Refactor Enum in codegen
#676
Conversation
lilizoey
added
the
quality-of-life
|
API docs are being generated and will be shortly available at: https://godot-rust.github.io/docs/gdext/pr-676 |
There was a problem hiding this comment.
Thanks a lot!
Several smaller changes are hidden in the moved-around blocks, I already commented on some. If there are some I've missed and you consider worth reviewing, please mention them 🙂
This will likely break #627, but as it stands, that PR doesn't have the highest chances to get merged (Godot decision still pending). So that's probably OK. And I'd expect potential rebase effort to be reasonable.
godot-codegen/src/generator/enums.rs
Outdated
| let mut enumerators = Vec::with_capacity(rust_enumerators.len()); | ||
| quote! { | ||
| #[repr(transparent)] | ||
| #[derive(#( #derives ),*)] |
There was a problem hiding this comment.
Here I wonder if the whole #[derive(...)] attribute should be stored in a variable... Or do you see a benefit in having individual derives in a list?
For example, this approach would not work in case the #derives list is empty, whereas the other could emit an empty token-stream.
There was a problem hiding this comment.
Actually this would work if the #derives list is empty, since this is perfectly legal to do:
#[derive()]
struct Foo;and i think the code is simpler if i dont store it in a variable.
lilizoey
force-pushed
the
refactor/enum-codegen-2
branch
3 times, most recently
from
294e187 to
abd726f
Compare
Bromeon
force-pushed
the
refactor/enum-codegen-2
branch
from
abd726f to
b7f19bc
Compare
lilizoey
force-pushed
the
refactor/enum-codegen-2
branch
from
b7f19bc to
de223d0
Compare
Like #673 but the codegen is kept as global functions and not inherent methods to
Enum/Enumerator.