Emilio Cobos Álvarez comments

Results 777 comments of


                                            Emilio Cobos Álvarez

Remove unused variable and types

The expected way to do this is `--allowlist-function 'CompressBlocks.*'`, then it will only generate the types used from those functions, does that not work?

Remove unused variable and types

Well, it's hard to automatically know what you want to consider unused vs. not. For example there are lots of functions in C that take ints but expect some sort...

Using `syn` for code generation instead of `quote`

Can you elaborate what use cases do you envision for this? And why are they not addressable for example if you use `parse_quote!` on the output of bindgen?

renaming of the library `C` was specified, however this crate contains no `#[link(...)]` attributes referencing this library

Yeah this seems unlikely to be related to bindgen, bindgen intentionally doesn't mess with anything compiler-related, just generates code. That said, maybe the `CargoCallbacks` are causing trouble? That'd be a...

C Atomics are not converted to equivalent atomics in Rust

Hmm, yeah, this is a bit tricky. Part of the issue here is that they're not quite the same, ABI wise. An `AtomicBool` is a struct with ``#[repr(C)]`, while I'm...

C Atomics are not converted to equivalent atomics in Rust

https://github.com/rust-lang/rust-bindgen/pull/2153 deals with the crash. What to do with atomics more generally needs more nuance.

Generating bindings is extremely slow

Hard to say without a test-case, but `cargo check` shouldn't re-run the `build.rs` file if their inputs haven't changed.

Assertion failure `!isValueDependent()` triggered

Do you have a full stack by any chance?

Map size_t to usize by default and check compatibility (fixes #1901, #1903)

This seems ok if tests pass, sorry for missing it. I can try to rebase and merge if you don't have the cycles to do it.

Allow modifying macro tokens at parse time

Err, sorry, should've read the PR message more carefully. Rather than modifying the tokens, would it be reasonable to pass the tokens to `will_parse_macro` (without modifying them)?