rustc: Remove `used_mut_nodes` from `TyCtxt`
This updates the borrowck query to return a result, and this result is then used
to incrementally check for unused mutable nodes given sets of all the used
mutable nodes.
Closes#42384
While an inner attribute here is in fact erroneous, that error ("inner
attribute is not permitted in this context") successfully gets set earlier;
this further admonition is nonsensical.
Resolves#45296.
rustbuild: Support specifying archiver and linker explicitly
With this patch `x.py test` passes without toolchain being in `PATH` if `cc`, `cxx`, `ar`, `linker` and `gdb` are specified in `config.toml` (except for a few `run-make` tests using `nm`).
Fixes https://github.com/rust-lang/rust/issues/41821
r? @Mark-Simulacrum
rustc: Fix some ThinLTO internalization
First the `addPreservedGUID` function forgot to take care of "alias" summaries.
I'm not 100% sure what this is but the current code now matches upstream. Next
the `computeDeadSymbols` return value wasn't actually being used, but it needed
to be used! Together these should...
Closes#45195
There was a bit of code shared between `try_get_with` and `ensure`, after I
added `ensure`. I refactored that shared code into a query-agnostic method
called `read_node_index`.
The new method `read_node_index` will attempt to find the node
index (`DepNodeIndex`) of a query. When `read_node_index` finds the
`DepNodeIndex`, it marks the current query as a reader of the node it's
requesting the index of.
This is used by `try_get_with` and `ensure` as it elides the unimportant (to
them) details of if the query is invalidated by previous changed computation (Red)
or new and if they had to mark the query green. For both `try_get_with` and
`ensure`, they just need to know if they can lookup the results or have to
reevaluate.
some low-hanging rustdoc optimizations
There were a few discussions earlier today in #rust-internals about the syscall usage and overall performance of rustdoc. This PR is intended to pick some low-hanging fruit and try to rein in some of the performance issues of rustdoc.
First the `addPreservedGUID` function forgot to take care of "alias" summaries.
I'm not 100% sure what this is but the current code now matches upstream. Next
the `computeDeadSymbols` return value wasn't actually being used, but it needed
to be used! Together these should...
Closes#45195
incr.comp.: Introduce `ensure` and `ensure` typeck_tables_of
Resolves#45210
In this Pull Request we introduce the `ensure` query/function. `ensure` has the
semantics and type of the function `Q1` below:
```rust
fn Q1::ensure(K){
Q(K);
}
```
Further, `ensure` avoids the need to load the result from disk (or execute the
provider, if we are not storing the results of Q to disk).
@nikomatsakis
Provide the full span of method calls to `check_argument_types`
... so that it includes the span of the passed arguments, not just the name of the called method.
Fixes#44760.
Fix TcpStream::connect_timeout on linux
Linux appears to set POLLOUT when a conection's refused, which is pretty
weird. Invert the check to look for an error explicitly. Also add an
explict test for this case.
Closes#45265.
r? @alexcrichton
Do some cleanups for hashmaps
@mystor noticed some things whilst reading through the hashmap RawTable code.
Firstly, in RawTable we deal with this hash_offset value that is the offset of the list of hashes from the buffer start. This is always zero, and this isn't consistently used (which means that we would have bugs if we set it to something else). We should just remove this since it doesn't help us at all.
Secondly, the probing length tag is not copied when cloning a raw table. This is minor and basically means we do a bit more work than we need on further inserts on a cloned hashmap.
r? @Gankro
compiletest/runtest: format ErrorKind with Display
The strings are nouns for the most part, so we give ErrorKind::Help a
more sensible string. This reduces quote hiccups in failure output.
unexpected "error": '...'
↓
unexpected error: '...'
Moved details of unstable non-ascii identifiers from the Reference
Moved details of unstable non-ascii identifiers from the Reference to the Unstable book
r? @steveklabnik
Add x86_64-unknown-linux-gnux32 target
This adds X32 ABI support for Linux on X86_64. Let's package and dist it so we can star testing libc, libstd, etc.
Fixes https://github.com/rust-lang/rfcs/issues/1339
doc-test: In Markdown tests, Use all of `<h1>` to `<h6>` as the test name
This mainly simplifies debugging error index tests, as the error codes are `<h2>`s in the huge document containing all codes.
Linux appears to set POLLOUT when a conection's refused, which is pretty
weird. Invert the check to look for an error explicitly. Also add an
explict test for this case.
Closes#45265.
Enable building clippy in CI
r? @alexcrichton
As discussed at Rustfest. Measured additional time is 4 minutes on my machine if no dependencies are shared with other tools. In reality most dependencies are shared (especially the slow to compile ones like serde).
cc @Manishearth
Does not run clippy's test suite, since
a) it is nontrivial in the rustc build system
b) it breaks more frequently but the breakage is negligible
If clippy breaks, the procedure to follow is documented under https://github.com/rust-lang/rust/blob/master/CONTRIBUTING.md#external-dependencies
This updates the borrowck query to return a result, and this result is then used
to incrementally check for unused mutable nodes given sets of all the used
mutable nodes.
Closes#42384
rustdoc is built separately to rustc now so the docs would need to be
generated separately as well. Also rustdoc doesn't build at stage 1
which prevented the compiler docs being built at stage 1.
This isn't strictly necessary for hashmap cloning to work. The tag is
used to hint for an upcoming resize, so it's good to copy this
information over.
(We can do cleverer things like actually resizing the hashmap when we
see the tag, or even cleaning up the entry order, but this requires
more thought and might not be worth it)
This offset is always zero, and we don't consistently take it into
account. This is okay, because it's zero, but if it ever changes we're
going to have bugs (e.g. in the `dealloc` call, where we don't take it
into account).
It's better to remove this for now; if we ever have a need for a
nonzero offset we can add it back, and handle it properly when we do so.
MIR-borrowck: Migrate remaining ast diagnostics
This PR migrates all of the remaining diagnostics in `rustc_borrowck` over to `rustc_mir`, exposing them for use by both AST-borrowck and MIR-borrowck.
This should hopefully resolve all remaining cases of diagnostic messages emitted from borrowck under `-Z borrowck-mir` without an origin annotation.
Modify MIR testing to require consecutive lines
MIR testing now requires that lines be consecutive. To achive this,
instead of collecting the expected mir as a string, it is now wrapped in
an `ExpectedLine` enum, that is either `Elision` or `Text(T)` where `T:
AsRef<str>`. `Text` lines must be matched in order, unless separated by
`Elision` lines. Elision occurs lazily, that is, an Elision will skip
as few lines as possible.
To add a new elision marker. Put a comment containing only "..." and
whitespace in any MIR testing block. Like so:
```
// fn write_42(_1: *mut i32) -> bool {
// ...
// bb0: {
// Validate(Acquire, [_1: *mut i32]);
// Validate(Release, [_1: *mut i32]);
// ...
// return;
// }
// }
```
Right now, all input before the line right after `// START` is elided,
and all input after the line right before `// END` is also not tested.
Many tests need to be updated. That will follow in the next commit.
cc #45153
r? @nikomatsakis
