This PR adds rules for negative implementations. It follows pretty much what the [RFC](https://github.com/rust-lang/rfcs/blob/master/text/0019-opt-in-builtin-traits.md) says with 1 main difference:
Instead of positive implementations override negative implementations, this have been implemented in a way that a negative implementation of `Trait` for `T` will overlap with a positive implementation, causing a coherence error.
@nikomatsakis r?
cc #13231
[breaking-change]
If `a.method();` can't be resolved, we first look for implemented traits
globally and suggest those. If there are no such traits found, we only
then fall back to suggesting from the unfiltered list of traits.
This seems to match what clang does on arm, but I cannot do any
experimentation with mips, but it matches how the LLVM intrinsics are
defined in any case...
Unlike the intrinics in C, this types the SSE values base on integer
size. This matches the LLVM intrinsics which have concrete vector types
(`<4 x i32>` etc.), and is no loss of expressivity: if one is using a C
function that really takes an untyped integral SSE value, just give it
whatever Rust type makes most sense.
Currently, we build a closure that does nothing but pass its argument
through to another function, this is rather wasteful and creates lots of
unnecessary closures.
While it's unstable and will probably be replaced or "reformed" at some point, it's useful in the mean time to be able to introspect the type system when debugging, and not be limited to sized types.
Fixes#21058
...to make it slightly clearer that there's not much point in boxing a vec.
On a different note, I read the contribution guidelines (https://github.com/rust-lang/rust/blob/master/CONTRIBUTING.md#pull-request-procedure) which say I should update the copyright date for this file. But I can see that nobody else has done this so far this year, despite there being a fair number of commits.
Does that instruction need removing?
Really small correction.
This anti-example in the Closures section is supposed to fail because of a borrow, but it was failing at the type inference because of insufficient type information.
This makes it fail for the expected reason.
Just a couple of tweaks to improve the appearance of pages like [The Rust Reference](http://doc.rust-lang.org/reference.html) on mobile to advance the progress of #20850. Changing the viewport wasn't quite good enough to make it mobile-friendly, as large tables forced the page to become wider than the screen width. Using `overflow-x`, tables that are too large become horizontally scrollable instead of forcing the *entire page* to become horizontally scrollable.
Also, there was at least one case where an extremely long piece of inline code was wider than the actual width of the screen (`std::ptr::copy_nonoverlapping_memory`), so I changed `word-wrap` to allow inline code to break in the middle of words instead of only text within `pre` tags.
This isn't a change included in this PR, but rather something up for discussion. Currently, code blocks that contain long lines simply wrap onto the next line. What if rather than wrapping, they were horizontally scrollable instead? I'm not sure what the general preference is regarding this.
r? @steveklabnik
