Allow optional capital letters and numbers in code Beta

[adamwalz]

This commit modifies the source code generation grammars to allow the user to select whether capital letters and numbers should be added to a lesson while maintaining syntactically accurate generated code. It allows language-specific code conventions to be maintained such as only generating rust structs when capital letters are included, and allows for language-specific numeric literals.

[adamwalz]

This PR is a proof-of-concept just to start a conversation around the idea. It is not meant to be merged in this state.

Personally I'm not overly happy with with how this refactor turned out. I was hoping that more of the grammar would be agnostic to capital letters and numbers, but it grew into what is basically 4 grammars per language. If this grammar approach ends up being used for optional capital letters and numbers it might be easier to separate the grammars into 4 files rather than have a single file for all versions.

What I do like about this approach is that it maintains language idiocies in a way that I think would be hard or impossible if the capitalization and adding numbers were performed as a post-transformation.

[aradzie]

Omg, this feature is awesome!
Please let me finish another feature that I am working on right now, the custom theme designer. It will take a day or two.
Then let's work together on this PR.

[aradzie]

I am sorry for the late reply. I was having a really good streak while working on themes, did not want to interrupt myself.

I was thinking about adding conditional grammar branches, similar to Rust conditional compilation.

In Rust we can write

#[cfg(target_os = "macos")]
fn macos_only() {
  // ...
}

In Keybr we could write

#[feature "upper-case"]
c_func_id_upper_case ->
    "Next"
  | "Prev
   ...

Or, where this is not enough, we could mark a portion of grammar with a feature flag:

c_func_id_suffix ->
    "abort"
  | "bsearch"
  | "calloc"
  | { :feature("upper-case") // new syntax!
      "Next"
    | "Prev
    }

It is an experimental syntax straight out of my head, it does not exist yet:

   "{" _ ":feature" _ "(" _ feature_id _ ")" ... optional rules ... "}"

(The underscore characters indicates optional whitespace.)

Here the curly braces are used to wrap an optional list of rules. Each opening curly brace is followed by a feature flag name, then a list of rules, then a closing brace. The algorithm should scan all such feature flags in a grammar and present them to the user in the settings. If the user chooses to disable a feature, then the algorithm should just remove that fragment from the grammar. It should be as simple as removing a branch from a tree.

The same syntax could be used for syntax highlighting:

c_func_id_suffix -> { :syntax("id")
  | "bsearch"
  | "calloc"
}

or

c_type -> { :syntax("keyword") ( "char" | "int" | "long" | "short" | "void" ) } [ "*" [ "*" ] ] ;

Here we use curly braces with with different flags to mark a generated text fragment with color attributes.

[aradzie]

I am really eager to add more languages to keybr. That is a good PR! However, you mixed two different features in the same PR:

1. New languages syntaxes fro Rust and Go
2. Optional uppercase and number characters.

If you give me a PR with just new syntaxes, I will merge it right away. Then we could iterate on adding the remaining features.

Thank you!