nullable annotations would be useful

[litherum]

It would help implementations know what's allowed to be null and what isn't.

https://clang.llvm.org/docs/analyzer/developer-docs/nullability.html

[litherum]

Bump. It's unclear if a create method fails due to a validation failure, if it should return nullptr or an invalid object.

[Kangz]

See #157 for arguments to the API.

In general return values that are object are never null, and instead are invalid objects, like the JS API.

[litherum]

I was thinking of __nullable from https://clang.llvm.org/docs/analyzer/developer-docs/nullability.html ...

[Kangz]

Erhm, that's non-standard and just an LLVM thing. That header also needs to be compatible with other C compilers like GCC or MSVC. At best we could do a WGPU_NULLABLE define that default to being an empty list of tokens but can be overriden.

[kainino0x]

for future reference: __nullable isn't the correct syntax, that appears to be some internal thing. Actual names are _Nullable and _Nonnull. Here's the reference page:
https://clang.llvm.org/docs/AttributeReference.html#nullability-attributes

[litherum]

WGPU_NULLABLE sounds like the perfect solution.

[kainino0x]

Partially addressed in #182, but there were some open follow-up questions:
#182 (review)

[kainino0x]

Added the important item from there to #190. Closing this issue, so we can track there.