Fix errors with parsing indented code-blocks

[Alhadis]

Last night I begun "translating" a document into reStructuredText, and I noticed an issue (which was already reported). I fixed it for you.

Notes:

• The fix for indented :: blocks prevented any inline patterns (such as *this* or this__) from being picked up in a line which ended in :::

  This was remedied by grouping all inline-patterns into a new group (#inlines) and embedding it within the respective beginCaptures group:

  

  The reason for this is because when TextMate starts a match, it doesn't apply other matches to its content unless it's told to (with a patterns statement).
  
  I'm pointing this out in case it helps explain future (or past) matching errors of a similar nature.
  
  

• The pattern to match field-names used (:)[A-z][A-z0-9 =\\s\\t_.]*(:) as a matching sequence. I've amended it so it's now (:)[A-Za-z][\\w\\s=.]*(:) instead. Because of ASCII ordering, [A‑z] actually matches this:
  
  

  ABCDEFGHIJKLMNOPQRSTUVWXYZ[]^_`abcdefghijklmnopqrstuvwxyz

  Note what's in-between; I doubt \ and ` are accepted in field-name identifiers.

[Alhadis]

In case it wasn't clear, this PR patches both issues mentioned in #22. There were two separate (but related) bugs.

[Lukasa]

Thanks for this @Alhadis! :sparkles:

[Alhadis]

Crap, you already pushed a release. Uh, I'm in the middle of adding Doctest Block highlighting too... just smoothing out some potential edge-cases:

This how it's meant to go? I'm no fan of Python, so I wouldn't know much about this >>> stuff.

[Lukasa]

Heh, that's ok, we can do lots of releases. Releases are good. =D

Per docutils documentation:

Doctest blocks are text blocks which begin with ">>> ", the Python interactive interpreter main prompt, and end with a blank line.

So that highlighting isn't quite right, I'm afraid.

[Alhadis]

How's this?

[Lukasa]

That looks great, I'd merge that. =D

[Alhadis]

Done too.

Sorry about prompting a double-release, haha.

[Lukasa]

No worries, pushing releases is easy. =D