Implement FixedSizeArray for all fixed size arrays
#28088
Conversation
|
r? @pcwalton (rust_highfive has picked a reviewer for you, use r? to override) |
|
Code from this reddit post by @eddyb. |
|
This is quite amazing. Ideally the element type should be an associated type, but I know the impl isn't allowed that way. |
| impl<T, A: Unsize<[T]>> FixedSizeArray<T> for A { | ||
| #[inline] | ||
| fn as_slice(&self) -> &[T] { | ||
| unsafe { &*(self as *const A as *const [T]) } |
There was a problem hiding this comment.
This can be just self as &[T] or in fact just self, and let coercion do the rest?
|
(It'd be great to have a more descriptive commit message or maybe some comments: I can definitely see people looking at this and wondering what's going on in future.) |
|
How well does it work with coherence? |
|
@eddyb No, it unfortunately doesn't work with coherence at all. |
|
@tbu- I tried making it smarter at some point: it should be possible to teach coherence that |
|
Using this for other blanket impls would make me a little uneasy, but this is an unstable trait anyway so I think this is fine. Could you squash the commits together and expand the first commit message like @huonw mentioned? Otherwise looks good to me! |
Do so by using the fact that fixed size arrays (like `[u8; 8]` can be coerced to slices `&[u8]`, this is expressed through the trait `Unsize<[T]>` that all fixed size arrays implement.
tbu-
force-pushed
the
pr_fixed_size_array
branch
from
5174841 to
4d2709d
Compare
|
@alexcrichton Squashed and expanded the first commit message. |
|
Yeah I don't believe this is what we'd want in the future (it's not how trait support for fixed size arrays would look if we had designed for it). |
|
Does anybody actually use like it was initially planned due to coherence restrictions, and without it it's almost useless, probably I should have killed it in the same PR that introduced it. |
|
@petrochenkov The trait is useful for abstractions over arrays, such as stack-allocated vectors, hybrid vectors, etc. |
|
⌛ Testing commit 4d2709d with merge d9a55d0... |
|
💔 Test failed - auto-linux-64-nopt-t |
|
@bors:retry On Mon, Aug 31, 2015 at 2:34 PM, bors [email protected] wrote:
|
|
@alexcrichton I believe there's a space missing after the colon in |
|
@bors retry |
|
⌛ Testing commit 4d2709d with merge 8f8b129... |
|
💔 Test failed - auto-win-gnu-64-nopt-t |
|
@bors: retry On Tue, Sep 1, 2015 at 4:58 AM, bors [email protected] wrote:
|
No description provided.