bors [Mon, 25 Nov 2019 01:20:38 +0000 (01:20 +0000)]
Auto merge of #66671 - matthewjasper:ast-address-of, r=Centril
Ast address-of
This is the parts of #64588 that don't affect MIR. If an address-of expression makes it to MIR lowering we error and lower to the best currently expressible approximation to limit further errors.
bors [Sun, 24 Nov 2019 10:52:12 +0000 (10:52 +0000)]
Auto merge of #66393 - ssomers:hash_benches, r=dtolnay
introduce benchmarks of HashSet operations
To avoid goofs such as corrected by #66280, I added benchmarks of binary HashSet operations.
Due to the fact x.py keeps recompiling the whole shebang (or at least a big part of it) whenever you touch the test code, and because piling up all tests in one file does not strike me as future proof, I tried moving the hash benches to the separate place they are for liballoc/collections/btree. But it turns out that, in a cleaned checkout, x.py still recompiles the whole shebang whenever you touch the test code (PS or when you add or delete any irrelevant file). So I'm not going to add more tests, and I doubt others will, and these tests have proven their point already, so this PR is kind of pointless
Rollup merge of #66594 - Aaron1011:fix/debug-print-cycle, r=matthewjasper
Fix cycle when debug-printing opaque types
Fixes #61577
When printing an opaque type in non-verbose mode, we use the
`tcx.predicates_of` query to retrieve the opaque type's bounds for
pretty-printing. However, the pervasiveness of logging within librustc
means that we may already be executing `tcx.predicates_of` for the
opaque type we're trying to print, leading to a cycle error.
This commit adds a new 'no queries' thread-local flag to the pretty
printer. This flag is enabled during the computation of `predicates_of`
for opaque types, and causes us to print the opaque type in 'verbose'
mode (which does not require computing any additinal queries). This
should only affect debug logging for highly nested log messages, not any
user-visible output.
Rollup merge of #64856 - jonhoo:format-temporaries, r=sfackler
Scope format! temporaries
This places the temporaries that `format!` generates to refer to its arguments (through `&dyn Trait`) in a short-lived scope surrounding just the invocation of `format!`. This enables `format!` to be used in generators without the temporaries preventing the generator from being `Send` (due to `dyn Trait` not being `Sync`).
bors [Sat, 23 Nov 2019 22:23:20 +0000 (22:23 +0000)]
Auto merge of #66680 - Centril:rollup-1ke3svj, r=Centril
Rollup of 5 pull requests
Successful merges:
- #61351 (Stabilize cfg(doc))
- #66539 (Point at type in `let` assignment on type errors)
- #66655 (rustdoc: Mark `--extern-private` as unstable)
- #66657 (rustdoc: Don't panic when failing to write .lock file)
- #66673 (Move def collector from `rustc` to `rustc_resolve`)
Rollup merge of #66673 - petrochenkov:toresolve, r=Centril
Move def collector from `rustc` to `rustc_resolve`
It's used only from `rustc_resolve`, so we can move it there, thus reducing the size of `rustc` (https://github.com/rust-lang/rust/issues/65031).
It's quite possible that we can merge the def collector pass into the "build reduced graph" pass (they are always run together and do similar things), but it's some larger work.
Rollup merge of #66583 - Phlosioneer:patch-2, r=Dylan-DPC
Clarify Step Documentation
While the redesign is in progress (#62886), clarify the purpose of replace_zero and replace_one.
First, "returning itself" is technically impossible due to the function signature of &mut self -> Self. A clone or copy operation must be used. So this is now explicitly stated in the documentation.
Second, the added docs give some guidance about the actual contract around implementation of replace_zero and replace one. Specifically, the only usage is to create a range with no more steps, by setting start to replace_one and end to replace_zero. So the only property that is actually used is `replace_one > replace_zero`. See https://github.com/rust-lang/rust/issues/42168#issuecomment-489554232
The new documentation does not say that is the *only* contract, and so it should not be considered an api change. It just highlights the most important detail for implementors.
The redesign doesn't seem to be landing any time soon, so this is a stopgap measure to reduce confusion in the meantime.
That post claims that there is no clean way to check for garbage pointers, and
so this PR adopts the same solution of tentatively attempting to convert a
dererence to a string, which throws a clean exception on garbage that we can
catch and recover from.
I only made the change to vec and not the other pretty printers because I wanted
to focus my effort on the simplest thing that would resolve issue #64343. In
particular, I *considered* generalizing this fix to work on the other datatypes
in the pretty-printing support library, but I don't want to invest effort in
that until after we resolve our overall debugging support strategy; see also
issues #60826 and #65564.
Rollup merge of #66574 - GuillaumeGomez:update-tidy-err-code-check, r=Mark-Simulacrum
Update tidy check for error codes testing
Now that all the error codes have been grouped into one file, we moved the long error explanations into markdown files, which means that we need to read those markdown file to also count their code blocks as well.
So 2 "big" things happened here:
* No more need to keep the current error code when reading the `error_codes.rs` file since it's been put into the equivalent markdown file.
* Need to instead read the markdown file (but it's simpler since I can just look for code blocks directly).
bors [Sat, 23 Nov 2019 01:13:41 +0000 (01:13 +0000)]
Auto merge of #66507 - ecstatic-morse:const-if-match, r=oli-obk
Enable `if` and `match` in constants behind a feature flag
This PR is an initial implementation of #49146. It introduces a `const_if_match` feature flag and does the following if it is enabled:
- Allows `Downcast` projections, `SwitchInt` terminators and `FakeRead`s for matched places through the MIR const-checker.
- Allows `if` and `match` expressions through the HIR const-checker.
- Stops converting `&&` to `&` and `||` to `|` in `const` and `static` items.
As a result, the following operations are now allowed in a const context behind the feature flag:
- `if` and `match`
- short circuiting logic operators (`&&` and `||`)
- the `assert` and `debug_assert` macros (if the `const_panic` feature flag is also enabled)
However, the following operations remain forbidden:
- `while`, `loop` and `for` (see #52000)
- the `?` operator (calls `From::from` on its error variant)
- the `assert_eq` and `assert_ne` macros, along with their `debug` variants (calls `fmt::Debug`)
This PR is possible now that we use dataflow for const qualification (see #64470 and #66385).
bors [Fri, 22 Nov 2019 22:03:11 +0000 (22:03 +0000)]
Auto merge of #66640 - Centril:rollup-862009l, r=Centril
Rollup of 8 pull requests
Successful merges:
- #66183 (*Syntactically* permit visibilities on trait items & enum variants)
- #66566 (Document pitfall with `impl PartialEq<B> for A`)
- #66575 (Remove pretty printing of specific nodes in AST)
- #66587 (Handle statics in MIR as const pointers)
- #66619 (follow the convention in this file to use third-person singular verbs)
- #66633 (Error code's long explanation cleanup)
- #66637 (fix reoccuring typo: dereferencable -> dereferenceable)
- #66639 (resolve: more declarative `fresh_binding`)
Rollup merge of #66639 - Centril:simplify-fresh-binding, r=petrochenkov
resolve: more declarative `fresh_binding`
Following up on https://github.com/rust-lang/rust/pull/64111, this PR redefines `fresh_binding` wrt. `already_bound_and` and `already_bound_or` in a more declarative and simplified fashion.
Rollup merge of #66587 - matthewjasper:handle-static-as-const, r=oli-obk
Handle statics in MIR as const pointers
This is the first PR towards the goal of removing `PlaceBase::Static`. In this PR:
* Statics are lowered to dereferencing a const pointer.
* The temporaries holding such pointers are tracked in MIR, for the most part this is only used for diagnostics. There are two exceptions:
* The borrow checker has some checks for thread-locals that directly use this data.
* Const checking will suppress "cannot dereference raw pointer" diagnostics for pointers to `static mut`/`extern static`. This is to maintain the current behaviour (12 tests fail otherwise).
The following are left to future PRs (I think that @spastorino will be working on the first 3):
* Applying the same treatments to promoted statics.
* Removing `PlaceBase::Static`.
* Replacing `PlaceBase` with `Local`.
* Moving the ever growing collection of metadata that we have for diagnostics in MIR passes somewhere more appropriate.
Rollup merge of #66575 - Mark-Simulacrum:no-uii, r=petrochenkov
Remove pretty printing of specific nodes in AST
The ability to print a specific item as identified by NodeId or path
seems not particularly useful, and certainly carries quite a bit of
complexity with it.
This is intended to simplify our CLI parsing a bit and remove a
non-uncomplicated piece of it; I largely did this to remove the
dependency on NodeId from librustc/session but it's not really
necessary to do so in this invasive a way. The alternative is
moving it to librustc_interface or driver, probably.
Rollup merge of #66183 - Centril:empty-vis-trait-decl, r=petrochenkov
*Syntactically* permit visibilities on trait items & enum variants
Fixes #65041
Suppose we have `$vis trait_item` or `$vis enum_variant` and `$vis` is a `:vis` macro fragment. Before this PR, this would fail to parse. This is now instead allowed as per language team consensus in https://github.com/rust-lang/rust/issues/65041#issuecomment-538105286. (See added tests for elaboration.)
Moreover, we now also permit visibility modifiers on trait items & enum variants *syntactically* but reject them with semantic checks (in `ast_validation`):
bors [Fri, 22 Nov 2019 13:54:41 +0000 (13:54 +0000)]
Auto merge of #66460 - cjgillot:hashstable_generic, r=Zoxc
Add a proc-macro to derive HashStable in librustc dependencies
A second proc-macro is added to derive HashStable for crates librustc depends on.
This proc-macro HashStable_Generic (to bikeshed) allows to decouple code and some librustc's boilerplate.
Not everything is migrated, because `Span` and `TokenKind` require to be placed inside librustc.
Types using them stay there too.
bors [Fri, 22 Nov 2019 10:33:29 +0000 (10:33 +0000)]
Auto merge of #66565 - Mark-Simulacrum:syntax-cfg-mod, r=petrochenkov
Move process_configure_mod to rustc_parse
This removes the hack in favor of perhaps a less principled, but less painful, approach.
This also supports my work to decouple `Session` from librustc, as `ParseSess` currently has `Attribute` as "part" of it but after this PR will no longer do so.
bors [Fri, 22 Nov 2019 04:06:52 +0000 (04:06 +0000)]
Auto merge of #66524 - ecstatic-morse:compiletest-multiple-revisions, r=Centril
Support multiple revisions in `compiletest`
The `//[X]~` syntax filters errors for tests that are run across multiple cfgs with `// revisions:`. This commit extends that syntax to accept `//[X,Y]~`, which will match multiple cfgs to the same error annotation. This is functionally the same as writing two comments, `//[X]~` and `//[Y]~`, but can fit on a single line.
While refactoring `compiletest` to support this, I also uncovered a small bug that was causing an incremental test to always pass, despite no errors being emitted.
Dylan MacKenzie [Mon, 18 Nov 2019 05:11:42 +0000 (21:11 -0800)]
Reorganize, bless and add tests for const control flow
This creates a new test directory, `ui/consts/control-flow` to hold
tests related to control flow in a const context. It also blesses all
existing tests with the new error messages, and adds new tests for the
`const_if_match` feature.
bors [Fri, 22 Nov 2019 00:16:40 +0000 (00:16 +0000)]
Auto merge of #66282 - Centril:simplify-try, r=oli-obk
[mir-opt] asking `?`s in a more optimized fashion
This PR works towards https://github.com/rust-lang/rust/issues/66234 by providing two optimization passes meant to run in sequence:
- `SimplifyArmIdentity` which transforms something like:
```rust
_LOCAL_TMP = ((_LOCAL_1 as Variant ).FIELD: TY );
((_LOCAL_0 as Variant).FIELD: TY) = move _LOCAL_TMP;
discriminant(_LOCAL_0) = VAR_IDX;
```
into:
```rust
_LOCAL_0 = move _LOCAL_1
```
- `SimplifyBranchSame` which transforms `SwitchInt`s to identical basic blocks into a `goto` to the first reachable target.
Together, these are meant to simplify the following into just `res`:
```rust
match res {
Ok(x) => Ok(x),
Err(x) => Err(x),
}
```
It should be noted however that the desugaring of `?` includes a function call and so the first pass in this PR relies on inlining to substitute that function call for identity on `x`. Inlining requires `mir-opt-level=2` so this might not have any effect in perf-bot but let's find out.
r? @oli-obk -- This is WIP, but I'd appreciate feedback. :)
Dylan MacKenzie [Mon, 18 Nov 2019 04:41:34 +0000 (20:41 -0800)]
Allow `Downcast` projections if `const_if_match` enabled
These are generated when matching on enum variants to extract the value
within. We should have no problem evaluating these, but care should be
taken that we aren't accidentally allowing some other operation.
Dylan MacKenzie [Mon, 18 Nov 2019 18:06:21 +0000 (10:06 -0800)]
Fix broken incremental test
This test does not actually emit any warnings, since
`#![allow(warnings)]` was specified. `compiletest` was erroneously
ignoring `//~` tests and looking only for `//[X]~` ones. As a result of
the changes in the previous commit, we now look for `//~` comments in
incremental tests and expect them to appear in *all* revisions.
Dylan MacKenzie [Sat, 16 Nov 2019 05:40:05 +0000 (21:40 -0800)]
Allow multiple cfgs per comment in "revisions:" tests
The `//[X]~` syntax filters errors for tests that are run across
multiple cfgs with `// revisions:`. This commit extends that syntax to
accept `//[X,Y]~`, which will match multiple cfgs to the same error
annotation. This is functionally the same as writing two comments,
`//[X]~` and `//[Y]~`, but can fit on a single line.
Introduce MIR optimizations for simplifying `x?` on `Result`s.
This optimization depends on inlining for the identity
conversions introduced by the lowering of the `?`.
To take advantage of `SimplifyArmIdentity`, `-Z mir-opt-level=2`
is required because that triggers the inlining MIR optimization.