Auto merge of #59211 - nox:refcell-borrow-state, r=KodrAus
Introduce RefCell::try_borrow_unguarded
*Come sit next to the fireplace with me, this is going to be a long story.*
So, you may already be aware that Servo has weird design constraints that forces us developers working on it to do weird things. The thing that interests us today is that we do layout on a separate thread with its own thread pool to do some things in parallel, whereas the data it uses comes from the script thread, which implements the entire DOM and related pieces, with `!Sync` data types such as `RefCell<T>`.
The invariant we maintain is that script does not do anything ever with the DOM data as long as layout is doing its job. That's all nice and all, but one thing we don't ensure is that we don't actually know if script was currently mutably borrowing some `RefCell<T>` prior to starting layout, which may lead to aliasing mutable memory and obviously undefined behaviour.
This PR reinstates `RefCell::borrow_state` so that [this method](https://github.com/servo/servo/blob/master/components/script/dom/bindings/cell.rs#L23-L30) can make use of it and return `None` if the cell was mutably borrowed.
Auto merge of #58972 - QuietMisdreavus:intra-doc-link-imports, r=GuillaumeGomez
rustdoc: don't process `Crate::external_traits` when collecting intra-doc links
Part of https://github.com/rust-lang/rust/issues/58745, closes https://github.com/rust-lang/rust/pull/58917
The `collect-intra-doc-links` pass keeps track of the modules it recurses through as it processes items. This is used to know what module to give the resolver when looking up links. When looking through the regular items of the crate, this works fine, but the `DocFolder` trait as written doesn't just process the main crate hierarchy - it also processes the trait items in the `external_traits` map. This is useful for other passes (so they can strip out `#[doc(hidden)]` items, for example), but here it creates a situation where we're processing items "outside" the regular module hierarchy. Since everything in `external_traits` is defined outside the current crate, we can't fall back to finding its module scope like we do with local items.
Skipping this collection saves us from emitting some spurious warnings. We don't even lose anything by skipping it, either - the docs loaded from here are only ever rendered through `html::render::document_short` which strips any links out, so the fact that the links haven't been loaded doesn't matter. Hopefully this removes most of the remaining spurious resolution warnings from intra-doc links.
Changes:
````
Account for changed format range in a test
Adapt format integration test to use new range
Add unit tests for Rustfmt::calc_text_edits
Prefer saturating_sub
Return only partial text edits on formatting request
Redo errors in format.rs
Rephrase unstable features helper function
Work around Rustfmt bug
Adjust range-formatting test
Update Rustfmt to 1.2
Allow unstable features under "dev" channel
Updated clippy_lints in to rev c81e43b92c83bca75337ac26a296b909a6a98ab1 to fix build breaking
Fix lint pass API changes
Remove a redundant import
Bump Clippy
Add more helpful compiler messages when the features are improperly specified
Silence some flaky tests.
Update Cargo.
Rename read_analysis to decode_buf
Update Clippy
Fix pointer to pointer silliness
Add serde implementation to complement rustc-serialize
````
Changes:
````
Update src/fn_call.rs
Fix calloc test
Add calloc test
Tidy up calloc code
Update compiletest
Fix tagging order
Add `calloc`
rename MIRI_TARGET -> MIRI_COMPILETEST_TARGET to clarify that this affects compiletest only
compiletest can just propagate MIRI_SYSROOT from the outside
````
Changes:
````
Remove force-host and explain no-prefer-dynamic
Escape a single quote in single_char_pattern hint
cargo fmt
Re-add tmp feature to compiletest
Remove libtest from deps
Re-allow clippy::identity_conversion in compiletest
Use latest compiletest-rs
Revert tests/compile-test.rs to 61aa5c957c219abe1fb2d1b5e51db8b365b4f565
Fix ICE in suspicious_else_formatting
use a multispan for MANY_SINGLE_CHAR_NAMES
Add missing `// run-pass` annotations to ICE tests
Remove clippy_dev as dev-dependency
NFC: fix typos
rustup https://github.com/rust-lang/rust/pull/59657
Add TransmutingNull Lint * Late Lint pass, catches: * One liner: 0 -> null -> transmute * One liner: std:null() -> transmute * Const (which resolves to null) -> transmute * UI Test case for Lint * Updated test for issue 3849, because now the lint that code generated is in Clippy. * Expanded `const.rs` miri-based Constant Folding code, to cover raw pointers
Run rustfmt
Set level of identity_conversion FP to warn
Rustup to https://github.com/rust-lang/rust/pull/58805
rustup 41316f0449025394fdca6606d3fdb3b8f37a9872
Updated source to match with recent rustc `master` toolchain changes
Fix dogfood error of question_mark lint fix
Fix question_mark lint+test
use `span_lint_and_sugg` in `explicit_counter_loop`
Fix some test failures
Hacky rustup
run cargo fmt
rustup https://github.com/rust-lang/rust/pull/59096/
Change explicit_counter_loop's message to add parentheses if necessary
Change explicit_counter_loop's message to reflect original variable name
cargo fmt
Add rustfix tests for mistyped_literal_suffix lint
Move some `unreadable_literal` ui tests to correct file
Add implementation for the EarlyLintPass trait into the Adding Lints documentation.
Add rust-toolchain for clippy_dev
````
This is, as best I can tell, a clean revert of #57842. It retains some interim changes, like moving `black_box`, and otherwise brings libtest back in as it was before removal.
Auto merge of #59770 - pietroalbini:pin-android-emulator, r=kennytm
ci: pin android emulator to 28.0.23
Apparently Android Emulator 28.0.25 (the latest one) is broken for us, and `sdkmanager` doesn't have a way to pin a package to a specific version. This PR stops downloading the emulator from `sdkmanager`, downloading it manually instead.
Tested this locally and confirmed it works. I'll open an issue on Google's issue tracker later today.
Rollup merge of #59740 - cuviper:folded-extend, r=scottmcm
Use for_each to extend collections
This updates the `Extend` implementations to use `for_each` for many
collections: `BinaryHeap`, `BTreeMap`, `BTreeSet`, `LinkedList`, `Path`,
`TokenStream`, `VecDeque`, and `Wtf8Buf`.
Folding with `for_each` enables better performance than a `for`-loop for
some iterators, especially if they can just forward to internal
iterators, like `Chain` and `FlatMap` do.
Auto merge of #59710 - alexcrichton:llvm-9-compat, r=sanxiyn
rustc: Start implementing compat with LLVM 9
This commit doesn't actually migrate to LLVM 9, but it brings our own
C++ bindings in line with LLVM 9 and able to compile against tip of
tree. The changes made were:
* The `MainSubprogram` flag for debuginfo moved between flag types.
* Iteration of archive members was tweaked slightly and we have to
construct the two iterators before constructing the returned
`RustArchiveIterator` value.
* The `getOrInsertFunction` binding now returns a wrapper which we use
`getCallee()` on to get the value we're interested in.
Rollup merge of #59727 - alexcrichton:wasi-apis, r=fitzgen
wasi: Use shared API for preopened fds
This commit updates the wasi target with supported added in
CraneStation/wasi-sysroot#10. That function allows both C and Rust to
cooperate in how preopened files are managed, enabling us to learn about
propened files through the same interface. The `open_parent` function in
the wasi `fs` module was updated to avoid its own initialization of a
global preopened map and instead delegate to libc to perform this
functionality.
This should both be more robust into the future in terms of handling
path logic as well as ensuring the propened map is correctly set up at
process boot time. This does currently require some unfortunate
allocations on our side, but if that becomes an issue we can always
paper over those in time!
This removes a misplaced assertion. The function containing the assertion is actually only ever called for upstream crates that are not considered for LTO, so we don't care whether upstream code has been merged in by LTO or not.
Rollup merge of #58894 - GuillaumeGomez:invalid-lifetime-bounds, r=estebank
Fix invalid bounds string generation in rustdoc
Fixes #58737.
Very weird and I'm not sure this is the best fix around. However, trying to fix it beforehand seems overly complicated compared to the gain (in `clean`, it wouldn't change anything since we **have to** return something so that wouldn't work, and in `hir`, I'm afraid I'd break something else for very little gain).
Also, I wasn't able to make a small code to reproduce the issue. The only way to test is to document `crossbeam` directly and check the `Scope` struct...
Josh Stone [Fri, 5 Apr 2019 21:51:07 +0000 (14:51 -0700)]
Use for_each to extend collections
This updates the `Extend` implementations to use `for_each` for many
collections: `BinaryHeap`, `BTreeMap`, `BTreeSet`, `LinkedList`, `Path`,
`TokenStream`, `VecDeque`, and `Wtf8Buf`.
Folding with `for_each` enables better performance than a `for`-loop for
some iterators, especially if they can just forward to internal
iterators, like `Chain` and `FlatMap` do.
This is a rebase of S-blocked-closed PR #55431 to current master. LLVM has moved forward since then, so maybe we can check whether the new LLVM 8.0 version unblocked this work.
Alex Crichton [Fri, 5 Apr 2019 13:37:03 +0000 (06:37 -0700)]
wasi: Use shared API for preopened fds
This commit updates the wasi target with supported added in
CraneStation/wasi-sysroot#10. That function allows both C and Rust to
cooperate in how preopened files are managed, enabling us to learn about
propened files through the same interface. The `open_parent` function in
the wasi `fs` module was updated to avoid its own initialization of a
global preopened map and instead delegate to libc to perform this
functionality.
This should both be more robust into the future in terms of handling
path logic as well as ensuring the propened map is correctly set up at
process boot time. This does currently require some unfortunate
allocations on our side, but if that becomes an issue we can always
paper over those in time!
Alex Crichton [Thu, 4 Apr 2019 20:05:41 +0000 (13:05 -0700)]
rustc: Start implementing compat with LLVM 9
This commit doesn't actually migrate to LLVM 9, but it brings our own
C++ bindings in line with LLVM 9 and able to compile against tip of
tree. The changes made were:
* The `MainSubprogram` flag for debuginfo moved between flag types.
* Iteration of archive members was tweaked slightly and we have to
construct the two iterators before constructing the returned
`RustArchiveIterator` value.
* The `getOrInsertFunction` binding now returns a wrapper which we use
`getCallee()` on to get the value we're interested in.
Auto merge of #59076 - dtolnay:comma, r=alexcrichton
Include trailing comma in multiline Debug representation
This PR changes the behavior of [`Formatter::debug_struct`](https://doc.rust-lang.org/std/fmt/struct.Formatter.html#method.debug_struct), [`debug_tuple`](https://doc.rust-lang.org/std/fmt/struct.Formatter.html#method.debug_tuple), [`debug_list`](https://doc.rust-lang.org/std/fmt/struct.Formatter.html#method.debug_list), [`debug_set`](https://doc.rust-lang.org/std/fmt/struct.Formatter.html#method.debug_set), and [`debug_map`](https://doc.rust-lang.org/std/fmt/struct.Formatter.html#method.debug_map) to render trailing commas in `{:#?}` mode, which is the dominant style in modern Rust code.
#### Before:
```console
Language {
name: "Rust",
trailing_commas: false
}
```
#### After:
```console
Language {
name: "Rust",
trailing_commas: true,
}
```
David Tolnay [Wed, 3 Apr 2019 14:30:08 +0000 (07:30 -0700)]
Include trailing comma in multiline Debug representation
This commit changes the behavior of Formatter::debug_struct,
debug_tuple, debug_list, debug_set, and debug_map to render trailing
commas in {:#?} mode, which is the dominant style in modern Rust code.
Auto merge of #59721 - Centril:rollup-ieam9ke, r=Centril
Rollup of 5 pull requests
Successful merges:
- #59665 (improve worst-case performance of HashSet.is_subset)
- #59687 (cleanup shebang handling in the lexer)
- #59690 (Mark unix::ffi::OsStrExt methods as inline)
- #59702 (Use declare_lint_pass! and impl_lint_pass! in more places)
- #59712 (wasm32: Default to a "static" relocation model)
Rollup merge of #59712 - alexcrichton:wasm-static-not-pic, r=eddyb
wasm32: Default to a "static" relocation model
LLVM 9 is adding support for a "pic" relocation model for wasm code,
which is quite different than the current model. In order to preserve
the mode of compilation that we have today default to "static" to ensure
that we don't accidentally start creating experimental relocatable
binaries.
Rollup merge of #59665 - ssomers:hashset_revisited, r=KodrAus
improve worst-case performance of HashSet.is_subset
One more simple optimization opportunity for HashSet that was applied in BTreeSet in #59186 (and wasn't in #57043). Already covered by the existing unit test.
Auto merge of #59681 - dtolnay:cargo, r=alexcrichton
Update cargo
20 commits in 63231f438a2b5b84ccf319a5de22343ee0316323..6f3e9c367abb497c64f360c3839dab5e74928d5c
2019-03-27 12:26:45 +0000 to 2019-04-04 14:11:33 +0000
- Fix Init for Fossil SCM project (rust-lang/cargo#6792)
- Fix member_manifest_version_error accessing the network (rust-lang/cargo#6799)
- Don't include email if it is empty (rust-lang/cargo#6802)
- Fix unused import warning (rust-lang/cargo#6807)
- Add some help and documentation for unstable flags (rust-lang/cargo#6791)
- Allow `cargo doc --open` with multiple packages (rust-lang/cargo#6803)
- Allow `cargo install --path P` to load config from P (rust-lang/cargo#6804)
- Add more suggestions on how to deal with excluding a package from a workspace (rust-lang/cargo#6805)
- Warn on version req with metadata (rust-lang/cargo#6806)
- cargo install: Be more restrictive about cli flags (rust-lang/cargo#6801)
- Support force-pushed repos with git-fetch-with-cli (rust-lang/cargo#6800)
- Cargo clippy (rust-lang/cargo#6759)
- Don't include metadata in wasm binary examples (rust-lang/cargo#6812)
- Update glossary for `feature` (rust-lang/cargo#6809)
- Include proc-macros in `build-override` (rust-lang/cargo#6811)
- Resolver: A dep is equivalent to one of the things it can resolve to (rust-lang/cargo#6776)
- Add some docs for `Downloads` (rust-lang/cargo#6815)
- Resolve: Be less strict while offline (rust-lang/cargo#6814)
- Accept trailing comma in test of impl Debug for PackageId (rust-lang/cargo#6818)
- Fix doc link (rust-lang/cargo#6820)
<br>
I specifically care about "Accept trailing comma in test of impl Debug for PackageId (rust-lang/cargo#6818)" to unblock https://github.com/rust-lang/rust/pull/59076.
Auto merge of #59643 - alexcrichton:wasi-symbols, r=sanxiyn
std: Upgrade `compiler_builtins` to fix wasi linkage
Turns out we needed to exclude a number of math functions on the
`wasm32-unknown-wasi` target, and this was fixed in 0.1.9 of
compiler-builtins and this is pulling in the fix to libstd's own build.
Alex Crichton [Thu, 4 Apr 2019 20:23:42 +0000 (13:23 -0700)]
wasm32: Default to a "static" relocation model
LLVM 9 is adding support for a "pic" relocation model for wasm code,
which is quite different than the current model. In order to preserve
the mode of compilation that we have today default to "static" to ensure
that we don't accidentally start creating experimental relocatable
binaries.
Alex Crichton [Tue, 2 Apr 2019 20:38:51 +0000 (13:38 -0700)]
std: Upgrade `compiler_builtins` to fix wasi linkage
Turns out we needed to exclude a number of math functions on the
`wasm32-unknown-wasi` target, and this was fixed in 0.1.9 of
compiler-builtins and this is pulling in the fix to libstd's own build.
Auto merge of #59676 - alexcrichton:osx-deadlock, r=sfackler
std: Avoid usage of `Once` in `Instant`
This commit removes usage of `Once` from the internal implementation of
time utilities on OSX and Windows. It turns out that we accidentally hit
a deadlock today (#59020) via events that look like:
* A thread invokes `park_timeout`
* Internally, only on OSX, `park_timeout` calls `Instant::elapsed`
* Inside of `Instant::elapsed` on OSX we enter a `Once` to initialize
global timer data
* Inside of `Once`, it attempts to `park`
This means on the same stack frame, when there's contention, we're
calling `park` from inside `park_timeout`, causing a deadlock!
The solution implemented in this commit was to remove usage of `Once`
and instead just do a small dance with atomics. There's no real need we
need to guarantee that the global information is only learned once, only
that it's only *stored* once. This implementation may have multiple
threads invoke `mach_timebase_info`, but only one will store the global
information which will amortize the cost for all other threads.
A similar fix has been applied to windows to be uniform across our
implementations, but looking at the code on Windows no deadlock was
possible. This is purely just a consistency update for Windows and in
theory a slightly leaner implementation.
Felix S Klock II [Tue, 26 Mar 2019 12:18:17 +0000 (13:18 +0100)]
Fix out-of-date comment
A comment in one match arm make a blanket statement about "reads/reservations", but in fact the whole point of this PR is that reservations are *not* handled by that particular arm anymore.