| Age | Commit message (Collapse) | Author | Lines | |
|---|---|---|---|---|
| 2024-06-18 | Uplift the new trait solver | Michael Goulet | -1686/+0 | |
| 2024-06-18 | SolverDelegate | Michael Goulet | -28/+31 | |
| 2024-06-16 | Uplift OpaqueTypeKey too, use it in response | Michael Goulet | -9/+3 | |
| 2024-06-16 | Uplift ExternalConstraintData | Michael Goulet | -15/+27 | |
| 2024-06-16 | Make ExternalConstraints just carry outlives | Michael Goulet | -18/+22 | |
| 2024-06-16 | Move InferCtxtSelectExt out of eval_ctxt module | Michael Goulet | -186/+0 | |
| 2024-06-16 | Auto merge of #126505 - compiler-errors:no-vtable, r=lcnr | bors | -38/+2 | |
| Only compute vtable information during codegen This PR removes vtable information from the `Object` and `TraitUpcasting` candidate sources in the trait solvers, and defers the computation of relevant information to `Instance::resolve`. This is because vtables really aren't a thing in the trait world -- they're an implementation detail in codegen. Previously it was just easiest to tangle this information together since we were already doing the work of looking at all the supertraits in the trait solver, and specifically because we use traits to represent when it's possible to call a method via a vtable (`Object` candidate) and do upcasting (`Unsize` candidate). but I am somewhat suspicious we're doing a *lot* of extra work, especially in polymorphic contexts, so let's see what perf says. | ||||
| 2024-06-14 | Only compute vtable information during codegen | Michael Goulet | -38/+2 | |
| 2024-06-14 | Make proof tree probing generic | Michael Goulet | -38/+45 | |
| 2024-06-14 | Make Candidate generic over interner | Michael Goulet | -1/+1 | |
| 2024-06-13 | Fix some TODOs | Michael Goulet | -4/+3 | |
| 2024-06-13 | Finish uplifting all of structural_traits | Michael Goulet | -157/+136 | |
| 2024-06-11 | Try not to make obligations in handle_opaque_type | Michael Goulet | -8/+7 | |
| 2024-06-06 | Uplift TypeRelation and Relate | Michael Goulet | -4/+4 | |
| 2024-06-05 | Add `Ty` to `ConstKind::Value` | Boxy | -2/+7 | |
| 2024-06-05 | Basic removal of `Ty` from places (boring) | Boxy | -8/+8 | |
| 2024-05-31 | Stop using translate_args in the new solver | Michael Goulet | -11/+0 | |
| 2024-05-29 | Rollup merge of #125664 - compiler-errors:trace-tweaks, r=lcnr | 许杰友 Jieyou Xu (Joe) | -7/+2 | |
| Tweak relations to no longer rely on `TypeTrace` Remove `At::trace`, and inline all of the `Trace::equate`,etc methods into `At`. The only nontrivial change is that we use `AliasTerm` to relate two unevaluated consts in the old-solver impl of `ConstEquate`, since `AliasTerm` does implement `ToTrace` and will relate the args structurally (shallowly). r? lcnr | ||||
| 2024-05-28 | Get rid of manual Trace calls | Michael Goulet | -7/+2 | |
| 2024-05-28 | Make ProofTreeBuilder actually generic over interner | Michael Goulet | -15/+21 | |
| 2024-05-28 | EvalCtxt::tcx() -> EvalCtxt::interner() | Michael Goulet | -15/+17 | |
| 2024-05-28 | eagerly normalize when adding goals | lcnr | -2/+75 | |
| 2024-05-27 | Auto merge of #125413 - lcnr:ambig-drop-region-constraints, r=compiler-errors | bors | -25/+39 | |
| drop region constraints for ambiguous goals See the comment in `compute_external_query_constraints`. While the underlying issue is preexisting, this fixes a bug introduced by #125343. It slightly weakens the leak chec, even if we didn't have any test which was affected. I want to write such a test before merging this PR. r? `@compiler-errors` | ||||
| 2024-05-24 | drop region constraints for ambiguous goals | lcnr | -25/+39 | |
| 2024-05-24 | remove proof tree formatter, make em shallow | lcnr | -22/+5 | |
| 2024-05-20 | Rework var resolution in InferCtxtLike, uplift EagerResolver | Michael Goulet | -1/+1 | |
| 2024-05-19 | Make EvalCtxt generic over interner | Michael Goulet | -22/+27 | |
| 2024-05-19 | Make search graph generic over interner | Michael Goulet | -3/+3 | |
| 2024-05-18 | Move NormalizesTo back down | Michael Goulet | -5/+7 | |
| I tried to rebase this down into the first commit but it is WAY too annoying x | ||||
| 2024-05-18 | Make proof tree building generic | Michael Goulet | -3/+3 | |
| 2024-05-18 | Uplift inspect into rustc_type_ir | Michael Goulet | -19/+27 | |
| 2024-05-18 | Uplift GenericArgKind, CanonicalVarValues, QueryInput | Michael Goulet | -13/+17 | |
| and make NestedGoals generic | ||||
| 2024-05-13 | Use a proper probe for shadowing impl | Michael Goulet | -1/+2 | |
| 2024-05-13 | Auto merge of #125076 - compiler-errors:alias-term, r=lcnr | bors | -5/+5 | |
| Split out `ty::AliasTerm` from `ty::AliasTy` Splitting out `AliasTerm` (for use in project and normalizes goals) and `AliasTy` (for use in `ty::Alias`) r? lcnr | ||||
| 2024-05-13 | split out AliasTy -> AliasTerm | Michael Goulet | -5/+5 | |
| 2024-05-13 | Remove `extern crate rustc_middle` from `rustc_trait_selection`. | Nicholas Nethercote | -0/+3 | |
| 2024-05-12 | structurally important functions to `debug` | lcnr | -3/+4 | |
| 2024-05-12 | solve: replace all `debug` with `trace` | lcnr | -21/+21 | |
| 2024-05-10 | Rollup merge of #124955 - nnethercote:next_ty_var, r=lcnr | Matthias Krüger | -14/+4 | |
| Use fewer origins when creating type variables. To reduce lots of repetitive boilerplate code. Details in the individual commit messages. r? ``@lcnr`` | ||||
| 2024-05-10 | Use fewer origins when creating type variables. | Nicholas Nethercote | -14/+4 | |
| `InferCtxt::next_{ty,const}_var*` all take an origin, but the `param_def_id` is almost always `None`. This commit changes them to just take a `Span` and build the origin within the method, and adds new methods for the rare cases where `param_def_id` might not be `None`. This avoids a lot of tedious origin building. Specifically: - next_ty_var{,_id_in_universe,_in_universe}: now take `Span` instead of `TypeVariableOrigin` - next_ty_var_with_origin: added - next_const_var{,_in_universe}: takes Span instead of ConstVariableOrigin - next_const_var_with_origin: added - next_region_var, next_region_var_in_universe: these are unchanged, still take RegionVariableOrigin The API inconsistency (ty/const vs region) seems worth it for the large conciseness improvements. | ||||
| 2024-05-09 | analyse visitor: build proof tree in probe | lcnr | -0/+1 | |
| 2024-05-07 | Don't ICE when we cannot eval a const to a valtree in the new solver | Michael Goulet | -3/+3 | |
| 2024-05-06 | Rollup merge of #124809 - lcnr:prepopulate-opaques, r=compiler-errors | Matthias Krüger | -9/+6 | |
| borrowck: prepopulate opaque storage more eagerly otherwise we ICE due to ambiguity when normalizing while computing implied bounds. r? ``@compiler-errors`` | ||||
| 2024-05-06 | Rollup merge of #124759 - compiler-errors:impl-args, r=lcnr | Matthias Krüger | -59/+26 | |
| Record impl args in the proof tree in new solver Rather than rematching them during select. Also use `ImplSource::Param` instead of `ImplSource::Builtin` for alias-bound candidates, so we don't ICE in `Instance::resolve`. r? lcnr | ||||
| 2024-05-06 | Use correct ImplSource for alias bounds | Michael Goulet | -2/+1 | |
| 2024-05-06 | Record impl args in the InsepctCandiate rather than rematching during select | Michael Goulet | -59/+27 | |
| 2024-05-06 | Rollup merge of #124724 - compiler-errors:prefer-lower, r=lcnr | Matthias Krüger | -5/+13 | |
| Prefer lower vtable candidates in select in new solver Also, adjust the select visitor to only winnow when the *parent* goal is `Certainty::Yes`. This means that we won't winnow in cases when we have any ambiguous inference guidance from two candidates. r? lcnr | ||||
| 2024-05-06 | switch new solver to directly inject opaque types | lcnr | -9/+6 | |
| 2024-05-06 | Prefer lower vtable candidates in select in new solver | Michael Goulet | -5/+13 | |
| 2024-05-04 | Record impl args in the proof tree | Michael Goulet | -2/+6 | |
 |
index : rust | |
| https://github.com/rust-lang/rust |
| about summary refs log tree commit diff |