1
Fork 0
Commit graph

236 commits

Author SHA1 Message Date
lcnr
938434ab82 enable intercrate in the solver InferCtxt 2023-03-21 16:34:04 +01:00
The 8472
7cce618d18 Fast path that skips over unchanged obligations in process_obligations
- only borrow the refcell once per loop
- avoid complex matches to reduce branch paths in the hot loop
- use a by-ref fast path that avoids mutations at the expense of having false negatives
2023-03-17 19:56:03 +01:00
Matthias Krüger
f0205d55ce
Rollup merge of #109166 - lcnr:define_opaque_types-explicit, r=oli-obk
make `define_opaque_types` fully explicit

based on the idea of #108389. Moved `define_opaque_types` into the actual operations, e.g. `eq`, instead of `infcx.at` because normalization doesn't use `define_opaque_types` and even creates it's own `At` with a different `define_opaque_types` internally.

Somewhat surprisingly, coherence actually relies on `DefineOpaqueTypes::Yes` for soundness which was revealed because I've incorrectly used `DefineOpaqueTypes::No` in `equate_impl_headers`. It feels concerning that even though this is the case, we still sometimes use `DefineOpaqueTypes::No` in coherence. I did not look into this as part of this PR as it is purely changing the structure of the code without changing behavior in any way.

r? ```@oli-obk```
2023-03-16 08:57:07 +01:00
lcnr
d2b7604db9 always make define_opaque_types explicit 2023-03-15 14:00:15 +01:00
bors
e84e5ff04a Auto merge of #107376 - aliemjay:remove-givens, r=lcnr
remove obsolete `givens` from regionck

Fixes #106567

r? `@lcnr` (feel free to reassign)
2023-03-15 02:50:58 +00:00
Michael Goulet
cf6424e881 Don't make projection tys out of anon consts 2023-03-14 23:12:42 +00:00
Michael Goulet
1b6d6f92d1 Fix canonicalizer bug for int/float vars too 2023-03-09 17:24:07 +00:00
Ali MJ Al-Nasrawy
95bcca85ff remove obsolete givens from regionck 2023-03-09 11:01:33 +03:00
Matthias Krüger
2428083558
Rollup merge of #108839 - compiler-errors:canonicalize-the-root-var, r=lcnr
Canonicalize root var when making response from new solver

During trait solving, if we equate two inference variables `?0` and `?1` but don't equate them with any rigid types, then `InferCtxt::probe_ty_var` will return `Err` for both of these. The canonicalizer code will then canonicalize the variables independently(!), and the response will not reflect the fact that these two variables have been made equal.

This hinders inference and I also don't think it's sound? I haven't thought too much about it past that, so let's talk about it.

r? ``@lcnr``
2023-03-08 21:26:51 +01:00
Michael Goulet
8d13454498 Canonicalize the ROOT VAR 2023-03-07 17:11:35 +00:00
Yuki Okushi
2a196db1f2
Rollup merge of #108728 - tshepang:fix-and-improve, r=michaelwoerister
infer: fix and improve comments
2023-03-07 23:06:22 +09:00
Tshepang Mbambo
34063c5706 infer: fix and improve comments 2023-03-04 07:38:37 +02:00
Tshepang Mbambo
4b92024cf1 field is not used outside the crate
See b61a28b2a1
2023-03-04 07:01:08 +02:00
Nilstrieb
e326777cca Some infer/mod.rs cleanups 2023-02-28 11:49:05 +00:00
Alan Egerton
695072daa6
Remove type-traversal trait aliases 2023-02-22 17:04:58 +00:00
Matthias Krüger
314fe4d170
Rollup merge of #104239 - b-naber:sccs-info, r=jackh726
Better debug logs for borrowck constraint graph

It's really cumbersome to work with `RegionVar`s when trying to debug borrowck code or when trying to understand how the borrowchecker works. This PR collects some region information (behind `cfg(debug_assertions)`) for created `RegionVar`s (NLL region vars, this PR doesn't touch canonicalization) and prints the nodes and edges of the strongly connected constraints graph using representatives that use that region information (either lifetime names, locations in MIR or spans).
2023-02-21 23:01:58 +01:00
Oli Scherer
88a7b6803b Make hidden type registration opt-in, so that each site can be reviewed on its own and we have the right defaults for trait solvers 2023-02-21 14:50:51 +00:00
b-naber
2f79f73821 collect region contexts during mir renumbering 2023-02-19 22:11:27 +00:00
Kyle Matsuda
8e92849cbb changes from review 2023-02-16 17:05:59 -07:00
Kyle Matsuda
c183110cc2 remove bound_type_of query; make type_of return EarlyBinder; change type_of in metadata 2023-02-16 17:05:56 -07:00
Kyle Matsuda
d822b97a27 change usages of type_of to bound_type_of 2023-02-16 17:01:52 -07:00
bors
c5d1b3ea96 Auto merge of #108020 - nnethercote:opt-mk_region, r=compiler-errors
Optimize `mk_region`

PR #107869 avoiding some interning under `mk_ty` by special-casing `Ty` variants with simple (integer) bodies. This PR does something similar for regions.

r? `@compiler-errors`
2023-02-16 16:11:54 +00:00
Matthias Krüger
8f65e25aec
Rollup merge of #108010 - compiler-errors:can_eq-returns-bool, r=lcnr
Make `InferCtxt::can_eq` and `InferCtxt::can_sub` return booleans

Nobody matches on the result, nor does the result return anything useful...
2023-02-15 21:30:56 +01:00
Nicholas Nethercote
cef9004f5a Add specialized variants of mk_region.
Much like there are specialized variants of `mk_ty`. This will enable
some optimization in the next commit.

Also rename the existing `re_error*` functions as `mk_re_error*`, for
consistency.
2023-02-15 09:02:44 +11:00
Michael Goulet
eb286dd070 Make can_eq and can_sub return booleans 2023-02-13 19:29:02 +00:00
Alan Egerton
63ad5d0522
Rename folder traits' tcx method to interner 2023-02-13 10:24:51 +00:00
Alan Egerton
9783fcc13b
Make folding traits generic over the Interner 2023-02-13 10:24:49 +00:00
Alan Egerton
ba55a453eb
Alias folding/visiting traits instead of re-export 2023-02-13 10:24:46 +00:00
Nicholas Nethercote
7a72560154 Reduce direct mk_ty usage.
We use more specific `mk_*` functions in most places, might as well use
them as much as possible.
2023-02-13 09:32:48 +11:00
Boxy
1f89e2aef2 emit AliasEq when relating type and const aliases 2023-02-10 14:33:13 +00:00
Michael Goulet
03a8a4ff3e Replacing bound vars is actually instantiating a binder 2023-02-07 23:13:54 +00:00
Nicholas Nethercote
fb8e6819aa Split and inline ShallowResolver::fold_ty. 2023-02-06 08:52:04 +11:00
Nicholas Nethercote
f08a3371b0 Improve early bailout test in resolve_vars_if_possible.
`!t.has_non_region_infer()` is the test used in
`OpportunisticVarResolver`, and catches a few cases that
`!t.needs_infer()` misses.
2023-02-06 08:52:04 +11:00
Michael Goulet
0e98a162c8 Track bound types like bound regions 2023-01-30 22:18:20 +00:00
Maybe Waffle
6a28fb42a8 Remove double spaces after dots in comments 2023-01-17 08:09:33 +00:00
nils
c61f29ca52
Rollup merge of #106714 - Ezrashaw:remove-e0490, r=davidtwco
remove unreachable error code `E0490`

AFAIK, the untested and undocumented error code `E0490` is now unreachable, it was from the days of the original borrow checker.

cc ``@GuillaumeGomez`` #61137
2023-01-12 15:44:52 +01:00
Ezra Shaw
02005e9f22
remove unreachable error code E0490 2023-01-12 14:15:21 +13:00
Michael Goulet
c8334ce60c Move autoderef to rustc_hir_analysis 2023-01-11 20:12:57 +00:00
Yuki Okushi
e5e116dca9
Rollup merge of #106204 - compiler-errors:no-take-opaques-in-compare, r=oli-obk
No need to take opaques in `check_type_bounds`

`InferCtxt` already has its defining use anchor set to err

r? ``@oli-obk``
2023-01-10 08:05:34 +09:00
Michael Goulet
f769d34291 Assert defining anchor is set in take_opaque_types 2023-01-09 18:14:28 +00:00
fee1-dead
7779386a3a
Rollup merge of #106164 - compiler-errors:check-region-tweak, r=oli-obk
Move `check_region_obligations_and_report_errors` to `TypeErrCtxt`

Makes sense for this function to live with its sibling `resolve_regions_and_report_errors`, around which it's basically just a wrapper.
2023-01-09 23:35:29 +08:00
Michael Goulet
6afd16171d
Rollup merge of #106131 - compiler-errors:not-ptrs, r=davidtwco
Mention "signature" rather than "fn pointer" when impl/trait methods are incompatible

Fixes #80929
Fixes #67296
2023-01-08 19:57:53 -08:00
Michael Goulet
49f849a3d2 Mention signature rather than fn pointers when comparing impl/trait methods 2023-01-08 03:49:24 +00:00
Ezra Shaw
93c0d8d5d5
remove unreachable error code E0313 2023-01-08 14:47:12 +13:00
Michael Goulet
c104ee9f6d Move check_region_obligations_and_report_errors to TypeErrCtxt 2023-01-03 23:58:12 +00:00
Michael Goulet
8973b3e3cc Bubble up ErrorGuaranteed from region constraints in method item compare 2022-12-24 21:37:00 +00:00
Michael Goulet
96154d7fa7 Add IMPLIED_BOUNDS_ENTAILMENT lint 2022-12-19 18:16:22 +00:00
Matthias Krüger
0aa4cde747 avoid .into() conversion to identical types 2022-12-18 16:20:32 +01:00
Boxy
dd19656df3 fold instead of obliterating args 2022-12-12 14:30:01 +00:00
Matthias Krüger
8e0d83a70c
Rollup merge of #105185 - compiler-errors:normalize_fn_sig-in-err-ctxt, r=lcnr
Move `normalize_fn_sig` to `TypeErrCtxt`

r? `@lcnr`
2022-12-02 21:22:50 +01:00