bors [Tue, 29 Mar 2022 07:27:14 +0000 (07:27 +0000)]
Auto merge of #95257 - compiler-errors:fn-borrow, r=lcnr
Add suggestion to borrow `Fn` and `FnMut` params/opaque/closures instead of move
I think that Closure/ParamTy/Opaque are all "opaque" enough that it's meaningful to suggest borrowing them instead of moving them at their usage sites when we see a move error. See the attached issue for example.
Is this suggestion too general? I could perhaps use the move site information to limit this to places like fn calls, but I don't know enough about mir borrowck to know if that's an easy change.
bors [Tue, 29 Mar 2022 05:15:58 +0000 (05:15 +0000)]
Auto merge of #95423 - ehuss:update-books, r=ehuss
Update books
## nomicon
3 commits in f6d6126fc96ecf4a7f7d22da330df9506293b0d0..11f1165e8a2f5840467e748c8108dc53c948ee9a
2022-02-26 02:21:21 +0900 to 2022-03-19 16:02:00 -0400
- Make the Vec impl be slightly more careful with ZSTs and alignment.
- implement `IntoIterator` for `Vec` (rust-lang/nomicon#337)
- Add an explanation shared to exclusive transmute (rust-lang/nomicon#344)
23 commits in 036e88a4f135365de85358febe5324976a56030a..ea90bbaf53ba64ef4e2da9ac2352b298aec6bec8
2022-03-04 21:53:33 -0500 to 2022-03-28 21:59:34 -0400
- Fix nostarch snapshot
- Snapshot of chapter 7 for nostarch
- Add a forward reference to chapter 14, another example of pub use
- Clarify pub use example. Fixes rust-lang/book#2716.
- Fancy quotes
- Fix incorrectly worded sentence. Fixes rust-lang/book#3086.
- Reword description of how a listing came to be
- Call out binary+library crate practices
- Define binary and library crates more explicitly
- Clarify when a path is a crate name and when it should be literal crate
- Make it clearer the outer `mod` doesn't move to the file
- Don't wrap this example in main when copying. Fixes rust-lang/book#2930.
- Try to make clearer that `mod` is not an `import`
- Mention mod.rs file naming scheme
- Explain why submodule subdirectories are needed more
- Rename a separate example of serve_order to deliver_order
- Show an example that `use` only applies in its own scope
- quick modules guide
- Tweak a snippet of ch18-03
- Propagating edits to chapter 10 back
- Responses to nostarch questions of chapter 10
- Update src/ch04-01-what-is-ownership.md
- Add Danish translation link. Connects to rust-lang/book#3079.
bors [Tue, 29 Mar 2022 01:09:22 +0000 (01:09 +0000)]
Auto merge of #95417 - ehuss:doc-no_std-error, r=Dylan-DPC
bootstrap: better error message for no_std docs
Currently if one tries to build std documentation for a no_std target, you get a confusing error message:
`error: The argument '--package [<SPEC>...]' was provided more than once, but cannot be used multiple times`
This is because [`std_cargo`](https://github.com/rust-lang/rust/blob/600ec284838c52d1f6657c2cf0097b58970b133b/src/bootstrap/compile.rs#L299-L305) has a built-in `-p alloc` argument that conflicts with the `cargo rustdoc` command used in the Std doc step.
This just adds a better error message in this scenario. It may be possible to fix this correctly, but that would likely be a bit more of an invasive change that I don't have time for right now.
By reversing the arguments we achieve several clarifications:
- The function closely resembles `cast` with an argument to
initialize the metadata. This is easier to teach and answers a long
outstanding question that had restricted cast to `Sized` pointee
targets. See multiples reviews of
<https://github.com/rust-lang/rust/pull/47631>
- The 'object identity', in the form of provenance, is now preserved
from the receiver argument to the result. This helps explain the method as
a builder-style, instead of some kind of setter that would modify
something in-place. Ensuring that the result has the identity of the
`self` argument is also beneficial for an intuition of effects.
- An outstanding concern, 'Correct argument type', is avoided by not
committing to any specific argument type. This is consistent with cast
which does not require its receiver to be a 'raw address'.
Hopefully the usage examples in `sync/rc.rs` serve as sufficient examples of the style to convince the reader of the readability improvements of this style, when compared to the previous order of arguments.
I want to take the opportunity to motivate inclusion of this method _separate_ from metadata API, separate from `feature(ptr_metadata)`. It does _not_ involve the `Pointee` trait in any form. This may be regarded as a very, very light form that does not commit to any details of the pointee trait, or its associated metadata. There are several use cases for which this is already sufficient and no further inspection of metadata is necessary.
- Storing the coercion of `*mut T` into `*mut dyn Trait` as a way to dynamically cast some an arbitrary instance of the same type to a dyn trait instance. In particular, one can have a field of type `Option<*mut dyn io::Seek>` to memorize if a particular writer is seekable. Then a method `fn(self: &T) -> Option<&dyn Seek>` can be provided, which does _not_ involve the static trait bound `T: Seek`. This makes it possible to create an API that is capable of utilizing seekable streams and non-seekable streams (instead of a possible less efficient manner such as more buffering) through the same entry-point.
- Enabling more generic forms of unsizing for no-`std` smart pointers. Using the stable APIs only few concrete cases are available. One can unsize arrays to `[T]` by `ptr::slice_from_raw_parts` but unsizing a custom smart pointer to, e.g., `dyn Iterator`, `dyn Future`, `dyn Debug`, can't easily be done generically. Exposing `with_metadata_of` would allow smart pointers to offer their own `unsafe` escape hatch with similar parameters where the caller provides the unsized metadata. This is particularly interesting for embedded where `dyn`-trait usage can drastically reduce code size.
When compiled with beta or nightly compiler on Godbolt with `-C opt-level=3` flag it prints the following assembly.
```asm
example::safe_substr_to:
push r15
push r14
push r12
push rbx
push rax
mov r14, rdi
test rdx, rdx
je .LBB0_8
mov rbx, rdx
mov r15, rsi
mov r12, qword ptr [rip + core::num::<impl u8>::is_utf8_char_boundary@GOTPCREL]
jmp .LBB0_4
.LBB0_2:
je .LBB0_9
.LBB0_3:
add rbx, -1
je .LBB0_8
.LBB0_4:
cmp rbx, r15
jae .LBB0_2
movzx edi, byte ptr [r14 + rbx]
call r12
test al, al
je .LBB0_3
mov r15, rbx
jmp .LBB0_9
.LBB0_8:
xor r15d, r15d
.LBB0_9:
mov rax, r14
mov rdx, r15
add rsp, 8
pop rbx
pop r12
pop r14
pop r15
ret
```
`qword ptr [rip + core::num::<impl u8>::is_utf8_char_boundary@GOTPCREL]` is not inlined. `-C remark=all` outputs the following message:
```
note: /rustc/7bccde19767082c7865a12902fa614ed4f8fed73/library/core/src/str/mod.rs:214:25: inline: _ZN4core3num20_$LT$impl$u20$u8$GT$21is_utf8_char_boundary17hace9f12f5ba07a7fE will not be inlined into _ZN4core3str21_$LT$impl$u20$str$GT$16is_char_boundary17hf2587e9a6b8c5e43E because its definition is unavailable
```
Stable compiler outputs more reasonable code:
```asm
example::safe_substr_to:
mov rcx, rdx
mov rax, rdi
test rdx, rdx
je .LBB0_9
mov rdx, rsi
jmp .LBB0_4
.LBB0_2:
cmp rdx, rcx
je .LBB0_7
.LBB0_3:
add rcx, -1
je .LBB0_9
.LBB0_4:
cmp rcx, rdx
jae .LBB0_2
cmp byte ptr [rax + rcx], -64
jl .LBB0_3
mov rdx, rcx
.LBB0_7:
ret
.LBB0_9:
xor edx, edx
ret
```
Dylan DPC [Mon, 28 Mar 2022 18:41:52 +0000 (20:41 +0200)]
Rollup merge of #95397 - dtolnay:disclaimer, r=m-ou-se
Link to std::io's platform-specific behavior disclaimer
This PR adds some links in standard library documentation to point to https://doc.rust-lang.org/std/io/index.html#platform-specific-behavior.
> ### Platform-specific behavior
>
> Many I/O functions throughout the standard library are documented to indicate what various library or syscalls they are delegated to. This is done to help applications both understand what’s happening under the hood as well as investigate any possibly unclear semantics. Note, however, that this is informative, not a binding contract. The implementation of many of these functions are subject to change over time and may call fewer or more syscalls/library functions.
Many of the `std::fs` APIs already link to this disclaimer when discussing system calls.
Separate dependencies for `parallel_compiler` feature, so they will not be compiled if feature not selected, reducing number of compiled crates from 238 to 224.
bors [Mon, 28 Mar 2022 16:15:17 +0000 (16:15 +0000)]
Auto merge of #95403 - Dylan-DPC:rollup-9on30mg, r=Dylan-DPC
Rollup of 6 pull requests
Successful merges:
- #95301 (Remove `Nonterminal::NtTT`.)
- #95314 (Tell users that `||` operators are not currently supported in let chain expressions)
- #95350 (resolve: Simplify some diagnostic code to avoid an ICE)
- #95370 ([bootstrap] Don't print `Suite not skipped` unless `--verbose` is set)
- #95390 (Ignore doc comments in a declarative macro matcher.)
- #95401 (Remove duplicated and unused test files)
Dylan DPC [Mon, 28 Mar 2022 14:08:10 +0000 (16:08 +0200)]
Rollup merge of #95370 - jyn514:less-verbose-skips, r=Dylan-DPC
[bootstrap] Don't print `Suite not skipped` unless `--verbose` is set
This was so verbose before that it made it hard to see what effect the flag actually had.
Before:
```
Set({test::src/tools/tidy}) not skipped for "bootstrap::test::Tidy" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/ui) because it is excluded
Suite(test::src/test/run-pass-valgrind) not skipped for "bootstrap::test::RunPassValgrind" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/mir-opt) because it is excluded
Suite(test::src/test/codegen) not skipped for "bootstrap::test::Codegen" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/codegen-units) not skipped for "bootstrap::test::CodegenUnits" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/assembly) not skipped for "bootstrap::test::Assembly" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/incremental) not skipped for "bootstrap::test::Incremental" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/debuginfo) because it is excluded
Skipping Suite(test::src/test/ui-fulldeps) because it is excluded
... about 100 more lines ...
```
After:
```
Skipping Suite(test::src/test/ui) because it is excluded
Skipping Suite(test::src/test/mir-opt) because it is excluded
Skipping Suite(test::src/test/debuginfo) because it is excluded
Skipping Suite(test::src/test/ui-fulldeps) because it is excluded
```
Dylan DPC [Mon, 28 Mar 2022 14:08:08 +0000 (16:08 +0200)]
Rollup merge of #95314 - c410-f3r:aqui-vamos-nos, r=lcnr
Tell users that `||` operators are not currently supported in let chain expressions
Tells that `||` operators are not currently supported instead of not allowed. See https://github.com/rust-lang/rust/issues/53667#issuecomment-1066075876
Dylan DPC [Mon, 28 Mar 2022 14:08:07 +0000 (16:08 +0200)]
Rollup merge of #95301 - nnethercote:rm-NtTT, r=petrochenkov
Remove `Nonterminal::NtTT`.
It's only needed for macro expansion, not as a general element in the
AST. This commit removes it, adds `NtOrTt` for the parser and macro
expansion cases, and renames the variants in `NamedMatch` to better
match the new type.
bors [Mon, 28 Mar 2022 13:49:22 +0000 (13:49 +0000)]
Auto merge of #95024 - koehlma:rustdoc-private-items, r=GuillaumeGomez,camelid,jsha
rustdoc: add 🔒 to items with restricted visibility
This change marks items with restricted visibility with 🔒 when building with `--document-private-items`:
<img width="278" alt="Screen Shot 2022-03-20 at 23 50 24" src="https://user-images.githubusercontent.com/509209/159189513-9e4b09bb-6785-41a5-bfe2-df02f83f8641.png">
There also appears a “Restricted Visibility” tooltip when hovering over the emoji.
---
The original PR for reference:
This change makes private items slightly transparent (similar to `unstable` items in rustc):
<img width="272" alt="Screen Shot 2022-03-16 at 22 17 43" src="https://user-images.githubusercontent.com/509209/158692627-a1f6f5ec-e043-4aa2-9352-8d2b15c31c08.png">
I found myself using `--document-private-items` a lot recently because I find the documentation of private internals quite helpful when working on a larger project. However, not being able to distinguish private from public items (see #87785) when looking at the documentation makes this somewhat cumbersome.
This PR addresses the third suggestion of issue #87785 by marking private items typographically. It seems to me that the other suggestions are more involved but this is at least a first step.
A private item is also made slightly transparent in the path displayed in the header of a page:
<img width="467" alt="Screen Shot 2022-03-16 at 22 19 51" src="https://user-images.githubusercontent.com/509209/158692885-0bbd3417-3c0b-486f-b8ab-99c05c6fa7ca.png">
bors [Mon, 28 Mar 2022 05:48:25 +0000 (05:48 +0000)]
Auto merge of #95300 - workingjubilee:less-bitsets, r=eddyb
Skip needless bitset for debuginfo
Found this while digging around looking at the inlining logic.
Seemed obvious enough so I decided to try to take care of it.
Is this what you had in mind, `@eddyb?`
bors [Mon, 28 Mar 2022 02:29:06 +0000 (02:29 +0000)]
Auto merge of #95393 - Dylan-DPC:rollup-l72f39g, r=Dylan-DPC
Rollup of 4 pull requests
Successful merges:
- #88375 (Clarify that ManuallyDrop<T> has same layout as T)
- #93755 (Allow comparing `Vec`s with different allocators using `==`)
- #95016 (Docs: make Vec::from_raw_parts documentation less strict)
- #95098 (impl From<&[T; N]> and From<&mut [T; N]> for Vec<T>)
Dylan DPC [Mon, 28 Mar 2022 02:12:10 +0000 (04:12 +0200)]
Rollup merge of #95016 - janpaul123:patch-1, r=dtolnay
Docs: make Vec::from_raw_parts documentation less strict
This is my first PR; be gentle!
In https://users.rust-lang.org/t/why-does-vec-from-raw-parts-require-same-size-and-not-same-size-capacity/73036/2?u=janpaul123 it was suggested to me that I should make a PR to make the documentation of `Vec::from_raw_parts` less strict, since we don't require `T` to have the same size, just `size_of::<T>() * capacity` to be the same, since that is what results in `Layout::size` being the same in `dealloc`, which is really what matters.
Also in https://users.rust-lang.org/t/why-does-vec-from-raw-parts-require-same-size-and-not-same-size-capacity/73036/8?u=janpaul123 it was suggested that it's better to use `slice::from_raw_parts`, which I think is useful advise that could also be mentioned in the docs, so I added that too.
I did not changed the `PartialOrd` impl too because it was not generic already (didn't support `Vec<T> <=> Vec<U> where T: PartialOrd<U>`).
Does it needs tests?
I don't think this will hurt type inference much because the default allocator is usually not inferred (`new()` specifies it directly, and even with other allocators, you pass the allocator to `new_in()` so the compiler usually knows the type).
I think this requires FCP since the impls are already stable.
Dylan DPC [Mon, 28 Mar 2022 02:12:09 +0000 (04:12 +0200)]
Rollup merge of #88375 - joshlf:patch-3, r=dtolnay
Clarify that ManuallyDrop<T> has same layout as T
This PR implements the documentation change under discussion in https://github.com/rust-lang/unsafe-code-guidelines/issues/302. It should not be approved or merged until the discussion there is resolved.
It's only needed for macro expansion, not as a general element in the
AST. This commit removes it, adds `NtOrTt` for the parser and macro
expansion cases, and renames the variants in `NamedMatch` to better
match the new type.
bors [Sun, 27 Mar 2022 21:36:42 +0000 (21:36 +0000)]
Auto merge of #95382 - Dylan-DPC:rollup-bebyfd1, r=Dylan-DPC
Rollup of 5 pull requests
Successful merges:
- #94939 (diagnostics: suggest missing comma in bad FRU syntax)
- #95120 (Implement `apply_switch_int_edge_effects` for backward analyses)
- #95364 (Add long error explanation for E0667)
- #95366 (Remove test files with duplicated checksums)
- #95368 (Fix typo in `String::try_reserve_exact` docs)
Dylan DPC [Sun, 27 Mar 2022 20:51:42 +0000 (22:51 +0200)]
Rollup merge of #95368 - lopopolo:lopopolo/string-try-reserve-exact-doc-typo, r=Dylan-DPC
Fix typo in `String::try_reserve_exact` docs
Copying the pattern from `Vec::try_reserve_exact` and `String::try_reserve`,
it looks like this doc comment is intending to refer to the currently-being-documented
function.
bors [Sun, 27 Mar 2022 18:55:58 +0000 (18:55 +0000)]
Auto merge of #94495 - estebank:missing-closing-gt, r=jackh726
Provide suggestion for missing `>` in a type parameter list
When encountering an inproperly terminated type parameter list, provide
a suggestion to close it after the last non-constraint type parameter
that was successfully parsed.
bors [Sun, 27 Mar 2022 16:36:05 +0000 (16:36 +0000)]
Auto merge of #95345 - dtolnay:escape0, r=Dylan-DPC
Debug print char 0 as '\0' rather than '\u{0}'
```rust
println!("{:?}", "foo\0");
```
- **Before:** `"foo\u{0}"`
- **After:** `"foo\0"`
```rust
println!("{:?}", '\0');
```
- **Before:** `'\u{0}'`
- **After:** `'\0'`
`'\0'` will be more recognizable to everyone than `'\u{0}'` because it's how we talk about character 0 in all of our docs and example code, such as https://doc.rust-lang.org/std/ffi/index.html, https://doc.rust-lang.org/std/ffi/struct.CStr.html, https://doc.rust-lang.org/std/ffi/struct.CString.html.
Joshua Nelson [Sun, 27 Mar 2022 15:11:02 +0000 (10:11 -0500)]
[bootstrap] Don't print `Suite not skipped` unless `--verbose` is set
This was so verbose before that it made it hard to see what effect the flag actually had.
Before:
```
Set({test::src/tools/tidy}) not skipped for "bootstrap::test::Tidy" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/ui) because it is excluded
Suite(test::src/test/run-pass-valgrind) not skipped for "bootstrap::test::RunPassValgrind" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/mir-opt) because it is excluded
Suite(test::src/test/codegen) not skipped for "bootstrap::test::Codegen" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/codegen-units) not skipped for "bootstrap::test::CodegenUnits" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/assembly) not skipped for "bootstrap::test::Assembly" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Suite(test::src/test/incremental) not skipped for "bootstrap::test::Incremental" -- not in [src/test/ui, src/test/mir-opt/, src/test/debuginfo, src/test/ui-fulldeps]
Skipping Suite(test::src/test/debuginfo) because it is excluded
Skipping Suite(test::src/test/ui-fulldeps) because it is excluded
... about 100 more lines ...
```
After:
```
Skipping Suite(test::src/test/ui) because it is excluded
Skipping Suite(test::src/test/mir-opt) because it is excluded
Skipping Suite(test::src/test/debuginfo) because it is excluded
Skipping Suite(test::src/test/ui-fulldeps) because it is excluded
```
Ryan Lopopolo [Sun, 27 Mar 2022 13:53:55 +0000 (06:53 -0700)]
Fix typo in `String::try_reserve_exact` docs
Copying the pattern from `Vec::try_reserve_exact` and `String::try_reserve`,
it looks like this doc comment is intending to refer to the currently-being-documented
function.
I need this in order to pick up https://github.com/BurntSushi/ripgrep/pull/1722, which picked up https://github.com/BurntSushi/bstr/pull/58, which unblocks https://github.com/rust-lang/rust/pull/95345. Ripgrep uses the Debug representation of a `BStr` in some of its tests. In old versions of bstr, that used to just use the standard library's `escape_debug()` implementation, so the output ends up being sensitive to whether the standard library renders character 0 as `\u{0}` or as `\0`. The newer bstr always renders character 0 as `\0` and ripgrep's test suite has been correspondingly updated.
bors [Sun, 27 Mar 2022 04:20:07 +0000 (04:20 +0000)]
Auto merge of #95351 - Dylan-DPC:rollup-o1il7tx, r=Dylan-DPC
Rollup of 5 pull requests
Successful merges:
- #91981 (Recover suggestions and useful information lost in previous PR)
- #93469 (Skip pointing out ambiguous impls in alloc/std crates too in inference errors)
- #95335 (Move resolve_path to rustc_builtin_macros and make it private)
- #95340 (interpret: with enforce_number_validity, ensure integers are truly Scalar::Int (i.e., no pointers))
- #95341 (ARMv6K Horizon OS has_thread_local support)
Dylan DPC [Sun, 27 Mar 2022 03:36:11 +0000 (05:36 +0200)]
Rollup merge of #95341 - Meziu:armv6k-3ds-target, r=nagisa
ARMv6K Horizon OS has_thread_local support
cc. ```@ian-h-chamberlain```
cc. ```@AzureMarker```
Being an ARM target, it has always had built-in support for `#[thread_local]`. This PR comes in just now because we were testing `std::thread` support with `thread_local_dtor`s. This will hopefully be the last PR for the target specification, unless anymore features will be needed as time goes on.
Dylan DPC [Sun, 27 Mar 2022 03:36:08 +0000 (05:36 +0200)]
Rollup merge of #93469 - compiler-errors:issue-93450, r=estebank
Skip pointing out ambiguous impls in alloc/std crates too in inference errors
This generalizes the logic in `annotate_source_of_ambiguity` to skip printing ambiguity errors traits in `alloc` and `std` as well, not just `core`.
While this does spot-fix the issue mentioned below, it would be nicer to generalize this logic, for example to detect when the trait predicate's `self_ty` has any numerical inference variables. Is it worthwhile to scrap this solution for one like that?
Esteban Kuber [Tue, 1 Mar 2022 19:40:48 +0000 (19:40 +0000)]
Provide suggestion for missing `>` in a type parameter list
When encountering an inproperly terminated type parameter list, provide
a suggestion to close it after the last non-constraint type parameter
that was successfully parsed.