This commit introduces two main changes:
1. Refactors the variable names used during the creation of
`NapiStruct` in `crates/macro/src/parser/mod.rs`. The goal is to
make the source of `NapiStruct.name` (the Rust identifier) and
`NapiStruct.js_name` (the JavaScript name, potentially from a
`#[napi(js_name = "...")]` attribute) more explicit. This change
is primarily for code readability and maintainability. The core
logic remains the same.
2. Adds a new test case to `examples/napi` to specifically verify
the behavior of `#[napi(js_name = "...")]` on structs. This
includes:
- A new Rust struct `OriginalRustNameForJsNamedStruct` in
`examples/napi/src/class.rs` decorated with
`#[napi(js_name = "MyJsNamedClass")]`.
- Corresponding tests in `examples/napi/__tests__/values.spec.ts`
that check for correct class instantiation, method calls, and
type alias generation (`OriginalRustNameForJsNamedStruct` as an
alias for `MyJsNamedClass`).
Note: I encountered some issues running the full build and tests for the example due to pre-commit hooks in the testing environment that I couldn't bypass. However, I believe the changes are correct based on static analysis and focused checks.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
- also refactor the enum codegen, now we will not emit #[derive(Copy,
Clone)] for the original enums.
- also refacotr the fn codegen, now #[napi] fn can accept env: &Env as
argument, this is useful when the return type contains lifetime
This is mandated by the Rust doc of the BufWriter. We've also seen bugs where
the data doesn't seem to be flushed properly, and thus the CLI fails to parse
JSONs emitted by napi-derive. This might be it, or not, but flushing shouldn't
hurt.
This allows us to customize the enum value for string_enum. So far, only the case attribute was allowing us to customize the value of those variants, this is a next step in allowing the value to be different from the actual name.
This has a new test, and snapshots have been updated accordingly.
@Brooooooklyn - could we include that in napi v2 and tag after it?
* fix: make sure env without exception pending before throw error
* fix(napi-macro): gengerate type file should align with type-def featrue
* fix(napi-macro): gengerate type file should align with type-def featrue
