While going over various problems signaled by valgrind when running
`make check` on a build configured with `--enable-valgrind`, I
discovered a bug in this test case.
Namely, the test case was previously creating an `i32` (originally an
`int` aka `isize` but then we changed the name and the fallback
rules), and then reading from a `*const isize`. Valgrind rightly
complains about this, since we are reading an 8 byte value on 64-bit
systems, but in principle only 4 bytes have been initialized.
(I wish this was the only valgrind unclean test, but unfortunately
there are a bunch more. This was just the easiest/first one that I
dissected.)
The methods gave wrong results for TyIs and TyUs, whose suffix len
should be 5 nowadays. But since they were only used for parsing,
and unneeded for that since 606a309d, remove them rather than fixing.
Returning a primitive bool results in a somewhat confusing API - does
`true` indicate success - i.e. no timeout, or that a timeout has
occurred? An explicitly named enum makes it clearer.
[breaking-change]
* Add previously omitted function `Arc::try_unwrap(Self) -> Result<T, Self>`
* Move `arc.downgrade()` to `Arc::downgrade(&Self)` per conventions.
* Deprecate `Arc::weak_count` and `Arc::strong_count` for raciness. It is almost
impossible to correctly act on these results without a CAS loop on the actual
fields.
* Rename `Arc::make_unique` to `Arc::make_mut` to avoid uniqueness terminology
and to clarify relation to `Arc::get_mut`.
* Add `Rc::would_unwrap(&Self) -> bool` to introspect whether try_unwrap would succeed,
because it's destructive (unlike get_mut).
* Move `rc.downgrade()` to `Rc::downgrade(&Self)` per conventions.
* Deprecate `Rc::weak_count` and `Rc::strong_count` for questionable utility.
* Deprecate `Rc::is_unique` for questionable semantics (there are two kinds of
uniqueness with Weak pointers in play).
* Rename `rc.make_unique()` to `Rc::make_mut(&mut Self)` per conventions, to
avoid uniqueness terminology, and to clarify the relation to `Rc::get_mut`.
Currently `f32 % f32` will generate a link error on 32-bit MSVC because LLVM
will lower the operation to a call to the nonexistent function `fmodf`. Work
around in this in the backend by lowering to a call to `fmod` instead with
necessary extension/truncation between floats/doubles.
Closes#27859
In order to test the validity of identifiers, exposing the name resolution module is necessary. Other changes mostly comprise of exposing modules publicly like parts of save-analysis, so they can be called appropriately.
This commit renames the `CString::{into_ptr, from_ptr}` methods to `into_raw`
and `from_raw` to mirror the corresponding methods on `Box` and the naming of
"raw" for `from_raw_parts` on slices and vectors.
cc #27769
