Issue #46589 - Kill borrows on a local variable whenever we assign ov…
…er this variable
This is a first patch for the issue, handling the simple case while I figure out the data structures involved in the more complex cases.
CloudABI is a sandboxed UNIX-like runtime environment. It is a
programming environment that uses a capability-based security model. In
practice this means that many POSIX interfaces are present, except for
ones that try to access resources out of thin air. For example, open()
is gone, but openat() is present.
Right now I'm at the point where I can compile very basic CloudABI
applications on all four supported architectures (ARM and x86, 32 and 64
bits). The next step will be to get libstd to work. Patches for that are
outside the scope of this change.
More info: https://nuxi.nl/cloudabi/https://github.com/NuxiNL/cloudlibc/
The code inside this conditional will not work on FreeBSD 10+ because
those versions use clang and libc++ rather than libstdc++.
Since FreeBSD comes with libc++ in the base, presumably all 10+ systems
will have it present.
Searching for libstdc++.a will not work if it is not present. As a
result, this would previously have set `LLVM_STATIC_STDCPP=libstdc++.a`,
which isn't a valid path and caused problems later on when building
`librustc_llvm`.
This could possibly be updated in the future to look for `libc++.a` on
FreeBSD, by expanding the code inside the conditional. In one attempt
to run this on x86_64-freebsd, I found that libc++ was not compiled with
PIC, so it failed anyway.
The main goal here is to use FreeBSD's normal libc++, instead of
statically linking the libstdc++ packaged with GCC, because that
libstdc++ has bugs that cause rustc to deadlock inside LLVM.
But the easiest way to use libc++ is to switch the build from GCC to
Clang, and the Clang package in the Ubuntu image already knows how to
cross-compile (given a sysroot and preferably cross-binutils), so the
toolchain script now uses that instead of building a custom compiler.
This also de-duplicates the `build-toolchain.sh` script.
Do not emit type errors on recovered blocks
When a parse error occurs on a block, the parser will recover and create
a block with the statements collected until that point. Now a flag
stating that a recovery has been performed in this block is propagated
so that the type checker knows that the type of the block (which will be
identified as `()`) shouldn't be checked against the expectation to
reduce the amount of irrelevant diagnostic errors shown to the user.
Fix#44579.
When a parse error occurs on a block, the parser will recover and create
a block with the statements collected until that point. Now a flag
stating that a recovery has been performed in this block is propagated
so that the type checker knows that the type of the block (which will be
identified as `()`) shouldn't be checked against the expectation to
reduce the amount of irrelevant diagnostic errors shown to the user.
rustc: Sort CGUs before merging
This commit fixes some nondeterminism in compilation when using multiple codegen
units. The algorithm for splitting codegen units currently takes the
otherwise-would-be-for-incremental partitioning and then continuously merges the
two smallest codegen units until the desired number of codegen units are
reached.
We want to be sure to merge the same codegen units each time a compilation is
run but there's some subtle reorderings amongst all the items which was causing
this step to be slightly buggy. Notably this step involves sorting codegen units
by size, but if two codegen units had the same size they would appear in
different locations in the list each time.
This commit fixes this issue by sorting codegen units by name before doing the
loop to merge the two smallest. This means that we've got a deterministic
order going in and since we're using a stable sort this should mean that we're
always now getting a deterministic merging of codegen units.
Closes#46846
Generics refactoring (groundwork for const generics)
These changes were suggested by @eddyb.
After this change, the `Generics` contain one `Vec` of an enum for the generic parameters, rather than two separate `Vec`s for lifetime and type parameters. Type params and const params will need to be in a shared `Vec` to preserve their ordering, and moving lifetimes into the same `Vec` should simplify the code that processes `Generics`.
docs: do not call integer overflows as underflows
In the API docs, integer overflow is sometimes called underflow. Underflow is really when the magnitude of a floating-point number is too small so the number underflows to subnormal or zero. With integers it is always overflow, even if the expected result is less than the minimum number that can be represented.
tweaks and fixes for doc(include)
This PR makes a handful of changes around `#[doc(include="file.md")]` (https://github.com/rust-lang/rust/issues/44732):
* Turns errors when loading files into full errors. This matches the original RFC text.
* Makes the `missing_docs` lint check for `#[doc(include="file.md")]` as well as regular `#[doc="text"]` attributes.
* Loads files included by `#[doc(include="file.md")]` into dep-info, mirroring the behavior of `include_str!()` and friends.
* Adds or modifies tests to check for all of these.
incr.comp.: Precompute small hash for filenames to save some work.
For each span we hash the filename of the file it points to. Since filenames can be quite long, especially with absolute paths, this PR pre-computes a hash of the filename and we then only hash the hash.
r? @nikomatsakis
Prevent rustc overwriting input files
If rustc is invoked on a file that would be overwritten by the
compilation, the compilation now fails, to avoid accidental loss. This
resolves#13019. Kudos to @estebank, whose patch I finished off.
rustc: do not raise the alignment of optimized enums to the niche's alignment.
This is the improved fix for #46769 that does not increase the size of any types (see also #46808).
Fix ICE when calling non-functions within closures
The visitor for walking function bodies did not previously properly
handle error-cases for function calls. These are now ignored,
preventing the panic. This fixes#46771.
This commit fixes some nondeterminism in compilation when using multiple codegen
units. The algorithm for splitting codegen units currently takes the
otherwise-would-be-for-incremental partitioning and then continuously merges the
two smallest codegen units until the desired number of codegen units are
reached.
We want to be sure to merge the same codegen units each time a compilation is
run but there's some subtle reorderings amongst all the items which was causing
this step to be slightly buggy. Notably this step involves sorting codegen units
by size, but if two codegen units had the same size they would appear in
different locations in the list each time.
This commit fixes this issue by sorting codegen units by name before doing the
loop to merge the two smallest. This means that we've got a deterministic
order going in and since we're using a stable sort this should mean that we're
always now getting a deterministic merging of codegen units.
Closes#46846
MIR: terminate unreachable blocks in construct_const
Fixes#46843.
#45821 added unreachable blocks in matches, which were terminated in
construct_fn but not in construct_const, causing a panic due to "no
terminator on block" when constants involved matching on enums.
The "unimplemented expression type" error may go away in the future, the
key is that we see the E0015 about using a non-const function and then
don't ICE.
The Generics now contain one Vec of an enum for the generic parameters,
rather than two separate Vec's for lifetime and type parameters.
Additionally, places that previously used Vec<LifetimeDef> now use
Vec<GenericParam> instead.
