Commit Graph

6967 Commits

Author SHA1 Message Date
Jade
3cf820e17d rfc3052: Remove authors field from Cargo manifests
Since RFC 3052 soft deprecated the authors field anyway, hiding it from
crates.io, docs.rs, and making Cargo not add it by default, and it is
not generally up to date/useful information, we should remove it from
crates in this repo.
2021-07-29 14:56:05 -07:00
bors
6e0a8bf790 Auto merge of #86998 - m-ou-se:const-panic-fmt-as-str, r=oli-obk
Make const panic!("..") work in Rust 2021.

During const eval, this replaces calls to core::panicking::panic_fmt and std::panicking::being_panic_fmt with a call to a new const fn: core::panicking::const_panic_fmt. That function uses fmt::Arguments::as_str() to get the str and calls panic_str with that instead.

panic!() invocations with formatting arguments are still not accepted, as the creation of such a fmt::Arguments cannot be done in constant functions right now.

r? `@RalfJung`
2021-07-29 07:12:07 +00:00
bors
581b1664c9 Auto merge of #87416 - Aaron1011:query-side-effect, r=cjgillot
Create `QuerySideEffects` and use it for diagnostics

The code for saving and loading diagnostics during execution is generalized to handle a new `QuerySideEffects` struct. Currently, this struct just holds diagnostics - in a follow-up PR, I plan to add support for storing attriutes marked as used during query execution.

This is a pure refactor, with no intended behavior changes.
2021-07-29 02:51:19 +00:00
Yuki Okushi
9e94d7bbe6
Rollup merge of #87527 - LeSeulArtichaut:no-mir-unsafeck, r=oli-obk
Don't run MIR unsafeck at all when using `-Zthir-unsafeck`

I don't know how I missed this :D
r? ``@oli-obk``
2021-07-29 06:11:48 +09:00
Yuki Okushi
1d5f15f9fb
Rollup merge of #87521 - midgleyc:long-E0498, r=GuillaumeGomez
Add long explanation for E0498

Helps with #61137
2021-07-29 06:11:47 +09:00
Yuki Okushi
9222984636
Rollup merge of #87435 - ibraheemdev:patch-4, r=JohnTitor
fix example code for E0617

Closes #86908
2021-07-29 06:11:44 +09:00
Ibraheem Ahmed
00198dd902 fix example code for E0617 2021-07-28 16:33:02 -04:00
bors
a28109a767 Auto merge of #87540 - JohnTitor:rollup-8xc6bl5, r=JohnTitor
Rollup of 10 pull requests

Successful merges:

 - #87315 (Add docs for raw-dylib to unstable book)
 - #87330 (Use hashbrown's `extend_reserve()` in `HashMap`)
 - #87443 (Don't treat git repos as non-existent when `ignore_git` is set)
 - #87453 (Suggest removing unnecessary &mut as help message)
 - #87500 (Document math behind MIN/MAX consts on integers)
 - #87501 (Remove min_type_alias_impl_trait in favor of type_alias_impl_trait)
 - #87507 (SGX mutex is *not* moveable)
 - #87513 (bootstrap.py: change `git log` option to indicate desired behavior)
 - #87523 (Stop creating a reference then immediately dereferencing it.)
 - #87524 (Fix ICE in `diagnostic_hir_wf_check`)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup
2021-07-28 16:59:00 +00:00
Mara Bos
0b8033ad8d Improve comments about const panic handling
Co-authored-by: Ralf Jung <post@ralfj.de>
2021-07-28 16:18:38 +02:00
Mara Bos
4e6356188f Check that const_panic_fmt is const too. 2021-07-28 16:18:38 +02:00
Mara Bos
b64c4f9560 Add new const_format_args!() macro and use it in panics. 2021-07-28 16:12:25 +02:00
Mara Bos
f827d3e285 Make const panic!("..") work in Rust 2021.
During const eval, this replaces calls to core::panicking::panic_fmt and
std::panicking::being_panic_fmt with a call to a new const fn:
core::panicking::const_panic_fmt. That function uses
fmt::Arguments::as_str() to get the str and calls panic_str with that
instead.

panic!() invocations with formatting arguments are still not accepted,
as the creation of such a fmt::Arguments cannot be done in constant
functions right now.
2021-07-28 16:10:41 +02:00
Yuki Okushi
7985e4c7c3
Rollup merge of #87524 - FabianWolff:issue-87495, r=Aaron1011
Fix ICE in `diagnostic_hir_wf_check`

Fixes #87495. The [documentation for `ObligationCauseCode::WellFormed`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_infer/traits/enum.ObligationCauseCode.html#variant.WellFormed) says that

> it is always correct [...] to specify `WellFormed(None)`

instead of `WellFormed(Some(...))`, which seems to have caused the issue here, as `diagnostic_hir_wf_check` does not expect to be called with an associated constant and will ICE:

fd853c00e2/compiler/rustc_typeck/src/hir_wf_check.rs (L131-L134)

Therefore, I have changed `check_associated_item()` to pass a `WellFormed(None)` for associated constants.

r? ``@Aaron1011``
2021-07-28 18:28:23 +09:00
Yuki Okushi
23479f716a
Rollup merge of #87501 - spastorino:remove-min-tait, r=oli-obk
Remove min_type_alias_impl_trait in favor of type_alias_impl_trait

r? ``@oli-obk``
2021-07-28 18:28:19 +09:00
Yuki Okushi
954137ea0e
Rollup merge of #87453 - ibraheemdev:i-68697, r=wesleywiser
Suggest removing unnecessary &mut as help message

Closes #68697
2021-07-28 18:28:17 +09:00
bors
eba3228b2a Auto merge of #86251 - Smittyvb:thir-tree-again, r=oli-obk
Support -Z unpretty=thir-tree again

Currently `-Z unpretty=thir-tree` is broken after some THIR refactorings. This re-implements it, making it easier to debug THIR-related issues.

We have to do analyzes before getting the THIR, since trying to create THIR from invalid HIR can ICE. But doing those analyzes requires the THIR to be built and stolen. We work around this by creating a separate query to construct the THIR tree string representation.

Closes https://github.com/rust-lang/project-thir-unsafeck/issues/8, fixes #85552.
2021-07-28 09:01:11 +00:00
bors
aea2e446f0 Auto merge of #86735 - jhpratt:rfc-3107, r=petrochenkov
Implement RFC 3107: `#[derive(Default)]` on enums with a `#[default]` attribute

This PR implements RFC 3107, which permits `#[derive(Default)]` on enums where a unit variant has a `#[default]` attribute. See comments for current status.
2021-07-28 05:59:30 +00:00
bors
8b50cc9a2c Auto merge of #85769 - jhpratt:stabilize-const-transmute-union, r=RalfJung
Stabilize `const_fn_transmute`, `const_fn_union`

This PR stabilizes the `const_fn_transmute` and `const_fn_union` features. It _does not_ stabilize any methods (obviously aside from `transmute`) that are blocked on only these features.

Closes #53605. Closes #51909.
2021-07-28 00:53:17 +00:00
Fabian Wolff
dbd0fd2c2a Fix ICE in diagnostic_hir_wf_check 2021-07-28 01:41:52 +02:00
LeSeulArtichaut
40b57be547 Don't run MIR unsafeck at all when using -Zthir-unsafeck 2021-07-27 23:26:55 +02:00
Jacob Pratt
7bf791d162
Stabilize const_fn_union 2021-07-27 16:03:33 -04:00
Jacob Pratt
36f02f3523
Stabilize const_fn_transmute 2021-07-27 16:03:09 -04:00
bors
2faabf5793 Auto merge of #80367 - camelid:check_match-combine-loop, r=Nadrieril
Combine two loops in `check_match`

Suggested by Nadrieril in
https://github.com/rust-lang/rust/pull/79051#discussion_r548778186.

Opening to get a perf run. Hopefully this code doesn't require everything in the
first loop to be done before running the second! (It shouldn't though.)

cc `@Nadrieril`
2021-07-27 19:56:18 +00:00
Jacob Pratt
72465b0a32
Prohibit #[default] in invalid places 2021-07-27 15:47:49 -04:00
Jacob Pratt
5ae2371ceb
Indicate E0665 is no longer emitted 2021-07-27 15:47:49 -04:00
Jacob Pratt
eef2856eaf
Add machine-applicable suggestions
This avoids the need for tools like rust-analyzer to implement these
suggestions themselves.
2021-07-27 15:47:49 -04:00
Jacob Pratt
c70147fd66
Permit deriving default on enums with #[default] 2021-07-27 15:47:47 -04:00
Chris Midgley
bc70252f50 Mark code as ignored due to requiring external plugin 2021-07-27 19:44:50 +01:00
Chris Midgley
7f563f7f4e Add long explanation for E0498 2021-07-27 18:06:34 +01:00
bors
fd853c00e2 Auto merge of #83484 - JulianKnodt:infer, r=oli-obk,lcnr
Add hir::GenericArg::Infer

In order to extend inference to consts, make an Infer type on hir::GenericArg.
2021-07-27 16:24:45 +00:00
Santiago Pastorino
9629d798ca
Remove min_type_alias_impl_trait feature 2021-07-27 12:33:14 -03:00
Santiago Pastorino
66dbeb5f23
existential_type was removed in favor of type_alias_impl_trait 2021-07-27 12:33:13 -03:00
Santiago Pastorino
38bdd0e019
Make all tests use type_alias_impl_trait feature instead of min 2021-07-27 12:33:02 -03:00
Santiago Pastorino
5bff8429a0
Use type_alias_impl_trait instead of min in compiler and lib 2021-07-27 12:27:08 -03:00
Yuki Okushi
54367b90fc
Rollup merge of #87499 - ibraheemdev:patch-6, r=dtolnay
Remove ASCII fast path from `rustc_lexer::{is_id_continue, is_id_start}`

`unicode_xid` now has a fast path built-in: 122b38775c
2021-07-27 19:52:50 +09:00
Yuki Okushi
a981cffb23
Rollup merge of #87497 - midgleyc:long-E0544, r=GuillaumeGomez
Add long explanation for E0544.

Helps with #61137
2021-07-27 19:52:49 +09:00
Yuki Okushi
d25439481c
Rollup merge of #87427 - RalfJung:no-mir-for, r=oli-obk
get rid of NoMirFor error variant

The only place where we throw that error, it is very quickly caught again and turned into a different error. So raise that other error immediately.
2021-07-27 19:52:46 +09:00
Yuki Okushi
4e1ebf23cd
Rollup merge of #86764 - estebank:issue-86756, r=pnkfelix
Avoid ICE on type error recovery

Fix #86756
2021-07-27 19:52:41 +09:00
Yuki Okushi
99a6474bc4
Rollup merge of #86450 - tmiasko:move-size-limit, r=pnkfelix
Add flag to configure `large_assignments` lint

The `large_assignments` lints detects moves over specified limit.  The
limit is configured through `move_size_limit = "N"` attribute placed at
the root of a crate. When attribute is absent, the lint is disabled.

Make it possible to enable the lint without making any changes to the
source code, through a new flag `-Zmove-size-limit=N`.  For example, to
detect moves exceeding 1023 bytes in a cargo crate, including all
dependencies one could use:

```
$ env RUSTFLAGS=-Zmove-size-limit=1024 cargo build -vv
```

Lint tracking issue #83518.
2021-07-27 19:52:40 +09:00
bors
998cfe5aad Auto merge of #85305 - MarcusDunn:master, r=pnkfelix
Stabilize bindings_after_at

attempting to stabilze bindings_after_at [#65490](https://github.com/rust-lang/rust/issues/65490), im pretty new to the whole thing so any pointers are greatly appreciated.
2021-07-27 05:53:31 +00:00
bors
7d6bf861f8 Auto merge of #83491 - jyn514:remove-pretty, r=pnkfelix
Remove unstable `--pretty` flag

It doesn't do anything `--unpretty` doesn't, and due to a bug, also
didn't show up in `--help`. I don't think there's any reason to keep it
around, I haven't seen anyone using it.

Closes https://github.com/rust-lang/rust/issues/36473.
2021-07-27 03:12:40 +00:00
Ibraheem Ahmed
a397fdcc38
Remove ASCII fast path from rustc_lexer::{is_id_continue, is_id_start} 2021-07-26 20:17:28 -04:00
Chris Midgley
b21024f7b9 Add long explanation for E0544. 2021-07-26 22:59:16 +01:00
bors
8bebfe5cc2 Auto merge of #87480 - GuillaumeGomez:rollup-3ly8t5d, r=GuillaumeGomez
Rollup of 8 pull requests

Successful merges:

 - #87436 (Suggest `;` on parse error where applicable)
 - #87444 (Flatten nested `format!` calls)
 - #87447 (Miri: santiy check that null pointer can never have an AllocId)
 - #87457 (freebsd remove compiler workaround.)
 - #87458 (Fix help message for modification to &T created by &{t})
 - #87464 (Remove unnecessary `structhead` parameter from `render_union`)
 - #87473 (Notify the Rust 2021 edition working group in zulip of edition bugs)
 - #87474 (Add missing whitespace after attribute in HTML template)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup
2021-07-26 21:50:24 +00:00
kadmin
8759f00c73 Actually infer args in visitors 2021-07-26 21:15:18 +00:00
bors
08095fc1f8 Auto merge of #87469 - sexxi-goose:union, r=nikomatsakis
2229: Don't capture preicese paths on top of a union

- Accessing fields of a union require unsafe block
- As part of 2229 we don't allow precision where we need an unsafe block
to capture.

Fixes: #87378

r? `@nikomatsakis`
2021-07-26 19:04:01 +00:00
Guillaume Gomez
f9d00b4a78
Rollup merge of #87458 - ibraheemdev:help-msg-block-borrow, r=oli-obk
Fix help message for modification to &T created by &{t}

Previous:
```rust
error[E0594]: cannot assign to `*x` which is behind a `&` reference
 --> src/main.rs:3:5
  |
2 |     let x: &usize = &mut{0};
  |                     ------- help: consider changing this to be a mutable reference: `&mut mut{0}`
3 |     *x = 1;
  |     ^^^^^^ `x` is a `&` reference, so the data it refers to cannot be written
```
2021-07-26 16:04:36 +02:00
Guillaume Gomez
e3c6cd23cf
Rollup merge of #87457 - devnexen:fbsd_remove_workaround, r=petrochenkov
freebsd remove compiler workaround.

related issue #43575
2021-07-26 16:04:35 +02:00
Guillaume Gomez
fc0b4848cb
Rollup merge of #87447 - RalfJung:not-null, r=oli-obk
Miri: santiy check that null pointer can never have an AllocId
2021-07-26 16:04:34 +02:00
Guillaume Gomez
4205077537
Rollup merge of #87436 - ebobrow:suggest-semicolon, r=oli-obk
Suggest `;` on parse error where applicable

fixes #87197
2021-07-26 16:04:26 +02:00