add flag to split sourcemap and span encoding out of main rmeta file

Author: yaahc

compiler/rustc_metadata/src/rmeta/decoder.rs

@@ -546,6 +546,10 @@ impl<'a, 'tcx> SpanDecoder for DecodeContext<'a, 'tcx> {
     }
 
     fn decode_span(&mut self) -> Span {
+        if !self.cdata().has_rmeta_extras() {
+            return DUMMY_SP;
+        }
+
         let start = self.position();
         let tag = SpanTag(self.peek_byte());
         let data = if tag.kind() == SpanKind::Indirect {
@@ -1628,8 +1632,8 @@ impl<'a> CrateMetadataRef<'a> {
             path.filter(|_| {
                 // Only spend time on further checks if we have what to translate *to*.
                 real_source_base_dir.is_some()
-                // Some tests need the translation to be always skipped.
-                && sess.opts.unstable_opts.translate_remapped_path_to_local_path
+                    // Some tests need the translation to be always skipped.
+                    && sess.opts.unstable_opts.translate_remapped_path_to_local_path
             })
             .filter(|virtual_dir| {
                 // Don't translate away `/rustc/$hash` if we're still remapping to it,
@@ -1999,6 +2003,10 @@ impl CrateMetadata {
         self.root.has_default_lib_allocator
     }
 
+    pub(crate) fn has_rmeta_extras(&self) -> bool {
+        self.root.has_rmeta_extras
+    }
+
     pub(crate) fn is_proc_macro_crate(&self) -> bool {
         self.root.is_proc_macro_crate()
     }

compiler/rustc_metadata/src/rmeta/encoder.rs

@@ -168,6 +168,10 @@ impl<'a, 'tcx> SpanEncoder for EncodeContext<'a, 'tcx> {
     }
 
     fn encode_span(&mut self, span: Span) {
+        if self.tcx.sess.is_split_rmeta_enabled() {
+            return;
+        }
+
         match self.span_shorthands.entry(span) {
             Entry::Occupied(o) => {
                 // If an offset is smaller than the absolute position, we encode with the offset.
@@ -436,7 +440,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
                 assert!(
                     last_pos <= position,
                     "make sure that the calls to `lazy*` \
-                     are in the same order as the metadata fields",
+                    are in the same order as the metadata fields",
                 );
                 position.get() - last_pos.get()
             }
@@ -603,6 +607,13 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
         adapted.encode(&mut self.opaque)
     }
 
+    fn dont_encode_source_map(
+        &mut self,
+    ) -> LazyTable<u32, Option<LazyValue<rustc_span::SourceFile>>> {
+        let adapted = TableBuilder::default();
+        adapted.encode(&mut self.opaque)
+    }
+
     fn encode_crate_root(&mut self) -> LazyValue<CrateRoot> {
         let tcx = self.tcx;
         let mut stats: Vec<(&'static str, usize)> = Vec::with_capacity(32);
@@ -708,7 +719,11 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
 
         // Encode source_map. This needs to be done last, because encoding `Span`s tells us which
         // `SourceFiles` we actually need to encode.
-        let source_map = stat!("source-map", || self.encode_source_map());
+        let source_map = if tcx.sess.is_split_rmeta_enabled() {
+            stat!("source-map", || self.dont_encode_source_map())
+        } else {
+            stat!("source-map", || self.encode_source_map())
+        };
         let target_modifiers = stat!("target-modifiers", || self.encode_target_modifiers());
 
         let root = stat!("final", || {
@@ -733,6 +748,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
                     attrs,
                     sym::default_lib_allocator,
                 ),
+                has_rmeta_extras: !tcx.sess.is_split_rmeta_enabled(),
                 proc_macro_data,
                 debugger_visualizers,
                 compiler_builtins: ast::attr::contains_name(attrs, sym::compiler_builtins),
@@ -1706,7 +1722,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
 
             record_array!(self.tables.module_children_non_reexports[def_id] <-
                 module_children.iter().filter(|child| child.reexport_chain.is_empty())
-                    .map(|child| child.res.def_id().index));
+                .map(|child| child.res.def_id().index));
 
             record_defaulted_array!(self.tables.module_children_reexports[def_id] <-
                 module_children.iter().filter(|child| !child.reexport_chain.is_empty()));
@@ -1755,7 +1771,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
             if matches!(rpitit_info, ty::ImplTraitInTraitData::Trait { .. }) {
                 record_array!(
                     self.tables.assumed_wf_types_for_rpitit[def_id]
-                        <- self.tcx.assumed_wf_types_for_rpitit(def_id)
+                    <- self.tcx.assumed_wf_types_for_rpitit(def_id)
                 );
                 self.encode_precise_capturing_args(def_id);
             }
@@ -1838,7 +1854,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
             for &local_def_id in tcx.mir_keys(()) {
                 if let DefKind::AssocFn | DefKind::Fn = tcx.def_kind(local_def_id) {
                     record_array!(self.tables.deduced_param_attrs[local_def_id.to_def_id()] <-
-                        self.tcx.deduced_param_attrs(local_def_id.to_def_id()));
+                                self.tcx.deduced_param_attrs(local_def_id.to_def_id()));
                 }
             }
         }

compiler/rustc_metadata/src/rmeta/mod.rs

@@ -260,6 +260,7 @@ pub(crate) struct CrateRoot {
     has_alloc_error_handler: bool,
     has_panic_handler: bool,
     has_default_lib_allocator: bool,
+    has_rmeta_extras: bool,
 
     crate_deps: LazyArray<CrateDep>,
     dylib_dependency_formats: LazyArray<Option<LinkagePreference>>,

compiler/rustc_session/src/options.rs

@@ -2537,6 +2537,8 @@ written to standard error output)"),
                   by the linker"),
     split_lto_unit: Option<bool> = (None, parse_opt_bool, [TRACKED],
         "enable LTO unit splitting (default: no)"),
+    split_rmeta: Option<bool> = (None, parse_opt_bool, [TRACKED],
+        "split extra rmeta data that isn't relevant rlib ABI into a separate file"),
     src_hash_algorithm: Option<SourceFileHashAlgorithm> = (None, parse_src_file_hash, [TRACKED],
         "hash algorithm of source files in debug info (`md5`, `sha1`, or `sha256`)"),
     #[rustc_lint_opt_deny_field_access("use `Session::stack_protector` instead of this field")]

compiler/rustc_session/src/session.rs

@@ -400,6 +400,10 @@ impl Session {
         self.opts.unstable_opts.split_lto_unit == Some(true)
     }
 
+    pub fn is_split_rmeta_enabled(&self) -> bool {
+        self.opts.unstable_opts.split_rmeta == Some(true)
+    }
+
     /// Check whether this compile session and crate type use static crt.
     pub fn crt_static(&self, crate_type: Option<CrateType>) -> bool {
         if !self.target.crt_static_respected {

tests/run-make/rdr-ignores-comments/rmake.rs

@@ -1,21 +1,23 @@
-// tests that changing the content of a comment will cause a change in the rmeta of a file
-use run_make_support::{rfs, rustc};
+// tests that changing the content of a comment will not cause a change in the rmeta of a library if
+// -Zsplit-rmeta is enabled
 use std::hash::{Hash, Hasher};
 use std::path::Path;
 
+use run_make_support::{rfs, rustc};
+
 fn main() {
     let before = check_and_hash("before.rs");
     let after = check_and_hash("after.rs");
     dbg!(before, after);
-    assert_neq!(before, after);
+    assert_eq!(before, after);
 }
 
 fn check_and_hash<P>(filename: P) -> u64
 where
     P: AsRef<Path>,
 {
     rfs::rename(filename, "foo.rs");
-    rustc().input("foo.rs").emit("metadata").run();
+    rustc().input("foo.rs").emit("metadata").arg("-Zsplit-rmeta").run();
     // hash the output
     let bytes = rfs::read("libfoo.rmeta");
     let mut hasher = std::hash::DefaultHasher::new();