Emilio Cobos Álvarez
Emilio Cobos Álvarez
Sorry for the lag replying here. > All of these things require libclang 8.0, so my first question is how would we do the version checks? Do we need to...
For now, you can also compile it with `-fno-inline-functions`, I think.
> Won't that prevent inlining inside xed? Yup, I think so, which is unfortunate, I agree. > Further, -fno-inline-functions and .generate_inline_functions(true) doesn't actually seem to generate bindings for those functions....
For the `JNIEnv` case it's easy, blocklist the JNIEnv type and add a raw `use jni::JNIEnv;`, we do that pretty extensively. The `*mut __va_list_tag` needs some more thought probably.
This is breaking a couple existing test-cases unfortunately. That needs looking into. Depending on how much of an edge case they are, it may be ok to change their arguments...
I think changing those tests to use `.*` and landing this should be ok, I'll make sure to document this change prominently for the next release if so. what are...
Ugh, this never worked though... I know how to fix it for this case, but it's not pretty, and it's definitely hard to make it work for the general case.
If only clang gave us a way of evaluating an expression in a given context...
Yeah, so I looked into patching libclang so it works, but I can't see a clear way to do this. Basically, clang needs to expand the macro so it knows...
FWIW, if those are switched to constants, they'll work properly with #260, even though it's a complex expression.