ci: move most of the prepare config into scripts
This PR moves most of the configuration from the CI yamls into bash scripts, driven by a small Python script (which understands and emulates the two `##vso[` commands we use).
There are two reasons why we'd want to do this:
* Being able to prepare the build environment locally by just running `src/ci/prepare.py` simplifies a lot setting up a local VM similar to CI (software pre-installed in the CI images won't be prepared, but it's a start anyway).
* When we'll switch to GitHub Actions we'll need to either duplicate code in multiple workflows or write a preprocessor. Having all the prepare steps in a single one is going to simplify the implementation of both options.
Along with the move I did a few changes to the actual scripts:
* Mirrored all the remaining external URLs we download (except chocolatey) to the `rust-lang-ci-mirrors` bucket, to increase reliability and reduce the chance of supply chain attacks. I didn't audit and mirror the CI scripts outside this PR though.
* Extracted CI-specific behavior (like issuing `##vso[` commands and detecting the host platform) into `shared.sh` and included it in most of the scripts. This way a switch to another CI provider will be less painful.
It's possible (and easier) to review this commit-by-commit.
r? @alexcrichton
cc @rust-lang/infra
Before this commit, we had an event that would only track the compression step
for proc-macros and Rust dylibs. After the commit we measure the time for
acutally generating the crate metadata bytes.
Stabilize `const_constructor`
# Stabilization proposal
I propose that we stabilize `#![feature(const_constructor)]`.
Tracking issue: https://github.com/rust-lang/rust/issues/61456
Version target: 1.40 (2019-11-05 => beta, 2019-12-19 => stable).
## What is stabilized
### User guide
Tuple struct and tuple variant constructors are now considered to be constant functions. As such a call expression where the callee has a tuple struct or variant constructor "function item" type can be called:
```rust
const fn make_options() {
// These already work because they are special cased:
Some(0);
(Option::Some)(1);
// These also work now:
let f = Option::Some;
f(2);
{Option::Some}(3);
<Option<_>>::Some(5);
}
```
### Motivation
Consistency with other `const fn`. Consistency between syntactic path forms.
This should also ensure that constructors implement `const Fn` traits and can be coerced to `const fn` function pointers, if they are introduced.
## Tests
* [ui/consts/const_constructor/const-construct-call.rs](0d75ab2293/src/test/ui/consts/const_constructor/const-construct-call.rs) - Tests various syntactic forms, use in both `const fn` and `const` items, and constructors in both the current and extern crates.
* [ui/consts/const_constructor/const_constructor_qpath.rs](1850dfcdab/src/test/ui/consts/const_constructor/const_constructor_qpath.rs) - Tests that type qualified paths to enum variants are also considered to be `const fn`.(#64247)
r? @oli-obk
Closes#61456
Closes #64247
Gather together usefulness tests
I took most tests that were testing only for match exhaustiveness, pattern refutability or match arm reachability, and put them in the same test folder. I found it helpful to have them all in the same place when working on the usefulness algorithm.
librustc_lexer: Enhance documentation
This PR enhances documentation state of the `librustc_lexer` (as initiative caused by [rustc-guide#474](https://github.com/rust-lang/rustc-guide/issues/474)), by adding:
- Module documentation.
- Doc-comments (and a bit of usual comments) in non-obvious (as for me) places.
r? @petrochenkov
cc @Centril
Stabilize `Option::flatten`
- PR: https://github.com/rust-lang/rust/pull/60256
- Tracking issue: https://github.com/rust-lang/rust/issues/60258
@elahn
> I was trying to `flat_map()` and found `map().flatten()` does the trick. This has been on nightly for 4 months, can we stabilise it?
@ethanboxx
> @Centril Helped me get this merged. What is the stabilization process?
@Centril
> @ethanboxx I'd just file a PR to stabilize it and we'll ask T-libs to FCP.
So here I am.
I am was unsure what number to put in `since = "-"` so I copied what someone had done in a recent PR.
I took most tests that were testing only for match exhaustiveness,
pattern refutability or match arm reachability, and put them in
the same test folder.
This allows us to remove `static_panic_msg` from the SSA<->LLVM
boundary, along with its fat pointer representation for &str.
Also changes the signature of PanicInfo::internal_contructor to
avoid copying.
Closes#65856.
Point at the span for the definition of ADTs internal to the current
crate.
Look at the leading char of the ident to determine whether we're
expecting a likely fn or any of a fn, a tuple struct or a tuple variant.
Turn fn `add_typo_suggestion` into a `Resolver` method.
Rollup of 6 pull requests
Successful merges:
- #65566 (Use heuristics to suggest assignment)
- #65738 (Coherence should allow fundamental types to impl traits when they are local)
- #65777 (Don't ICE for completely unexpandable `impl Trait` types)
- #65834 (Remove lint callback from driver)
- #65839 (Clean up `check_consts` now that new promotion pass is implemented)
- #65855 (Add long error explaination for E0666)
Failed merges:
r? @ghost
Clean up `check_consts` now that new promotion pass is implemented
`check_consts::resolver` contained a layer of abstraction (`QualifResolver`) to allow the existing, eager style of qualif propagation to work with either a dataflow results cursor or by applying the transfer function directly (if dataflow was not needed e.g. for promotion). However, #63812 uses a different, lazy paradigm for checking promotability, which makes this unnecessary. This PR cleans up `check_consts::validation` to use `FlowSensitiveResolver` directly, instead of through the now obselete `QualifResolver` API.
Also, this contains a few commits (the first four) that address some FIXMEs in #63812 regarding code duplication. They could be split out, but I think they will be relatively noncontroversial? Notably, `validation::Mode` is renamed to `ConstKind` and used in `promote_consts` to denote what kind of item we are in.
This is best reviewed commit-by-commit and is low priority.
r? @eddyb
Remove lint callback from driver
This is leftover from a restructuring of lint registration for drivers; it should now happen via the register_lints field on Config rather than this function.
This is not used by anyone to my knowledge (including the compiler itself); it was introduced in an abandoned refactor in #65193.
Don't ICE for completely unexpandable `impl Trait` types
Save the resolution of these types (to themselves) to the typeck tables so that they will eventually reach E0720.
closes#65561
Coherence should allow fundamental types to impl traits when they are local
After #64414, `impl<T> Remote for Box<T> { }` is disallowed, but it is also disallowed in liballoc, where `Box` is a local type!
Enabling `#![feature(re_rebalance_coherence)]` in `liballoc` results in:
```
error[E0210]: type parameter `F` must be used as the type parameter for some local type (e.g., `MyStruct<F>`)
--> src\liballoc\boxed.rs:1098:1
|
1098 | impl<F: ?Sized + Future + Unpin> Future for Box<F> {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type parameter `F` must be used as the type parameter for some local type
```
This PR relaxes `uncover_fundamental_ty` to skip local fundamental types.
I didn't add a test since `liballoc` already fails to compile, but I can add one if needed.
r? @nikomatsakis
cc #63599
Use heuristics to suggest assignment
When detecting a possible `=` -> `:` typo in a `let` binding, suggest
assigning instead of setting the type.
Partially address #57828.
rustc: add `Span`s to `inferred_outlives_of` predicates.
This would simplify #59789, and I suspect it has some potential in diagnostics (although we don't seem to use the predicate `Span`s much atm).
trait-based structural match implementation
Moves from using a `#[structural_match]` attribute to using a marker trait (or pair of such traits, really) instead.
Fix#63438.
(This however does not remove the hacks that I believe were put into place to support the previous approach of injecting the attribute based on the presence of both derives... I have left that for follow-on work.)
