Rollup merge of #122811 - nnethercote:mv-SourceMap-init, r=WaffleLapkin

Move `SourceMap` initialization

So it happens at the same time as `SessionGlobals` initialization, rather than shortly afterward.

r? `@WaffleLapkin`

Author: GuillaumeGomez

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()], "");
 

compiler/rustc_interface/src/interface.rs

@@ -5,6 +5,7 @@ use rustc_ast::{LitKind, MetaItemKind};
 use rustc_codegen_ssa::traits::CodegenBackend;
 use rustc_data_structures::defer;
 use rustc_data_structures::fx::{FxHashMap, FxHashSet};
+use rustc_data_structures::jobserver;
 use rustc_data_structures::stable_hasher::StableHasher;
 use rustc_data_structures::sync::Lrc;
 use rustc_errors::registry::Registry;
@@ -21,7 +22,7 @@ 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;
@@ -323,6 +324,18 @@ pub struct Config {
     pub expanded_args: Vec<String>,
 }
 
+/// Initialize jobserver before getting `jobserver::client` and `build_session`.
+pub(crate) fn initialize_checked_jobserver(early_dcx: &EarlyDiagCtxt) {
+    jobserver::initialize_checked(|err| {
+        #[allow(rustc::untranslatable_diagnostic)]
+        #[allow(rustc::diagnostic_outside_of_impl)]
+        early_dcx
+            .early_struct_warn(err)
+            .with_note("the build environment is likely misconfigured")
+            .emit()
+    });
+}
+
 // JUSTIFICATION: before session exists, only config
 #[allow(rustc::bad_opt_access)]
 #[allow(rustc::untranslatable_diagnostic)] // FIXME: make this translatable
@@ -334,20 +347,25 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
 
     // Check jobserver before run_in_thread_pool_with_globals, which call jobserver::acquire_thread
     let early_dcx = EarlyDiagCtxt::new(config.opts.error_format);
-    early_dcx.initialize_checked_jobserver();
+    initialize_checked_jobserver(&early_dcx);
+
+    crate::callbacks::setup_callbacks();
+
+    let sysroot = filesearch::materialize_sysroot(config.opts.maybe_sysroot.clone());
+    let target = config::build_target_config(&early_dcx, &config.opts, &sysroot);
+    let file_loader = config.file_loader.unwrap_or_else(|| Box::new(RealFileLoader));
+    let path_mapping = config.opts.file_path_mapping();
+    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 { file_loader, path_mapping, hash_kind },
         |current_gcx| {
-            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,
@@ -372,9 +390,7 @@ 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);
@@ -393,7 +409,6 @@ 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"),
@@ -440,45 +455,43 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
                 current_gcx,
             };
 
-            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
         },
     )
 }

compiler/rustc_interface/src/tests.rs

@@ -1,5 +1,5 @@
 #![allow(rustc::bad_opt_access)]
-use crate::interface::parse_cfg;
+use crate::interface::{initialize_checked_jobserver, parse_cfg};
 use rustc_data_structures::profiling::TimePassesFormat;
 use rustc_errors::{emitter::HumanReadableErrorType, registry, ColorConfig};
 use rustc_session::config::{
@@ -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};
@@ -25,42 +26,52 @@ use std::num::NonZero;
 use std::path::{Path, PathBuf};
 use std::sync::Arc;
 
-fn mk_session(matches: getopts::Matches) -> (Session, Cfg) {
+fn sess_and_cfg<F>(args: &[&'static str], f: F)
+where
+    F: FnOnce(Session, Cfg),
+{
     let mut early_dcx = EarlyDiagCtxt::new(ErrorOutputType::default());
-    early_dcx.initialize_checked_jobserver();
+    initialize_checked_jobserver(&early_dcx);
 
-    let registry = registry::Registry::new(&[]);
+    let matches = optgroups().parse(args).unwrap();
     let sessopts = build_session_options(&mut early_dcx, &matches);
-    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 sysroot = filesearch::materialize_sysroot(sessopts.maybe_sysroot.clone());
-
     let target = rustc_session::config::build_target_config(&early_dcx, &sessopts, &sysroot);
+    let hash_kind = sessopts.unstable_opts.src_hash_algorithm(&target);
+    let sm_inputs = Some(SourceMapInputs {
+        file_loader: Box::new(RealFileLoader) as _,
+        path_mapping: sessopts.file_path_mapping(),
+        hash_kind,
+    });
 
-    let sess = build_session(
-        early_dcx,
-        sessopts,
-        io,
-        None,
-        registry,
-        vec![],
-        Default::default(),
-        None,
-        target,
-        sysroot,
-        "",
-        None,
-        Arc::default(),
-        Default::default(),
-    );
-    let cfg = parse_cfg(&sess.dcx(), matches.opt_strs("cfg"));
-    (sess, cfg)
+    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,
+            io,
+            None,
+            registry::Registry::new(&[]),
+            vec![],
+            Default::default(),
+            target,
+            sysroot,
+            "",
+            None,
+            Arc::default(),
+            Default::default(),
+        );
+        let cfg = parse_cfg(&sess.dcx(), matches.opt_strs("cfg"));
+        let cfg = build_configuration(&sess, cfg);
+        f(sess, cfg)
+    });
 }
 
 fn new_public_extern_entry<S, I>(locations: I) -> ExternEntry
@@ -125,21 +136,15 @@ fn assert_non_crate_hash_different(x: &Options, y: &Options) {
 // When the user supplies --test we should implicitly supply --cfg test
 #[test]
 fn test_switch_implies_cfg_test() {
-    rustc_span::create_default_session_globals_then(|| {
-        let matches = optgroups().parse(&["--test".to_string()]).unwrap();
-        let (sess, cfg) = mk_session(matches);
-        let cfg = build_configuration(&sess, cfg);
+    sess_and_cfg(&["--test"], |_sess, cfg| {
         assert!(cfg.contains(&(sym::test, None)));
-    });
+    })
 }
 
 // When the user supplies --test and --cfg test, don't implicitly add another --cfg test
 #[test]
 fn test_switch_implies_cfg_test_unless_cfg_test() {
-    rustc_span::create_default_session_globals_then(|| {
-        let matches = optgroups().parse(&["--test".to_string(), "--cfg=test".to_string()]).unwrap();
-        let (sess, cfg) = mk_session(matches);
-        let cfg = build_configuration(&sess, cfg);
+    sess_and_cfg(&["--test", "--cfg=test"], |_sess, cfg| {
         let mut test_items = cfg.iter().filter(|&&(name, _)| name == sym::test);
         assert!(test_items.next().is_some());
         assert!(test_items.next().is_none());
@@ -148,22 +153,15 @@ fn test_switch_implies_cfg_test_unless_cfg_test() {
 
 #[test]
 fn test_can_print_warnings() {
-    rustc_span::create_default_session_globals_then(|| {
-        let matches = optgroups().parse(&["-Awarnings".to_string()]).unwrap();
-        let (sess, _) = mk_session(matches);
+    sess_and_cfg(&["-Awarnings"], |sess, _cfg| {
         assert!(!sess.dcx().can_emit_warnings());
     });
 
-    rustc_span::create_default_session_globals_then(|| {
-        let matches =
-            optgroups().parse(&["-Awarnings".to_string(), "-Dwarnings".to_string()]).unwrap();
-        let (sess, _) = mk_session(matches);
+    sess_and_cfg(&["-Awarnings", "-Dwarnings"], |sess, _cfg| {
         assert!(sess.dcx().can_emit_warnings());
     });
 
-    rustc_span::create_default_session_globals_then(|| {
-        let matches = optgroups().parse(&["-Adead_code".to_string()]).unwrap();
-        let (sess, _) = mk_session(matches);
+    sess_and_cfg(&["-Adead_code"], |sess, _cfg| {
         assert!(sess.dcx().can_emit_warnings());
     });
 }

compiler/rustc_interface/src/util.rs

@@ -14,6 +14,7 @@ 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};
@@ -65,8 +66,9 @@ fn init_stack_size() -> usize {
     })
 }
 
-pub(crate) fn run_in_thread_with_globals<F: FnOnce(CurrentGcx) -> R + Send, R: Send>(
+fn run_in_thread_with_globals<F: FnOnce(CurrentGcx) -> R + Send, R: Send>(
     edition: Edition,
+    sm_inputs: SourceMapInputs,
     f: F,
 ) -> R {
     // The "thread pool" is a single spawned thread in the non-parallel
@@ -84,7 +86,9 @@ pub(crate) fn run_in_thread_with_globals<F: FnOnce(CurrentGcx) -> R + Send, R: S
         // name contains null bytes.
         let r = builder
             .spawn_scoped(s, move || {
-                rustc_span::create_session_globals_then(edition, || f(CurrentGcx::new()))
+                rustc_span::create_session_globals_then(edition, Some(sm_inputs), || {
+                    f(CurrentGcx::new())
+                })
             })
             .unwrap()
             .join();
@@ -100,15 +104,17 @@ pub(crate) fn run_in_thread_with_globals<F: FnOnce(CurrentGcx) -> R + Send, R: S
 pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce(CurrentGcx) -> 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(CurrentGcx) -> R + Send, R: Send>(
     edition: Edition,
     threads: usize,
+    sm_inputs: SourceMapInputs,
     f: F,
 ) -> R {
     use rustc_data_structures::{defer, jobserver, sync::FromDyn};
@@ -120,7 +126,7 @@ pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce(CurrentGcx) -> 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, |current_gcx| {
+        return run_in_thread_with_globals(edition, sm_inputs, |current_gcx| {
             // Register the thread for use with the `WorkerLocal` type.
             registry.register();
 
@@ -169,7 +175,7 @@ pub(crate) fn run_in_thread_pool_with_globals<F: FnOnce(CurrentGcx) -> 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