Auto merge of #101342 - saethlin:update-backtrace, r=Mark-Simulacrum
Update backtrace
Most notably, this pulls in https://github.com/rust-lang/backtrace-rs/commit/ebc9a85466506e4fd0ba9cb72b15c42e320a0867, which is both a bugfix and a significant performance improvement for Miri, since fixing the bug makes `RUST_BACKTRACE=1` backtraces much smaller: https://github.com/rust-lang/miri/issues/2273, https://github.com/rust-lang/miri-test-libstd/issues/8
This also pulls in other commits which turn the backtrace-rs CI green. That's nice.
Auto merge of #101378 - matthiaskrgr:rollup-s1awa47, r=matthiaskrgr
Rollup of 4 pull requests
Successful merges:
- #101335 (rustdoc: remove old CSS selector that causes weird spacing)
- #101347 (ffx component run should provide a collection)
- #101364 (Shrink suggestion span of argument mismatch error)
- #101365 (remove redundant clones)
Failed merges:
- #101349 (rustdoc: remove `.impl-items { flex-basis }` CSS, not in flex container)
Rollup merge of #101348 - GuillaumeGomez:cleanup-css-theme, r=notriddle
Cleanup css theme
Follow-up of https://github.com/rust-lang/rust/pull/100494.
The change for the border color of the search input in the dark mode was actually a weird case: the search input border was unique, it didn't share the same variable with other items with borders. This weird case being unique to the dark theme, I removed it, hence the modification in the GUI test.
Live demo is [here](https://rustdoc.crud.net/imperio/cleanup-css-theme/std/index.html).
Rollup merge of #101338 - diminishedprime:patch-2, r=tmandry
Fix unsupported syntax in .manifest file
Fuchsia .manifest files do not support a `#` comment syntax. Because of this, if you copy and paste the current example code for this file, and then remove the line you don't need, you still see an error. To make this a bit easier to follow, split this into two code blocks, one for rustc, and one for cargo.
Rollup merge of #101325 - ChrisDenton:BCRYPT_RNG_ALG_HANDLE, r=thomcc
Windows RNG: Use `BCRYPT_RNG_ALG_HANDLE` by default
This only changes a small amount of actual code, the rest is documentation outlining the history of this module as I feel it will be relevant to any future issues that might crop up.
The code change is to use the `BCRYPT_RNG_ALG_HANDLE` [pseudo-handle](https://docs.microsoft.com/en-us/windows/win32/seccng/cng-algorithm-pseudo-handles) by default, which simply uses the default RNG. Previously we used `BCRYPT_USE_SYSTEM_PREFERRED_RNG` which has to load the system configuration and then find and load that RNG. I suspect this was the cause of failures on some systems (e.g. due to corrupted config). However, this is admittedly speculation as I can't reproduce the issue myself (and it does seem quite rare even in the wild). Still, removing a possible point of failure is likely worthwhile in any case.
Rollup merge of #101217 - eholk:drop-tracking-73137, r=jyn514
[drop tracking] Use parent expression for scope, not parent node
Previously we were just using the parent node as the scope for a temporary value, but it turns out this is too narrow. For example, in an expression like
Foo {
b: &42,
a: async { 0 }.await,
}
the scope for the &42 was set to the ExprField node for `b: &42`, when we actually want to use the Foo struct expression.
We fix this by recursively searching through parent nodes until we find a Node::Expr. It may be that we don't find one, and if so that's okay, we will just fall back on the enclosing temporary scope which is always sufficient.
will print an error if used with a short-lived pipe, e.g.
% ./main | head -n 1
hello world
thread 'main' panicked at 'failed printing to stdout: Broken pipe (os error 32)', library/std/src/io/stdio.rs:1016:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
by enabling `#[unix_sigpipe = "sig_dfl"]` like this
there is no error, because `SIGPIPE` will not be ignored and thus the program will be killed appropriately:
% ./main | head -n 1
hello world
The current libstd behaviour of ignoring `SIGPIPE` before `fn main()` can be explicitly requested by using `#[unix_sigpipe = "sig_ign"]`.
With `#[unix_sigpipe = "inherit"]`, no change at all is made to `SIGPIPE`, which typically means the behaviour will be the same as `#[unix_sigpipe = "sig_dfl"]`.
See https://github.com/rust-lang/rust/issues/62569 and referenced issues for discussions regarding the `SIGPIPE` problem itself
See the [this](https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/Proposal.3A.20First.20step.20towards.20solving.20the.20SIGPIPE.20problem) Zulip topic for more discussions, including about this PR.
But with the switch to `<details>`, the code has changed drastically out from
under it, to the point where you have to go out of your way to actually get
it to render this way, and the result looks overly-tight and weird alongside
the normal version where this code is not reachable.
Auto merge of #101333 - matthiaskrgr:rollup-qpf1otj, r=matthiaskrgr
Rollup of 6 pull requests
Successful merges:
- #100121 (Try normalizing types without RevealAll in ParamEnv in MIR validation)
- #100200 (Change implementation of `-Z gcc-ld` and `lld-wrapper` again)
- #100814 ( Porting 'compiler/rustc_trait_selection' to translatable diagnostics - Part 1)
- #101215 (Also replace the version placeholder in rustc_attr)
- #101260 (Use `FILE_ATTRIBUTE_TAG_INFO` to get reparse tag)
- #101323 (Remove unused .toggle-label CSS rule)
Matt Hamrick [Fri, 2 Sep 2022 18:14:22 +0000 (11:14 -0700)]
Fix unsupported syntax in .manifest file
Fuchsia .manifest files do not support a `#` comment syntax. Because of this, if you copy and paste the current example code for this file, and then remove the line you don't need, you still see an error. To make this a bit easier to follow, split this into two code blocks, one for rustc, and one for cargo.
Rollup merge of #101260 - ChrisDenton:attribute-tag, r=thomcc
Use `FILE_ATTRIBUTE_TAG_INFO` to get reparse tag
I've been looking at this code recently and it just occurred to me we don't actually use the full reparse data at this point, only the tag. [`GetFileInformationByHandleEx`](https://docs.microsoft.com/en-us/windows/win32/api/winbase/nf-winbase-getfileinformationbyhandleex) can do exactly that by filling a [`FILE_ATTRIBUTE_TAG_INFO`](https://docs.microsoft.com/en-us/windows/win32/api/winbase/ns-winbase-file_attribute_tag_info) struct.
r? ```````@thomcc``````` since you've made changes here recently (which is why I have this code on my mind atm)
Rollup merge of #101215 - est31:rustdoc_version_placeholder, r=Mark-Simulacrum
Also replace the version placeholder in rustc_attr
Replace the version placeholder with the current version in the rustc_attr crate too so that users won't see the placeholder but instead the explicit version. This especially fixes the bug for rustdoc not showing it but instead the placeholder.
Originally reported [here](https://rust-lang.zulipchat.com/#narrow/stream/241545-t-release/topic/libs.20stabilization.20placeholder/near/296057188).
Rollup merge of #100200 - petrochenkov:zgccld2, r=lqd,Mark-Simulacrum
Change implementation of `-Z gcc-ld` and `lld-wrapper` again
This PR partially reverts https://github.com/rust-lang/rust/pull/97375 and uses the strategy described in https://github.com/rust-lang/rust/issues/97402#issuecomment-1147404520 instead, thus fixes https://github.com/rust-lang/rust/issues/97755.
Rollup merge of #100121 - Nilstrieb:mir-validator-param-env, r=oli-obk
Try normalizing types without RevealAll in ParamEnv in MIR validation
Before, the MIR validator used RevealAll in its ParamEnv for type
checking. This could cause false negatives in some cases due to
RevealAll ParamEnvs not always use all predicates as expected here.
Since some MIR passes like inlining use RevealAll as well, keep using
it in the MIR validator too, but when it fails usign RevealAll, also
try the check without it, to stop false negatives.
Fixes #99866
cc ````````@compiler-errors```````` who nicely helped me on zulip
Auto merge of #101318 - GuillaumeGomez:rollup-qsr51z4, r=GuillaumeGomez
Rollup of 9 pull requests
Successful merges:
- #97739 (Uplift the `let_underscore` lints from clippy into rustc.)
- #99583 (Add additional methods to the Demand type)
- #100147 (optimization of access level table construction)
- #100552 (rustc_target: Add a compatibility layer to separate internal and user-facing linker flavors)
- #100827 (Simplify MIR opt tests)
- #101166 (Generate error index with mdbook instead of raw HTML pages)
- #101294 (Fix #100844 rebase accident)
- #101298 (rustdoc: remove unused CSS `#main-content > .since`)
- #101304 (Add autolabels for `A-query-system`)
This rule was added (actually, it was called `#main > .since` back then) with cdca0843779eed0b9046e9fee48c91458ad51605 and you can see an example of the bug it's intended to fix in <https://doc.rust-lang.org/1.9.0/std/fmt/fn.write.html> by looking at the `1.0.0` version marker.
However, a5a2f2b951ea982a666eaf52b1874d8f1b17290b changed it so that `<span class="since">` is always placed in an out-of-band wrapper, so it's never nested directly below `#main` / `#main-content` any more.
Rollup merge of #101166 - GuillaumeGomez:error-index-mdbook, r=notriddle
Generate error index with mdbook instead of raw HTML pages
This is a follow-up of https://github.com/rust-lang/rust/pull/100922.
This comes from a remark from ````@estebank```` who said that the search was a nice thing on the previous version and that it wasn't possible anymore. An easy way to come around this limitation was to use `mdbook`, which is what I did here.
Now some explanations on the code. I'll explain how I developed this and why I reached this solution. First I did it very basically by simply setting the source directory and the output directory, generated a `SUMMARY.md` manually which listed all error codes and that was it. Two problems arose from this:
1. A lot of new HTML files were generated at the top level
2. An `index.html` file was generated at the top-level (the summary in short).
So for `1.`, it's not great to have too many files at the top-level as it could create file conflicts more easily. And for `2.`, this is actually a huge issue because <doc.rust-lang.org> generates an `index.html` file with a links to a few different resources, so it should never be overwritten. <s>Unfortunately, `mdbook` **always** generates an `index.html` file so the only solution I could see (except for sending them a contribution, I'll maybe do that later) was to temporaly move a potentially existing `index.html` file and then puts it back once done. For this last part, to ensure that we don't return *before* it has been put back, I wrapped the `mdbook` generation code inside `render_html_inner` which is called from `render_html` which in turn handle the "save" of `index.html`.</s>
EDIT: `mdbook` completely deletes ALL the content in the target directory so I instead generate into a sub directory and then I move the files to the real target directory.
To keep compatibility with the old version, I also put the `<script>` ````@notriddle```` nicely provided in the previous PR only into the `error-index.html` file to prevent unneeded repetition. I didn't use `mdbook` `additional-js` option because the JS is included at the end of all HTML files, which we don't want for two reasons:
1. It's slow.
2. We only want it to be run in `error-index.html` (even if we also ensure in the JS itself too!).
<s>Now the last part: why we generate the summary twice. We actually generate it once to tell `mdbook` what the book will look like and a second time because a create a new chapter content which will actually list all the error codes (with the updated paths).</s>
EDIT: I removed the need for two summaries.
You can test it [here](https://rustdoc.crud.net/imperio/error-index-mdbook/error-index.html).
Rollup merge of #100827 - JakobDegen:better-tests, r=wesleywiser
Simplify MIR opt tests
This commit removes many cases of MIR opt tests emitting `.diff`s for more than one pass. These tests cannot be `unit-test`s, and so they are easy to break, and they also provide little value due to having excessively strong opinions over *how* a piece of code should be optimized.
Where reasonable, we instead add separate test files that only emit the `PreCodegen.after` MIR for code where we want to track what the end to end effect of the optimization pipeline on the example code is.
Rollup merge of #100552 - petrochenkov:flavorcompat, r=lqd
rustc_target: Add a compatibility layer to separate internal and user-facing linker flavors
I want to do some refactorings in `rustc_target` - merge `lld_flavor` and `linker_is_gnu` into `linker_flavor`, support combination gcc+lld (https://github.com/rust-lang/rust/pull/96827).
This PR adds some compatibility infra that makes that possible without making any changes to user-facing interfaces - `-Clinker-flavor` values and json target specs. (For json target specs this infra may eventually go away since they are not very stable.)
The second commit does some light refactoring of internal linker flavors (applies changes from https://github.com/petrochenkov/rust/commit/53eca42973b7e379b9fa0469c33f08680b57c35f that don't require mass-editing target specs).
Rollup merge of #97739 - a2aaron:let_underscore, r=estebank
Uplift the `let_underscore` lints from clippy into rustc.
This PR resolves #97241.
This PR adds three lints from clippy--`let_underscore_drop`, `let_underscore_lock`, and `let_underscore_must_use`, which are meant to capture likely-incorrect uses of `let _ = ...` bindings (in particular, doing this on a type with a non-trivial `Drop` causes the `Drop` to occur immediately, instead of at the end of the scope. For a type like `MutexGuard`, this effectively releases the lock immediately, which is almost certainly the wrong behavior)
In porting the lints from clippy I had to copy over a bunch of utility functions from `clippy_util` that these lints also relied upon. Is that the right approach?
Note that I've set the `must_use` and `drop` lints to Allow by default and set `lock` to Deny by default (this matches the same settings that clippy has). In talking with `@estebank` he informed me to do a Crater run (I am not sure what type of Crater run to request here--I think it's just "check only"?)
On the linked issue, there's some discussion about using `must_use` and `Drop` together as a heuristic for when to warn--I did not implement this yet.
Auto merge of #100943 - jyn514:query-system-2, r=cjgillot
Simplify the `define_query` macro
This moves a bunch of control flow out of the macro into generic functions, leaving the macro just to call the function with a new generic parameter for each query.
It may be possible to improve compile-times / icache by instantiating the generic functions only with the query key, not the query type itself, but I'm going to leave that for a follow-up PR.
Helps with https://github.com/rust-lang/rust/issues/96524.
Jakob Degen [Sun, 21 Aug 2022 04:47:53 +0000 (21:47 -0700)]
Simplify MIR opt tests
This commit removes many cases of MIR opt tests emitting `.diff`s for more than one pass. These
tests cannot be `unit-test`s, and so they are easy to break, and they also provide little value due
to having excessively strong opinions over *how* a piece of code should be optimized.
Where reasonable, we instead add separate test files that only emit the `PreCodegen.after` MIR for
code where we want to track what the result of the net result of the optimization pipeline's output
is.
Auto merge of #100935 - cuviper:upgrade-android-ci, r=Mark-Simulacrum
ci: Upgrade android containers from ubuntu:16.04 to 22.04
The main goal of updating to 22.04 is to get away from `llvm.allow-old-toolchain`.
These containers are not building LLVM for android, so only the host version matters.
A side benefit is that they can also use the system `cmake` instead of building one.
This rule was added (actually, it was called `#main > .since` back then) with cdca0843779eed0b9046e9fee48c91458ad51605 and you can see an example of the
bug it's intended to fix in <https://doc.rust-lang.org/1.9.0/std/fmt/fn.write.html>
by looking at the `1.0.0` version marker.
However, a5a2f2b951ea982a666eaf52b1874d8f1b17290b changed it so that
`<span class="since">` is always placed in an out-of-band wrapper, so it's
never nested directly below `#main` / `#main-content` any more.
Auto merge of #101295 - matthiaskrgr:rollup-046o38p, r=matthiaskrgr
Rollup of 14 pull requests
Successful merges:
- #94467 (Add `special_module_name` lint)
- #100852 (Use `getuid` to check instead of `USER` env var in rustbuild)
- #101072 (bootstrap: Add llvm-has-rust-patches target option)
- #101190 (Make docs formulation more consistent for NonZero{int})
- #101245 (Remove unneeded where whitespace)
- #101251 (Fix bad target name in Walkthrough)
- #101254 (rustdoc: remove unused `.docblock .impl-items` CSS)
- #101256 (Fixes/adjustments to Fuchsia doc walkthrough)
- #101270 (Update outdated comment about output capturing in print_to.)
- #101271 (Fix filename of armv4t-none-eabi.md)
- #101274 (Fix typo in comment)
- #101279 (Fix doc_auto_cfg for impl blocks in different modules with different `cfg`)
- #101285 (Do not suggest adding `move` to closure when `move` is already used)
- #101292 (rustdoc: remove unneeded CSS `.content table td:first-child > a`)
`Builder::expr_into_pattern` has a single call site. Currently the
`pattern` argument at the call site is always cloned.
This commit changes things so that we instead do a clone within
`expr_into_pattern`, but only if the pattern has the
`PatKind::AscribeUserType` kind, and we only clone the annotation within
the pattern instead of the entire pattern.
`thir::Pat::kind` is a `Box<PatKind>`, which doesn't follow the usual
pattern in AST/HIR/THIR which is that the "kind" enum for a node is
stored inline within the parent struct.
This commit makes the `PatKind` directly inline within the `Pat`. This
requires using `Box<Pat>` in all the types that hold a `Pat.
Ideally, `Pat` would be stored in `Thir` like `Expr` and `Stmt` and
referred to with a `PatId` rather than `Box<Pat>`. But this is hard to
do because lots of `Pat`s get created after the destruction of the `Cx`
that does normal THIR building. But this does get us a step closer to
`PatId`, because all the `Box<Pat>` occurrences would be replaced with
`PatId` if `PatId` ever happened.
At 128 bytes, `Pat` is large. Subsequent commits will shrink it.
Rollup merge of #101279 - GuillaumeGomez:doc_auto_cfg_nested_impl, r=notriddle
Fix doc_auto_cfg for impl blocks in different modules with different `cfg`
Fixes #101129.
Just like reexports, impl blocks don't necessarily share the same "space" as the item they implement so we need to merge attributes from its parents as well.
This is so you can check out an upstream commit in src/llvm-project and
have everything just work.
This simplifies the logic in `is_rust_llvm` a bit; it doesn't need to
check for download-ci-llvm because we would have already errored if both
that and llvm-config were specified on the host platform.
Rollup merge of #94467 - ibraheemdev:master, r=pnkfelix
Add `special_module_name` lint
Declaring `lib` as a module is one of the most common beginner mistakes when trying to setup a binary and library target in the same crate. `special_module_name` lints against it, as well as `mod main;`
```
warning: found module declaration for main.rs
--> $DIR/special_module_name.rs:4:1
|
LL | mod main;
| ^^^^^^^^^
|
= note: a binary crate cannot be used as library
warning: found module declaration for lib.rs
--> $DIR/special_module_name.rs:1:1
|
LL | mod lib;
| ^^^^^^^^
|
= note: `#[warn(special_module_name)]` on by default
= note: lib.rs is the root of this crate's library target
= help: to refer to it from other targets, use the library's name as the path
```
Note that the help message is not the best in that it doesn't provide an example of an import path (`the_actual_crate_name::`), and doesn't check whether the current file is part of a library/binary target to provide more specific error messages. I'm not sure where this lint would have to be run to access that information.
Auto merge of #98960 - cjgillot:entry-kind, r=estebank
Remove EntryKind from metadata.
This PR continues the refactor of metadata emission to be more systematic, iterating on definitions and filtering based on each definition's `DefKind`. This allows to remove the large `EntryKind` enum, replaced by linear tables in metadata.
Allow deriving multiple subdiagnostics using one SessionSubdiagnostic
This reimplements ac638c1, which had to be reverted in the previous
commit because it contains a rebase accident that itself reverted
significant unrelated changes to SessionSubdiagnostic.
During rebase, this commit accidentally reverted unrelated changes to
the subdiagnostic derive (those allowing multipart_suggestions to be
derived). This commit reverts all changes to the subdiagnostic code made
in ac638c1f5fc, the next commit will reintroduce the actually intended
changes.