diff options
| author | Nicholas Nethercote <n.nethercote@gmail.com> | 2023-11-28 09:36:09 +1100 |
|---|---|---|
| committer | Nicholas Nethercote <n.nethercote@gmail.com> | 2023-11-28 12:14:25 +1100 |
| commit | d9fef774e3ba9e00f0ea143f632f66b1d87afff6 (patch) | |
| tree | ce5a45137b4be5bab2d133434cece647a5653762 /tests/rustdoc-js-std/reference-shrink.js | |
| parent | 705b484922697b3dfc73ea7cb58dc12cca963ecf (diff) | |
| download | rust-d9fef774e3ba9e00f0ea143f632f66b1d87afff6.tar.gz rust-d9fef774e3ba9e00f0ea143f632f66b1d87afff6.zip | |
Remove `hir::BinOp`, `hir::BinOpKind`, and `hir::UnOp`.
They're identical to the same-named types from `ast`. I find it silly (and inefficient) to have all this boilerplate code to convert one type to an identical type. There is already a small amount of type sharing between the AST and HIR, e.g. `Attribute`, `MacroDef`. The commit adds a `pub use` to `rustc_hir` so that, for example, `ast::BinOp` can also be referred to as `hir::BinOp`. This is so the many existing `hir`-qualified mentions of these types don't need to change. The commit also moves a couple of operations from the (removed) HIR types to the AST types, e.g. `is_by_value`.
Diffstat (limited to 'tests/rustdoc-js-std/reference-shrink.js')
0 files changed, 0 insertions, 0 deletions
