I tend to think it should be kept the same because a) is easier to write the code (I have a similar script that allows access to the wx Erlang constants) and b) searching the WxWidget docs is easier when the names are the same.

@kerryb what’s the point of the WxObject wrapper if everything else apart from the constants is Erlang syntax?

For those interested I did recently do this, for all 25,000 lines of related constants, mostly through some multi-cursor editing and find/replace regexes. src/ files here.

In my opinion you should follow Elixir’s naming convention, so all of the macros and functions should be snake case.

The main reason I didn’t do that was … well, laziness to be honest, but also I’ve used tools before (eg Java and Ruby bridges to Apple’s Cocoa) where having to keep mentally translating the names of things when reading documentation was confusing.

I understand that it’s for a reason, so instead of converting it to snake case I would rather fit it in Elixir’s standards, for example a call like:

wxALIGN_RIGHT()

could be changed to:

wx(:ALIGN_RIGHT)

I do like that idea – it seems nicer than polluting the namespace by importing hundreds of functions into a module.

Maybe you could use more clear example usage here.

Yes, good point – I just pasted some snippets from a code sample I’d converted, assuming that if anyone was interested they’d already be far more familiar with wx than I am. I ought to replace it with a working “hello world” example.