Conversation
inickles
force-pushed
the
inickles/export-replaced
branch
from
596e4e7 to
b91980d
Compare
ahl
left a comment
ahl
left a comment
There was a problem hiding this comment.
I think it would be helpful to understand the utility of this change, and I'd like to see how this is materialized in the generated code.
Also the style of the code here doesn't particularly match the code around it.
| let mut sorted_replace: Vec<_> = self.settings.replace.iter().collect(); | ||
| sorted_replace.sort_by_key(|(name, _)| name.as_str()); | ||
| let replaced_type_reexports = | ||
| sorted_replace.iter().map(|(type_name, (replace_path, _))| { | ||
| let type_ident = quote::format_ident!("{}", type_name); | ||
| let replace_tokens: TokenStream = | ||
| replace_path.parse().expect("invalid replacement type path"); | ||
| quote! { | ||
| #[allow(unused_imports)] | ||
| pub use #replace_tokens as #type_ident; | ||
| } | ||
| }); | ||
| let replaced_type_reexports = quote! { #(#replaced_type_reexports)* }; |
There was a problem hiding this comment.
as discussed, I think this may make more sense in typify.
Do we think this should be the only behavior or should be opt-in (or opt-out)?
What if there's a replace imperative that's never used? Is that caught elsewhere? Would we want to check for that?
Would we want to do something for the other ways to replace a type such as x-rust-type.
ahl
left a comment
ahl
left a comment
There was a problem hiding this comment.
I mean to say: let's change this (not approved)
|
Closing in place of oxidecomputer/typify#1007 |
2 participants
This exports the types that someone includes in
replacewhen using thegeneratie_api!macro, making them available to users.