Capture precise paths in THIR and MIR

Cargo.lock

@@ -4011,6 +4011,7 @@ dependencies = [
  "rustc_errors",
  "rustc_hir",
  "rustc_index",
+ "rustc_lexer",
  "rustc_middle",
  "rustc_serialize",
  "rustc_session",

compiler/rustc_ast_lowering/src/expr.rs

@@ -347,7 +347,7 @@ impl<'hir> LoweringContext<'_, 'hir> {
         // `_ => else_block` where `else_block` is `{}` if there's `None`:
         let else_pat = self.pat_wild(span);
         let (else_expr, contains_else_clause) = match else_opt {
-            None => (self.expr_block_empty(span), false),
+            None => (self.expr_block_empty(span.shrink_to_hi()), false),
             Some(els) => (self.lower_expr(els), true),
         };
         let else_arm = self.arm(else_pat, else_expr);

compiler/rustc_codegen_llvm/src/attributes.rs

@@ -127,9 +127,6 @@ fn set_probestack(cx: &CodegenCx<'ll, '_>, llfn: &'ll Value) {
         return;
     }
 
-    // FIXME(richkadel): Make sure probestack plays nice with `-Z instrument-coverage`
-    // or disable it if not, similar to above early exits.
-
     // Flag our internal `__rust_probestack` function as the stack probe symbol.
     // This is defined in the `compiler-builtins` crate for each architecture.
     llvm::AddFunctionAttrStringValue(

compiler/rustc_codegen_llvm/src/coverageinfo/mapgen.rs

@@ -3,11 +3,14 @@ use crate::coverageinfo;
 use crate::llvm;
 
 use llvm::coverageinfo::CounterMappingRegion;
-use rustc_codegen_ssa::coverageinfo::map::{Counter, CounterExpression};
+use rustc_codegen_ssa::coverageinfo::map::{Counter, CounterExpression, FunctionCoverage};
 use rustc_codegen_ssa::traits::ConstMethods;
-use rustc_data_structures::fx::FxIndexSet;
+use rustc_data_structures::fx::{FxHashMap, FxHashSet, FxIndexSet};
+use rustc_hir::def_id::{DefId, DefIdSet, LOCAL_CRATE};
 use rustc_llvm::RustString;
 use rustc_middle::mir::coverage::CodeRegion;
+use rustc_middle::ty::{Instance, TyCtxt};
+use rustc_span::Symbol;
 
 use std::ffi::CString;
 
@@ -26,14 +29,17 @@ use tracing::debug;
 /// undocumented details in Clang's implementation (that may or may not be important) were also
 /// replicated for Rust's Coverage Map.
 pub fn finalize<'ll, 'tcx>(cx: &CodegenCx<'ll, 'tcx>) {
+    let tcx = cx.tcx;
     // Ensure LLVM supports Coverage Map Version 4 (encoded as a zero-based value: 3).
     // If not, the LLVM Version must be less than 11.
     let version = coverageinfo::mapping_version();
     if version != 3 {
-        cx.tcx.sess.fatal("rustc option `-Z instrument-coverage` requires LLVM 11 or higher.");
+        tcx.sess.fatal("rustc option `-Z instrument-coverage` requires LLVM 11 or higher.");
     }
 
-    let function_coverage_map = match cx.coverage_context() {
+    debug!("Generating coverage map for CodegenUnit: `{}`", cx.codegen_unit.name());
+
+    let mut function_coverage_map = match cx.coverage_context() {
         Some(ctx) => ctx.take_function_coverage_map(),
         None => return,
     };
@@ -42,14 +48,15 @@ pub fn finalize<'ll, 'tcx>(cx: &CodegenCx<'ll, 'tcx>) {
         return;
     }
 
+    add_unreachable_coverage(tcx, &mut function_coverage_map);
+
     let mut mapgen = CoverageMapGenerator::new();
 
     // Encode coverage mappings and generate function records
     let mut function_data = Vec::new();
     for (instance, function_coverage) in function_coverage_map {
-        debug!("Generate coverage map for: {:?}", instance);
-
-        let mangled_function_name = cx.tcx.symbol_name(instance).to_string();
+        debug!("Generate function coverage for {}, {:?}", cx.codegen_unit.name(), instance);
+        let mangled_function_name = tcx.symbol_name(instance).to_string();
         let function_source_hash = function_coverage.source_hash();
         let (expressions, counter_regions) =
             function_coverage.get_expressions_and_counter_regions();
@@ -228,3 +235,156 @@ fn save_function_record(
     let is_used = true;
     coverageinfo::save_func_record_to_mod(cx, func_name_hash, func_record_val, is_used);
 }
+
+/// When finalizing the coverage map, `FunctionCoverage` only has the `CodeRegion`s and counters for
+/// the functions that went through codegen; such as public functions and "used" functions
+/// (functions referenced by other "used" or public items). Any other functions considered unused,
+/// or "Unreachable" were still parsed and processed through the MIR stage.
+///
+/// We can find the unreachable functions by the set different of all MIR `DefId`s (`tcx` query
+/// `mir_keys`) minus the codegenned `DefId`s (`tcx` query `collect_and_partition_mono_items`).
+///
+/// *HOWEVER* the codegenned `DefId`s are partitioned across multiple `CodegenUnit`s (CGUs), and
+/// this function is processing a `function_coverage_map` for the functions (`Instance`/`DefId`)
+/// allocated to only one of those CGUs. We must NOT inject any "Unreachable" functions's
+/// `CodeRegion`s more than once, so we have to pick which CGU's `function_coverage_map` to add
+/// each "Unreachable" function to.
+///
+/// Some constraints:
+///
+/// 1. The file name of an "Unreachable" function must match the file name of the existing
+///    codegenned (covered) function to which the unreachable code regions will be added.
+/// 2. The function to which the unreachable code regions will be added must not be a genaric
+///    function (must not have type parameters) because the coverage tools will get confused
+///    if the codegenned function has more than one instantiation and additional `CodeRegion`s
+///    attached to only one of those instantiations.
+fn add_unreachable_coverage<'tcx>(
+    tcx: TyCtxt<'tcx>,
+    function_coverage_map: &mut FxHashMap<Instance<'tcx>, FunctionCoverage<'tcx>>,
+) {
+    // FIXME(#79622): Can this solution be simplified and/or improved? Are there other sources
+    // of compiler state data that might help (or better sources that could be exposed, but
+    // aren't yet)?
+
+    // Note: If the crate *only* defines generic functions, there are no codegenerated non-generic
+    // functions to add any unreachable code to. In this case, the unreachable code regions will
+    // have no coverage, instead of having coverage with zero executions.
+    //
+    // This is probably still an improvement over Clang, which does not generate any coverage
+    // for uninstantiated template functions.
+
+    let has_non_generic_def_ids =
+        function_coverage_map.keys().any(|instance| instance.def.attrs(tcx).len() == 0);
+
+    if !has_non_generic_def_ids {
+        // There are no non-generic functions to add unreachable `CodeRegion`s to
+        return;
+    }
+
+    let all_def_ids: DefIdSet =
+        tcx.mir_keys(LOCAL_CRATE).iter().map(|local_def_id| local_def_id.to_def_id()).collect();
+
+    let (codegenned_def_ids, _) = tcx.collect_and_partition_mono_items(LOCAL_CRATE);
+
+    let mut unreachable_def_ids_by_file: FxHashMap<Symbol, Vec<DefId>> = FxHashMap::default();
+    for &non_codegenned_def_id in all_def_ids.difference(codegenned_def_ids) {
+        // Make sure the non-codegenned (unreachable) function has a file_name
+        if let Some(non_codegenned_file_name) = tcx.covered_file_name(non_codegenned_def_id) {
+            let def_ids = unreachable_def_ids_by_file
+                .entry(*non_codegenned_file_name)
+                .or_insert_with(|| Vec::new());
+            def_ids.push(non_codegenned_def_id);
+        }
+    }
+
+    if unreachable_def_ids_by_file.is_empty() {
+        // There are no unreachable functions with file names to add (in any CGU)
+        return;
+    }
+
+    // Since there may be multiple `CodegenUnit`s, some codegenned_def_ids may be codegenned in a
+    // different CGU, and will be added to the function_coverage_map for each CGU. Determine which
+    // function_coverage_map has the responsibility for publishing unreachable coverage
+    // based on file name:
+    //
+    // For each covered file name, sort ONLY the non-generic codegenned_def_ids, and if
+    // covered_def_ids.contains(the first def_id) for a given file_name, add the unreachable code
+    // region in this function_coverage_map. Otherwise, ignore it and assume another CGU's
+    // function_coverage_map will be adding it (because it will be first for one, and only one,
+    // of them).
+    let mut sorted_codegenned_def_ids: Vec<DefId> =
+        codegenned_def_ids.iter().map(|def_id| *def_id).collect();
+    sorted_codegenned_def_ids.sort_unstable();
+
+    let mut first_covered_def_id_by_file: FxHashMap<Symbol, DefId> = FxHashMap::default();
+    for &def_id in sorted_codegenned_def_ids.iter() {
+        // Only consider non-generic functions, to potentially add unreachable code regions
+        if tcx.generics_of(def_id).count() == 0 {
+            if let Some(covered_file_name) = tcx.covered_file_name(def_id) {
+                // Only add files known to have unreachable functions
+                if unreachable_def_ids_by_file.contains_key(covered_file_name) {
+                    first_covered_def_id_by_file.entry(*covered_file_name).or_insert(def_id);
+                }
+            }
+        }
+    }
+
+    // Get the set of def_ids with coverage regions, known by *this* CoverageContext.
+    let cgu_covered_def_ids: DefIdSet =
+        function_coverage_map.keys().map(|instance| instance.def.def_id()).collect();
+
+    let mut cgu_covered_files: FxHashSet<Symbol> = first_covered_def_id_by_file
+        .iter()
+        .filter_map(
+            |(&file_name, def_id)| {
+                if cgu_covered_def_ids.contains(def_id) { Some(file_name) } else { None }
+            },
+        )
+        .collect();
+
+    // Find the first covered, non-generic function (instance) for each cgu_covered_file. Take the
+    // unreachable code regions for that file, and add them to the function.
+    //
+    // There are three `for` loops here, but (a) the lists have already been reduced to the minimum
+    // required values, the lists are further reduced (by `remove()` calls) when elements are no
+    // longer needed, and there are several opportunities to branch out of loops early.
+    for (instance, function_coverage) in function_coverage_map.iter_mut() {
+        if instance.def.attrs(tcx).len() > 0 {
+            continue;
+        }
+        // The covered function is not generic...
+        let covered_def_id = instance.def.def_id();
+        if let Some(covered_file_name) = tcx.covered_file_name(covered_def_id) {
+            if !cgu_covered_files.remove(&covered_file_name) {
+                continue;
+            }
+            // The covered function's file is one of the files with unreachable code regions, so
+            // all of the unreachable code regions for this file will be added to this function.
+            for def_id in
+                unreachable_def_ids_by_file.remove(&covered_file_name).into_iter().flatten()
+            {
+                // Note, this loop adds an unreachable code regions for each MIR-derived region.
+                // Alternatively, we could add a single code region for the maximum span of all
+                // code regions here.
+                //
+                // Observed downsides of this approach are:
+                //
+                // 1. The coverage results will appear inconsistent compared with the same (or
+                //    similar) code in a function that is reached.
+                // 2. If the function is unreachable from one crate but reachable when compiling
+                //    another referencing crate (such as a cross-crate reference to a
+                //    generic function or inlined function), actual coverage regions overlaid
+                //    on a single larger code span of `Zero` coverage can appear confusing or
+                //    wrong. Chaning the unreachable coverage from a `code_region` to a
+                //    `gap_region` can help, but still can look odd with `0` line counts for
+                //    lines between executed (> 0) lines (such as for blank lines or comments).
+                for &region in tcx.covered_code_regions(def_id) {
+                    function_coverage.add_unreachable_region(region.clone());
+                }
+            }
+            if cgu_covered_files.is_empty() {
+                break;
+            }
+        }
+    }
+}

compiler/rustc_codegen_ssa/src/lib.rs

@@ -2,6 +2,7 @@
 #![feature(bool_to_option)]
 #![feature(option_expect_none)]
 #![feature(box_patterns)]
+#![feature(drain_filter)]
 #![feature(try_blocks)]
 #![feature(in_band_lifetimes)]
 #![feature(nll)]

compiler/rustc_driver/src/lib.rs

@@ -223,7 +223,6 @@ fn run_compiler(
             file_loader: None,
             diagnostic_output,
             stderr: None,
-            crate_name: None,
             lint_caps: Default::default(),
             register_lints: None,
             override_queries: None,
@@ -307,7 +306,6 @@ fn run_compiler(
         file_loader,
         diagnostic_output,
         stderr: None,
-        crate_name: None,
         lint_caps: Default::default(),
         register_lints: None,
         override_queries: None,

compiler/rustc_error_codes/src/error_codes.rs

@@ -283,6 +283,7 @@ E0537: include_str!("./error_codes/E0537.md"),
 E0538: include_str!("./error_codes/E0538.md"),
 E0539: include_str!("./error_codes/E0539.md"),
 E0541: include_str!("./error_codes/E0541.md"),
+E0546: include_str!("./error_codes/E0546.md"),
 E0550: include_str!("./error_codes/E0550.md"),
 E0551: include_str!("./error_codes/E0551.md"),
 E0552: include_str!("./error_codes/E0552.md"),
@@ -603,7 +604,6 @@ E0779: include_str!("./error_codes/E0779.md"),
     E0543, // missing 'reason'
     E0544, // multiple stability levels
     E0545, // incorrect 'issue'
-    E0546, // missing 'feature'
     E0547, // missing 'issue'
 //  E0548, // replaced with a generic attribute input check
     // rustc_deprecated attribute must be paired with either stable or unstable

compiler/rustc_error_codes/src/error_codes/E0546.md

@@ -0,0 +1,27 @@
+A feature name is missing.
+
+Erroneous code example:
+
+```compile_fail,E0546
+#![feature(staged_api)]
+#![stable(since = "1.0.0", feature = "test")]
+
+#[unstable(issue = "none")] // invalid
+fn unstable_fn() {}
+
+#[stable(since = "1.0.0")] // invalid
+fn stable_fn() {}
+```
+
+To fix the issue you need to provide a feature name.
+
+```
+#![feature(staged_api)]
+#![stable(since = "1.0.0", feature = "test")]
+
+#[unstable(feature = "unstable_fn", issue = "none")] // ok!
+fn unstable_fn() {}
+
+#[stable(feature = "stable_fn", since = "1.0.0")] // ok!
+fn stable_fn() {}
+```

compiler/rustc_expand/src/base.rs

@@ -235,12 +235,10 @@ impl Annotatable {
     pub fn derive_allowed(&self) -> bool {
         match *self {
             Annotatable::Stmt(ref stmt) => match stmt.kind {
-                ast::StmtKind::Item(ref item) => match item.kind {
-                    ast::ItemKind::Struct(..)
-                    | ast::ItemKind::Enum(..)
-                    | ast::ItemKind::Union(..) => true,
-                    _ => false,
-                },
+                ast::StmtKind::Item(ref item) => matches!(
+                    item.kind,
+                    ast::ItemKind::Struct(..) | ast::ItemKind::Enum(..) | ast::ItemKind::Union(..)
+                ),
                 _ => false,
             },
             Annotatable::Item(ref item) => match item.kind {

compiler/rustc_expand/src/expand.rs

@@ -1134,7 +1134,9 @@ impl<'a, 'b> MutVisitor for InvocationCollector<'a, 'b> {
             if let Some(attr) = self.take_first_attr_no_derive(&mut expr) {
                 // Collect the invoc regardless of whether or not attributes are permitted here
                 // expansion will eat the attribute so it won't error later.
-                attr.0.as_ref().map(|attr| self.cfg.maybe_emit_expr_attr_err(attr));
+                if let Some(attr) = attr.0.as_ref() {
+                    self.cfg.maybe_emit_expr_attr_err(attr)
+                }
 
                 // AstFragmentKind::Expr requires the macro to emit an expression.
                 return self
@@ -1231,7 +1233,9 @@ impl<'a, 'b> MutVisitor for InvocationCollector<'a, 'b> {
             self.cfg.configure_expr_kind(&mut expr.kind);
 
             if let Some(attr) = self.take_first_attr_no_derive(&mut expr) {
-                attr.0.as_ref().map(|attr| self.cfg.maybe_emit_expr_attr_err(attr));
+                if let Some(attr) = attr.0.as_ref() {
+                    self.cfg.maybe_emit_expr_attr_err(attr)
+                }
 
                 return self
                     .collect_attr(attr, Annotatable::Expr(P(expr)), AstFragmentKind::OptExpr)

compiler/rustc_hir/src/hir.rs

@@ -2401,7 +2401,7 @@ impl StructField<'_> {
     // Still necessary in couple of places
     pub fn is_positional(&self) -> bool {
         let first = self.ident.as_str().as_bytes()[0];
-        first >= b'0' && first <= b'9'
+        (b'0'..=b'9').contains(&first)
     }
 }
 

compiler/rustc_infer/src/infer/higher_ranked/mod.rs

@@ -24,9 +24,9 @@ impl<'a, 'tcx> CombineFields<'a, 'tcx> {
         // as-is, we need to do some extra work here in order to make sure
         // that function subtyping works correctly with respect to regions
         //
-        // Note: this is a subtle algorithm.  For a full explanation,
-        // please see the large comment at the end of the file in the (inlined) module
-        // `doc`.
+        // Note: this is a subtle algorithm.  For a full explanation, please see
+        // the rustc dev guide:
+        // <https://rustc-dev-guide.rust-lang.org/borrow_check/region_inference/placeholders_and_universes.html>
 
         let span = self.trace.cause.span;