bors [Sat, 19 Nov 2022 12:44:57 +0000 (12:44 +0000)]
Auto merge of #104571 - clubby789:remove-vec-rc-opt, r=the8472
Revert Vec/Rc storage reuse opt
Remove the optimization for using storage added by #104205.
The perf wins were pretty small, and it relies on non-guarenteed behaviour. On platforms that don't implement shrinking in place, the performance will be significantly worse.
While it could be gated to platforms that do this (such as GNU), I don't think it's worth the overhead of maintaining it for very small gains. (#104565, #104563)
Dylan DPC [Sat, 19 Nov 2022 06:24:45 +0000 (11:54 +0530)]
Rollup merge of #104553 - mwillsey:asinh-acosh-accuracy, r=thomcc
Improve accuracy of asinh and acosh
This PR addresses the inaccuracy of `asinh` and `acosh` identified by the [Herbie](http://herbie.uwplse.org/) tool, `@pavpanchekha,` `@finnbear` in #104548. It also adds a couple tests that failed in the existing implementations and now pass.
Dylan DPC [Sat, 19 Nov 2022 06:24:44 +0000 (11:54 +0530)]
Rollup merge of #104411 - lcnr:bivariance-nll, r=compiler-errors
nll: correctly deal with bivariance
fixes #104409
when in a bivariant context, relating stuff should always trivially succeed. Also changes the mir validator to correctly deal with higher ranked regions.
Dylan DPC [Sat, 19 Nov 2022 06:24:43 +0000 (11:54 +0530)]
Rollup merge of #104001 - Ayush1325:custom-entry, r=bjorn3
Improve generating Custom entry function
This commit is aimed at making compiler-generated entry functions (Basically just C `main` right now) more generic so other targets can do similar things for custom entry. This was initially implemented as part of https://github.com/rust-lang/rust/pull/100316.
Currently, this moves the entry function name and Call convention to the target spec.
bors [Fri, 18 Nov 2022 23:20:53 +0000 (23:20 +0000)]
Auto merge of #104591 - Manishearth:rollup-b3ser4e, r=Manishearth
Rollup of 8 pull requests
Successful merges:
- #102977 (remove HRTB from `[T]::is_sorted_by{,_key}`)
- #103378 (Fix mod_inv termination for the last iteration)
- #103456 (`unchecked_{shl|shr}` should use `u32` as the RHS)
- #103701 (Simplify some pointer method implementations)
- #104047 (Diagnostics `icu4x` based list formatting.)
- #104338 (Enforce that `dyn*` coercions are actually pointer-sized)
- #104498 (Edit docs for `rustc_errors::Handler::stash_diagnostic`)
- #104556 (rustdoc: use `code-header` class to format enum variants)
Rollup merge of #104338 - compiler-errors:pointer-sized, r=eholk
Enforce that `dyn*` coercions are actually pointer-sized
Implement a perma-unstable, rudimentary `PointerSized` trait to enforce `dyn*` casts are `usize`-sized for now, at least to prevent ICEs and weird codegen issues from cropping up after monomorphization since currently we enforce *nothing*.
This probably can/should be removed in favor of a more sophisticated trait for handling `dyn*` conversions when we decide on one, but I just want to get something up for discussion and experimentation for now.
r? ```@eholk``` cc ```@tmandry``` (though feel free to claim/reassign)
Rollup merge of #104047 - crlf0710:icu_based_list_formatting, r=davidtwco
Diagnostics `icu4x` based list formatting.
This adds a new kind of `DiagnosticArg` and add the ability to convert it to a `FluentValue::Custom`. When emitting fluent output, it makes use of `ListFormatter` from `icu4x` project to convert it to localized version of list following the fluent locale, as a kind of eager translation.
Tested locally with locales like `en`, `ja`, etc, and they work fine. <del>Though neither `zh-CN` nor `zh-Hans` works correctly, it seems they fallback to `und` locale somehow, emitting only comma-based list. I believe this is an internal issue of `icu4x` itself.</del>(Works fine after https://github.com/rust-lang/rust/pull/104047#issuecomment-1304860940)
Rollup merge of #103456 - scottmcm:fix-unchecked-shifts, r=scottmcm
`unchecked_{shl|shr}` should use `u32` as the RHS
The other shift methods, such as https://doc.rust-lang.org/nightly/std/primitive.u64.html#method.checked_shr and https://doc.rust-lang.org/nightly/std/primitive.i16.html#method.wrapping_shl, use `u32` for the shift amount. That's consistent with other things, like `count_ones`, which also always use `u32` for a bit count, regardless of the size of the type.
This PR changes `unchecked_shl` and `unchecked_shr` to also use `u32` for the shift amount (rather than Self).
Rollup merge of #103378 - nagisa:fix-infinite-offset, r=scottmcm
Fix mod_inv termination for the last iteration
On usize=u64 platforms, the 4th iteration would overflow the `mod_gate` back to 0. Similarly for usize=u32 platforms, the 3rd iteration would overflow much the same way.
I tested various approaches to resolving this, including approaches with `saturating_mul` and `widening_mul` to a double usize. Turns out LLVM likes `mul_with_overflow` the best. In fact now, that LLVM can see the iteration count is limited, it will happily unroll the loop into a nice linear sequence.
You will also notice that the code around the loop got simplified somewhat. Now that LLVM is handling the loop nicely, there isn’t any more reasons to manually unroll the first iteration out of the loop (though looking at the code today I’m not sure all that complexity was necessary in the first place).
Rollup merge of #102977 - lukas-code:is-sorted-hrtb, r=m-ou-se
remove HRTB from `[T]::is_sorted_by{,_key}`
Changes the signature of `[T]::is_sorted_by{,_key}` to match `[T]::binary_search_by{,_key}` and make code like https://github.com/rust-lang/rust/issues/53485#issuecomment-885393452 compile.
bors [Fri, 18 Nov 2022 20:26:58 +0000 (20:26 +0000)]
Auto merge of #104573 - matthiaskrgr:rollup-k36ybtp, r=matthiaskrgr
Rollup of 8 pull requests
Successful merges:
- #101162 (Migrate rustc_resolve to use SessionDiagnostic, part # 1)
- #103386 (Don't allow `CoerceUnsized` into `dyn*` (except for trait upcasting))
- #103405 (Detect incorrect chaining of if and if let conditions and recover)
- #103594 (Fix non-associativity of `Instant` math on `aarch64-apple-darwin` targets)
- #104006 (Add variant_name function to `LangItem`)
- #104494 (Migrate GUI test to use functions)
- #104516 (rustdoc: clean up sidebar width CSS)
- #104550 (fix a typo)
Matthias Krüger [Fri, 18 Nov 2022 13:13:38 +0000 (14:13 +0100)]
Rollup merge of #104516 - notriddle:notriddle/flex-basis-sidebar-width, r=GuillaumeGomez
rustdoc: clean up sidebar width CSS
This commit takes advantage of the ability to set [flex-basis] to a specific length instead of setting it to `auto` and changing min-/max-width, and setting flex-grow/-shrink both to 0.
Matthias Krüger [Fri, 18 Nov 2022 13:13:37 +0000 (14:13 +0100)]
Rollup merge of #104006 - flip1995:lang-items-clippy, r=oli-obk
Add variant_name function to `LangItem`
Clippy has an internal lint that checks for the usage of hardcoded def paths and suggests to replace them with a lang or diagnostic item, if possible. This was implemented with a hack, by getting all the variants of the `LangItem` enum and then index into it with the position of the `LangItem` in the `items` list. This is no longer possible, because the `items` list can't be accessed anymore.
Follow up to #103603
cc `@camsteffen`
r? `@oli-obk`
This is blocking the sync between Clippy and Rust. I'm not sure if this is the best solution here, or if I should add a method `items()` to `LanguageItems` and keep the code in Clippy unchanged.
Matthias Krüger [Fri, 18 Nov 2022 13:13:36 +0000 (14:13 +0100)]
Rollup merge of #103594 - maniwani:fix-issue-91417, r=thomcc
Fix non-associativity of `Instant` math on `aarch64-apple-darwin` targets
This is a duplicate of #94100 (since the original author is unresponsive), which resolves #91417.
On `aarch64-apple-darwin` targets, the internal resolution of `Instant` is lower than that of `Duration`, so math between them becomes non-associative with small-enough durations.
This PR makes this target use the standard Unix implementation (where `Instant` has 1ns resolution), but with `CLOCK_UPTIME_RAW` so it still returns the same values as `mach_absolute_time`[^1].
(Edit: I need someone to confirm that this still works, I do not have access to an M1 device.)
Matthias Krüger [Fri, 18 Nov 2022 13:13:35 +0000 (14:13 +0100)]
Rollup merge of #103386 - compiler-errors:no-coerceunsized-to-dynstar, r=eholk
Don't allow `CoerceUnsized` into `dyn*` (except for trait upcasting)
This makes sure we don't accidentally allow coercions like `Box<T>` -> `Box<dyn* Trait>`, or in the case of this ICE, `&T` to `&dyn* Trait`. These coercions don't make sense, at least not via the `CoerceUnsized` trait.
bors [Fri, 18 Nov 2022 06:54:37 +0000 (06:54 +0000)]
Auto merge of #103179 - ferrocene:pa-run-in-run-make, r=jyn514
Fix `src/test/run-make/issue-36710` on cross-compiled targets
This PR fixes the `src/test/run-make/issue-36710` test not working on cross-compiled targets by telling the make infra how to run tests remotely with `remote-test-server`.
This PR includes a revert of #102723 (cc `@pcc),` which disabled that test on all cross-compiled targets.
Matthias Krüger [Thu, 17 Nov 2022 21:33:22 +0000 (22:33 +0100)]
Rollup merge of #104545 - flip1995:diag_item_matches_macro, r=compiler-errors
Readd the matches_macro diag item
This is now used by Clippy
r? `@compiler-errors`
This was removed in #104383. But in the meantime Clippy now makes use of it https://github.com/rust-lang/rust-clippy/blob/dac600e32fce89cb4b05ae6edb0c47982b99eb48/clippy_lints/src/manual_is_ascii_check.rs#L153
---
This is blocking the Clippy sync. (kinda. I could work around it, but I don't want to play ping-pong with this change.)
Matthias Krüger [Thu, 17 Nov 2022 21:33:21 +0000 (22:33 +0100)]
Rollup merge of #104508 - compiler-errors:dyn-return, r=oli-obk
Check `dyn*` return type correctly
In `check_fn`, if the declared return type is `dyn Trait`, then we check the return type separately to produce better diagnostics, because this is never valid -- however, when `dyn*` was introduced, this check was never adjusted to only account for *unsized* `dyn Trait` and not *sized* `dyn* Trait`.
Matthias Krüger [Thu, 17 Nov 2022 21:33:20 +0000 (22:33 +0100)]
Rollup merge of #104503 - notriddle:notriddle/where, r=GuillaumeGomez
rustdoc: remove redundant font-color CSS on `.where`
Before 7f6ce7dddd49f453da15bb4d586a5990985814d8, light-theme where clauses had color `#4E4C4C` while the main color was `#000`. One of that commit's simplifications made it so that everything used the same black.
Matthias Krüger [Thu, 17 Nov 2022 21:33:20 +0000 (22:33 +0100)]
Rollup merge of #104496 - djkoloski:dont_normalize_compiler_backtraces, r=jackh726
Don't attempt to normalize compiler backtraces
Backtraces can very significantly depending on environment and cause spurious test suite failures. Ensuring a proper failure-status should be sufficient to keep a crash properly documented.
This caused a failure in Fuchsia's test suite CI due to an extra newline between "stack" and "error" appearing after normalization.
Matthias Krüger [Thu, 17 Nov 2022 21:33:19 +0000 (22:33 +0100)]
Rollup merge of #104483 - oli-obk:santa-clauses-make-goals, r=compiler-errors
Convert predicates into Predicate in the Obligation constructor
instead of having almost all callers do that.
This reduces a bit of boilerplate, and also paves the way for my work towards https://github.com/rust-lang/compiler-team/issues/531 (as it makes it easier to accept both goals and clauses where right now it only accepts predicates).
Matthias Krüger [Thu, 17 Nov 2022 21:33:18 +0000 (22:33 +0100)]
Rollup merge of #104366 - GuillaumeGomez:simplify-settings-theme-choice, r=notriddle
Simplify settings theme choice
I removed the storage changes from https://github.com/rust-lang/rust/pull/98765 and only kept the UI changes.
You can test it [here](https://rustdoc.crud.net/imperio/simplify-settings-theme-choice/foo/index.html).
Discussion about this still in progress [on zulip](https://rust-lang.zulipchat.com/#narrow/stream/266220-rustdoc/topic/Last.20part.20of.20settings.20simplification).
Matthias Krüger [Thu, 17 Nov 2022 21:33:17 +0000 (22:33 +0100)]
Rollup merge of #103852 - compiler-errors:rpitit-early-from-impl, r=lcnr
Don't remap early-bound regions for return-position `impl Trait` in trait originating from `impl`
long title :sweat:
We don't want to remap early-bound regions that originate from the `impl`s themselves, since they have no corresponding region in the trait. Not sure if there's a better condition than checking if the EBR's def-id's parent is the impl -- maybe we should be checking if the region comes from the method or RPITIT... :shrug:
Philipp Krones [Sat, 5 Nov 2022 14:08:37 +0000 (15:08 +0100)]
Add variant_name function to `LangItem`
Clippy has an internal lint that checks for the usage of hardcoded def
paths and suggests to replace them with a lang or diagnostic item, if
possible. This was implemented with a hack, by getting all the variants
of the `LangItem` enum and then index into it with the position of the
`LangItem` in the `items` list. This is no longer possible, because the
`items` list can't be accessed anymore.
bors [Thu, 17 Nov 2022 17:15:31 +0000 (17:15 +0000)]
Auto merge of #104361 - vladimir-ea:watchos_fix_linking, r=oli-obk
[watchos] Dynamic linking is not allowed for watchos targets
Dynamic linking of all apple targets was (re-) enabled in PR #100636. However, dynamic linking is not allowed on WatchOS so this broke the build of standard library for WatchOS.
This change disables dynamic linking for WatchOS non-simulator targets.
bors [Thu, 17 Nov 2022 13:47:03 +0000 (13:47 +0000)]
Auto merge of #104219 - bryangarza:async-track-caller-dup, r=eholk
Support `#[track_caller]` on async fns
Adds `#[track_caller]` to the generator that is created when we desugar the async fn.
Fixes #78840
Open questions:
- What is the performance impact of adding `#[track_caller]` to every `GenFuture`'s `poll(...)` function, even if it's unused (i.e., the parent span does not set `#[track_caller]`)? We might need to set it only conditionally, if the indirection causes overhead we don't want.
bors [Thu, 17 Nov 2022 10:48:22 +0000 (10:48 +0000)]
Auto merge of #104205 - clubby789:grow-rc, r=thomcc
Attempt to reuse `Vec<T>` backing storage for `Rc/Arc<[T]>`
If a `Vec<T>` has sufficient capacity to store the inner `RcBox<[T]>`, we can just reuse the existing allocation and shift the elements up, instead of making a new allocation.
Pietro Albini [Thu, 3 Nov 2022 15:04:07 +0000 (16:04 +0100)]
run tests on the remote device even when the default address is used
When running tests inside the Android emulator, bootstrap doesn't set
the TEST_DEVICE_ADDR environment variable, as the default address
(127.0.0.1:12345) is used.
Instead, REMOTE_TEST_CLIENT is set all the times when remote testing is
needed, and in no other cases. To ensure Android tests are executed in
the emulator, change the check.
bors [Thu, 17 Nov 2022 07:42:27 +0000 (07:42 +0000)]
Auto merge of #104170 - cjgillot:hir-def-id, r=fee1-dead
Record `LocalDefId` in HIR nodes instead of a side table
This is part of an attempt to remove the `HirId -> LocalDefId` table from HIR.
This attempt is a prerequisite to creation of `LocalDefId` after HIR lowering (https://github.com/rust-lang/rust/pull/96840), by controlling how `def_id` information is accessed.
This first part adds the information to HIR nodes themselves instead of a table.
The second part is https://github.com/rust-lang/rust/pull/103902
The third part will be to make `hir::Visitor::visit_fn` take a `LocalDefId` as last parameter.
The fourth part will be to completely remove the side table.
Michael Howell [Thu, 17 Nov 2022 05:52:30 +0000 (22:52 -0700)]
rustdoc: clean up sidebar width CSS
This commit takes advantage of the ability to set [flex-basis] to a specific
length instead of setting it to `auto` and changing min-/max-width, and
setting flex-grow/-shrink both to 0.
bors [Thu, 17 Nov 2022 04:47:11 +0000 (04:47 +0000)]
Auto merge of #103779 - the8472:simd-str-contains, r=thomcc
x86_64 SSE2 fast-path for str.contains(&str) and short needles
Based on Wojciech Muła's [SIMD-friendly algorithms for substring searching](http://0x80.pl/articles/simd-strfind.html#sse-avx2)
The two-way algorithm is Big-O efficient but it needs to preprocess the needle
to find a "critical factorization" of it. This additional work is significant
for short needles. Additionally it mostly advances needle.len() bytes at a time.
The SIMD-based approach used here on the other hand can advance based on its
vector width, which can exceed the needle length. Except for pathological cases,
but due to being limited to small needles the worst case blowup is also small.
benchmarks taken on a Zen2, compiled with `-Ccodegen-units=1`: