incr.comp.: Implement query result cache and use it to cache type checking tables.
This is a spike implementation of caching more than LLVM IR and object files when doing incremental compilation. At the moment, only the `typeck_tables_of` query is cached but MIR and borrow-check will follow shortly. The feature is activated by running with `-Zincremental-queries` in addition to `-Zincremental`, it is not yet active by default.
r? @nikomatsakis
Replace hir::TyImplTrait with TyImplTraitUniversal and
TyImplTraitExistential.
Add an ImplTraitContext enum to rustc::hir::lowering to track the kind
and allowedness of an impl Trait.
Significantly alter lowering to thread ImplTraitContext and one other
boolean parameter described below throughought much of lowering.
The other parameter is for tracking if lowering a function is in a trait
impl, as there is not enough information to otherwise know this
information during lowering otherwise.
This change also removes the checks from ast_ty_to_ty for impl trait
allowedness as they are now all taking place in HIR lowering.
RFC 2008: Future-proofing enums/structs with #[non_exhaustive] attribute
This work-in-progress pull request contains my changes to implement [RFC 2008](https://github.com/rust-lang/rfcs/pull/2008). The related tracking issue is #44109.
As of writing, enum-related functionality is not included and there are some issues related to tuple/unit structs. Enum related tests are currently ignored.
WIP PR requested by @nikomatsakis [in Gitter](https://gitter.im/rust-impl-period/WG-compiler-middle?at=59e90e6297cedeb0482ade3e).
DefaultImpl is a highly confusing name for what we now call auto impls,
as in `impl Send for ..`. The name auto impl is not formally decided
but for sanity anything is better than `DefaultImpl` which refers
neither to `default impl` nor to `impl Default`.
incr.comp.: Implement compiler diagnostic persistence.
This PR implements storing and loading diagnostics that the compiler generates and thus allows for emitting warnings during incremental compilation without actually re-evaluating the thing the warning originally came from. It also lays some groundwork for storing and loading type information and MIR in the incr. comp. cache.
~~It is still work in progress:~~
- ~~There's still some documentation to be added.~~
- ~~The way anonymous queries are handled might lead to duplicated emissions of warnings. Not sure if there is a better way or how frequent such duplication would be in practice.~~
Diagnostic message duplication is addressed separately in #45519.
r? @nikomatsakis
Evaluate fixed-length array length expressions lazily.
This is in preparation for polymorphic array lengths (aka `[T; T::A]`) and const generics.
We need deferred const-evaluation to break cycles when array types show up in positions which require knowing the array type to typeck the array length, e.g. the array type is in a `where` clause.
The final step - actually passing bounds in scope to array length expressions from the parent - is not done because it still produces cycles when *normalizing* `ParamEnv`s, and @nikomatsakis' in-progress lazy normalization work is needed to deal with that uniformly.
However, the changes here are still useful to unlock work on const generics, which @EpicatSupercell manifested interest in, and I might be mentoring them for that, but we need this baseline first.
r? @nikomatsakis cc @oli-obk
This'll allow us to reconstruct query parameters purely from the `DepNode`
they're associated with. Some queries could move straight to `HirId` but others
that don't always have a correspondance between `HirId` and `DefId` moved to
two-level maps where the query operates over a `DefIndex`, returning a map,
which is then keyed off `ItemLocalId`.
Closes#44414
This commit moves the `crates` method to a query and then migrates all callers
to use a query instead of the now-renamed `crates_untracked` method where
possible.
Closes#41417
This commit moves the calculation of the `LanguageItems` structure into a
query rather than being calculated before the `TyCtxt` exists, with the eventual
end goal of removing some `CrateStore` methods.
The main use of `CrateStore` *before* the `TyCtxt` is created is during
resolution, but we want to be sure that any methods used before resolution are
not used after the `TyCtxt` is created. This commit starts moving the methods
used by resolve to all be named `{name}_untracked` where the rest of the
compiler uses just `{name}` as a query.
During this transition a number of new queries were added to account for
post-resolve usage of these methods.
This commit started by moving methods from `CrateStore` to queries, but it ended
up necessitating some deeper refactorings to move more items in general to
queries.
Before this commit the *resolver* would walk over the AST and process foreign
modules (`extern { .. }` blocks) and collect `#[link]` annotations. It would
then also process the command line `-l` directives and such. This information
was then stored as precalculated lists in the `CrateStore` object for iterating
over later.
After this, commit, however, this pass no longer happens during resolution but
now instead happens through queries. A query for the linked libraries of a crate
will crawl the crate for `extern` blocks and then process the linkage
annotations at that time.
This map, like `trait_map`, is calculated in resolve, but we want to be sure to
track it for incremental compliation. Hide it behind a query to get more
refactorings later.
The spans previously encoded only span the first token after the opening
brace, up to the closing brace of inline `mod` declarations. Thus, when
examining exports from an external crate, the spans don't include the
header of inline `mod` declarations.
rustc: Implement the #[global_allocator] attribute
This PR is an implementation of [RFC 1974] which specifies a new method of
defining a global allocator for a program. This obsoletes the old
`#![allocator]` attribute and also removes support for it.
[RFC 1974]: https://github.com/rust-lang/rfcs/pull/1974
The new `#[global_allocator]` attribute solves many issues encountered with the
`#![allocator]` attribute such as composition and restrictions on the crate
graph itself. The compiler now has much more control over the ABI of the
allocator and how it's implemented, allowing much more freedom in terms of how
this feature is implemented.
cc #27389
This PR is an implementation of [RFC 1974] which specifies a new method of
defining a global allocator for a program. This obsoletes the old
`#![allocator]` attribute and also removes support for it.
[RFC 1974]: https://github.com/rust-lang/rfcs/pull/197
The new `#[global_allocator]` attribute solves many issues encountered with the
`#![allocator]` attribute such as composition and restrictions on the crate
graph itself. The compiler now has much more control over the ABI of the
allocator and how it's implemented, allowing much more freedom in terms of how
this feature is implemented.
cc #27389
incr.comp.: Track expanded spans instead of FileMaps.
This PR removes explicit tracking of FileMaps in response to #42101. The reasoning behind being able to just *not* track access to FileMaps is similar to why we don't track access to the `DefId->DefPath` map:
1. One can only get ahold of a `Span` value by accessing the HIR (for local things) or a `metadata::schema::Entry` (for things from external crates).
2. For both of these things we compute a hash that incorporates the *expanded spans*, that is, what we hash is in the (FileMap independent) format `filename:line:col`.
3. Consequently, everything that emits a span should already be tracked via its dependency to something that has the span included in its hash and changes would be detected via that hash.
One caveat here is that we have to be conservative when exporting things in metadata. A crate can be built without debuginfo and would thus by default not incorporate most spans into the metadata hashes. However, a downstream crate can make an inline copy of things in the upstream crate and span changes in the upstream crate would then go undetected, even if the downstream uses them (e.g. by emitting debuginfo for an inlined function). For this reason, we always incorporate spans into metadata hashes for now (there might be more efficient ways to handle this safely when red-green tracking is implemented).
r? @nikomatsakis
make *most* maps private
Currently we access the `DepTrackingMap` fields directly rather than using the query accessors. This seems bad. This branch removes several such uses, but not all, and extends the macro so that queries can hide their maps (so we can prevent regressions). The extension to the macro is kind of ugly :/ but couldn't find a simple way to do it otherwise (I guess I could use a nested macro...). Anyway I figure it's only temporary.
r? @eddyb
Implement a file-path remapping feature in support of debuginfo and reproducible builds
This PR adds the `-Zremap-path-prefix-from`/`-Zremap-path-prefix-to` commandline option pair and is a more general implementation of #41419. As opposed to the previous attempt, this implementation should enable reproducible builds regardless of the working directory of the compiler.
This implementation of the feature is more general in the sense that the re-mapping will affect *all* paths the compiler emits, including the ones in error messages.
r? @alexcrichton
#37653 support `default impl` for specialization
this commit implements the first step of the `default impl` feature:
> all items in a `default impl` are (implicitly) `default` and hence
> specializable.
In order to test this feature I've copied all the tests provided for the
`default` method implementation (in run-pass/specialization and
compile-fail/specialization directories) and moved the `default` keyword
from the item to the impl.
See [referenced](https://github.com/rust-lang/rust/issues/37653) issue for further info
r? @aturon
