Properly parse '--extern-private' with name and path
It turns out that https://github.com/rust-lang/rust/pull/57586 didn't properly parse `--extern-private name=path`.
This PR properly implements the `--extern-private` option. I've added a new `extern-private` option to `compiletest`, which causes an `--extern-private` option to be passed to the compiler with the proper path.
Part of https://github.com/rust-lang/rust/issues/44663
On musl (and some other platforms), compiletest ends up creating a static rlib
(instead of a dylib) when building 'aux-build' crates.
This commit changes the '--extern-private' path computed by compiletest
to properly take this into account
Fix cross-crate visibility of fictive variant constructors
After merging https://github.com/rust-lang/rust/pull/59376 I realized that the code in the decoder wasn't entirely correct - we "decoded" fictive variant constructors with their variant's visibility, which could be public, rather than demoted to `pub(crate)`.
Fictive constructors are not directly usable in expression/patterns, but the effect still can be observed with imports.
r? @davidtwco
Make BufWriter use get_mut instead of manipulating inner in Write implementation
`get_mut` allows us to abstract over the implementation detail of inner being optional.
update polonius-engine
This updates polonius-engine to [version 0.7.0](https://github.com/rust-lang/polonius/blob/master/RELEASES.md#v070), which adds a hybrid algorithm that starts off with performing a cheaper, location-insensitive analysis before proceeding with the full analysis.
r? @nikomatsakis
Increase `Span` from 4 bytes to 8 bytes.
This increases the size of some important types, such as `ast::Expr` and
`mir::Statement`. However, it drastically reduces how much the interner
is used, and the fields are more natural sizes that don't require bit
operations to extract.
As a result, instruction counts drop across a range of workloads, by as
much as 10% for `script-servo` incremental builds.
Peak memory usage goes up a little for some cases, but down by more for
some other cases -- as much as 18% for non-incremental builds of
`packed-simd`.
The commit also:
- removes the `repr(packed)`, because it has negligible effect, but can
cause undefined behaviour;
- replaces explicit impls of common traits (`Copy`, `PartialEq`, etc.)
with derived ones.
r? @petrochenkov
Revert "compile crates under test w/ -Zemit-stack-sizes"
Revert PR #59401 to fix issue #59652 (a stable-to-beta regression).
This is result of squashing two revert commits:
Revert "compile all crates under test w/ -Zemit-stack-sizes"
This reverts commit 7d365cf27f.
Revert "bootstrap: build compiler-builtins with -Z emit-stack-sizes"
This reverts commit 8b8488ce8f.
----
(My intention is that someone can re-add this code again later, either after the `ld.gold` issue itself is fixed, or with safe-guards to check whether `ld.gold` is in use and then issuing warnings about the problems here when they arise.)
Clean up handling of `-Z pgo-gen` commandline option.
This PR adapts the `-Z pgo-gen` flag to how Clang and GCC handle the corresponding `-fprofile-generate` flag. In particular, the flag now optionally takes a directory to place the profiling data in and allows to omit the argument (instead of having to pass an empty string).
Exclude profiler-generated symbols from MSVC __imp_-symbol workaround.
LLVM's profiling instrumentation adds a few symbols that are used by the profiler runtime. Since these show up as globals in the LLVM IR, the compiler generates `dllimport`-related `__imp_` stubs for them. This can lead to linker errors because the instrumentation symbols have weak linkage or are in a comdat section, but the `__imp_` stubs aren't.
Instead of trying to replicate the linkage/comdat setup for the stubs, this PR just excludes the profiler-related symbols from stub-generation since they aren't supposed to be referenced via `__declspec(dllimport)` anywhere anyway.
r? @alexcrichton
EDIT: I considered making this more general, i.e. inferring from the symbol name if it is a Rust symbol or not. But then I figured out that that would yield false negatives for `#[no_mangle]` et al, so I went with a blacklist approach.
