Implement unquoting

[limansky]

Hi,

I've tried to implement unquoting. This code is definitely not production ready, but just to show my vision on how it might be implemented. My approach is quite different than it is done in #12. I tried to avoid hacking strings. It is easy for content, but I'm still thinking how to deal with attributes, because both Metadata and NamespaceBinding expect string values at the construction time. I think, I can inherit these classes to show that here is a hole and it will be substituted later (of course it requires quite a lot of coding because of how scala-xml parsers implemented) . Another approach is to put strings into the Hole, but I'm not sure how to decompose AST properly to cover all cases (literals, variables, calls, etc), if it possible.

Please, add your comments and suggestions.

[limansky]

I've slept on it and I think that we need to implement own model classes for XML, and the parser inherited from MarkupParserBase. Of course these classes will be converted to scala-xml Nodes on lifting stage.

Even though it requires some extra coding, it will save us from hacking in future when we will implement error handling and possible some other stuff. I'll try to prototype it this week.

[limansky]

Actually I was able to make the rest of the tests working within this approach.

[allanrenucci]

test("reconstruct unquote within namespaced elem") {
  val foo = xml"""<foo xmlns:pre=${"foo" + "bar"}/>"""
  val bar = <foo xmlns:pre={"foo" + "bar"}/>
  assert(foo.scope == bar.scope)
}

You'll still fail this test case though. As you said previously, we'll probably have to implement our own xml syntax tree

[limansky]

@allanrenucci Good catch!

The problem is what in case of using of another model, we'll need to have our own parser, because MarkupParser is too much coupled with scala.xml model.

[allanrenucci]

I am actually writing a parser based on https://github.com/lihaoyi/fastparse/blob/master/scalaparse/shared/src/main/scala/scalaparse/Xml.scala