bors [Tue, 18 Jan 2022 14:32:52 +0000 (14:32 +0000)]
Auto merge of #92731 - bjorn3:asm_support_changes, r=nagisa
Avoid unnecessary monomorphization of inline asm related functions
This should reduce build time for codegen backends by avoiding duplicated monomorphization of certain inline asm related functions for each passed in closure type.
bors [Tue, 18 Jan 2022 06:07:58 +0000 (06:07 +0000)]
Auto merge of #93021 - matthiaskrgr:rollup-o7z8zoe, r=matthiaskrgr
Rollup of 14 pull requests
Successful merges:
- #92629 (Pick themes on settings page, not every page)
- #92640 (Fix ICEs related to `Deref<Target=[T; N]>` on newtypes)
- #92701 (Add some more attribute validation)
- #92803 (Hide mobile sidebar on some clicks)
- #92830 (Rustdoc style cleanups)
- #92866 ("Does exists" typos fix)
- #92870 (add `rustc_diagnostic_item` attribute to `AtomicBool` type)
- #92914 (htmldocck: Add support for `/text()` in ``@snapshot`)`
- #92923 (Abstract the pretty printer's ringbuffer to be infinitely sized)
- #92946 (Exclude llvm-libunwind from the self-contained set on s390x-musl targets)
- #92947 (rustdoc: Use `intersperse` in a `visit_path` function)
- #92997 (Add `~const` bound test for negative impls)
- #93004 (update codegen test for LLVM 14)
- #93016 (Stabilize vec_spare_capacity)
Matthias Krüger [Tue, 18 Jan 2022 03:42:08 +0000 (04:42 +0100)]
Rollup merge of #92947 - vacuus:rustdoc-core-visit-path, r=camelid
rustdoc: Use `intersperse` in a `visit_path` function
(~~Is there a better way to word the title?~~ Eh, this works, I guess.)
I'm surprised that the compiler didn't complain when I left out the `.to_string()`, but hey, if it works then it works.
Matthias Krüger [Tue, 18 Jan 2022 03:42:06 +0000 (04:42 +0100)]
Rollup merge of #92923 - dtolnay:ringbuffer, r=petrochenkov
Abstract the pretty printer's ringbuffer to be infinitely sized
This PR backports https://github.com/dtolnay/prettyplease/commit/8e5e83c3ff7faaa272115a81b76f07a1d4a91969 from the `prettyplease` crate into `rustc_ast_pretty`.
Using a dedicated RingBuffer type with non-wrapping indices, instead of manually `%`-ing indices into a capped sized buffer, unlocks a number of simplifications to the pretty printing algorithm implementation in followup commits such as https://github.com/dtolnay/prettyplease/commit/fcb5968b1e1357497c7d7e205d22bb4e8c1dbd38 and https://github.com/dtolnay/prettyplease/commit/4427cedcb80294e89b823f923fdb16544c737bd9.
This change also greatly reduces memory overhead of the pretty printer. The old implementation always grows its buffer to 205920 bytes even for files without deeply nested code, because it only wraps its indices when they hit the maximum tolerable size of the ring buffer (the size after which the pretty printer will crash if there are that many tokens buffered). In contrast, the new implementation uses memory proportional to the peak number of simultaneously buffered tokens only, not the number of tokens that have ever been in the buffer.
Speaking of crashing the pretty printer and "maximum tolerable size", the constant used for that in the old implementation is a lie:
It was raised from 3 to 55 in https://github.com/rust-lang/rust/pull/33934 because that was empirically the size that avoided crashing on one particular test crate, but according to https://github.com/rust-lang/rust/pull/33934#issuecomment-226700470 other syntax trees still crash at that size. There is no reason to believe that any particular size is good enough for arbitrary code, and using a large number like 55 adds overhead to inputs that never need close to that much of a buffer. The new implementation eliminates this tradeoff.
Matthias Krüger [Tue, 18 Jan 2022 03:42:02 +0000 (04:42 +0100)]
Rollup merge of #92830 - jsha:style-cleanups, r=GuillaumeGomez
Rustdoc style cleanups
- Make "since" version numbers grey again (regressed in #92602).
- Remove unneeded selectors for when crate filter dropdown is a
sibling of search-input.
- Crate filter dropdown doesn't need to be 100% width on mobile.
- Only build crate filter dropdown when there is more than one crate.
- Remove unused addCrateDropdown
Matthias Krüger [Tue, 18 Jan 2022 03:41:58 +0000 (04:41 +0100)]
Rollup merge of #92640 - compiler-errors:array-deref-on-newtype, r=lcnr
Fix ICEs related to `Deref<Target=[T; N]>` on newtypes
1. Stash a const infer's type into the canonical var during canonicalization, so we can recreate the fresh const infer with that same type.
For example, given `[T; _]` we know `_` is a `usize`. If we go from infer => canonical => infer, we shouldn't forget that variable is a usize.
Fixes #92626
Fixes #83704
2. Don't stash the autoderef'd slice type that we get from method lookup, but instead recreate it during method confirmation. We need to do this because the type we receive back after picking the method references a type variable that does not exist after probing is done.
Fixes #92637
... A better solution for the second issue would be to actually _properly_ implement `Deref` for `[T; N]` instead of fixing this autoderef hack to stop leaking inference variables. But I actually looked into this, and there are many complications with const impls.
Matthias Krüger [Tue, 18 Jan 2022 03:41:58 +0000 (04:41 +0100)]
Rollup merge of #92629 - jsha:theme-picker-local-only-2, r=GuillaumeGomez
Pick themes on settings page, not every page
This hides the paintbrush icon on most pages by default, in preference for the settings on the settings page. When loading from a local file, and not in mobile view, continue to show the theme picker. That's because some browsers limit access to localStorage from file:/// URLs, so choosing a theme from settings.html doesn't take effect.
bors [Tue, 18 Jan 2022 02:32:36 +0000 (02:32 +0000)]
Auto merge of #93001 - flip1995:clippyup, r=Manishearth
Out of cycle Clippy update
I want to do an out-of-cycle sync for rust-lang/rust-clippy#8295, and possibly backport this to stable together with https://github.com/rust-lang/rust/issues/92938. If this doesn't get backported to stable, then I at least want to backport it to beta.
bors [Mon, 17 Jan 2022 23:01:35 +0000 (23:01 +0000)]
Auto merge of #93009 - matthiaskrgr:rollup-3fkxg6i, r=matthiaskrgr
Rollup of 10 pull requests
Successful merges:
- #90498 (Clarifications in the target tier policy)
- #92164 (Implement `#[rustc_must_implement_one_of]` attribute)
- #92729 (rustc_codegen_llvm: Remove (almost) unused span parameter from many functions in metadata.rs)
- #92752 (Correct minor typos in some long error code explanations)
- #92801 (Enable wrapping words by default)
- #92825 (Rename environment variable for overriding rustc version)
- #92877 (Remove LLVMRustMarkAllFunctionsNounwind)
- #92936 (rustdoc: Remove `collect` in `html::markdown::parse`)
- #92956 (Add `log2` and `log10` to `NonZeroU*`)
- #92960 (Use `carrying_{mul|add}` in `num::bignum`)
kadmin [Mon, 10 Jan 2022 23:39:21 +0000 (23:39 +0000)]
Update term for use in more places
Replace use of `ty()` on term and use it in more places. This will allow more flexibility in the
future, but slightly worried it allows items which are consts which only accept types.
- Make "since" version numbers grey again (regressed in #92602).
- Remove unneeded selectors for when crate filter dropdown is a
sibling of search-input.
- Crate filter dropdown doesn't need to be 100% width on mobile.
- Only build crate filter dropdown when there is more than one crate.
- Remove unused addCrateDropdown.
Matthias Krüger [Mon, 17 Jan 2022 19:07:07 +0000 (20:07 +0100)]
Rollup merge of #92877 - Amanieu:remove_llvm_nounwind, r=Mark-Simulacrum
Remove LLVMRustMarkAllFunctionsNounwind
This was originally introduced in #10916 as a way to remove all landing
pads when performing LTO. However this is no longer necessary today
since rustc properly marks all functions and call-sites as nounwind
where appropriate.
In fact this is incorrect in the presence of `extern "C-unwind"` which
must create a landing pad when compiled with `-C panic=abort` so that
foreign exceptions are caught and properly turned into aborts.
Matthias Krüger [Mon, 17 Jan 2022 19:07:05 +0000 (20:07 +0100)]
Rollup merge of #92801 - jsha:overflow-wrap, r=GuillaumeGomez
Enable wrapping words by default
Faced with a very long word, browsers will let it overflow its
box horizontally rather than break it in the middle. We essentially
never want that behavior. We would rather break the word and keep it
inside its horizontal limits. So we apply a default overflow-wrap:
break-word/anywhere to the document as a while.
In some contexts we would rather add a horizontal scrollbar (code
blocks), or elide the excess text with an ellipsis (sidebar). Those
still work as expected.
Fixes #92771
[Some related discussion](https://rust-lang.zulipchat.com/#narrow/stream/266220-rustdoc/topic/wrap.20.2F.20overflow.20.2F.20scroll) and a related issue: #92421.
Matthias Krüger [Mon, 17 Jan 2022 19:07:03 +0000 (20:07 +0100)]
Rollup merge of #92729 - michaelwoerister:remove-unused-span-debuginfo, r=petrochenkov
rustc_codegen_llvm: Remove (almost) unused span parameter from many functions in metadata.rs
Many functions and intermediate data structures in `rustc_codegen_llvm/src/debuginfo/metadata.rs` take a span parameter that is only used for providing a span to a `span_bug!()` invocation, in case the debuginfo typemap gets corrupted. However, this span does not really convey useful information as it just points to the first point a type is used -- and half of the time is initialized to `DUMMY_SP`.
This PR removes this span parameter from the module.
It also removes the following unused parameters from `composite_type_metadata()` together with an outdated comment:
```rust
// Ignore source location information as long as it
// can't be reconstructed for non-local crates.
_file_metadata: &'ll DIFile,
_definition_span: Span,
```
This PR adds a new attribute — `#[rustc_must_implement_one_of]` that allows changing the "minimal complete definition" of a trait. It's similar to GHC's minimal `{-# MINIMAL #-}` pragma, though `#[rustc_must_implement_one_of]` is weaker atm.
Such attribute was long wanted. It can be, for example, used in `Read` trait to make transitions to recently added `read_buf` easier:
```rust
#[rustc_must_implement_one_of(read, read_buf)]
pub trait Read {
fn read(&mut self, buf: &mut [u8]) -> Result<usize> {
let mut buf = ReadBuf::new(buf);
self.read_buf(&mut buf)?;
Ok(buf.filled_len())
}
impl Read for Ty0 {}
//^ This will fail to compile even though all `Read` methods have default implementations
// Both of these will compile just fine
impl Read for Ty1 {
fn read(&mut self, buf: &mut [u8]) -> Result<usize> { /* ... */ }
}
impl Read for Ty2 {
fn read_buf(&mut self, buf: &mut ReadBuf<'_>) -> Result<()> { /* ... */ }
}
```
For now, this is implemented as an internal attribute to start experimenting on the design of this feature. In the future we may want to extend it:
- Allow arbitrary requirements like `a | (b & c)`
- Allow multiple requirements like
- ```rust
#[rustc_must_implement_one_of(a, b)]
#[rustc_must_implement_one_of(c, d)]
```
- Make it appear in rustdoc documentation
- Change the syntax?
- Etc
Eventually, we should make an RFC and make this (or rather similar) attribute public.
---
I'm fairly new to compiler development and not at all sure if the implementation makes sense, but at least it passes tests :)
Matthias Krüger [Mon, 17 Jan 2022 19:07:01 +0000 (20:07 +0100)]
Rollup merge of #90498 - joshtriplett:target-tier-policy-draft-updates, r=Mark-Simulacrum
Clarifications in the target tier policy
We've added several targets since the introduction of the target tier policy. Based on experiences of those adding such targets, and discussions around such additions, clarify the target tier policy to make it easier to follow and work with.
None of these changes substantively change the requirements on targets. (In some cases the changes do direct target submitters to follow specific process requirements for the addition of a target, such as how to respond to requirements, where to put target-specific documentation, or what should appear in that documentation. Those changes are procedural in nature and document the procedures we already direct people to follow.)
- Clarify how to quote and respond to the target tier policy requirements. Several times, people have seemed unclear on how to respond to some of the policy requirements, particularly those that just state things the target developers must *not* do (e.g. not posting to PRs that break the target). Add a note that such requirements just need acknowledgement, nothing more.
- Clarify dependency requirements in the face of cross-compilation. I previously phrased this confusingly in terms of "host tools", since that is the case where an exception applies (allowing proprietary target libraries commonly used by binaries for the target). Rephrase it to apply equally to cross-compilation. This doesn't change the net effect of the requirements, since other requirements already cover the dependencies of the Rust toolchain.
- Clarify documentation about running binaries. The requirement for target documentation talks about "running tests", but tier 3 targets often don't support running the full testsuite, and in practice the documentation for how to run an individual binary may be more useful. Change "running tests" to "running binaries, or running tests".
- Explain where to place target-specific documentation (a subdirectory of platform-support, with a link from the platform-support entry for the target).
- Add a template for target-specific documentation.
kadmin [Sat, 8 Jan 2022 09:28:12 +0000 (09:28 +0000)]
Use Term in ProjectionPredicate
ProjectionPredicate should be able to handle both associated types and consts so this adds the
first step of that. It mainly just pipes types all the way down, not entirely sure how to handle
consts, but hopefully that'll come with time.
bors [Mon, 17 Jan 2022 14:50:50 +0000 (14:50 +0000)]
Auto merge of #90986 - camsteffen:nested-filter, r=cjgillot
Replace `NestedVisitorMap` with generic `NestedFilter`
This is an attempt to make the `intravisit::Visitor` API simpler and "more const" with regard to nested visiting.
With this change, `intravisit::Visitor` does not visit nested things by default, unless you specify `type NestedFilter = nested_filter::OnlyBodies` (or `All`). `nested_visit_map` returns `Self::Map` instead of `NestedVisitorMap<Self::Map>`. It panics by default (unreachable if `type NestedFilter` is omitted).
One somewhat trixty thing here is that `nested_filter::{OnlyBodies, All}` live in `rustc_middle` so that they may have `type Map = map::Map` and so that `impl Visitor`s never need to specify `type Map` - it has a default of `Self::NestedFilter::Map`.
bors [Mon, 17 Jan 2022 09:40:29 +0000 (09:40 +0000)]
Auto merge of #92816 - tmiasko:rm-llvm-asm, r=Amanieu
Remove deprecated LLVM-style inline assembly
The `llvm_asm!` was deprecated back in #87590 1.56.0, with intention to remove
it once `asm!` was stabilized, which already happened in #91728 1.59.0. Now it
is time to remove `llvm_asm!` to avoid continued maintenance cost.
Matthias Krüger [Mon, 17 Jan 2022 05:08:19 +0000 (06:08 +0100)]
Rollup merge of #92981 - RalfJung:const_ptr_offset_from, r=dtolnay
fix const_ptr_offset_from tracking issue
The old tracking issue #41079 was for exposing those functions at all, and got closed when they were stabilized. We had nothing tracking their `const`ness so I opened a new tracking issue: https://github.com/rust-lang/rust/issues/92980.
Matthias Krüger [Mon, 17 Jan 2022 05:08:18 +0000 (06:08 +0100)]
Rollup merge of #92977 - kornelski:popdoc, r=dtolnay
Docs: recommend VecDeque instead of Vec::remove(0)
Suggestion based on a [discussion](https://internals.rust-lang.org/t/should-vec-have-a-try-remove-mut-self-usize-option-t-function/15964/9?u=kornel) where user needlessly struggled with `remove(0)` and accidentally created a quadratic cost.
Matthias Krüger [Mon, 17 Jan 2022 05:08:17 +0000 (06:08 +0100)]
Rollup merge of #92953 - azdavis:azdavis-copy-example, r=dtolnay
Copy an example to PartialOrd as well
In https://github.com/rust-lang/rust/pull/88202 I added an example for deriving PartialOrd on enums, but only later did I realize that I actually put the example on Ord.
This copies the example to PartialOrd as well, which is where I intended for it to be.
We could also delete the example on Ord, but I see there's already some highly similar examples shared between Ord and PartialOrd, so I figured we could leave it.
I also changed some type annotations in an example from `x : T` to the more common style (in Rust) of `x: T`.
Matthias Krüger [Mon, 17 Jan 2022 05:08:13 +0000 (06:08 +0100)]
Rollup merge of #92808 - compiler-errors:wrap-struct-shorthand-field-in-variant, r=davidtwco
Fix `try wrapping expression in variant` suggestion with struct field shorthand
Fixes a broken suggestion: [playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2021&gist=83fe2dbfe1485f8cfca1aef2a6582e77)
before:
```
error[E0308]: mismatched types
--> src/main.rs:7:19
|
7 | let x = Foo { bar };
| ^^^ expected enum `Option`, found integer
|
= note: expected enum `Option<i32>`
found type `{integer}`
help: try wrapping the expression in `Some`
|
7 | let x = Foo { Some(bar) };
| +++++ +
```
after:
```
error[E0308]: mismatched types
--> src/main.rs:7:19
|
7 | let x = Foo { bar };
| ^^^ expected enum `Option`, found integer
|
= note: expected enum `Option<i32>`
found type `{integer}`
help: try wrapping the expression in `Some`
|
7 | let x = Foo { bar: Some(bar) };
| ~~~~~~~~~~~~~~
```
r? ``@m-ou-se``
since you touched the code last in #91080
bors [Mon, 17 Jan 2022 02:06:54 +0000 (02:06 +0000)]
Auto merge of #92473 - petrochenkov:ltrattr2, r=Aaron1011
expand: Pick `cfg`s and `cfg_attrs` one by one, like other attributes
This is a rebase of https://github.com/rust-lang/rust/pull/83354, but without any language-changing parts ~(except for https://github.com/rust-lang/rust/pull/84110)~, i.e. the attribute expansion order is the same.
This is a pre-requisite for any other changes making cfg attributes closer to regular macro attributes
- Possibly changing their expansion order (https://github.com/rust-lang/rust/issues/83331)
- Keeping macro backtraces for cfg attributes, or otherwise making them visible after expansion without keeping them in place literally (https://github.com/rust-lang/rust/pull/84110).
Two exceptions to the "one by one" behavior are:
- cfgs eagerly expanded by `derive` and `cfg_eval`, they are still expanded in a batch, that's by design.
- cfgs at the crate root, they are currently expanded not during the main expansion pass, but before that, during `#![feature]` collection. I'll try to disentangle that logic later in a separate PR.
Matthias Krüger [Sun, 16 Jan 2022 15:58:18 +0000 (16:58 +0100)]
Rollup merge of #92746 - estebank:question-mark-in-type, r=davidtwco
Parse `Ty?` as `Option<Ty>` and provide structured suggestion
Swift has specific syntax that desugars to `Option<T>` similar to our
`?` operator, which means that people might try to use it in Rust. Parse
it and gracefully recover.
Matthias Krüger [Sun, 16 Jan 2022 15:58:17 +0000 (16:58 +0100)]
Rollup merge of #92710 - jackh726:issue-92280, r=nikomatsakis
Include Projections when elaborating TypeOutlives
Fixes #92280
In `Elaborator`, we elaborate that `Foo<<Bar as Baz>::Assoc>: 'a` -> `<Bar as Baz>::Assoc: 'a`. This is the same rule that would be applied to any other `Param`. If there are escaping vars, we continue to do nothing.
Matthias Krüger [Sun, 16 Jan 2022 15:58:16 +0000 (16:58 +0100)]
Rollup merge of #92706 - umanwizard:btree, r=dtolnay
Clarify explicitly that BTree{Map,Set} are ordered.
One of the main reasons one would want to use a BTree{Map,Set} rather than a Hash{Map,Set} is because they maintain their keys in sorted order; but this was never explicitly stated in the top-level docs (it was only indirectly alluded to there, and stated explicitly in the docs for `iter`, `values`, etc.)
This PR states the ordering guarantee more prominently.
Matthias Krüger [Sun, 16 Jan 2022 15:58:14 +0000 (16:58 +0100)]
Rollup merge of #92619 - Alexendoo:macro-diagnostic-items, r=matthewjasper
Add diagnostic items for macros
For use in Clippy, it adds diagnostic items to all the stable public macros
Clippy has lints that look for almost all of these (currently by name or path), but there are a few that aren't currently part of any lint, I could remove those if it's preferred to add them as needed rather than ahead of time
bors [Sun, 16 Jan 2022 14:25:25 +0000 (14:25 +0000)]
Auto merge of #92245 - petrochenkov:cmrval, r=nagisa
rustc_metadata: Switch all decoder methods from vectors to iterators
To avoid allocations in some cases.
Also remove unnecessary `is_proc_macro_crate` checks from decoder, currently the general strategy is to shift all the work to the encoder and assume that all the encoded data is correct and can be decoded unconditionally in the decoder.