Implement methods to extract Ruby and HTML individually from ERB

[marcoroth]

This pull request implements methods to extract the Ruby and HTML source individually from the HTML+ERB source.

It implements the erbx_extract_ruby_to_buffer and erbx_extract_html_to_buffer functions in C and exposes them in the ruby gem via ERBX.extract_ruby and ERBX.extract_html.

ERBX.extract_ruby("<h1><% RUBY_VERSION %></h1>")
# => "       RUBY_VERSION        "

ERBX.extract_html("<h1><% RUBY_VERSION %></h1>")
# => "<h1>                  </h1>"

Maybe we should rewrite them later when we have a proper AST so we can also filter out the ERB tokens after the start tag (=, -, etc.). Or we introduce/expand the TOKEN_ERB_START to include any ERB start tag.