Aaron Hill pointed out that unnecessary parens around a macro call
(paradigmatically, `format!`) yielded a suggestion of hideous
macro-expanded code. (The slightly unusual choice of using the
pretty-printer to compose suggestions was quite recently commented on
in the commit message for 1081bbbfc ("abolish ICE when pretty-printing
async block"), but without any grounds to condemn it as a 𝘣𝘢𝘥
choice. Hill's report provides the grounds.) `span_to_snippet` is
fallable as far as the type system is concerned (because, who knows,
macros or something), so the pretty-printing can live on in the
oft-neglected `else` branch.
Resolves#55109.
resolve: Scale back hard-coded extern prelude additions on 2015 edition
https://github.com/rust-lang/rust/pull/54404 stabilized `feature(extern_prelude)` on 2015 edition, including the hard-coded parts not passed with `--extern`.
First of all, I'd want to confirm that this is intended stabilization, rather than a part of the "extended beta" scheme that's going to be reverted before releasing stable.
(EDIT: to clarify - this is a question, I'm \*asking\* for confirmation, rather than give it.)
Second, on 2015 edition extern prelude is not so fundamentally tied to imports and is a mere convenience, so this PR scales them back to the uncontroversial subset.
The "uncontroversial subset" means that if libcore is injected it brings `core` into prelude, if libstd is injected it brings `std` and `core` into prelude.
On 2015 edition this can be implemented through the library prelude (rather than hard-coding in the compiler) right now, I'll do it in a follow-up PR.
UPDATE: The change is done for both 2015 and 2018 editions now as discussed below.
Closes https://github.com/rust-lang/rust/issues/53166
nll type annotations in multisegment path
This turned out to be sort of tricky. The problem is that if you have a path like
```
<Foo<&'static u32>>::bar
```
and it comes from an impl like `impl<T> Foo<T>` then the self-type the user gave doesn't *directly* map to the substitutions that the impl wants. To handle this, then, we have to preserve not just the "user-given substs" we used to do, but also a "user-given self-ty", which we have to apply later. This PR makes those changes.
It also removes the code from NLL relate-ops that handled canonical variables and moves to use normal inference variables instead. This simplifies a few things and gives us a bit more flexibility (for example, I predict we are going to have to start normalizing at some point, and it would be easy now).
r? @matthewjasper -- you were just touching this code, do you feel comfortable reviewing this?
Fixes#54574
This commit updates the test output for the updated NLL compare mode
that uses `-Z borrowck=migrate` rather than `-Z borrowck=mir`. The
previous commit changes `compiletest` and this commit only updates
`.nll.stderr` files.
This commit changes the NLL compare mode to pass `-Z borrowck=migrate`
rather than `-Z borrowck=nll` to better test what will be deployed. It
does not include the test output updates, as separation of these commits
makes reviewing simpler.
When lints are emitted from the AST borrow checker, they do not signal
an error as it is not known at that time whether, due to attributes,
that lint will error or warn. This means that when lints are buffered
in the MIR they will always be downgraded, as the AST borrowck will not
have been marked as having errored, even if a lint was upgraded to
an error after being emitted from the AST borrowck. The simple solution
to this is to not buffer any lints from the MIR borrowck.
Make `bad_style` a silent alias for `nonstandard_style`
Now only `nonstandard_style` is suggested in `rustc -W help`, but `bad_style` will not produce a warning. Closes#41646.
r? @Manishearth
Under the semantics of #54986 (our short term plan), the partial
initialization itself will signal an error. We don't need to add noise
to the output by also complaining about `mut`. (In particular, the
user may well revise their code in a way that does not require `mut`.)
(This makes it a little easier to add instrumentation of the entry and
exit by adding `debug!` at the beginning and end, though note that the
function body *does* use the `?` operator...)
A few iterator-related improvements
- typeck: don't collect into a vector when unnecessary
- create only one vector when winnowing candidates
- change a cloning map to `into_iter`
1. Extract the tests for whether or not we have workable localStorage out into
a helper method, so it can be more easily reused
2. Use it in getCurrentValue() too, for the same reasons, as suggested in code
review
Exit with code 101 on fatal codegen errors
Fixes#54992.
This PR installs a custom fatal error handler that prints the error from LLVM and exits with 101. There should be no visible change in the output from LLVM. This allows distinguishing a fatal LLVM error with a compilation error by exit code.
This PR also modifies the LLVM codegen backend to ICE instead of emitting a fatal error when encountering a LLVM worker thread panic for the same reason.
r? @cuviper
