put empty generic lists behind a pointer
This reduces the size of hir::Expr from 128 to 88 bytes (!) and shaves
200MB out of #36799.
This is a performance-sensitive PR so please don't roll it up.
r? @eddyb
On "the parameter type `T` may not live long enough" error, point to the
parameter type suggesting lifetime bindings:
```
error[E0310]: the parameter type `T` may not live long enough
--> $DIR/lifetime-doesnt-live-long-enough.rs:28:5
|
27 | struct Foo<T> {
| - help: consider adding an explicit lifetime bound `T: 'static`...
28 | foo: &'static T
| ^^^^^^^^^^^^^^^
|
note: ...so that the reference type `&'static T` does not outlive the data it points at
--> $DIR/lifetime-doesnt-live-long-enough.rs:28:5
|
28 | foo: &'static T
| ^^^^^^^^^^^^^^^
```
The convention for suggesting close matches is to provide at most one match (the
closest one). Change the suggestions for misspelt method names to obey that.
typeck::check::coercion - roll back failed unsizing type vars
This wraps unsizing coercions within an additional level of
`commit_if_ok`, which rolls back type variables if the unsizing coercion
fails. This prevents a large amount of type-variables from accumulating
while type-checking a large function, e.g. shaving 2GB off one of the
4GB peaks in #36799.
This is a performance-sensitive PR so please don't roll it up.
r? @eddyb
cc @nikomatsakis
Now that region::Scope is no longer interned, its size is more
important. This PR encodes region::Scope in 8 bytes instead of 12, which
should speed up region inference somewhat (perf testing needed) and
should improve the margins on #36799 by 64MB (that's not a lot, I did
this PR mostly to speed up region inference).
add comparison operators to must-use lint (under `fn_must_use` feature)
Although RFC 1940 is about annotating functions with `#[must_use]`, a
key part of the motivation was linting unused equality operators.
(See
https://github.com/rust-lang/rfcs/pull/1812#issuecomment-265695898—it
seems to have not been clear to discussants at the time that marking the
comparison methods as `must_use` would not give us the lints on
comparison operators, at least in (what the present author understood
as) the most straightforward implementation, as landed in #43728
(3645b062).)
To rectify the situation, we here lint unused comparison operators as
part of the unused-must-use lint (feature gated by the `fn_must_use`
feature flag, which now arguably becomes a slight (tolerable in the
opinion of the present author) misnomer).
This is in the matter of #43302.
cc @crumblingstatue
This wraps unsizing coercions within an additional level of
`commit_if_ok`, which rolls back type variables if the unsizing coercion
fails. This prevents a large amount of type-variables from accumulating
while type-checking a large function, e.g. shaving 2GB off one of the
4GB peaks in #36799.
Improve diagnostics when attempting to match tuple enum variant with struct pattern
Adds an extra note as below to explain that a tuple pattern was probably intended.
```
error[E0026]: variant `X::Y` does not have a field named `data`
--> src/main.rs:18:16
|
18 | X::Y { data } => println!("The data is {}", data)
| ^^^^ variant `X::Y` does not have field `data`
error[E0027]: pattern does not mention field `0`
--> src/main.rs:18:9
|
18 | X::Y { data } => println!("The data is {}", data)
| ^^^^^^^^^^^^^ missing field `0`
|
= note: trying to match a tuple variant with a struct variant pattern
```
Fixes#41314.
You can otherwise end up in a situation where you don't actually resize
but still call into handle_cap_increase which then corrupts head/tail.
Closes#44800
incr.comp.: Add new DepGraph implementation.
This commits does a few things:
1. It adds the new dep-graph implementation -- *in addition* to the old one. This way we can start testing the new implementation without switching all tests at once.
2. It persists the new dep-graph (which includes query result fingerprints) to the incr. comp. caching directory and also loads this data.
3. It removes support for loading fingerprints of metadata imported from other crates (except for when running autotests). This is not needed anymore with red/green. It could provide a performance advantage but that's yet to be determined. For now, as red/green is not fully implemented yet, the cross-crate incremental tests are disabled.
Note, this PR is based on top of soon-to-be-merged #44696 and only the last 4 commits are new:
```
- incr.comp.: Initial implemenation of append-only dep-graph. (c90147c)
- incr.comp.: Do some various cleanup. (8ce20c5)
- incr.comp.: Serialize and deserialize new DepGraph. (0e13c1a)
- incr.comp.: Remove support for loading metadata fingerprints. (270a134)
EDIT 2:
- incr.comp.: Make #[rustc_dirty/clean] test for fingerprint equality ... (d8f7ff9)
```
(EDIT: GH displays the commits in the wrong order for some reason)
Also note that this PR is expected to certainly result in performance regressions in the incr. comp. test cases, since we are adding quite a few things (a whole additional dep-graph, for example) without removing anything. End-to-end performance measurements will only make sense again after red/green is enabled and all the legacy tracking has been turned off.
EDIT 2: Pushed another commit that makes the `#[rustc_dirty]`/`#[rustc_clean]` based autotests compared query result fingerprints instead of testing `DepNode` existence.
This verifies that TrustedRandomAccess has no side effects when the
iterator item implements Copy. This also implements TrustedLen and
TrustedRandomAccess for str::Bytes.
