Add check for [T;N]/usize mismatch in astconv

[JulianKnodt]

Helps clarify the issue in #80506
by adding a specific check for mismatches between [T;N] and usize.

r? @lcnr

[rust-log-analyzer]

The job mingw-check failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

configure: rust.channel         := nightly
configure: rust.debug-assertions := True
configure: llvm.assertions      := True
configure: dist.missing-tools   := True
configure: build.configure-args := ['--enable-sccache', '--disable-manage-submodu ...
configure: writing `config.toml` in current directory
configure: 
configure: run `python /checkout/x.py --help`
configure: 
---
skip untracked path cpu-usage.csv during rustfmt invocations
skip untracked path src/doc/book/ during rustfmt invocations
skip untracked path src/doc/rust-by-example/ during rustfmt invocations
skip untracked path src/llvm-project/ during rustfmt invocations
Diff in /checkout/compiler/rustc_typeck/src/astconv/generics.rs at line 7:
 use rustc_errors::{pluralize, struct_span_err, Applicability, DiagnosticId, ErrorReported};
 use rustc_hir as hir;
 use rustc_hir::def_id::DefId;
-use rustc_hir::{GenericArg};
+use rustc_hir::GenericArg;
 use rustc_middle::ty::{
     self, subst, subst::SubstsRef, GenericParamDef, GenericParamDefKind, Ty, TyCtxt,
 };
Diff in /checkout/compiler/rustc_typeck/src/astconv/generics.rs at line 61:
Running `"/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/rustfmt" "--config-path" "/checkout" "--edition" "2018" "--unstable-features" "--skip-children" "--check" "/checkout/compiler/rustc_typeck/src/astconv/generics.rs"` failed.
If you're running `tidy`, try again with `--bless`. Or, if you just want to format code, run `./x.py fmt` instead.
 
         // Specific suggestion set for diagnostics
         if let Some(param) = param {
-          match (arg, &param.kind) {
-              (
-                  GenericArg::Type(hir::Ty { kind: hir::TyKind::Path { .. }, .. }),
-                  GenericParamDefKind::Const { .. },
-              ) => {
-                  let suggestions = vec![
-                      (arg.span().shrink_to_lo(), String::from("{ ")),
-                      (arg.span().shrink_to_hi(), String::from(" }")),
-                  ];
-                  err.multipart_suggestion(
-                      "if this generic argument was intended as a const parameter, \
+            match (arg, &param.kind) {
+                (
+                    GenericArg::Type(hir::Ty { kind: hir::TyKind::Path { .. }, .. }),
+                    GenericParamDefKind::Const { .. },
+                ) => {
+                    let suggestions = vec![
+                        (arg.span().shrink_to_lo(), String::from("{ ")),
+                        (arg.span().shrink_to_hi(), String::from(" }")),
+                    ];
+                    err.multipart_suggestion(
+                        "if this generic argument was intended as a const parameter, \
                     try surrounding it with braces:",
-                      suggestions,
-                      Applicability::MaybeIncorrect,
-                  );
-              (
-              (
-                  GenericArg::Type(hir::Ty { kind: hir::TyKind::Array(_, len), .. }),
-                  GenericParamDefKind::Const { .. },
-              ) => if tcx.type_of(param.def_id).is_ptr_sized_integral() {
-                  let snippet = sess
-                      .source_map()
-                      .span_to_snippet(tcx.hir().span(len.hir_id))
-                      .unwrap_or(String::from("N"));
-                  err.span_suggestion(
-                      arg.span(),
-                      "`[T; N]` provided where `usize` was expected, try",
-                      format!("{}", snippet),
-                      Applicability::MaybeIncorrect,
-                  );
-              _ => {}
-          }
+                        suggestions,
+                        suggestions,
+                        Applicability::MaybeIncorrect,
+                    );
+                (
+                (
+                    GenericArg::Type(hir::Ty { kind: hir::TyKind::Array(_, len), .. }),
+                    GenericParamDefKind::Const { .. },
+                ) => {
+                    if tcx.type_of(param.def_id).is_ptr_sized_integral() {
+                        let snippet = sess
+                            .source_map()
+                            .span_to_snippet(tcx.hir().span(len.hir_id))
+                            .unwrap_or(String::from("N"));
+                        err.span_suggestion(
+                            arg.span(),
+                            "`[T; N]` provided where `usize` was expected, try",
+                            format!("{}", snippet),
+                            Applicability::MaybeIncorrect,
+                        );
+                }
+                _ => {}
+            }
         }
         }
 
         // This note is only true when generic parameters are strictly ordered by their kind.
Build completed unsuccessfully in 0:00:14

[lcnr]

thanks ❤️

can you also add a test for [u8; 1 + 2]? We probably want to always recommend using braces for now as especially for complex expressions the error recovery when they are missing is fairly hard.

[rust-log-analyzer]

The job mingw-check failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

    Checking rustc_attr v0.0.0 (/checkout/compiler/rustc_attr)
    Checking rustc_hir v0.0.0 (/checkout/compiler/rustc_hir)
    Checking rustc_parse v0.0.0 (/checkout/compiler/rustc_parse)
    Checking chalk-engine v0.36.0
error[E0599]: no method named `features` found for reference `&Session` in the current scope
   --> compiler/rustc_hir/src/hir.rs:299:60
    |
299 |                 ast::ParamKindOrd::Const { unordered: sess.features().const_generics }
    |                                                            ^^^^^^^^ private field, not a method
error: aborting due to previous error

For more information about this error, try `rustc --explain E0599`.
error: could not compile `rustc_hir`
error: could not compile `rustc_hir`

To learn more, run the command again with --verbose.
warning: build failed, waiting for other jobs to finish...
error: build failed
command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "check" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "16" "--release" "--color" "always" "--features" " llvm max_level_info" "--manifest-path" "/checkout/compiler/rustc/Cargo.toml" "-p" "rustc-main" "-p" "rustc_codegen_ssa" "-p" "rustc_session" "-p" "rustc_lint_defs" "-p" "rustc_feature" "-p" "rustc_data_structures" "-p" "rustc_graphviz" "-p" "rustc_target" "-p" "rustc_index" "-p" "rustc_span" "-p" "rustc_arena" "-p" "rustc_macros" "-p" "rustc_symbol_mangling" "-p" "rustc_middle" "-p" "rustc_query_system" "-p" "rustc_type_ir" "-p" "rustc_incremental" "-p" "rustc_apfloat" "-p" "rustc_hir" "-p" "rustc_ast" "-p" "rustc_lexer" "-p" "rustc_errors" "-p" "rustc_attr" "-p" "rustc_ast_pretty" "-p" "rustc_fs_util" "-p" "rustc_serialize" "-p" "rustc_driver" "-p" "rustc_parse" "-p" "rustc_error_codes" "-p" "rustc_lint" "-p" "rustc_trait_selection" "-p" "rustc_infer" "-p" "rustc_parse_format" "-p" "rustc_mir" "-p" "coverage_test_macros" "-p" "rustc_hir_pretty" "-p" "rustc_metadata" "-p" "rustc_expand" "-p" "rustc_ast_passes" "-p" "rustc_interface" "-p" "rustc_builtin_macros" "-p" "rustc_mir_build" "-p" "rustc_resolve" "-p" "rustc_codegen_llvm" "-p" "rustc_llvm" "-p" "rustc_ty_utils" "-p" "rustc_traits" "-p" "rustc_passes" "-p" "rustc_ast_lowering" "-p" "rustc_privacy" "-p" "rustc_typeck" "-p" "rustc_save_analysis" "-p" "rustc_plugin_impl" "--message-format" "json-render-diagnostics"
failed to run: /checkout/obj/build/bootstrap/debug/bootstrap check
Build completed unsuccessfully in 0:01:50

[lcnr]

thanks 👍

@bors r+ rollup

[bors]

📌 Commit 54883e0 has been approved by lcnr

[rylev]

@JulianKnodt @lcnr it seems this change was responsible for a moderate performance regression (and a minor perf gain). You can see the results in #80798 where we reverted this PR and ran the perf suite.

The regression was in the deeply-nested-async benchmark which does not contain any const code it seems. Any ideas?

[lcnr]

is it possible that this loop is a lot hotter than I expected so storing arg with param here is quite costly?

Other than that i really don't know waht could be causing the perf regression.

[JulianKnodt]

I agree that this would be the most likely cause, it could be that the type for param for deeply-nested-async stuff is huge and thus takes up a lot of space on the stack, so even if it's not hot it might just be causing a lot more cache misses

edit: these are refs, so the size shouldn't matter, ignore above

Would it be good to change this to an index instead? I don't remember where the param was coming from at the moment

[Mark-Simulacrum]

It seems possible that this increases the size of force_infer_lt enough that it becomes stored on the stack instead of a register, I guess, but it's pretty odd that it would cause any problems (still just two pointers vs one). An index would still be a usize so should be equally costly I'd expect.

[JulianKnodt]

I added a print to the inner body, and during compilation it appears that the body is hit at least 1.5M times, so I guess that's fairly hot?

[lcnr]

you mean during the compilation of the compiler itself? so during ./x.py build --stage 2 or just while building std?

[JulianKnodt]

I was running ./x.py test src/test/ui, but it got to compiling stage 1 artifacts and I quit out