There should never be an `Active` but not initialized node in the
borrow tree. If such a node exists, and if it has a protector, then
on a foreign read, it could become disabled. This leads to some
problems when formally proving that read-reordering optimizations are
sound.
The root node is the only node for which this can happen, since all
other nodes can only become `Active` when actually used. But special-
casing the root node here is annoying to reason about, everything
becomes much nicer if we can simply say that *all* `Active` nodes
must be initialized. This requires making the root node default-
initialized.
This is also more intuitive, since the root arguably becomes ini-
tialized during the allocation, which can be considered a write.
phase_rustdoc: add a heuristic to make us more certain that this is really rustdoc
Also add anyhow to test-cargo-miri; it has a custom build probe and is widely used so let's make sure the build script does not fail.
CONTRIBUTING: vsocde settings: add --all-targets
Seems like when setting a custom command, one needs to pass this flag manually, it's no longer automatically added by RA.
miri script: build with stable toolchain
`./miri toolchain` sets up a `rustup override miri`. But then if something goes wrong and the `miri` toolchain doesn't work, one can't even run `./miri toolchain` again as building miri-script needs a toolchain...
So let's always use stable to build miri-script, making it override-independent. I assume everyone will have that installed.
Add libc direct test for shims and update CONTRIBUTING.md
Add more ``libc`` direct test for shims as mentioned in #3179.
Changes:
- Added ``libc`` direct tests for ``rename`` and ``ftruncate64``
- Added the command for running ``pass-dep`` test in ``CONTRIBUTING.md``
Rollup of 9 pull requests
Successful merges:
- #121281 (regression test for #103626)
- #122168 (Fix validation on substituted callee bodies in MIR inliner)
- #122217 (Handle str literals written with `'` lexed as lifetime)
- #122379 (transmute: caution against int2ptr transmutation)
- #122840 (`rustdoc --test`: Prevent reaching the maximum size of command-line by using files for arguments if there are too many)
- #122907 (Uniquify `ReError` on input mode in canonicalizer)
- #122942 (Add test in higher ranked subtype)
- #122943 (add a couple more ice tests)
- #122963 (core/panicking: fix outdated comment)
r? `@ghost`
`@rustbot` modify labels: rollup
Add test in higher ranked subtype
I'm a beginner in this repository, and there are some things I'm not sure about:
- Is it okay that there is a warning:
```
rustc_infer::infer::relate::generalize may incompletely handle alias type: AliasTy { args: [?1t, '^0.Named(DefId(0:15 ~ structually_relate_aliases[de75]::{impl#1}::'a), "'a")], def_id: DefId(0:5 ~ structually_relate_aliases[de75]::ToUnit::Unit) }
```
- Is it okay that there are two duplicate errors in the same line?
- Did I put the test in the right place?
Any suggestions would be appreciated.
Fixes#121649
`rustdoc --test`: Prevent reaching the maximum size of command-line by using files for arguments if there are too many
Fixes#122722.
Thanks to this I discovered that rust was using ``@`` to add arguments from a file, quite convenient.
If there are too many `cfg` arguments given to `rustdoc --test`, it'll now put them into a temporary file and passing it as argument to the rustc command.
I added a test with 100_000 `cfg` arguments to ensure it'll not break again.
r? `@notrid`
transmute: caution against int2ptr transmutation
This came up in https://github.com/rust-lang/rust/pull/121282.
Cc ```@saethlin``` ```@scottmcm```
Eventually we'll add a proper description of provenance that we can reference, but that's a bunch of work and it's unclear who will have the time to do that when. Meanwhile, let's at least do what we can without mentioning provenance explicitly.
Handle str literals written with `'` lexed as lifetime
Given `'hello world'` and `'1 str', provide a structured suggestion for a valid string literal:
```
error[E0762]: unterminated character literal
--> $DIR/lex-bad-str-literal-as-char-3.rs:2:26
|
LL | println!('hello world');
| ^^^^
|
help: if you meant to write a `str` literal, use double quotes
|
LL | println!("hello world");
| ~ ~
```
```
error[E0762]: unterminated character literal
--> $DIR/lex-bad-str-literal-as-char-1.rs:2:20
|
LL | println!('1 + 1');
| ^^^^
|
help: if you meant to write a `str` literal, use double quotes
|
LL | println!("1 + 1");
| ~ ~
```
Fix#119685.
Fix validation on substituted callee bodies in MIR inliner
When inlining a coroutine, we will substitute the MIR body with the args of the call. There is code in the MIR validator that attempts to prevent query cycles, and will use the coroutine body directly when it detects that's the body that's being validated. That means that when inlining a coroutine body that has been substituted, it may no longer be parameterized over the original args of the coroutine, which will lead to substitution ICEs.
Fixes#119064
refactor check_{lang,library}_ub: use a single intrinsic
This enacts the plan I laid out [here](https://github.com/rust-lang/rust/pull/122282#issuecomment-1996917998): use a single intrinsic, called `ub_checks` (in aniticpation of https://github.com/rust-lang/compiler-team/issues/725), that just exposes the value of `debug_assertions` (consistently implemented in both codegen and the interpreter). Put the language vs library UB logic into the library.
This makes it easier to do something like https://github.com/rust-lang/rust/pull/122282 in the future: that just slightly alters the semantics of `ub_checks` (making it more approximating when crates built with different flags are mixed), but it no longer affects whether these checks can happen in Miri or compile-time.
The first commit just moves things around; I don't think these macros and functions belong into `intrinsics.rs` as they are not intrinsics.
r? `@saethlin`
make CI failures easier to interpret
RUST_BACKTRACE=1 means we show 2 backtraces that are almost always irrelevant: from ui_test and from the miri script. Instead let's set the env var inside the test harness so we see backtraces of Miri ICEing but nothing else.
Rollup of 11 pull requests
Successful merges:
- #120577 (Stabilize slice_split_at_unchecked)
- #122698 (Cancel `cargo update` job if there's no updates)
- #122780 (Rename `hir::Local` into `hir::LetStmt`)
- #122915 (Delay a bug if no RPITITs were found)
- #122916 (docs(sync): normalize dot in fn summaries)
- #122921 (Enable more mir-opt tests in debug builds)
- #122922 (-Zprint-type-sizes: print the types of awaitees and unnamed coroutine locals.)
- #122927 (Change an ICE regression test to use the original reproducer)
- #122930 (add panic location to 'panicked while processing panic')
- #122931 (Fix some typos in the pin.rs)
- #122933 (tag_for_variant follow-ups)
r? `@ghost`
`@rustbot` modify labels: rollup
Report retags as distinct from real memory accesses for data races
This changes the error reporting for data races such that reference invariants are no longer reported as real read and writes.
Before:
```
Data race detected between (1) non-atomic write on thread `unnamed-6` and (2) non-atomic read on thread `unnamed-5` at alloc1034971+0x10c. (2) just happened here
```
After:
```
Data race detected between (1) non-atomic write on thread `unnamed-8` and (2) shared reference invariant on thread `unnamed-6` at alloc1018329+0x190. (2) just happened here
```
Non-atomic read accesses from the *other* thread don't have this information tracked so those are called `some potential non-atomic read access` here.
