rust-lang / rust-lang/rust-bindgen
[Feature request] non_exhaustive structs
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 5.3k
- Forks
- 829
- Avg merge
- 1d 1h
- Merged PRs (30d)
- 15
Description
From a C API point of view, adding a field to a struct is often not seen as a breaking change, for fields that are not mentioned when giving a struct by value are zeroed. The Rust equivalent would be to make all structs non_exhaustive (which, unless the type is not supposed to be constructed from Rust, only makes sense if Default is derived on it).
Therefore, I think it would make sense to enable having non_exhaustive on structs, so that more C non-breaking changes can be translated to Rust non-breaking changes. A -sys crate using this would merely need to make the decision of whether either
- the underlying C library only adds fields at the end (or in alignment-guaranteed gaps), or
- the -sys crate makes the concession that the
#[repr(C)]merely guarantees that it is aligned with the linked C code, but no concrete layout guarantee.
Is this something that you agree would make sense to have in bindgen (say, as an optional "make all (or selected) structs non_exhaustive")? Is there anything that a PR implementing this could do to make 2. go over better? (I don't think we can make the repr property non-pub, but maybe there are tricks I'm missing.)
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names no files or tests. Start by locating bindgen's struct-generation and configuration paths, then inspect how repr(C), Default, and other generated attributes are handled. Done means an optional setting can apply non_exhaustive to all or selected structs, with the layout and C-compatibility trade-offs covered by tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c, rust
- Domain
- devtools
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100