Change `SymbolName::name` to a `&str`.
This eliminates a bunch of `Symbol::intern()` and `Symbol::as_str()`
calls, which is good, because they require locking the interner.
Note that the unsafety in `from_cycle_error()` is identical to the
unsafety on other adjacent impls.
r? @eddyb
This now reuses `fn discriminant_ty` in project, removing
some code duplication. Doing so made me realize that
we previously had a disagreement about the discriminant
type of generators, with MIR using `u32` and codegen and
trait selection using `i32`.
We now always use `u32`.
build dist for x86_64-unknown-illumos
This change creates a new Docker image, "dist-x86_64-illumos", and sets
things up to build the full set of "dist" packages for illumos hosts, so
that illumos users can use "rustup" to install packages. It also
adjusts the manifest builder to expect complete toolchains for this
platform.
This eliminates a bunch of `Symbol::intern()` and `Symbol::as_str()`
calls, which is good, because they require locking the interner.
Note that the unsafety in `from_cycle_error()` is identical to the
unsafety on other adjacent impls.
When encountering a local binding with a type that isn't completed, the
parser will reach a `=` token. When this happen, consider the type
"complete" as far as the parser is concerned to avoid further errors
being emitted by parse recovery logic.
Note that the output of `unpretty-debug.stdout` has changed. In that
test the hash values are normalized from a symbol numbers to small
numbers like "0#0" and "0#1". The increase in the number of static
symbols must have caused the original numbers to contain more digits,
resulting in different pretty-printing prior to normalization.
lint: use `transparent_newtype_field` to avoid ICE
Fixes#73747.
This PR re-uses the `transparent_newtype_field` function instead of manually calling `is_zst` on normalized fields to determine which field in a transparent type is the non-zero-sized field, thus avoiding an ICE.
typeck: use `item_name` in cross-crate packed diag
Fixes#73112.
This PR replaces the use of `expect_local` and `hir().get` to fetch the identifier for a ADT with `item_name` - which works across crates.
pprust: support multiline comments within lines
Fixes#73626.
This PR adds support to `rustc_ast_pretty` for multiline comments that start and end within a line of source code.
Fun fact: [the commit which added this assert](d12ea39896) was from 2011!
d12ea39896/src/comp/pretty/pprust.rs (L1146-L1150)
Slight reorganization of sys/(fast_)thread_local
I was long confused by the `thread_local` and `fast_thread_local` modules in the `sys(_common)` part of libstd. The names make it *sound* like `fast_thread_local` is just a faster version of `thread_local`, but really these are totally different APIs: one provides thread-local "keys", which are non-addressable pointer-sized pieces of local storage with an associated destructor; the other (the "fast" one) provides just a destructor.
So I propose we rename `fast_thread_local` to `thread_local_dtor`, and `thread_local` to `thread_local_key`. That's what this PR does.
Don't allow `DESTDIR` to influence LLVM builds
When running a command like `DESTDIR=foo x.py install` in a completely
clean build directory, this will cause LLVM to be installed into
`DESTDIR`, which then causes the build to fail later when it attempts
to *use* those LLVM files.
This commit re-uses the `transparent_newtype_field` function instead of
manually calling `is_zst` on normalized fields to determine which field
in a transparent type is the non-zero-sized field, thus avoiding an ICE.
Signed-off-by: David Wood <david@davidtw.co>
