Revert "Disallow ambiguous attributes on expressions" on nightly
As discussed in [today's t-compiler meeting](https://rust-lang.zulipchat.com/#narrow/stream/238009-t-compiler.2Fmeetings/topic/.5Bweekly.5D.202024-06-06/near/443079505), this reverts PR #124099 to fix P-critical beta regressions #125199.
r? ``@wesleywiser``
Opening as draft so that ``@wesleywiser`` and ``@apiraino,`` you can tell me whether you wanted:
1. a `beta-accepted` revert of #124099 on nightly (this PR)? That will need to be backported to beta (even though #126093 may be the last of those)
2. a revert of #124099 on beta?
3. all of the above?
I also opened #126102, another draft PR to revert #124099 on beta, should you choose options 2 or 3.
Rollup of 12 pull requests
Successful merges:
- #125220 (Repair several `riscv64gc-unknown-linux-gnu` codegen tests)
- #126033 (CI: fix publishing of toolstate history)
- #126034 (Clarify our tier 1 Windows Server support)
- #126035 (Some minor query system cleanups)
- #126051 (Clarify an `x fmt` error.)
- #126059 (Raise `DEFAULT_MIN_STACK_SIZE` to at least 64KiB)
- #126064 (Migrate `run-make/manual-crate-name` to `rmake.rs`)
- #126072 (compiletest: Allow multiple `//@ run-flags:` headers)
- #126073 (Port `tests/run-make-fulldeps/obtain-borrowck` to ui-fulldeps)
- #126081 (Do not use relative paths to Rust source root in run-make tests)
- #126086 (use windows compatible executable name for libcxx-version)
- #126096 ([RFC-2011] Allow `core_intrinsics` when activated)
r? `@ghost`
`@rustbot` modify labels: rollup
compiletest: Allow multiple `//@ run-flags:` headers
While working on some tests, I was annoyed to find that multiple `// `@run-flags:`` headers do not combine with each other (as `//@ compile-flags:` headers do), and instead all but one are silently discarded.
This makes it impossible to split long flag lists into multiple lines.
Fortunately it's easy to just recycle the existing logic from the other command-line-flags headers.
Some minor query system cleanups
* Improves diagnostics on conflicting query flags
* removes unnecessary impls
* `track_caller`
pulled out of https://github.com/rust-lang/rust/pull/115613
Clarify our tier 1 Windows Server support
I've been asked a number of times about our minimum Windows Server support so this PR updates the documentation to be more explicit.
Note that this doesn't change our support at all, it just clarifies it. Windows Server 2016 is the first "Windows 10" server OS (specifically it has the same [build number](https://en.wikipedia.org/wiki/Windows_Server_2016#Release_to_manufacturing) as Windows 10 1607).
CI: fix publishing of toolstate history
Hopefully fixes the upload of toolstate history that I broke in https://github.com/rust-lang/rust/pull/125145. The problem is that the toolstate environment variables need to be available not just when updating `latest.json` through the Bash/Python script, but also in the main CI workflow, where `history` is updated in bootstrap (the toolstate logic is distributed in two places :/ ).
The only tricky thing is how to pass the token secret to the CI job, as I need to enable it only for privileged (`auto`/`try`) builds. I assume that the secret is filled only on `rust-lang-ci`, not on `rust-lang`, so I chose the easiest way of just configuring the environment variable globally. We'll see if that works on PR CI.
r? `@ehuss`
Previously, we were running bare commands for `ToolBuild` step and
were unable to utilize some of the flags which are already handled by
`compile::stream_cargo`.
This change makes `ToolBuild` to use `compile::stream_cargo`, allowing us
to benefit from the flags supported by the bootstrap cargo.
Signed-off-by: onur-ozkan <work@onurozkan.dev>
Update to LLVM 18.1.7
This release only contains a single commit in code that we build: 7e6ece9b4f It addresses a regression in LLVM 18.1.6 that may result in compiler crashes when targeting ppc.
check host's libstdc++ version when using ci llvm
If the host's libstdc++ version is too old using ci-llvm may result in an ABI mismatch between the local libstdc++ and libLLVM.so. This PR adds a sanity check to immediately fail at the beginning of the bootstrap before starting the actual build. I am not sure if '8' is the best threshold, but it should be a good start and we can increase it anytime if needed.
Fixes#123037
fix: do not resolve prelude within block modules
fix#17338 (continuing from #17251).
In #17251, we injected preludes into non-top-level modules, which leading to r-a to directly resolve names in preludes in block modules. This PR fix it by checking whether the module is a pseudo-module introduced by blocks. (similar to what we do for extern preludes)
Changed package.json so vscode extension settings have submenus
There are a lot of options that are a part of rust-analyzer, sometimes it can be hard to find an option that you are looking for. To fix this I have put all configurations into categories based on their names. I have also changed the schema in `crates/rust-analyzer/src/config.rs` to reflect this.
Currently for each generated entry the title is redeclared, this does function but I am prepared to change this if it is a problem.
fix: Highlight unlinked files consistently with inactive files
Currently, rust-analyzer highlights the entire region when a `cfg` is inactive (e.g. `#[cfg(windows)]` on a Linux machine). However, unlinked files only highlight the first three characters of the file.
This was introduced in #8444, but users have repeatedly found themselves with no rust-analyzer support for a file and unsure why (see e.g. #13226 and the intentionally prominent pop-up added in PR #14366).
(Anecdotally, we see this issue bite our users regularly, particularly people new to Rust.)
Instead, highlight the entire inactive file, but mark it as all as unused. This allows users to hover and run the quickfix from any line.
Whilst this is marginally more prominent, it's less invasive than a pop-up, and users do want to know why they're getting no rust-analyzer support in certain files.
Before (note the subtle grey underline is only at the beginning of the first line):
After (user can hover and fix from any line):
Remove the `ty` field from type system `Const`s
Fixes#125556Fixes#122908
Part of the work on `adt_const_params`/`generic_const_param_types`/`min_generic_const_exprs`/generally making the compiler nicer. cc rust-lang/project-const-generics#44
Please review commit-by-commit otherwise I wasted a lot of time not just squashing this into a giant mess (and also it'll be SO much nicer because theres a lot of fluff changes mixed in with other more careful changes if looking via File Changes
---
Why do this?
- The `ty` field keeps causing ICEs and weird behaviour due to it either being treated as "part of the const" or it being forgotten about leading to ICEs.
- As we move forward with `adt_const_params` and a potential `min_generic_const_exprs` it's going to become more complex to actually lower the correct `Ty<'tcx>`
- It muddles the idea behind how we check `Const` arguments have the correct type. By having the `ty` field it may seem like we ought to be relating it when we relate two types, or that its generally important information about the `Const`.
- Brings the compiler more in line with `a-mir-formality` as that also tracks the type of type system `Const`s via `ConstArgHasType` bounds in the env instead of on the `Const` itself.
- A lot of stuff is a lot nicer when you dont have to pass around the type of a const lol. Everywhere we construct `Const` is now significantly nicer 😅
See #125671's description for some more information about the `ty` field
---
General summary of changes in this PR:
- Add `Ty` to `ConstKind::Value` as otherwise there is no way to implement `ConstArgHasType` to ensure that const arguments are correctly typed for the parameter when we stop creating anon consts for all const args. It's also just incredibly difficult/annoying to thread the correct `Ty` around to a bunch of ctfe functions otherwise.
- Fully implement `ConstArgHasType` in both the old and new solver. Since it now has no reliance on the `ty` field it serves its originally intended purpose of being able to act as a double check that trait vs impls have correctly typed const parameters. It also will now be able to be responsible for checking types of const arguments to parameters under `min_generic_const_exprs`.
- Add `Ty` to `mir::Const::Ty`. I dont have a great understanding of why mir constants are setup like this to be honest. Regardless they need to be able to determine the type of the const and the easiest way to make this happen was to simply store the `Ty` along side the `ty::Const`. Maybe we can do better here in the future but I'd have to spend way more time looking at everywhere we use `mir::Const`.
- rustdoc has its own `Const` which also has a `ty` field. It was relatively easy to remove this.
---
r? `@lcnr` `@compiler-errors`
