Add "return value" topics to "Thinking in Nu"

[NotTheDr01ds]

Another fairly substantial update to the "Thinking in Nu" chapter. It adds two major topics that often confuse new users:

• Implicit return
• Only the last value of an expression is returned

Also:

• Major update to "Think of Nushell as a Compiled Language" and the linked "How Nushell Code Gets Run" chapter

  • There was a lot of duplication between the two chapters. The code-examples that fail parsing in Nushell have now been moved almost exclusively the to the "Thinking in Nu" Chapter and out of "How Nushell Code Gets Run". There are links between the two chapters where needed.
  • This is hopefully much more clear. I certainly understand it better ;-)
  • Previously, the pain point of lacking an eval was much more visible than the benefits. Hopefully the benefits are more clear now.

• Adds a Table-of-Contents to both chapters

• Each section closes with a "Thinking in Nushell" callout for consistency. (Side-note: This is listed as a "warning" block, but only for color/style to make it most noticeable).

• Additional rewrites for clarity and flow

[NotTheDr01ds]

@kubouch If you get a chance, I'd love your review, but especially the revisions to "Parse-time Evaluation" since this is based on your work in the first place.

[kubouch]

Looks good overall, thanks! I left a few quick comments from my phone, but I didn't find anything that would desperately need fixing, you can address / not address them as you see fit.

[NotTheDr01ds]

@kubouch Thanks! I've just pushed some changes based on your feedback. Greatly appreciated!

[kubouch]

OK, looks good, it can always be refined in other PRs. We might want to start talking about Nushell as statically compiled instead of just compiled to avoid the confusion with Python bytecode, see my comment above.