Tuesday I worked on generating constants, enums, and stubs for functions, interfaces, and classes. So, it was generating a skeleton of the whole Foundation and AppKit frameworks. Before I got into the meat of it, though, I found that the bridgesupport XML file doesn't have properties or their getters and setters. Given a known property like "sharedApplication", I couldn't even find it referenced in the XML. So, bridgesupport files aren't going to help.
This brought me back to my default strategy of leveraging documentation. In theory, I don't need to parse it. I can just have a documentation generator spit out structured data instead. Doxygen can parse Objective-C headers, so I'm thinking about that. Alternatively, if it doesn't work, we can parse the source directly, but I like the idea of not having to write a language lexer/parser for Objective-C, oh my gosh.
The problem is even this feels very unfun and I was already at the threshold of my unfun limit with this stuff. But it was a well defined problem so I posted it to Upwork. I actually found somebody that might be perfect. He does the kind of stuff a lot and works on a project called Babelfish that might be really useful for stuff like this in the future in other languages. He also wrote one of the DOM bindings for Go I was looking at with the WASM prototypes. Crazy coincidence!
So, hopefully I'll have some help with this now because there is a lot of it to do in the long term and I have enough work as it is.