The default_stat fn should be public #9637
Conversation
|
This seems like a good candidate for using a re-export instead of having to reach all the way into |
|
I also imagine that @kballard's upcoming path rewrite will obsolete this(?). |
|
No, I copied the stat modules as-is into the new Path. I figure they can be rewritten later. However, I don't think that we should make |
|
I submitted a new pull request (#9652) with a version that actually implements Default for stat. For most of the os/architectures I couldn't use Deriving because of the arrays, unfortunately. If that one gets merged, this one can be closed, or vice versa. |
|
Closing for reasons outlined in #9652 |
Add `unused_format_specs` lint
Currently catches two cases:
An empty precision specifier:
```rust
// the same as {}
println!("{:.}", x);
```
And using formatting specs on `format_args!()`:
```rust
// prints `x.`, not `x .`
println("{:5}.", format_args!("x"));
```
changelog: new lint: [`unused_format_specs`]
Anybody working with a C API that uses the stat structure will want
default_statto be accessible, since, as can be seen by the implementation of default_stat, all uses ofstatwithout it would need to be platform-specific.In this PR I have simply made it public--I would understand if there was a desire to put the public API somewhere else instead. (There is no reason it should be in
std::path.) Maybe an impl ofDefaultinstd::libcitself?