Matthias Krüger [Sun, 19 Dec 2021 09:45:51 +0000 (10:45 +0100)]
Rollup merge of #91984 - Aaron1011:rustc-middle-lifetime, r=oli-obk
Remove `in_band_lifetimes` from `rustc_middle`
See #91867
This was mostly straightforward. In several places, I take advantage
of the fact that lifetimes are non-hygenic: a macro declares the
'tcx' lifetime, which is then used in types passed in as macro
arguments.
bors [Sun, 19 Dec 2021 06:26:07 +0000 (06:26 +0000)]
Auto merge of #91871 - camelid:urlpartsbuilder, r=GuillaumeGomez
rustdoc: Add `UrlPartsBuilder`
This is a type for efficiently and easily constructing the part of a URL
after the domain: `nightly/core/str/struct.Bytes.html`.
It allows simplifying some code and avoiding some allocations in the
`href_*` functions.
It will also allow making `Cache.paths` et al. use `Symbol` without
having to allocate `String`s in the `href_*` functions. `String`s would
be necessary otherwise because `Symbol::as_str()` returns `SymbolStr`,
whose `Deref<Target = str>` impl requires the `str` to not outlive it.
This is the primary motivation for the addition of `UrlPartsBuilder`.
bors [Sun, 19 Dec 2021 03:33:19 +0000 (03:33 +0000)]
Auto merge of #89831 - Aaron1011:project-caching-speedup, r=jackh726
Re-introduce concept of projection cache 'completion'
Instead of clearing out the cache entirely, we store
the intermediate evaluation result into the cache entry.
This accomplishes several things:
* We avoid the performance hit associated with re-evaluating
the sub-obligations
* We avoid causing issues with incremental compilation, since
the final evaluation result is always the same
* We avoid affecting other uses of the same `InferCtxt` which
might care about 'side effects' from processing the sub-obligations
(e,g. region constraints). Only code that is specifically aware
of the new 'complete' code is affected
Aaron Hill [Sat, 9 Oct 2021 16:29:39 +0000 (11:29 -0500)]
Re-introduce concept of projection cache 'completion'
Instead of clearing out the cache entirely, we store
the intermediate evaluation result into the cache entry.
This accomplishes several things:
* We avoid the performance hit associated with re-evaluating
the sub-obligations
* We avoid causing issues with incremental compilation, since
the final evaluation result is always the same
* We avoid affecting other uses of the same `InferCtxt` which
might care about 'side effects' from processing the sub-obligations
(e,g. region constraints). Only code that is specifically aware
of the new 'complete' code is affected
Matthias Krüger [Sat, 18 Dec 2021 23:38:44 +0000 (00:38 +0100)]
Rollup merge of #92082 - jyn514:remove-from-iterator, r=jyn514
rustdoc: Write doc-comments directly instead of using FromIterator
The FromIterator impl made the code much harder to understand. The types
don't make sense until you realize there's a custom FromIterator impl.
This is the first commit from https://github.com/rust-lang/rust/pull/91305; since ``@camelid`` wrote it originally I don't feel bad unilaterally approving it.
r? ``@ghost``
``@bors`` r+
Note that this will conflict with https://github.com/rust-lang/rust/pull/92078.
Matthias Krüger [Sat, 18 Dec 2021 23:38:41 +0000 (00:38 +0100)]
Rollup merge of #91932 - Kixiron:randomize-seed, r=nagisa
Add user seed to `-Z randomize-layout`
Allows users of -`Z randomize-layout` to provide `-Z layout-seed=<seed>` in order to further randomizing type layout randomization. Extension of [compiler-team/#457](https://github.com/rust-lang/compiler-team/issues/457), allows users to change struct layouts without changing code and hoping that item path hashes change, aiding in detecting layout errors
bors [Sat, 18 Dec 2021 21:23:37 +0000 (21:23 +0000)]
Auto merge of #91837 - Kobzol:stable-hash-map-avoid-sort, r=the8472
Avoid sorting in hash map stable hashing
Suggested by `@the8472` [here](https://github.com/rust-lang/rust/pull/89404#issuecomment-991813333). I hope that I understood it right, I replaced the sort with modular multiplication, which should be commutative.
Can I ask for a perf. run? However, locally it didn't help at all. Creating the `StableHasher` all over again is probably slowing it down quite a lot. And using `FxHasher` is not straightforward, because the keys and values only implement `HashStable` (and probably they shouldn't be just hashed via `Hash` anyway for it to actually be stable).
Maybe the `StableHash` interface could be changed somehow to better suppor these scenarios where the hasher is short-lived. Or the `StableHasher` implementation could have variants with e.g. a shorter buffer for these scenarios.
bors [Sat, 18 Dec 2021 18:08:51 +0000 (18:08 +0000)]
Auto merge of #92065 - matthiaskrgr:rollup-qmpcsuj, r=matthiaskrgr
Rollup of 7 pull requests
Successful merges:
- #91566 (Apply path remapping to DW_AT_GNU_dwo_name when producing split DWARF)
- #91926 (Remove `in_band_lifetimes` from `rustc_metadata`)
- #91931 (Remove `in_band_lifetimes` from `rustc_codegen_llvm`)
- #92024 (rustc_codegen_llvm: Give each codegen unit a unique DWARF name on all platforms, not just Apple ones.)
- #92037 (Use a const ParamEnv when in default_method_body_is_const)
- #92047 (Set `RUST_BACKTRACE=0` when running location-detail tests)
- #92050 (Add a space and 2 grave accents )
bors [Sat, 18 Dec 2021 15:01:32 +0000 (15:01 +0000)]
Auto merge of #92064 - matthiaskrgr:rollup-tgj2pai, r=matthiaskrgr
Rollup of 7 pull requests
Successful merges:
- #91858 (pass -Wl,-z,origin to set DF_ORIGIN when using rpath)
- #91923 (Remove `in_band_lifetimes` from `rustc_query_impl`)
- #91925 (Remove `in_band_lifetimes` from `rustc_privacy`)
- #91977 (Clean up search code and unify function returned values)
- #92018 (Fix typo in "new region bound" suggestion)
- #92022 (Eliminate duplicate codes of expected_found_bool)
- #92032 (hir: Do not introduce dummy type names for `extern` blocks in def paths)
Matthias Krüger [Sat, 18 Dec 2021 13:49:42 +0000 (14:49 +0100)]
Rollup merge of #92037 - fee1-dead:fix_env_dmbic, r=oli-obk
Use a const ParamEnv when in default_method_body_is_const
r? `@oli-obk`
This PR fixes the param_env function to return `constness: Const` correctly for trait methods marked with `#[default_method_body_is_const]`. The snippet below is erroneously accepted by the compiler and has been fixed by this change. ([Playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2021&gist=12dc6681b2eeee5f604203d96259eeb4))
Matthias Krüger [Sat, 18 Dec 2021 13:49:41 +0000 (14:49 +0100)]
Rollup merge of #92024 - pcwalton:per-codegen-unit-names, r=davidtwco
rustc_codegen_llvm: Give each codegen unit a unique DWARF name on all platforms, not just Apple ones.
To avoid breaking split DWARF, we need to ensure that each codegen unit has a
unique `DW_AT_name`. This is because there's a remote chance that different
codegen units for the same module will have entirely identical DWARF entries
for the purpose of the DWO ID, which would violate Appendix F ("Split Dwarf
Object Files") of the DWARF 5 specification. LLVM uses the algorithm specified
in section 7.32 "Type Signature Computation" to compute the DWO ID, which does
not include any fields that would distinguish compilation units. So we must
embed the codegen unit name into the `DW_AT_name`.
Matthias Krüger [Sat, 18 Dec 2021 13:49:40 +0000 (14:49 +0100)]
Rollup merge of #91931 - LegionMammal978:less-inband-codegen_llvm, r=davidtwco
Remove `in_band_lifetimes` from `rustc_codegen_llvm`
See #91867 for more information.
This one took a while. This crate has dozens of functions not associated with any type, and most of them were using in-band lifetimes for `'ll` and `'tcx`.
Matthias Krüger [Sat, 18 Dec 2021 13:49:38 +0000 (14:49 +0100)]
Rollup merge of #91566 - cbeuw:remap-dwo-name, r=davidtwco
Apply path remapping to DW_AT_GNU_dwo_name when producing split DWARF
`--remap-path-prefix` doesn't apply to paths to `.o` (in case of packed) or `.dwo` (in case of unpacked) files in `DW_AT_GNU_dwo_name`. GCC also has this bug https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91888
Matthias Krüger [Sat, 18 Dec 2021 10:28:06 +0000 (11:28 +0100)]
Rollup merge of #92022 - woodenarrow:br_expected_bool, r=estebank
Eliminate duplicate codes of expected_found_bool
The function expected_found_bool is the same as ExpectedFound::new. So use ExpectedFound::new to replace expected_found_bool to eliminate duplicate codes.
Matthias Krüger [Sat, 18 Dec 2021 10:28:01 +0000 (11:28 +0100)]
Rollup merge of #91858 - semarie:runpath, r=petrochenkov
pass -Wl,-z,origin to set DF_ORIGIN when using rpath
DF_ORIGIN flag signifies that the object being loaded may make reference to the $ORIGIN substitution string.
Some implementations are just ignoring [DF_ORIGIN](http://www.sco.com/developers/gabi/latest/ch5.dynamic.html#df_flags) and do [substitution](http://www.sco.com/developers/gabi/latest/ch5.dynamic.html#substitution) for $ORIGIN if present (whatever DF_ORIGIN presence or not) like glibc. But some others mandate the present of DF_ORIGIN for the substitution (like OpenBSD).
Set the flag inconditionally if rpath is wanted.
One possible fallout is if the linker rejects `-z origin` option.
Matthias Krüger [Sat, 18 Dec 2021 09:26:35 +0000 (10:26 +0100)]
Rollup merge of #91439 - ecstatic-morse:const-cmp-trait-default-methods, r=oli-obk
Mark defaulted `PartialEq`/`PartialOrd` methods as const
WIthout it, `const` impls of these traits are unpleasant to write. I think this kind of change is allowed now. although it looks like it might require some Miri tweaks. Let's find out.
bors [Sat, 18 Dec 2021 07:20:50 +0000 (07:20 +0000)]
Auto merge of #92059 - matthiaskrgr:rollup-bainfyh, r=matthiaskrgr
Rollup of 7 pull requests
Successful merges:
- #87901 (Fix suggestion of additional `pub` when using `pub pub fn ...`)
- #89090 (Lint bare traits in AstConv.)
- #91818 (Show the unused type for `unused_results` lint)
- #91910 (miri: lift restriction on extern types being the only field in a struct)
- #91928 (Constify (most) `Option` methods)
- #91975 (Move generator check earlier in inlining.)
- #92016 (builtin_macros: allow external consumers for AsmArgs parsing)
Matthias Krüger [Sat, 18 Dec 2021 07:16:31 +0000 (08:16 +0100)]
Rollup merge of #92016 - calebcartwright:expose-asm-args-parsing, r=Amanieu
builtin_macros: allow external consumers for AsmArgs parsing
As discussed in Zulip (https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/parsing.20of.20AsmArgs.20for.20inline.20assembly), we need a function entry point that rustfmt can leverage from a pre-expansion context to get a more structured representation of the asm args without having to duplicate/maintain the token stream parsing
Matthias Krüger [Sat, 18 Dec 2021 07:16:30 +0000 (08:16 +0100)]
Rollup merge of #91975 - cjgillot:noinline-generator, r=jackh726
Move generator check earlier in inlining.
Inlining into generator may create references to other generators. For instance, inlining `Pin::<&mut from_generator::GenFuture<[generator1]>>::new_unchecked` into `generator2`. This cross reference can then create cycles when computing inlining for `generator1`.
In order to avoid this kind of surprises, we forbid all inlining into generators, and rely on LLVM to do the right thing. The existing `remove-zst-query-cycle` already ICEs in inline-mir mode, so we use it as test.
Matthias Krüger [Sat, 18 Dec 2021 07:16:27 +0000 (08:16 +0100)]
Rollup merge of #91818 - camelid:unused-result-type, r=jackh726
Show the unused type for `unused_results` lint
I think it's helpful to know what type was unused when looking at these
warnings. The type will likely determine whether the result *should* be
used, or whether it should just be ignored.
Including the type also matches the behavior of the `must_use` lint:
unused `SomeType` that must be used.
bors [Fri, 17 Dec 2021 22:12:34 +0000 (22:12 +0000)]
Auto merge of #89841 - cormacrelf:let-else-typed, r=nagisa
Implement let-else type annotations natively
Tracking issue: #87335
Fixes #89688, fixes #89807, edit: fixes #89960 as well
As explained in https://github.com/rust-lang/rust/issues/89688#issuecomment-940405082, the previous desugaring moved the let-else scrutinee into a dummy variable, which meant if you wanted to refer to it again in the else block, it had moved.
This introduces a new hir type, ~~`hir::LetExpr`~~ `hir::Let`, which takes over all the fields of `hir::ExprKind::Let(...)` and adds an optional type annotation. The `hir::Let` is then treated like a `hir::Local` when type checking a function body, specifically:
* `GatherLocalsVisitor` overrides a new `Visitor::visit_let_expr` and does pretty much exactly what it does for `visit_local`, assigning a local type to the `hir::Let` ~~(they could be deduplicated but they are right next to each other, so at least we know they're the same)~~
* It reuses the code in `check_decl_local` to typecheck the `hir::Let`, simply returning 'bool' for the expression type after doing that.
* ~~`FnCtxt::check_expr_let` passes this local type in to `demand_scrutinee_type`, and then imitates check_decl_local's pattern checking~~
* ~~`demand_scrutinee_type` (the blindest change for me, please give this extra scrutiny) uses this local type instead of of creating a new one~~
* ~~Just realised the `check_expr_with_needs` was passing NoExpectation further down, need to pass the type there too. And apparently this Expectation API already exists.~~
Some other misc notes:
* ~~Is the clippy code supposed to be autoformatted? I tried not to give huge diffs but maybe some rustfmt changes simply haven't hit it yet.~~
* in `rustc_ast_lowering/src/block.rs`, I noticed some existing `self.alias_attrs()` calls in `LoweringContext::lower_stmts` seem to be copying attributes from the lowered locals/etc to the statements. Is that right? I'm new at this, I don't know.
bors [Fri, 17 Dec 2021 19:17:29 +0000 (19:17 +0000)]
Auto merge of #91838 - scottmcm:array-slice-eq-via-arrays-not-slices, r=dtolnay
Do array-slice equality via array equality, rather than always via slices
~~Draft because it needs a rebase after #91766 eventually gets through bors.~~
This enables the optimizations from #85828 to be used for array-to-slice comparisons too, not just array-to-array.
For example, <https://play.rust-lang.org/?version=nightly&mode=release&edition=2021&gist=5f9ba69b3d5825a782f897c830d3a6aa>
```rust
pub fn demo(x: &[u8], y: [u8; 4]) -> bool {
*x == y
}
```
Currently writes the array to stack for no reason:
```nasm
sub rsp, 4
mov dword ptr [rsp], edx
cmp rsi, 4
jne .LBB0_1
mov eax, dword ptr [rdi]
cmp eax, dword ptr [rsp]
sete al
add rsp, 4
ret
.LBB0_1:
xor eax, eax
add rsp, 4
ret
```
Whereas with the change in this PR it just compares it directly:
```nasm
cmp rsi, 4
jne .LBB1_1
cmp dword ptr [rdi], edx
sete al
ret
bors [Fri, 17 Dec 2021 10:37:26 +0000 (10:37 +0000)]
Auto merge of #92023 - ehuss:update-cargo, r=ehuss
Update cargo
8 commits in a359ce16073401f28b84840da85b268aa3d37c88..fcef61230c3b6213b6b0d233a36ba4ebd1649ec3
2021-12-14 18:40:22 +0000 to 2021-12-17 02:30:38 +0000
- Minor docs change for `cargo test --help` (rust-lang/cargo#10210)
- Make clippy happy (rust-lang/cargo#10205)
- Enhance descriptions of issue templates (rust-lang/cargo#10202)
- Add workaround for sporadic kills when building on Macos (rust-lang/cargo#10196)
- Detect filesystem loop during walking the projects (rust-lang/cargo#10188)
- Error about not having any crates with documentation (rust-lang/cargo#10204)
- Don't document libs with doc=false (rust-lang/cargo#10201)
- Bumps up tar to 0.4.36 (rust-lang/cargo#10198)
Patrick Walton [Fri, 17 Dec 2021 04:39:41 +0000 (20:39 -0800)]
rustc_codegen_llvm: Give each codegen unit a unique DWARF name on all
platforms, not just Apple ones.
To avoid breaking split DWARF, we need to ensure that each codegen unit has a
unique `DW_AT_name`. This is because there's a remote chance that different
codegen units for the same module will have entirely identical DWARF entries
for the purpose of the DWO ID, which would violate Appendix F ("Split Dwarf
Object Files") of the DWARF 5 specification. LLVM uses the algorithm specified
in section 7.32 "Type Signature Computation" to compute the DWO ID, which does
not include any fields that would distinguish compilation units. So we must
embed the codegen unit name into the `DW_AT_name`.
bors [Thu, 16 Dec 2021 23:08:02 +0000 (23:08 +0000)]
Auto merge of #92003 - matthiaskrgr:rollup-obgv0rt, r=matthiaskrgr
Rollup of 7 pull requests
Successful merges:
- #91172 (Warn when a `#[test]`-like built-in attribute macro is present multiple times.)
- #91796 (Fix since attribute for const_manually_drop feature)
- #91879 (Remove `in_band_lifetimes` from `rustc_borrowck`)
- #91947 (Add `io::Error::other`)
- #91967 (Pull in libdevstat on FreeBSD)
- #91987 (Add module documentation for rustdoc passes)
- #92001 (Fix default_method_body_is_const when used across crates)
bors [Thu, 16 Dec 2021 19:56:45 +0000 (19:56 +0000)]
Auto merge of #91616 - saethlin:sort_unchecked-sb-fix, r=Mark-Simulacrum
Fix #91306 by deriving all access from a single *mut T
Fixes #91306.
The previous code is invalid because the first argument to `copy_nonoverlapping` is invalidated by the mutable borrow taken out to construct the second argument.
I believe this patch fixes that, and this code should now pass Miri with `-Ztag-raw-pointers`, ~~but I'm currently stuck trying to run my reproducer with a this patched version of the standard library (alternatively, running Miri on the standard library tests itself would suffice).~~ Ralf walked me through this on Zulip.
I've also added fixes for 7 more problems other than those I reported. Most of them are easy to hit by calling sort_unstable on random arrays. I don't have reproducers for every change, but they seem pretty clear-cut to me. But I did only start learning stacked borrows 2 days ago so that might be a large dash of Dunning-Kruger.
Matthias Krüger [Thu, 16 Dec 2021 16:23:12 +0000 (17:23 +0100)]
Rollup merge of #91987 - jsha:docdocgoose, r=jyn514
Add module documentation for rustdoc passes
These are currently documented at https://rustc-dev-guide.rust-lang.org/rustdoc-internals.html#hot-potato but can easily go out of date. We'd like to document them in place and link to https://doc.rust-lang.org/nightly/nightly-rustc/rustdoc/passes/index.html
[Discussed on Zulip](https://rust-lang.zulipchat.com/#narrow/stream/266220-rustdoc/topic/moving.20pass.20docs/near/265058351).
Matthias Krüger [Thu, 16 Dec 2021 16:23:10 +0000 (17:23 +0100)]
Rollup merge of #91947 - ibraheemdev:io-error-other, r=joshtriplett
Add `io::Error::other`
This PR adds a small utility constructor, `io::Error::other`, a shorthand for `io::Error::new(io::ErrorKind::Other, err)`, something I find myself writing often.
For some concrete stats, a quick search on [grep.app](https://grep.app) shows that more than half of the uses of `io::Error::new` use `ErrorKind::Other`:
```
Error::new\((?:std::)?(?:io::)?ErrorKind:: => 3,898 results
Error::new\((?:std::)?(?:io::)?ErrorKind::Other => 2,186 results
```
Ben Kimock [Tue, 7 Dec 2021 01:44:02 +0000 (20:44 -0500)]
Fix SB problems in slice sorting
Most of these problems originate in use of get_unchecked_mut.
When calling ptr::copy_nonoverlapping, using get_unchecked_mut for both
arguments causes the borrow created to make the second pointer to invalid the
first.
The pairs of identical MaybeUninit::slice_as_mut_ptr calls similarly
invalidate each other.
There was also a similar borrow invalidation problem with the use of
slice::get_unchecked_mut to derive the pointer for the CopyOnDrop.