Handle HTML files

[markfinger]

Would be useful to be able to handle HTML files during the build process. Processing should include finding dependencies and rewriting paths to point to hashed/versioned paths.

Benefits

• A frequent pain-point is transitioning a hard-coded html file to a production environment. Generally, I've been happy to add an extra minute's overhead of cloning the file and rewriting urls as it's a trade-off made to preserve a simple codebase. That being said, it's a pain when making frequent, incremental deployments. Like most repetitive tasks, it's ripe for automation.
• Might help to reduce the time necessary to go from an empty project to a wired up, viewable project.
• Moving the bootstrapping html-document out of the build script would be great.

Precedents

• https://github.com/webpack/html-loader
• https://github.com/ampedandwired/html-webpack-plugin

Problems to solve

• Handling index resolution, eg: foo/index.html is served from both foo/index.html and foo/. Might need to consider expanding the notion of urls to cover multiple urls per file.

[markfinger]

Heavily dependent on #113

[markfinger]

While most compilation units (js, css, etc) can be generated agnostic of the outcomes of their dependencies, HTML files require explicit links to their dependencies assets.

When compiling, will require a means to defer the completion of html files until urls can be generated for the linked assets.