bsloane1650 opened a new pull request #264: WIP - DAFFODIL-2169 URL: https://github.com/apache/incubator-daffodil/pull/264 The groundwork for adding type-safty to our DPath and infoset values. This doesn't actually effect any of the existing code yet, but should give any idea on what I am planning. Given the volume of effort needed to actually incorporate this into the codebase, I wanted to get feedback before I attempted to do so. When I do thread it through, I suspect I will run across other issues. I'll try to keep the "meaningful" changes separate from the boilerplate updates to type signatures. The bad: 1) We need to support having DINodes, but also need the DataValue type to be visible to Daffodil-lib (because we will want to use it for the number-conversion functions). I am accomplishing this by defining a trait it DataValue.scala that is instantiated only by DINode, along with an implicit to convert back to a DINode. I don't think this will cause additional memory allocations, but it might invoke an additional typecheck when we try to get extract a DINode. 2) We still need to thread this through all of Daffodil. The good: 1) DataValue is not AnyRef 2) By using phantom types, we can impose an arbitrary type hierarchy on our DataValue.I suspect the only useful useful distinction here is going to be Nullable vs NonNullable. We might also be able to take advantage of the seperate DataValue types for each supported underlying type. In particular, we might use them to tighten the return type when subclassing. In theory, we could have done that when using AnyRef, but not that the types are still marked as DataValue it may be clearer what is going on. I also included DataValuePrimitive to indicate types which are definined in Daffodil-lib. This is mostly just a demo that this is possible to define more complicated type hierarchies. I don't actually see a way to take advantage of it.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
