Generalize LLD usage in bootstrap
The current usage of using LLD (`rust.use-lld = true`) in bootstrap is a bit messy. What it claimed:
> Indicates whether LLD will be used to link Rust crates during bootstrap on
> supported platforms. The LLD from the bootstrap distribution will be used
> and not the LLD compiled during the bootstrap.
What it did:
1) On MSVC, it did indeed use the snapshot compiler's `rust-lld`, but at the same time it was invoking a global `lld` binary (since https://github.com/rust-lang/rust/pull/102101), therefore it wouldn't work if `lld` wasn't available.
2) On other targets, it was just straight up using a global `lld` linker. If it wasn't available, it would fail.
This PR (hopefully) cleans up handling of LLD in bootstrap. It introduces a new enum called `LldMode`, which explicitly distinguishes between no LLD, external LLD and self-contained LLD. Since it's non-trivial to provide a custom path to LLD, if an external `lld` is used, the linker binary has to be named exactly `lld` and it has to be available in PATH.
In addition, this PR also dog-foods [MCP510](https://github.com/rust-lang/compiler-team/issues/510) in bootstrap.
To keep backwards compatibility somewhat, I kept the original `use-lld` flag and mapped the `true` value to `"external"`, which is how it behaved before on Linux and other non-MSVC targets.
Having the option to use an external `lld` on Linux should come in handy for testing on CI once MCP510 sets the default linker on Linux to `lld`.
Note that thanks to MCP510, currently "self-contained" means that `lld` is used from the stage N-1 compiler (before, we always used `lld` from the snapshot/stage0 compiler).
Best reviewed commit by commit.
CC `@petrochenkov`
remove redundant imports
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and removing redundant imports code into two PR.
r? `@petrochenkov`
Don't print host effect param in pretty `path_generic_args`
Make `own_args_no_defaults` pass back the `GenericParamDef`, so that we can pass both the args *and* param definitions into `path_generic_args`. That allows us to use the `GenericParamDef` to filter out effect params.
This allows us to filter out the host param regardless of whether it's `sym::host` or `true`/`false`.
This also renames a couple of `const_effect_param` -> `host_effect_param`, and restores `~const` pretty printing to `TraitPredPrintModifiersAndPath`.
cc #118785
r? `@fee1-dead` cc `@oli-obk`
Stablize arc_unwrap_or_clone
Fixes: #93610
This likely needs FCP. I created this PR as it's stabilization is trivial and FCP can be just conducted here. Not sure how to ping the libs API team (last attempt didn't work apparently according to GH UI)
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and
removing redundant imports code into two PR.
Rollup of 5 pull requests
Successful merges:
- #117966 (add safe compilation options)
- #118747 (Remove extra check cfg handled by libc directly)
- #118774 (add test for inductive cycle hangs)
- #118775 (chore: add test case for type with generic)
- #118782 (use `&` instead of start-process in x.ps1)
r? `@ghost`
`@rustbot` modify labels: rollup
Don't warn an empty pattern unreachable if we're not sure the data is valid
Exhaustiveness checking used to be naive about the possibility of a place containing invalid data. This could cause it to emit an "unreachable pattern" lint on an arm that was in fact reachable, as in https://github.com/rust-lang/rust/issues/117119.
This PR fixes that. We now track whether a place that is matched on may hold invalid data. This also forced me to be extra precise about how exhaustiveness manages empty types.
Note that this now errs in the opposite direction: the following arm is truly unreachable (because the binding causes a read of the value) but not linted as such. I'd rather not recommend writing a `match ... {}` that has the implicit side-effect of loading the value. [Never patterns](https://github.com/rust-lang/rust/issues/118155) will solve this cleanly.
```rust
match union.value {
_x => unreachable!(),
}
```
I recommend reviewing commit by commit. I went all-in on the test suite because this went through a lot of iterations and I kept everything. The bit I'm least confident in is `is_known_valid_scrutinee` in `check_match.rs`.
Fixes https://github.com/rust-lang/rust/issues/117119.
As bootstrap locks its entire build directory, parallel bootstrapping
for anything becomes impossible. This change enables developers to bypass
the locking mechanism when it is unnecessary for their specific use case.
Signed-off-by: onur-ozkan <work@onurozkan.dev>
Build Fuchsia in CI
Fittingly, when I first put this up it was failing due to discovering an ICE in clippy (looks like fixed in https://github.com/rust-lang/rust-clippy/pull/11760), probably more fallout from recent type system changes. Other recent regressions this would have caught include
- #117455 and #117493
- #117602
Originally we discussed basing this on cargotest, but they ended up not sharing anything. Fuchsia has its own tool to manage checkouts and its own build system. What it requires is a fully "install"ed toolchain with a host and fuchsia target. We share logic from the dist-various-2 builder to build the fuchsia target.
Right now this runs clippy and skips linking a bunch of targets, since most issues we catch are in the frontend. In theory we could probably get the build CPU time down quite a bit with this approach, but right now some linked targets are creeping into the dependencies anyway and we don't have a good way of preventing that yet.
The approach is basically to get a checkout at a pinned commit and then run a [script](https://fuchsia-review.git.corp.google.com/c/fuchsia/+/943833/6/scripts/rust/build_fuchsia_from_rust_ci.sh) at a predetermined location. I would like to update that pin every few weeks. Partial checkouts are used to minimize clone time, but we don't filter out prebuilt packages.
r? `@Mark-Simulacrum`
Based on discussion in [this Zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/242791-t-infra/topic/Putting.20Fuchsia.20in.20crater).
use `&` instead of start-process in x.ps1
start-process has weird parsing rules and buggy behavior. we've already had to work around them several times, and the workarounds were not complete. i wonder who could have added it HMMMMMM
```
PS C:\Users\jyn\src\rust> git log --reverse -S Start-Process x.ps1
commit 775c3c0493e9a383a7f1c521b06d36f2e3d0d886
Author: Jynn Nelson <github@jyn.dev>
Date: Sun Jul 31 14:02:31 2022 -0500
Add `x.sh` and `x.ps1` shell scripts
```
the latest broken thing is trailing backslashes:
```
$ x.ps1 t .\tests\ui\error-emitter\
```
would be transformed into
```
['t', '.\\tests\\ui\\error-emitter"']
```
rather than trying to hack around that too, abandon start-process altogether and just use `&`.
r? `@ChrisDenton`
add test for inductive cycle hangs
the same pattern is already tested for coinductive cycles, but I now understand the underlying issue and want to make sure we also test it for inductive ones
r? `@compiler-errors`
Remove extra check cfg handled by libc directly
The `libc` crate has handle for quite some time now [check-cfg in it's own build script](497ac428bc/build.rs (L6-L32)).
We therefor no longer need to manually define them.
add safe compilation options
Add two options when building rustc : strip and stack protector.
If set `strip = true`, `rustc` will be stripped of symbols using `-Cstrip=symbols`.
Also can set `stack-protector` and then `rustc` will be compiled with stack protectors.
start-process has weird parsing rules and buggy behavior. we've already had to work around them several times, and the workarounds were not complete.
i wonder who could have added it HMMMMMM
```
PS C:\Users\jyn\src\rust> git log --reverse -S Start-Process x.ps1
commit 775c3c0493e9a383a7f1c521b06d36f2e3d0d886
Author: Jynn Nelson <github@jyn.dev>
Date: Sun Jul 31 14:02:31 2022 -0500
Add `x.sh` and `x.ps1` shell scripts
```
the latest broken thing is trailing backslashes:
```
$ x.ps1 t .\tests\ui\error-emitter\
```
would be transformed into
```
['t', '.\\tests\\ui\\error-emitter"']
```
rather than trying to hack around that too, abandon start-process altogether and just use `&`.
Lower some forgotten spans
I wrote a HIR visitor that visited all of the spans in the HIR, and made it ICE when we have a unlowered span. That led me to discover these unlowered spans.
Strengthen well known check-cfg names and values test
https://github.com/rust-lang/rust/pull/118494 is changing the implementation of how we expect well known check-cfg names and values, but we currently don't have a test that checks every well known only some of them.
This PR therefore strengthen our well known names/values test to include all of the configs to at least avoid unintended regressions and validate new entry.
*this PR also contains some drive-by consolidation of unexpected `target_os`, `target_arch` into a single file*
r? `@nnethercote` (maybe? feel free to re-assign)
Add more SIMD platform-intrinsics
- [x] simd_masked_load
- [x] LLVM codegen - llvm.masked.load
- [x] cranelift codegen - implemented but untested
- [ ] simd_masked_store
- [x] LLVM codegen - llvm.masked.store
- [ ] cranelift codegen
Also added a run-pass test to test both intrinsics, and additional build-fail & check-fail to cover validation for both intrinsics
as they unnecessarily clutter the diagnostic output and make the
experience of adding a new target to the compiler more painful than
it should be.
target_os and target_arch are still being tested in the
well-known-values.rs test, but in one place.
Make async generators fused by default
I actually changed my mind about this since the implementation PR landed. I think it's beneficial for `async gen` blocks to be "fused" by default -- i.e., for them to repeatedly return `Poll::Ready(None)` -- rather than panic.
We have [`FusedStream`](https://docs.rs/futures/latest/futures/stream/trait.FusedStream.html) in futures-rs to represent streams with this capability already anyways.
r? eholk
cc ```@rust-lang/wg-async,``` would like to know if anyone else has opinions about this.
coverage: Simplify the heuristic for ignoring `async fn` return spans
The code for extracting coverage spans from MIR has a special heuristic for dealing with `async fn`, so that the function's closing brace does not have a confusing double count.
The code implementing that heuristic is currently mixed in with the code for flushing remaining spans after the main refinement loop, making the refinement code harder to understand.
We can solve that by hoisting the heuristic to an earlier stage, after the spans have been extracted and sorted but before they have been processed by the refinement loop.
The coverage tests verify that the heuristic is still effective, so coverage mappings/reports for `async fn` have not changed.
---
This PR also has the side-effect of fixing the `None some_prev` panic that started appearing after #118525.
The old code assumed that `prev` would always be present after the refinement loop. That was only true if the list of collected spans was non-empty, but prior to #118525 that didn't seem to come up in practice. After that change, the list of collected spans could be empty in some specific circumstances, leading to panics.
The new code uses an `if let` to inspect `prev`, which correctly does nothing if there is no span present.
