Functional WEIDU

[lynx]

https://en.wikipedia.org/wiki/Expect#Examples

 

When the behaviour of the app is deterministic, you know what kind of prompts to expect and when, in which case Expect can handle automating it easily.

[Gimble]

Stupid question from a newbie: could you separate out this new functionality into a meta-tool that would translate your new user-readable stuff into raw Weidu-code, and then distribute that generated Weidu code as a way to reduce installation overhead?

 

I more or less did that in my user-scripting land: I have a "C" style preprocessor that parses a BAF file ( with extra #includes, #defines, etc.) and generates a .BAF file as output, that I then hand off to Weidu for compiling. It's not a perfect solution but it does 90% of what I want, and doesn't pollute the Weidu language namespace with a different model.

 

Even more cool would be something that would take an existing Weidu script and reverse-engineer it back into user-readable stuff. I don't do anything but AI scripting, but I can say if I were ever willing to take on modding, the stuff you have proposed is far more accessible than the raw data. I may be a good coder (I hope so, I get paid to do it), but I really don't relish grokking other people's Weidu ugliness to learn how to do basic modding.

 

Yes, I'm lazy. It's part of not having a lot of free time to work with.

[lynx]

The slowness is not due to the extra abstraction, but the unrolled weidu it produces. If speed was really an issue, extending weidu directly is the only way to go.