Auto merge of #100803 - klensy:do-not-encode-preinterned-symbols, r=bjorn3

Symbols: do not write string values of preinterned symbols into compiled artifacts

r? `@bjorn3`

Followup for #98851

https://github.com/rust-lang/rust/pull/98851#issuecomment-1215606291

compiler/rustc_macros/src/symbols.rs

@@ -195,10 +195,10 @@ fn symbols_with_errors(input: TokenStream) -> (TokenStream, Vec<syn::Error>) {
             #n,
         });
     }
-    let _ = counter; // for future use
 
     let output = quote! {
         const SYMBOL_DIGITS_BASE: u32 = #digits_base;
+        const PREINTERNED_SYMBOLS_COUNT: u32 = #counter;
 
         #[doc(hidden)]
         #[allow(non_upper_case_globals)]

compiler/rustc_metadata/src/rmeta/decoder.rs

@@ -631,6 +631,10 @@ impl<'a, 'tcx> Decodable<DecodeContext<'a, 'tcx>> for Symbol {
 
                 sym
             }
+            SYMBOL_PREINTERNED => {
+                let symbol_index = d.read_u32();
+                Symbol::new_from_decoded(symbol_index)
+            }
             _ => unreachable!(),
         }
     }

compiler/rustc_metadata/src/rmeta/encoder.rs

@@ -317,17 +317,24 @@ impl<'a, 'tcx> Encodable<EncodeContext<'a, 'tcx>> for Span {
 
 impl<'a, 'tcx> Encodable<EncodeContext<'a, 'tcx>> for Symbol {
     fn encode(&self, s: &mut EncodeContext<'a, 'tcx>) {
-        match s.symbol_table.entry(*self) {
-            Entry::Vacant(o) => {
-                s.opaque.emit_u8(SYMBOL_STR);
-                let pos = s.opaque.position();
-                o.insert(pos);
-                s.emit_str(self.as_str());
-            }
-            Entry::Occupied(o) => {
-                let x = o.get().clone();
-                s.emit_u8(SYMBOL_OFFSET);
-                s.emit_usize(x);
+        // if symbol preinterned, emit tag and symbol index
+        if self.is_preinterned() {
+            s.opaque.emit_u8(SYMBOL_PREINTERNED);
+            s.opaque.emit_u32(self.as_u32());
+        } else {
+            // otherwise write it as string or as offset to it
+            match s.symbol_table.entry(*self) {
+                Entry::Vacant(o) => {
+                    s.opaque.emit_u8(SYMBOL_STR);
+                    let pos = s.opaque.position();
+                    o.insert(pos);
+                    s.emit_str(self.as_str());
+                }
+                Entry::Occupied(o) => {
+                    let x = o.get().clone();
+                    s.emit_u8(SYMBOL_OFFSET);
+                    s.emit_usize(x);
+                }
             }
         }
     }

compiler/rustc_metadata/src/rmeta/mod.rs

@@ -448,6 +448,7 @@ const TAG_PARTIAL_SPAN: u8 = 2;
 // Tags for encoding Symbol's
 const SYMBOL_STR: u8 = 0;
 const SYMBOL_OFFSET: u8 = 1;
+const SYMBOL_PREINTERNED: u8 = 2;
 
 pub fn provide(providers: &mut Providers) {
     encoder::provide(providers);

compiler/rustc_query_impl/src/on_disk_cache.rs

@@ -42,6 +42,7 @@ const TAG_EXPN_DATA: u8 = 1;
 // Tags for encoding Symbol's
 const SYMBOL_STR: u8 = 0;
 const SYMBOL_OFFSET: u8 = 1;
+const SYMBOL_PREINTERNED: u8 = 2;
 
 /// Provides an interface to incremental compilation data cached from the
 /// previous compilation session. This data will eventually include the results
@@ -745,6 +746,10 @@ impl<'a, 'tcx> Decodable<CacheDecoder<'a, 'tcx>> for Symbol {
 
                 sym
             }
+            SYMBOL_PREINTERNED => {
+                let symbol_index = d.read_u32();
+                Symbol::new_from_decoded(symbol_index)
+            }
             _ => unreachable!(),
         }
     }
@@ -939,17 +944,24 @@ impl<'a, 'tcx> Encodable<CacheEncoder<'a, 'tcx>> for Span {
 // copy&paste impl from rustc_metadata
 impl<'a, 'tcx> Encodable<CacheEncoder<'a, 'tcx>> for Symbol {
     fn encode(&self, s: &mut CacheEncoder<'a, 'tcx>) {
-        match s.symbol_table.entry(*self) {
-            Entry::Vacant(o) => {
-                s.encoder.emit_u8(SYMBOL_STR);
-                let pos = s.encoder.position();
-                o.insert(pos);
-                s.emit_str(self.as_str());
-            }
-            Entry::Occupied(o) => {
-                let x = o.get().clone();
-                s.emit_u8(SYMBOL_OFFSET);
-                s.emit_usize(x);
+        // if symbol preinterned, emit tag and symbol index
+        if self.is_preinterned() {
+            s.encoder.emit_u8(SYMBOL_PREINTERNED);
+            s.encoder.emit_u32(self.as_u32());
+        } else {
+            // otherwise write it as string or as offset to it
+            match s.symbol_table.entry(*self) {
+                Entry::Vacant(o) => {
+                    s.encoder.emit_u8(SYMBOL_STR);
+                    let pos = s.encoder.position();
+                    o.insert(pos);
+                    s.emit_str(self.as_str());
+                }
+                Entry::Occupied(o) => {
+                    let x = o.get().clone();
+                    s.emit_u8(SYMBOL_OFFSET);
+                    s.emit_usize(x);
+                }
             }
         }
     }

compiler/rustc_span/src/symbol.rs

@@ -1804,6 +1804,11 @@ impl Symbol {
         Symbol(SymbolIndex::from_u32(n))
     }
 
+    /// for use in Decoder only
+    pub fn new_from_decoded(n: u32) -> Self {
+        Self::new(n)
+    }
+
     /// Maps a string to its interned representation.
     pub fn intern(string: &str) -> Self {
         with_session_globals(|session_globals| session_globals.symbol_interner.intern(string))
@@ -2028,6 +2033,11 @@ impl Symbol {
     pub fn can_be_raw(self) -> bool {
         self != kw::Empty && self != kw::Underscore && !self.is_path_segment_keyword()
     }
+
+    /// Is this symbol was interned in compiler's `symbols!` macro
+    pub fn is_preinterned(self) -> bool {
+        self.as_u32() < PREINTERNED_SYMBOLS_COUNT
+    }
 }
 
 impl Ident {