Implement the `!` type
This implements the never type (`!`) and hides it behind the feature gate `#[feature(never_type)]`. With the feature gate off, things should build as normal (although some error messages may be different). With the gate on, `!` is usable as a type and diverging type variables (ie. types that are unconstrained by anything in the code) will default to `!` instead of `()`.
Take commandline arguments into account for incr. comp.
Implements the conservative strategy described in https://github.com/rust-lang/rust/issues/33727.
From now one, every time a new commandline option is added, one has to specify if it influences the incremental compilation cache. I've tried to implement this as automatic as possible: One just has to added either the `[TRACKED]` or the `[UNTRACKED]` marker next to the field. The `Options`, `CodegenOptions`, and `DebuggingOptions` definitions in `session::config` show plenty of examples.
The PR removes some cruft from `session::config::Options`, mostly unnecessary copies of flags also present in `DebuggingOptions` or `CodeGenOptions` in the same struct.
One notable removal is the `cfg` field that contained the values passed via `--cfg` commandline arguments. I chose to remove it because (1) its content is only a subset of what later is stored in `hir::Crate::config` and it's pretty likely that reading the cfgs from `Options` would not be what you wanted, and (2) we could not incorporate it into the dep-tracking hash of the `Options` struct because of how the test framework works, leaving us with a piece of untracked but vital data.
It is now recommended (just as before) to access the crate config via the `krate()` method in the HIR map.
Because the `cfg` field is not present in the `Options` struct any more, some methods in the `CompilerCalls` trait now take the crate config as an explicit parameter -- which might constitute a breaking change for plugin authors.
[MIR] Add Storage{Live,Dead} statements to emit llvm.lifetime.{start,end}.
Storage live ranges are tracked for all MIR variables and temporaries with a drop scope.
`StorageLive` is lowered to `llvm.lifetime.start` and `StorageDead` to `llvm.lifetime.end`.
There are some improvements possible here, such as:
* pack multiple storage liveness statements by using the index of first local + `u64` bitset
* enforce that locals are not directly accessed outside their storage live range
* shrink storage live ranges for never-borrowed locals to initialization -> last use
* emit storage liveness statements for *all* temporaries
* however, the remaining ones are *always* SSA immediates, so they'd be noop in MIR trans
* could have a flag on the temporary that its storage is irrelevant (a la C's old `register`)
* would also deny borrows if necessary
* this seems like an overcompliation and with packing & optimizations it may be pointless
Even in the current state, it helps stage2 `rustc` compile `boiler` without overflowing (see #35408).
A later addition fixes#26764 and closes#27372 by emitting `.section` directives for dylib metadata to avoid them being allocated into memory or read as `.note`. For this PR, those bugs were tripping valgrind.
changed E0067 to new error format
Updated E0067 to new error format.
Part of #35233Fixes#35502
Passes all the tests when running:
`python src/bootstrap/bootstrap.py --step check-cfail --stage 1`
**This seems strange, given that the format for E0067 has been changed.**
It feels like it should fail some unit tests maybe?
Let me know if I'm mistaken. Otherwise I can create a unit test for it.
Thanks for letting me help!
r? @jonathandturner
Improve &-ptr printing
This PR replaces printing `&-ptr` with a more readable description. To do so it uses a few heuristics.
If the name of the type is unknown, too long (longer than just saying "reference"), or too complex (a type with explicit lifetime annotations), it will instead opt to print either "reference" or "mutable reference", depending on the mutability of the type.
Before:
```
error[E0308]: mismatched types
--> src/test/compile-fail/issue-7061.rs:14:46
|
14 | fn foo(&'a mut self) -> Box<BarStruct> { self }
| ^^^^ expected box, found &-ptr
|
= note: expected type `Box<BarStruct>`
= note: found type `&'a mut BarStruct`
error: aborting due to previous error
```
After:
```
error[E0308]: mismatched types
--> src/test/compile-fail/issue-7061.rs:14:46
|
14 | fn foo(&'a mut self) -> Box<BarStruct> { self }
| ^^^^ expected box, found mutable reference
|
= note: expected type `Box<BarStruct>`
= note: found type `&'a mut BarStruct`
error: aborting due to previous error
```
E0248, E0267 & E0268 Change into issue format
r? @jonathandturner Part of #35391, #35519 and #35520. I have squashed all changes into a single commit. Please review the changes.
E0248 Change in issue format
E0267 UT New Format
E0268 UT New Format
E0267 & E0268 New Error Format
Emscripten test fixes
This picks up parts of #31623 to disable certain tests that emscripten can't run, as threads/processes are not supported.
I re-applied @tomaka's changes manually, I can rebase those commits with his credentials if he wants.
It also disables jemalloc for emscripten (at least in Rustbuild, I have to check if there is another setting for the same thing in the old makefile approach).
This should not impact anything for normal builds.
Update E0138 to new format
Part of #35233Fix#35510
r? @jonathandturner
Question: How can I only underline the function name ? I have observed the debug output and the struct of item, but I can't find the `Span` for function name. Should I modify the struct I get to save function name's position or there is another way to get it ? (I can only find `Span`s for function attributes, inputs, outputs, blocks)
macros: Make metavariables hygienic
This PR makes metavariables hygienic. For example, consider:
```rust
macro_rules! foo {
($x:tt) => { // Suppose that this token tree argument is always a metavariable.
macro_rules! bar { ($x:expr, $y:expr) => { ($x, $y) } }
}
}
fn main() {
foo!($z); // This currently compiles.
foo!($y); // This is an error today but compiles after this PR.
}
```
Today, the `macro_rules! bar { ... }` definition is only valid when the metavariable passed to `foo` is not `$y` (since it unhygienically conflicts with the `$y` in the definition of `bar`) or `$x` (c.f. #35450).
After this PR, the definition of `bar` is always valid (and `bar!(a, b)` always expands to `(a, b)` as expected).
This can break code that was allowed in #34925 (landed two weeks ago). For example,
```rust
macro_rules! outer {
($t:tt) => {
macro_rules! inner { ($i:item) => { $t } }
}
}
outer!($i); // This `$i` should not interact with the `$i` in the definition of `inner!`.
inner!(fn main() {}); // After this PR, this is an error ("unknown macro variable `i`").
```
Due to the severe limitations on nested `macro_rules!` before #34925, this is not a breaking change for stable/beta.
Fixes#35450.
r? @nrc
These tests check for the old error messages "`return` in a function
declared as diverging" and "computation may converge in a function
declared as diverging". The first of these is now invalid as `return` is
permitted in functions that return `!`. The second of these is subsumed
by the "mismatched types" error.
[MIR] Add explicit SetDiscriminant StatementKind for deaggregating enums
cc #35186
To deaggregate enums, we need to be able to explicitly set the discriminant. This PR implements a new StatementKind that does that.
I think some of the places that have `panics!` now could maybe do something smarter.
