diff options
| author | Nicholas Nethercote <n.nethercote@gmail.com> | 2025-06-02 08:59:29 +1000 |
|---|---|---|
| committer | Nicholas Nethercote <n.nethercote@gmail.com> | 2025-06-30 20:42:27 +1000 |
| commit | 478f8287c0e2c35cda511fd3ac01b7ac78ee7cfe (patch) | |
| tree | 4d8f19b4e4e440ed8a22ee809ce2a565707d4c27 /compiler/rustc_metadata/src/rmeta/decoder.rs | |
| parent | ed2d759783dc9de134bbb3f01085b1e6dbf539f3 (diff) | |
Introduce `ByteSymbol`.
It's like `Symbol` but for byte strings. The interner is now used for both `Symbol` and `ByteSymbol`. E.g. if you intern `"dog"` and `b"dog"` you'll get a `Symbol` and a `ByteSymbol` with the same index and the characters will only be stored once. The motivation for this is to eliminate the `Arc`s in `ast::LitKind`, to make `ast::LitKind` impl `Copy`, and to avoid the need to arena-allocate `ast::LitKind` in HIR. The latter change reduces peak memory by a non-trivial amount on literal-heavy benchmarks such as `deep-vector` and `tuple-stress`. `Encoder`, `Decoder`, `SpanEncoder`, and `SpanDecoder` all get some changes so that they can handle normal strings and byte strings. This change does slow down compilation of programs that use `include_bytes!` on large files, because the contents of those files are now interned (hashed). This makes `include_bytes!` more similar to `include_str!`, though `include_bytes!` contents still aren't escaped, and hashing is still much cheaper than escaping.
Diffstat (limited to 'compiler/rustc_metadata/src/rmeta/decoder.rs')
| -rw-r--r-- | compiler/rustc_metadata/src/rmeta/decoder.rs | 60 |
1 files changed, 37 insertions, 23 deletions
diff --git a/compiler/rustc_metadata/src/rmeta/decoder.rs b/compiler/rustc_metadata/src/rmeta/decoder.rs index 2696c47c62c..065c261c194 100644 --- a/compiler/rustc_metadata/src/rmeta/decoder.rs +++ b/compiler/rustc_metadata/src/rmeta/decoder.rs @@ -32,7 +32,9 @@ use rustc_session::Session; use rustc_session::config::TargetModifier; use rustc_session::cstore::{CrateSource, ExternCrate}; use rustc_span::hygiene::HygieneDecodeContext; -use rustc_span::{BytePos, DUMMY_SP, Pos, SpanData, SpanDecoder, SyntaxContext, kw}; +use rustc_span::{ + BytePos, ByteSymbol, DUMMY_SP, Pos, SpanData, SpanDecoder, Symbol, SyntaxContext, kw, +}; use tracing::debug; use crate::creader::CStore; @@ -384,6 +386,28 @@ impl<'a, 'tcx> DecodeContext<'a, 'tcx> { fn read_raw_bytes(&mut self, len: usize) -> &[u8] { self.opaque.read_raw_bytes(len) } + + fn decode_symbol_or_byte_symbol<S>( + &mut self, + new_from_index: impl Fn(u32) -> S, + read_and_intern_str_or_byte_str_this: impl Fn(&mut Self) -> S, + read_and_intern_str_or_byte_str_opaque: impl Fn(&mut MemDecoder<'a>) -> S, + ) -> S { + let tag = self.read_u8(); + + match tag { + SYMBOL_STR => read_and_intern_str_or_byte_str_this(self), + SYMBOL_OFFSET => { + // read str offset + let pos = self.read_usize(); + + // move to str offset and read + self.opaque.with_position(pos, |d| read_and_intern_str_or_byte_str_opaque(d)) + } + SYMBOL_PREDEFINED => new_from_index(self.read_u32()), + _ => unreachable!(), + } + } } impl<'a, 'tcx> TyDecoder<'tcx> for DecodeContext<'a, 'tcx> { @@ -545,29 +569,19 @@ impl<'a, 'tcx> SpanDecoder for DecodeContext<'a, 'tcx> { } fn decode_symbol(&mut self) -> Symbol { - let tag = self.read_u8(); - - match tag { - SYMBOL_STR => { - let s = self.read_str(); - Symbol::intern(s) - } - SYMBOL_OFFSET => { - // read str offset - let pos = self.read_usize(); + self.decode_symbol_or_byte_symbol( + Symbol::new, + |this| Symbol::intern(this.read_str()), + |opaque| Symbol::intern(opaque.read_str()), + ) + } - // move to str offset and read - self.opaque.with_position(pos, |d| { - let s = d.read_str(); - Symbol::intern(s) - }) - } - SYMBOL_PREDEFINED => { - let symbol_index = self.read_u32(); - Symbol::new(symbol_index) - } - _ => unreachable!(), - } + fn decode_byte_symbol(&mut self) -> ByteSymbol { + self.decode_symbol_or_byte_symbol( + ByteSymbol::new, + |this| ByteSymbol::intern(this.read_byte_str()), + |opaque| ByteSymbol::intern(opaque.read_byte_str()), + ) } } |
