Disentangle InferCtxt, MemCategorizationContext and ExprUseVisitor. #42563
Conversation
eddyb
added
the
S-waiting-on-crater
|
cc @Mark-Simulacrum @tomprince This could use some cratering/cargobombing. EDIT: Although, if crater works again, I could do that now myself. |
|
Note that some lints use visitors because they need a second pass (e.g. to find usages of something). Those Lint's could be rewritten to store the relevant information, thus reducing them to single-pass, but this isn't exactly trivial. |
|
@llogiq EDIT: as in, it's not feasible to replace it without replicating large portions of it. |
|
Crater report only has network errors and |
eddyb
added
S-waiting-on-review
|
@bors try |
Disentangle InferCtxt, MemCategorizationContext and ExprUseVisitor. At some point in the past, `InferCtxt` started being used to replace an old "`Typer`" abstraction, which provided access to `TypeckTables` and had optionally type inference to account for. That didn't play so nicely with the `'gcx`/`'tcx` split and I had to introduce `borrowck_fake_infer_ctxt`. The situation wasn't great but it wasn't too painful inside `rustc` itself. Recently I've found that method being used in clippy, which does need EUV (before we make it plausible to run lints on HAIR or MIR), and set out to separate inference from tables, for the sake of lint authors. Also fixes #42435 to make it trivial to compute type layout or use EUV from lints. The remaining uses of `TypeckTables` in `InferCtxt` are for closure kinds and signatures, used in trait selection and projection normalization. The solution there is likely to add them as bounds to `ParamEnv`. r? @nikomatsakis cc @mcarton @llogiq @Manishearth
| self.infcx.is_tainted_by_errors() | ||
| } | ||
|
|
||
| fn closure_kind(&self, def_id: DefId, span: Span) -> ty::ClosureKind { |
There was a problem hiding this comment.
did this get copied from the version in infcx? I don't see any corresponding - lines...?
There was a problem hiding this comment.
Yes but it works a bit differently - ideally in_progress_tables would be used nowhere.
There was a problem hiding this comment.
I might be able to make it simpler and check tables then tcx without caring about infcx, since now I am always inserting the kinds in the tables ahead of time.
|
r=me modulo crater etc |
|
💥 Test timed out |
|
Second crater report doesn't have any real regressions either. |
|
@bors r=nikomatsakis |
|
📌 Commit fc5c31c has been approved by |
Disentangle InferCtxt, MemCategorizationContext and ExprUseVisitor. At some point in the past, `InferCtxt` started being used to replace an old "`Typer`" abstraction, which provided access to `TypeckTables` and had optionally type inference to account for. That didn't play so nicely with the `'gcx`/`'tcx` split and I had to introduce `borrowck_fake_infer_ctxt`. The situation wasn't great but it wasn't too painful inside `rustc` itself. Recently I've found that method being used in clippy, which does need EUV (before we make it plausible to run lints on HAIR or MIR), and set out to separate inference from tables, for the sake of lint authors. Also fixes #42435 to make it trivial to compute type layout or use EUV from lints. The remaining uses of `TypeckTables` in `InferCtxt` are for closure kinds and signatures, used in trait selection and projection normalization. The solution there is likely to add them as bounds to `ParamEnv`. r? @nikomatsakis cc @mcarton @llogiq @Manishearth
|
☀️ Test successful - status-appveyor, status-travis |
|
I let the cargobomb run I started complete, and the results are here. |
|
@tomprince Thanks! Those all look like false positives, so I'm confident this is fine. |
At some point in the past,
InferCtxtstarted being used to replace an old "Typer" abstraction, which provided access toTypeckTablesand had optionally type inference to account for.That didn't play so nicely with the
'gcx/'tcxsplit and I had to introduceborrowck_fake_infer_ctxt.The situation wasn't great but it wasn't too painful inside
rustcitself.Recently I've found that method being used in clippy, which does need EUV (before we make it plausible to run lints on HAIR or MIR), and set out to separate inference from tables, for the sake of lint authors.
Also fixes #42435 to make it trivial to compute type layout or use EUV from lints.
The remaining uses of
TypeckTablesinInferCtxtare for closure kinds and signatures, used in trait selection and projection normalization. The solution there is likely to add them as bounds toParamEnv.r? @nikomatsakis
cc @mcarton @llogiq @Manishearth