Construct SourceMap at the same time as SessionGlobals.

nnethercote · nnethercote · commit d55849c70102 · 2024-03-21T14:22:27.000+11:00
Currently `SourceMap` is constructed slightly later than
`SessionGlobals`, and inserted. This commit changes things so they are
done at the same time.

Benefits:
- `SessionGlobals::source_map` changes from
  `Lock&lt;Option&lt;Lrc&lt;SourceMap&gt;&gt;&gt;` to `Option&lt;Lrc&lt;SourceMap&gt;&gt;`. It's still
  optional, but mutability isn't required because it's initialized at
  construction.
- `set_source_map` is removed, simplifying `run_compiler`, which is
  good because that's a critical function and it's nice to make it
  simpler.

This requires moving things around a bit, so the
necessary inputs are available when `SessionGlobals` is created, in
particular the `loader` and `hash_kind`, which are no longer computed by
`build_session`. These inputs are captured by the new `SourceMapInputs`
type, which is threaded through various places. And there's a new
diff --git a/compiler/rustc_hir/src/tests.rs b/compiler/rustc_hir/src/tests.rs
@@ -14,7 +14,7 @@ fn def_path_hash_depends_on_crate_id() {
     // the crate by changing the crate disambiguator (e.g. via bumping the
     // crate's version number).
 
-    create_session_globals_then(Edition::Edition2024, || {
+    create_session_globals_then(Edition::Edition2024, None, || {
         let id0 = StableCrateId::new(Symbol::intern("foo"), false, vec!["1".to_string()], "");
         let id1 = StableCrateId::new(Symbol::intern("foo"), false, vec!["2".to_string()], "");
 
diff --git a/compiler/rustc_interface/src/interface.rs b/compiler/rustc_interface/src/interface.rs
@@ -19,7 +19,7 @@ use rustc_session::config::{self, Cfg, CheckCfg, ExpectedValues, Input, OutFileN
 use rustc_session::filesearch::{self, sysroot_candidates};
 use rustc_session::parse::ParseSess;
 use rustc_session::{lint, CompilerIO, EarlyDiagCtxt, Session};
-use rustc_span::source_map::FileLoader;
+use rustc_span::source_map::{FileLoader, RealFileLoader, SourceMapInputs};
 use rustc_span::symbol::sym;
 use rustc_span::FileName;
 use std::path::PathBuf;
@@ -331,18 +331,22 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
     let early_dcx = EarlyDiagCtxt::new(config.opts.error_format);
     early_dcx.initialize_checked_jobserver();
 
+    crate::callbacks::setup_callbacks();
+
+    let loader = config.file_loader.unwrap_or_else(|| Box::new(RealFileLoader));
+    let sysroot = filesearch::materialize_sysroot(config.opts.maybe_sysroot.clone());
+    let target = config::build_target_config(&early_dcx, &config.opts, &sysroot);
+    let hash_kind = config.opts.unstable_opts.src_hash_algorithm(&target);
+
     util::run_in_thread_pool_with_globals(
         config.opts.edition,
         config.opts.unstable_opts.threads,
+        SourceMapInputs(loader, config.opts.file_path_mapping(), hash_kind),
         || {
-            crate::callbacks::setup_callbacks();
-
+            // The previous `early_dcx` can't be reused here because it doesn't
+            // impl `Send`. Creating a new one is fine.
             let early_dcx = EarlyDiagCtxt::new(config.opts.error_format);
 
-            let sysroot = filesearch::materialize_sysroot(config.opts.maybe_sysroot.clone());
-
-            let target = config::build_target_config(&early_dcx, &config.opts, &sysroot);
-
             let codegen_backend = match config.make_codegen_backend {
                 None => util::get_codegen_backend(
                     &early_dcx,
@@ -367,9 +371,7 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
                 config.opts.unstable_opts.translate_directionality_markers,
             ) {
                 Ok(bundle) => bundle,
-                Err(e) => {
-                    early_dcx.early_fatal(format!("failed to load fluent bundle: {e}"));
-                }
+                Err(e) => early_dcx.early_fatal(format!("failed to load fluent bundle: {e}")),
             };
 
             let mut locale_resources = Vec::from(config.locale_resources);
@@ -388,7 +390,6 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
                 config.registry.clone(),
                 locale_resources,
                 config.lint_caps,
-                config.file_loader,
                 target,
                 sysroot,
                 util::rustc_version_str().unwrap_or("unknown"),
@@ -431,45 +432,43 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
             let compiler =
                 Compiler { sess, codegen_backend, override_queries: config.override_queries };
 
-            rustc_span::set_source_map(compiler.sess.psess.clone_source_map(), move || {
-                // There are two paths out of `f`.
-                // - Normal exit.
-                // - Panic, e.g. triggered by `abort_if_errors`.
-                //
-                // We must run `finish_diagnostics` in both cases.
-                let res = {
-                    // If `f` panics, `finish_diagnostics` will run during
-                    // unwinding because of the `defer`.
-                    let mut guar = None;
-                    let sess_abort_guard = defer(|| {
-                        guar = compiler.sess.finish_diagnostics(&config.registry);
-                    });
-
-                    let res = f(&compiler);
-
-                    // If `f` doesn't panic, `finish_diagnostics` will run
-                    // normally when `sess_abort_guard` is dropped.
-                    drop(sess_abort_guard);
-
-                    // If `finish_diagnostics` emits errors (e.g. stashed
-                    // errors) we can't return an error directly, because the
-                    // return type of this function is `R`, not `Result<R, E>`.
-                    // But we need to communicate the errors' existence to the
-                    // caller, otherwise the caller might mistakenly think that
-                    // no errors occurred and return a zero exit code. So we
-                    // abort (panic) instead, similar to if `f` had panicked.
-                    if guar.is_some() {
-                        compiler.sess.dcx().abort_if_errors();
-                    }
+            // There are two paths out of `f`.
+            // - Normal exit.
+            // - Panic, e.g. triggered by `abort_if_errors`.
+            //
+            // We must run `finish_diagnostics` in both cases.
+            let res = {
+                // If `f` panics, `finish_diagnostics` will run during
+                // unwinding because of the `defer`.
+                let mut guar = None;
+                let sess_abort_guard = defer(|| {
+                    guar = compiler.sess.finish_diagnostics(&config.registry);
+                });
+
+                let res = f(&compiler);
+
+                // If `f` doesn't panic, `finish_diagnostics` will run
+                // normally when `sess_abort_guard` is dropped.
+                drop(sess_abort_guard);
+
+                // If `finish_diagnostics` emits errors (e.g. stashed
+                // errors) we can't return an error directly, because the
+                // return type of this function is `R`, not `Result<R, E>`.
+                // But we need to communicate the errors' existence to the
+                // caller, otherwise the caller might mistakenly think that
+                // no errors occurred and return a zero exit code. So we
+                // abort (panic) instead, similar to if `f` had panicked.
+                if guar.is_some() {
+                    compiler.sess.dcx().abort_if_errors();
+                }
 
-                    res
-                };
+                res
+            };
 
-                let prof = compiler.sess.prof.clone();
-                prof.generic_activity("drop_compiler").run(move || drop(compiler));
+            let prof = compiler.sess.prof.clone();
+            prof.generic_activity("drop_compiler").run(move || drop(compiler));
 
-                res
-            })
+            res
         },
     )
 }
diff --git a/compiler/rustc_interface/src/tests.rs b/compiler/rustc_interface/src/tests.rs
@@ -16,6 +16,7 @@ use rustc_session::search_paths::SearchPath;
 use rustc_session::utils::{CanonicalizedPath, NativeLib, NativeLibKind};
 use rustc_session::{build_session, filesearch, getopts, CompilerIO, EarlyDiagCtxt, Session};
 use rustc_span::edition::{Edition, DEFAULT_EDITION};
+use rustc_span::source_map::{RealFileLoader, SourceMapInputs};
 use rustc_span::symbol::sym;
 use rustc_span::{FileName, SourceFileHashAlgorithm};
 use rustc_target::spec::{CodeModel, LinkerFlavorCli, MergeFunctions, PanicStrategy, RelocModel};
@@ -36,15 +37,22 @@ where
     let sessopts = build_session_options(&mut early_dcx, &matches);
     let sysroot = filesearch::materialize_sysroot(sessopts.maybe_sysroot.clone());
     let target = rustc_session::config::build_target_config(&early_dcx, &sessopts, &sysroot);
-
-    rustc_span::create_session_globals_then(DEFAULT_EDITION, || {
+    let hash_kind = sessopts.unstable_opts.src_hash_algorithm(&target);
+    let sm_inputs = Some(SourceMapInputs(
+        Box::new(RealFileLoader) as _,
+        sessopts.file_path_mapping(),
+        hash_kind,
+    ));
+
+    rustc_span::create_session_globals_then(DEFAULT_EDITION, sm_inputs, || {
         let temps_dir = sessopts.unstable_opts.temps_dir.as_deref().map(PathBuf::from);
         let io = CompilerIO {
             input: Input::Str { name: FileName::Custom(String::new()), input: String::new() },
             output_dir: None,
             output_file: None,
             temps_dir,
         };
+
         let sess = build_session(
             early_dcx,
             sessopts,
@@ -53,7 +61,6 @@ where
             registry::Registry::new(&[]),
             vec![],
             Default::default(),
-            None,
             target,
             sysroot,
             "",
diff --git a/compiler/rustc_interface/src/util.rs b/compiler/rustc_interface/src/util.rs
@@ -13,6 +13,7 @@ use rustc_session::lint::{self, BuiltinLintDiag, LintBuffer};
 use rustc_session::{filesearch, Session};
 use rustc_span::edit_distance::find_best_match_for_name;
 use rustc_span::edition::Edition;
+use rustc_span::source_map::SourceMapInputs;
 use rustc_span::symbol::sym;
 use rustc_target::spec::Target;
 use session::output::{categorize_crate_type, CRATE_TYPES};
@@ -56,8 +57,9 @@ fn get_stack_size() -> Option<usize> {
     env::var_os("RUST_MIN_STACK").is_none().then_some(STACK_SIZE)
 }
 
-pub(crate) fn run_in_thread_with_globals<F: FnOnce() -> R + Send, R: Send>(
+fn run_in_thread_with_globals<F: FnOnce() -> R + Send, R: Send>(
     edition: Edition,
+    sm_inputs: SourceMapInputs,
     f: F,
 ) -> R {
     // The "thread pool" is a single spawned thread in the non-parallel
@@ -77,7 +79,9 @@ pub(crate) fn run_in_thread_with_globals<F: FnOnce() -> R + Send, R: Send>(
         // `unwrap` is ok here because `spawn_scoped` only panics if the thread
         // name contains null bytes.
         let r = builder
-            .spawn_scoped(s, move || rustc_span::create_session_globals_then(edition, f))
+            .spawn_scoped(s, move || {
+                rustc_span::create_session_globals_then(edition, Some(sm_inputs), f)
+            })
             .unwrap()
             .join();
 
@@ -92,15 +96,17 @@ pub(crate) fn run_in_thread_with_globals<F: FnOnce() -> R + Send, R: Send>(
 pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce() -> R + Send, R: Send>(
     edition: Edition,
     _threads: usize,
+    sm_inputs: SourceMapInputs,
     f: F,
 ) -> R {
-    run_in_thread_with_globals(edition, f)
+    run_in_thread_with_globals(edition, sm_inputs, f)
 }
 
 #[cfg(parallel_compiler)]
 pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce() -> R + Send, R: Send>(
     edition: Edition,
     threads: usize,
+    sm_inputs: SourceMapInputs,
     f: F,
 ) -> R {
     use rustc_data_structures::{defer, jobserver, sync::FromDyn};
@@ -112,7 +118,7 @@ pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce() -> R + Send, R: Send>(
     let registry = sync::Registry::new(std::num::NonZero::new(threads).unwrap());
 
     if !sync::is_dyn_thread_safe() {
-        return run_in_thread_with_globals(edition, || {
+        return run_in_thread_with_globals(edition, sm_inputs, || {
             // Register the thread for use with the `WorkerLocal` type.
             registry.register();
 
@@ -153,7 +159,7 @@ pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce() -> R + Send, R: Send>(
     // pool. Upon creation, each worker thread created gets a copy of the
     // session globals in TLS. This is possible because `SessionGlobals` impls
     // `Send` in the parallel compiler.
-    rustc_span::create_session_globals_then(edition, || {
+    rustc_span::create_session_globals_then(edition, Some(sm_inputs), || {
         rustc_span::with_session_globals(|session_globals| {
             let session_globals = FromDyn::from(session_globals);
             builder
diff --git a/compiler/rustc_log/src/lib.rs b/compiler/rustc_log/src/lib.rs
@@ -17,7 +17,7 @@
 //!     rustc_log::init_logger(rustc_log::LoggerConfig::from_env("LOG")).unwrap();
 //!
 //!     let edition = rustc_span::edition::Edition::Edition2021;
-//!     rustc_span::create_session_globals_then(edition, || {
+//!     rustc_span::create_session_globals_then(edition, None, || {
 //!         /* ... */
 //!     });
 //! }
diff --git a/compiler/rustc_session/src/config.rs b/compiler/rustc_session/src/config.rs
@@ -1088,7 +1088,7 @@ impl Options {
             || self.unstable_opts.query_dep_graph
     }
 
-    pub(crate) fn file_path_mapping(&self) -> FilePathMapping {
+    pub fn file_path_mapping(&self) -> FilePathMapping {
         file_path_mapping(self.remap_path_prefix.clone(), &self.unstable_opts)
     }
 
@@ -1125,6 +1125,16 @@ impl UnstableOptions {
             track_diagnostics: self.track_diagnostics,
         }
     }
+
+    pub fn src_hash_algorithm(&self, target: &Target) -> SourceFileHashAlgorithm {
+        self.src_hash_algorithm.unwrap_or_else(|| {
+            if target.is_like_msvc {
+                SourceFileHashAlgorithm::Sha256
+            } else {
+                SourceFileHashAlgorithm::Md5
+            }
+        })
+    }
 }
 
 // The type of entry function, so users can have their own entry functions
diff --git a/compiler/rustc_session/src/session.rs b/compiler/rustc_session/src/session.rs
@@ -28,8 +28,8 @@ use rustc_errors::{
 use rustc_macros::HashStable_Generic;
 pub use rustc_span::def_id::StableCrateId;
 use rustc_span::edition::Edition;
-use rustc_span::source_map::{FileLoader, FilePathMapping, RealFileLoader, SourceMap};
-use rustc_span::{SourceFileHashAlgorithm, Span, Symbol};
+use rustc_span::source_map::{FilePathMapping, SourceMap};
+use rustc_span::{Span, Symbol};
 use rustc_target::asm::InlineAsmArch;
 use rustc_target::spec::{CodeModel, PanicStrategy, RelocModel, RelroLevel};
 use rustc_target::spec::{
@@ -1007,7 +1007,6 @@ pub fn build_session(
     registry: rustc_errors::registry::Registry,
     fluent_resources: Vec<&'static str>,
     driver_lint_caps: FxHashMap<lint::LintId, lint::Level>,
-    file_loader: Option<Box<dyn FileLoader + Send + Sync + 'static>>,
     target: Target,
     sysroot: PathBuf,
     cfg_version: &'static str,
@@ -1034,24 +1033,11 @@ pub fn build_session(
         early_dcx.early_warn(warning)
     }
 
-    let loader = file_loader.unwrap_or_else(|| Box::new(RealFileLoader));
-    let hash_kind = sopts.unstable_opts.src_hash_algorithm.unwrap_or_else(|| {
-        if target.is_like_msvc {
-            SourceFileHashAlgorithm::Sha256
-        } else {
-            SourceFileHashAlgorithm::Md5
-        }
-    });
-    let source_map = Lrc::new(SourceMap::with_file_loader_and_hash_kind(
-        loader,
-        sopts.file_path_mapping(),
-        hash_kind,
-    ));
-
     let fallback_bundle = fallback_fluent_bundle(
         fluent_resources,
         sopts.unstable_opts.translate_directionality_markers,
     );
+    let source_map = rustc_span::source_map::get_source_map().unwrap();
     let emitter = default_emitter(&sopts, registry, source_map.clone(), bundle, fallback_bundle);
 
     let mut dcx =
diff --git a/compiler/rustc_span/src/lib.rs b/compiler/rustc_span/src/lib.rs
diff --git a/compiler/rustc_span/src/source_map.rs b/compiler/rustc_span/src/source_map.rs
diff --git a/src/tools/clippy/clippy_lints/src/doc/needless_doctest_main.rs b/src/tools/clippy/clippy_lints/src/doc/needless_doctest_main.rs

Original file line number	Diff line number	Diff line change
`@@ -1088,7 +1088,7 @@ impl Options {`
`1088`	`1088`	`\|\| self.unstable_opts.query_dep_graph`
`1089`	`1089`	`}`
`1090`	`1090`
`1091`		`- pub(crate) fn file_path_mapping(&self) -> FilePathMapping {`
	`1091`	`+ pub fn file_path_mapping(&self) -> FilePathMapping {`
`1092`	`1092`	`file_path_mapping(self.remap_path_prefix.clone(), &self.unstable_opts)`
`1093`	`1093`	`}`
`1094`	`1094`
`@@ -1125,6 +1125,16 @@ impl UnstableOptions {`
`1125`	`1125`	`track_diagnostics: self.track_diagnostics,`
`1126`	`1126`	`}`
`1127`	`1127`	`}`
	`1128`	`+`
	`1129`	`+ pub fn src_hash_algorithm(&self, target: &Target) -> SourceFileHashAlgorithm {`
	`1130`	`+ self.src_hash_algorithm.unwrap_or_else(\|\| {`
	`1131`	`+ if target.is_like_msvc {`
	`1132`	`+ SourceFileHashAlgorithm::Sha256`
	`1133`	`+ } else {`
	`1134`	`+ SourceFileHashAlgorithm::Md5`
	`1135`	`+ }`
	`1136`	`+ })`
	`1137`	`+ }`
`1128`	`1138`	`}`
`1129`	`1139`
`1130`	`1140`	`// The type of entry function, so users can have their own entry functions`