map is being considered as a new wit type and reserved word. The latest wit-bindgen implemented the proposal and now fails to generate bindings if a reserved word is encountered in an invalid context. While this is a breaking change for the ecosystem it's understandable. However, I'm unable to use wkg to fetch this wit interface as the escaped keyword is replaced with the unescaped value in the emitted wit. This in turn causes the component to fail to build.
To ensure forwards compatibility, wkg should preserve all escaped wit identifiers with their escaped value.
Source wit:
https://github.com/componentized/valkey/blob/da24efeb1e4134971890c66b52748c316f9bc3e8/wit/resp.wit#L16
Problematic wkg fetch'd wit:
https://github.com/componentized/valkey/blob/da24efeb1e4134971890c66b52748c316f9bc3e8/components/wit/deps/componentized-valkey/package.wit#L22
In this case, the wit is not published, but defined as a local override within the same repository.
mapis being considered as a new wit type and reserved word. The latestwit-bindgenimplemented the proposal and now fails to generate bindings if a reserved word is encountered in an invalid context. While this is a breaking change for the ecosystem it's understandable. However, I'm unable to use wkg to fetch this wit interface as the escaped keyword is replaced with the unescaped value in the emitted wit. This in turn causes the component to fail to build.To ensure forwards compatibility, wkg should preserve all escaped wit identifiers with their escaped value.
Source wit:
https://github.com/componentized/valkey/blob/da24efeb1e4134971890c66b52748c316f9bc3e8/wit/resp.wit#L16
Problematic
wkg fetch'd wit:https://github.com/componentized/valkey/blob/da24efeb1e4134971890c66b52748c316f9bc3e8/components/wit/deps/componentized-valkey/package.wit#L22
In this case, the wit is not published, but defined as a local override within the same repository.