Formalize interpolation language into a real typed interpreted language

[mitchellh]

This PR turns the syntax of interpolated configuration values into a real parsed language with real types. Before this, it was just strings all the way down (somehow everything turned into strings). This was really weird because it didn't allow math operations, we actually support some concept of maps in variables, etc. It also led to less than stellar errors.

I'm going to do a bit more work on this PR, but opening it for any feedback and to show to some folks.

This PR is completely backwards compatible. It also ends up fixing a lot of issues or introducing new functionality:

• String concatenation just works, so you can do things like ${file("foo_${count.index}.txt")} now.
• Escape characters work properly: ${"foo\nbar"} (newline, example)
• Numbers are supported: ${function(42)}, also floating point
• All function argument counts are semantically verified
• Types are all verified (although they're still all strings for now since the terraform package still gives us all strings. That is the next step to fix after this PR. The language itself is fully typed and type checked).
• Super well tested. The last interpolation regexp stuff was reasonably tested but not as well as it could've been.

The language itself is a very standard interpreted language with lex -> parse -> semantic check -> execute phases. Lexer goes to a token stream parser goes to an AST tree, semantic check and execution is done using the visitor pattern on the AST tree.

Fixes #744

[mitchellh]

The tests pass, Travis is just sad for some reason.

[armon]

[lamdor]

💯 Thanks for doing this, it's appreciated!

[svanharmelen]

Nice... Really nice... Your clearly back from the holidays 😉

[knuckolls]

👍 Much needed. Thanks. I happened to find a bug and wrote a fix in the existing code late last week pertaining to doing element(resource.*.parameter, count.index) calls. Going to see if that still remains on this branch and report back.

[ceh]

Woah, a treat for the weekend. I will enjoy reading this. Cheers @mitchellh!

[knuckolls]

The bug does still exist. I'll get it written up this afternoon.

[mitchellh]

@knuckolls I'm guessing its not a parsing bug so much as a logic but in the function at this point. And also probably because we don't have proper types being sent into the configuration yet.

[knuckolls]

@mitchellh wrote it up in a new issue #795

[mitchellh]

Coming in!

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.