (e.g. issue jgm#8059 affects only the pre char setting).

The addition of tests for an alphanumeric pre/post char demonstrates
the aforementioned issue.

So far, orgStateLastPreCharPos wasn't updated appropriately after each
parsing to native Str (by the parser str). In addition to solving
this, the guard notAfterString in emphasisStart is removed to allow
emphasis after Str at the first place.

With the introduction of the settings for allowed pre/post chars
around emphasized text (and the resolve to jgm#8059), such markup-chars
may appear in the middle of the word according to the user's choice.