Rollup merge of #133545 - clubby789:symbol-intern-lit, r=jieyouxu

Lint against Symbol::intern on a string literal

Disabled in tests where this doesn't make much sense

compiler/rustc_ast_passes/src/feature_gate.rs

@@ -4,9 +4,9 @@ use rustc_ast::{NodeId, PatKind, attr, token};
 use rustc_feature::{AttributeGate, BUILTIN_ATTRIBUTE_MAP, BuiltinAttribute, Features, GateIssue};
 use rustc_session::Session;
 use rustc_session::parse::{feature_err, feature_err_issue, feature_warn};
+use rustc_span::Span;
 use rustc_span::source_map::Spanned;
-use rustc_span::symbol::sym;
-use rustc_span::{Span, Symbol};
+use rustc_span::symbol::{Symbol, sym};
 use rustc_target::spec::abi;
 use thin_vec::ThinVec;
 
@@ -695,6 +695,7 @@ fn check_new_solver_banned_features(sess: &Session, features: &Features) {
         .find(|feat| feat.gate_name == sym::generic_const_exprs)
         .map(|feat| feat.attr_sp)
     {
+        #[cfg_attr(not(bootstrap), allow(rustc::symbol_intern_string_literal))]
         sess.dcx().emit_err(errors::IncompatibleFeatures {
             spans: vec![gce_span],
             f1: Symbol::intern("-Znext-solver=globally"),

compiler/rustc_borrowck/src/universal_regions.rs

@@ -33,8 +33,8 @@ use rustc_middle::ty::{
     TyCtxt, TypeVisitableExt,
 };
 use rustc_middle::{bug, span_bug};
+use rustc_span::ErrorGuaranteed;
 use rustc_span::symbol::{kw, sym};
-use rustc_span::{ErrorGuaranteed, Symbol};
 use tracing::{debug, instrument};
 
 use crate::BorrowckInferCtxt;
@@ -524,7 +524,7 @@ impl<'cx, 'tcx> UniversalRegionsBuilder<'cx, 'tcx> {
 
                 let reg_vid = self
                     .infcx
-                    .next_nll_region_var(FR, || RegionCtxt::Free(Symbol::intern("c-variadic")))
+                    .next_nll_region_var(FR, || RegionCtxt::Free(sym::c_dash_variadic))
                     .as_var();
 
                 let region = ty::Region::new_var(self.infcx.tcx, reg_vid);
@@ -540,10 +540,8 @@ impl<'cx, 'tcx> UniversalRegionsBuilder<'cx, 'tcx> {
             }
         }
 
-        let fr_fn_body = self
-            .infcx
-            .next_nll_region_var(FR, || RegionCtxt::Free(Symbol::intern("fn_body")))
-            .as_var();
+        let fr_fn_body =
+            self.infcx.next_nll_region_var(FR, || RegionCtxt::Free(sym::fn_body)).as_var();
 
         let num_universals = self.infcx.num_region_vars();
 

compiler/rustc_codegen_cranelift/src/driver/jit.rs

@@ -74,7 +74,7 @@ fn create_jit_module(
     jit_builder.symbol("__clif_jit_fn", clif_jit_fn as *const u8);
     let mut jit_module = UnwindModule::new(JITModule::new(jit_builder), false);
 
-    let cx = crate::CodegenCx::new(tcx, jit_module.isa(), false, Symbol::intern("dummy_cgu_name"));
+    let cx = crate::CodegenCx::new(tcx, jit_module.isa(), false, sym::dummy_cgu_name);
 
     crate::allocator::codegen(tcx, &mut jit_module);
 
@@ -276,12 +276,7 @@ fn jit_fn(instance_ptr: *const Instance<'static>, trampoline_ptr: *const u8) ->
 
             jit_module.module.prepare_for_function_redefine(func_id).unwrap();
 
-            let mut cx = crate::CodegenCx::new(
-                tcx,
-                jit_module.isa(),
-                false,
-                Symbol::intern("dummy_cgu_name"),
-            );
+            let mut cx = crate::CodegenCx::new(tcx, jit_module.isa(), false, sym::dummy_cgu_name);
             codegen_and_compile_fn(tcx, &mut cx, &mut Context::new(), jit_module, instance);
 
             assert!(cx.global_asm.is_empty());

compiler/rustc_codegen_cranelift/src/lib.rs

@@ -189,18 +189,13 @@ impl CodegenBackend for CraneliftCodegenBackend {
         // FIXME return the actually used target features. this is necessary for #[cfg(target_feature)]
         if sess.target.arch == "x86_64" && sess.target.os != "none" {
             // x86_64 mandates SSE2 support
-            vec![Symbol::intern("fxsr"), sym::sse, Symbol::intern("sse2")]
+            vec![sym::fsxr, sym::sse, sym::sse2]
         } else if sess.target.arch == "aarch64" {
             match &*sess.target.os {
                 "none" => vec![],
                 // On macOS the aes, sha2 and sha3 features are enabled by default and ring
                 // fails to compile on macOS when they are not present.
-                "macos" => vec![
-                    sym::neon,
-                    Symbol::intern("aes"),
-                    Symbol::intern("sha2"),
-                    Symbol::intern("sha3"),
-                ],
+                "macos" => vec![sym::neon, sym::aes, sym::sha2, sym::sha3],
                 // AArch64 mandates Neon support
                 _ => vec![sym::neon],
             }

compiler/rustc_hir_analysis/src/collect/generics_of.rs

@@ -419,7 +419,7 @@ pub(super) fn generics_of(tcx: TyCtxt<'_>, def_id: LocalDefId) -> ty::Generics {
     if let Node::ConstBlock(_) = node {
         own_params.push(ty::GenericParamDef {
             index: next_index(),
-            name: Symbol::intern("<const_ty>"),
+            name: rustc_span::sym::const_ty_placeholder,
             def_id: def_id.to_def_id(),
             pure_wrt_drop: false,
             kind: ty::GenericParamDefKind::Type { has_default: false, synthetic: false },

compiler/rustc_hir_typeck/src/method/suggest.rs

@@ -664,7 +664,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
         let is_write = sugg_span.ctxt().outer_expn_data().macro_def_id.is_some_and(|def_id| {
             tcx.is_diagnostic_item(sym::write_macro, def_id)
                 || tcx.is_diagnostic_item(sym::writeln_macro, def_id)
-        }) && item_name.name == Symbol::intern("write_fmt");
+        }) && item_name.name == sym::write_fmt;
         let mut err = if is_write && let SelfSource::MethodCall(rcvr_expr) = source {
             self.suggest_missing_writer(rcvr_ty, rcvr_expr)
         } else {

compiler/rustc_lint/messages.ftl

@@ -775,6 +775,9 @@ lint_suspicious_double_ref_clone =
 lint_suspicious_double_ref_deref =
     using `.deref()` on a double reference, which returns `{$ty}` instead of dereferencing the inner type
 
+lint_symbol_intern_string_literal = using `Symbol::intern` on a string literal
+    .help = consider adding the symbol to `compiler/rustc_span/src/symbol.rs`
+
 lint_trailing_semi_macro = trailing semicolon in macro used in expression position
     .note1 = macro invocations at the end of a block are treated as expressions
     .note2 = to ignore the value produced by the macro, add a semicolon after the invocation of `{$name}`

compiler/rustc_lint/src/internal.rs

@@ -17,8 +17,9 @@ use tracing::debug;
 
 use crate::lints::{
     BadOptAccessDiag, DefaultHashTypesDiag, DiagOutOfImpl, LintPassByHand, NonExistentDocKeyword,
-    NonGlobImportTypeIrInherent, QueryInstability, QueryUntracked, SpanUseEqCtxtDiag, TyQualified,
-    TykindDiag, TykindKind, TypeIrInherentUsage, UntranslatableDiag,
+    NonGlobImportTypeIrInherent, QueryInstability, QueryUntracked, SpanUseEqCtxtDiag,
+    SymbolInternStringLiteralDiag, TyQualified, TykindDiag, TykindKind, TypeIrInherentUsage,
+    UntranslatableDiag,
 };
 use crate::{EarlyContext, EarlyLintPass, LateContext, LateLintPass, LintContext};
 
@@ -650,3 +651,33 @@ fn is_span_ctxt_call(cx: &LateContext<'_>, expr: &Expr<'_>) -> bool {
         _ => false,
     }
 }
+
+declare_tool_lint! {
+    /// The `symbol_intern_string_literal` detects `Symbol::intern` being called on a string literal
+    pub rustc::SYMBOL_INTERN_STRING_LITERAL,
+    // rustc_driver crates out of the compiler can't/shouldn't add preinterned symbols;
+    // bootstrap will deny this manually
+    Allow,
+    "Forbid uses of string literals in `Symbol::intern`, suggesting preinterning instead",
+    report_in_external_macro: true
+}
+
+declare_lint_pass!(SymbolInternStringLiteral => [SYMBOL_INTERN_STRING_LITERAL]);
+
+impl<'tcx> LateLintPass<'tcx> for SymbolInternStringLiteral {
+    fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx rustc_hir::Expr<'tcx>) {
+        if let ExprKind::Call(path, [arg]) = expr.kind
+            && let ExprKind::Path(ref qpath) = path.kind
+            && let Some(def_id) = cx.qpath_res(qpath, path.hir_id).opt_def_id()
+            && cx.tcx.is_diagnostic_item(sym::SymbolIntern, def_id)
+            && let ExprKind::Lit(kind) = arg.kind
+            && let rustc_ast::LitKind::Str(_, _) = kind.node
+        {
+            cx.emit_span_lint(
+                SYMBOL_INTERN_STRING_LITERAL,
+                kind.span,
+                SymbolInternStringLiteralDiag,
+            );
+        }
+    }
+}

compiler/rustc_lint/src/lib.rs

@@ -614,6 +614,8 @@ fn register_internals(store: &mut LintStore) {
     store.register_late_mod_pass(|_| Box::new(PassByValue));
     store.register_lints(&SpanUseEqCtxt::lint_vec());
     store.register_late_mod_pass(|_| Box::new(SpanUseEqCtxt));
+    store.register_lints(&SymbolInternStringLiteral::lint_vec());
+    store.register_late_mod_pass(|_| Box::new(SymbolInternStringLiteral));
     // FIXME(davidtwco): deliberately do not include `UNTRANSLATABLE_DIAGNOSTIC` and
     // `DIAGNOSTIC_OUTSIDE_OF_IMPL` here because `-Wrustc::internal` is provided to every crate and
     // these lints will trigger all of the time - change this once migration to diagnostic structs

compiler/rustc_lint/src/lints.rs

@@ -907,6 +907,11 @@ pub(crate) struct QueryUntracked {
 #[diag(lint_span_use_eq_ctxt)]
 pub(crate) struct SpanUseEqCtxtDiag;
 
+#[derive(LintDiagnostic)]
+#[diag(lint_symbol_intern_string_literal)]
+#[help]
+pub(crate) struct SymbolInternStringLiteralDiag;
+
 #[derive(LintDiagnostic)]
 #[diag(lint_tykind_kind)]
 pub(crate) struct TykindKind {

compiler/rustc_lint/src/unused.rs

@@ -1563,7 +1563,7 @@ impl UnusedImportBraces {
                     }
                     rename.unwrap_or(orig_ident).name
                 }
-                ast::UseTreeKind::Glob => Symbol::intern("*"),
+                ast::UseTreeKind::Glob => sym::asterisk,
                 ast::UseTreeKind::Nested { .. } => return,
             };
 

compiler/rustc_metadata/src/creader.rs

@@ -861,8 +861,10 @@ impl<'a, 'tcx> CrateLoader<'a, 'tcx> {
         // First up we check for global allocators. Look at the crate graph here
         // and see what's a global allocator, including if we ourselves are a
         // global allocator.
-        let mut global_allocator =
-            self.cstore.has_global_allocator.then(|| Symbol::intern("this crate"));
+        #[cfg_attr(not(bootstrap), allow(rustc::symbol_intern_string_literal))]
+        let this_crate = Symbol::intern("this crate");
+
+        let mut global_allocator = self.cstore.has_global_allocator.then_some(this_crate);
         for (_, data) in self.cstore.iter_crate_data() {
             if data.has_global_allocator() {
                 match global_allocator {
@@ -876,8 +878,7 @@ impl<'a, 'tcx> CrateLoader<'a, 'tcx> {
                 }
             }
         }
-        let mut alloc_error_handler =
-            self.cstore.has_alloc_error_handler.then(|| Symbol::intern("this crate"));
+        let mut alloc_error_handler = self.cstore.has_alloc_error_handler.then_some(this_crate);
         for (_, data) in self.cstore.iter_crate_data() {
             if data.has_alloc_error_handler() {
                 match alloc_error_handler {

compiler/rustc_metadata/src/rmeta/decoder.rs

@@ -872,6 +872,7 @@ impl MetadataBlob {
 
                         let def_kind = root.tables.def_kind.get(blob, item).unwrap();
                         let def_key = root.tables.def_keys.get(blob, item).unwrap().decode(blob);
+                        #[cfg_attr(not(bootstrap), allow(rustc::symbol_intern_string_literal))]
                         let def_name = if item == CRATE_DEF_INDEX {
                             rustc_span::symbol::kw::Crate
                         } else {

compiler/rustc_parse/src/parser/ty.rs

@@ -9,7 +9,7 @@ use rustc_ast::{
 };
 use rustc_errors::{Applicability, PResult};
 use rustc_span::symbol::{Ident, kw, sym};
-use rustc_span::{ErrorGuaranteed, Span, Symbol};
+use rustc_span::{ErrorGuaranteed, Span};
 use thin_vec::{ThinVec, thin_vec};
 
 use super::{Parser, PathStyle, SeqSep, TokenType, Trailing};
@@ -1136,7 +1136,7 @@ impl<'a> Parser<'a> {
                 Some(ast::Path {
                     span: fn_token_span.to(self.prev_token.span),
                     segments: thin_vec![ast::PathSegment {
-                        ident: Ident::new(Symbol::intern("Fn"), fn_token_span),
+                        ident: Ident::new(sym::Fn, fn_token_span),
                         id: DUMMY_NODE_ID,
                         args: Some(P(ast::GenericArgs::Parenthesized(ast::ParenthesizedArgs {
                             span: args_lo.to(self.prev_token.span),

compiler/rustc_resolve/src/late/diagnostics.rs

@@ -3113,6 +3113,7 @@ impl<'ast, 'ra: 'ast, 'tcx> LateResolutionVisitor<'_, 'ast, 'ra, 'tcx> {
             }
         }
 
+        #[cfg_attr(not(bootstrap), allow(rustc::symbol_intern_string_literal))]
         let existing_name = match &in_scope_lifetimes[..] {
             [] => Symbol::intern("'a"),
             [(existing, _)] => existing.name,

compiler/rustc_resolve/src/lib.rs

@@ -2257,7 +2257,7 @@ fn module_to_string(module: Module<'_>) -> Option<String> {
                 collect_mod(names, parent);
             }
         } else {
-            names.push(Symbol::intern("<opaque>"));
+            names.push(sym::opaque_module_name_placeholder);
             collect_mod(names, module.parent.unwrap());
         }
     }

compiler/rustc_session/src/output.rs

@@ -87,7 +87,7 @@ pub fn find_crate_name(sess: &Session, attrs: &[ast::Attribute]) -> Symbol {
         }
     }
 
-    Symbol::intern("rust_out")
+    sym::rust_out
 }
 
 pub fn validate_crate_name(sess: &Session, s: Symbol, sp: Option<Span>) {

compiler/rustc_span/src/symbol.rs

@@ -298,6 +298,7 @@ symbols! {
         Return,
         Right,
         Rust,
+        RustaceansAreAwesome,
         RustcDecodable,
         RustcEncodable,
         RwLock,
@@ -316,6 +317,7 @@ symbols! {
         StructuralPartialEq,
         SubdiagMessage,
         Subdiagnostic,
+        SymbolIntern,
         Sync,
         SyncUnsafeCell,
         T,
@@ -377,6 +379,7 @@ symbols! {
         adt_const_params,
         advanced_slice_patterns,
         adx_target_feature,
+        aes,
         aggregate_raw_ptr,
         alias,
         align,
@@ -440,6 +443,7 @@ symbols! {
         associated_types,
         assume,
         assume_init,
+        asterisk: "*",
         async_await,
         async_call,
         async_call_mut,
@@ -521,6 +525,7 @@ symbols! {
         btreeset_iter,
         builtin_syntax,
         c,
+        c_dash_variadic,
         c_str,
         c_str_literals,
         c_unwind,
@@ -651,6 +656,7 @@ symbols! {
         const_trait_bound_opt_out,
         const_trait_impl,
         const_try,
+        const_ty_placeholder: "<const_ty>",
         constant,
         constructor,
         convert_identity,
@@ -780,6 +786,7 @@ symbols! {
         drop_types_in_const,
         dropck_eyepatch,
         dropck_parametricity,
+        dummy_cgu_name,
         dylib,
         dyn_compatible_for_dispatch,
         dyn_metadata,
@@ -923,6 +930,7 @@ symbols! {
         fmuladdf32,
         fmuladdf64,
         fn_align,
+        fn_body,
         fn_delegation,
         fn_must_use,
         fn_mut,
@@ -963,6 +971,7 @@ symbols! {
         fs_create_dir,
         fsub_algebraic,
         fsub_fast,
+        fsxr,
         full,
         fundamental,
         fused_iterator,
@@ -1386,6 +1395,7 @@ symbols! {
         on,
         on_unimplemented,
         opaque,
+        opaque_module_name_placeholder: "<opaque>",
         open_options_new,
         ops,
         opt_out_copy,
@@ -1655,6 +1665,7 @@ symbols! {
         rust_eh_catch_typeinfo,
         rust_eh_personality,
         rust_logo,
+        rust_out,
         rustc,
         rustc_abi,
         rustc_allocator,
@@ -1777,6 +1788,8 @@ symbols! {
         self_in_typedefs,
         self_struct_ctor,
         semitransparent,
+        sha2,
+        sha3,
         sha512_sm_x86,
         shadow_call_stack,
         shallow,
@@ -1891,6 +1904,7 @@ symbols! {
         sreg,
         sreg_low16,
         sse,
+        sse2,
         sse4a_target_feature,
         stable,
         staged_api,
@@ -2178,6 +2192,7 @@ symbols! {
         wrapping_sub,
         wreg,
         write_bytes,
+        write_fmt,
         write_macro,
         write_str,
         write_via_move,
@@ -2407,6 +2422,7 @@ impl Symbol {
     }
 
     /// Maps a string to its interned representation.
+    #[rustc_diagnostic_item = "SymbolIntern"]
     pub fn intern(string: &str) -> Self {
         with_session_globals(|session_globals| session_globals.symbol_interner.intern(string))
     }

compiler/rustc_trait_selection/src/traits/dyn_compatibility.rs

@@ -18,7 +18,6 @@ use rustc_middle::ty::{
     TypeVisitableExt, TypeVisitor, TypingMode, Upcast,
 };
 use rustc_span::Span;
-use rustc_span::symbol::Symbol;
 use smallvec::SmallVec;
 use tracing::{debug, instrument};
 
@@ -679,7 +678,7 @@ fn receiver_is_dispatchable<'tcx>(
     // FIXME(mikeyhew) this is a total hack. Once dyn_compatible_for_dispatch is stabilized, we can
     // replace this with `dyn Trait`
     let unsized_self_ty: Ty<'tcx> =
-        Ty::new_param(tcx, u32::MAX, Symbol::intern("RustaceansAreAwesome"));
+        Ty::new_param(tcx, u32::MAX, rustc_span::sym::RustaceansAreAwesome);
 
     // `Receiver[Self => U]`
     let unsized_receiver_ty =

src/bootstrap/src/core/builder/cargo.rs

@@ -1033,6 +1033,15 @@ impl Builder<'_> {
 
         if mode == Mode::Rustc {
             rustflags.arg("-Wrustc::internal");
+            // cfg(bootstrap) - remove this check when lint is in bootstrap compiler
+            if stage != 0 {
+                // Lint is allow by default so downstream tools don't get a lit
+                // they can do nothing about
+                // We shouldn't be preinterning symbols used by tests
+                if cmd_kind != Kind::Test {
+                    rustflags.arg("-Drustc::symbol_intern_string_literal");
+                }
+            }
             // FIXME(edition_2024): Change this to `-Wrust_2024_idioms` when all
             // of the individual lints are satisfied.
             rustflags.arg("-Wkeyword_idents_2024");