Auto merge of #98840 - cjgillot:span-inline-ctxt, r=wesleywiser
Inline SyntaxContext in both encoded span representation. The current interned representation for spans does not use the `ctxt_or_zero: u16` field. This PR proposes to use this field to store the `SyntaxContext` of the interned span instead. When `ctxt_or_zero` and the interned span's `ctxt` don't match, the inlined one takes precedence. This allows to implement `Span::ctxt` and `Span::with_ctxt` with much less probability to access the interner. Those functions are used a lot for hygiene, so this may be worth it.
This commit is contained in:
commit
e7119a0300
@ -533,9 +533,6 @@ pub fn with_hi(self, hi: BytePos) -> Span {
|
||||
self.data().with_hi(hi)
|
||||
}
|
||||
#[inline]
|
||||
pub fn ctxt(self) -> SyntaxContext {
|
||||
self.data_untracked().ctxt
|
||||
}
|
||||
pub fn eq_ctxt(self, other: Span) -> bool {
|
||||
self.data_untracked().ctxt == other.data_untracked().ctxt
|
||||
}
|
||||
|
@ -28,12 +28,17 @@
|
||||
/// Inline (compressed) format:
|
||||
/// - `span.base_or_index == span_data.lo`
|
||||
/// - `span.len_or_tag == len == span_data.hi - span_data.lo` (must be `<= MAX_LEN`)
|
||||
/// - `span.ctxt == span_data.ctxt` (must be `<= MAX_CTXT`)
|
||||
/// - `span.ctxt_or_tag == span_data.ctxt` (must be `<= MAX_CTXT`)
|
||||
///
|
||||
/// Interned format with inline `SyntaxContext`:
|
||||
/// - `span.base_or_index == index` (indexes into the interner table)
|
||||
/// - `span.len_or_tag == LEN_TAG` (high bit set, all other bits are zero)
|
||||
/// - `span.ctxt_or_tag == span_data.ctxt` (must be `<= MAX_CTXT`)
|
||||
///
|
||||
/// Interned format:
|
||||
/// - `span.base_or_index == index` (indexes into the interner table)
|
||||
/// - `span.len_or_tag == LEN_TAG` (high bit set, all other bits are zero)
|
||||
/// - `span.ctxt == 0`
|
||||
/// - `span.ctxt_or_tag == CTXT_TAG`
|
||||
///
|
||||
/// The inline form uses 0 for the tag value (rather than 1) so that we don't
|
||||
/// need to mask out the tag bit when getting the length, and so that the
|
||||
@ -50,10 +55,10 @@
|
||||
/// at 3 or 4, and then it drops off quickly from 8 onwards. 15 bits is enough
|
||||
/// for 99.99%+ of cases, but larger values (sometimes 20+ bits) might occur
|
||||
/// dozens of times in a typical crate.
|
||||
/// - `ctxt` is 16 bits in `Span` and 32 bits in `SpanData`, which means that
|
||||
/// - `ctxt_or_tag` is 16 bits in `Span` and 32 bits in `SpanData`, which means that
|
||||
/// large `ctxt` values will cause interning. The number of bits needed for
|
||||
/// `ctxt` values depend partly on the crate size and partly on the form of
|
||||
/// the code. No crates in `rustc-perf` need more than 15 bits for `ctxt`,
|
||||
/// the code. No crates in `rustc-perf` need more than 15 bits for `ctxt_or_tag`,
|
||||
/// but larger crates might need more than 16 bits.
|
||||
///
|
||||
/// In order to reliably use parented spans in incremental compilation,
|
||||
@ -65,15 +70,16 @@
|
||||
pub struct Span {
|
||||
base_or_index: u32,
|
||||
len_or_tag: u16,
|
||||
ctxt_or_zero: u16,
|
||||
ctxt_or_tag: u16,
|
||||
}
|
||||
|
||||
const LEN_TAG: u16 = 0b1000_0000_0000_0000;
|
||||
const MAX_LEN: u32 = 0b0111_1111_1111_1111;
|
||||
const MAX_CTXT: u32 = 0b1111_1111_1111_1111;
|
||||
const CTXT_TAG: u32 = 0b1111_1111_1111_1111;
|
||||
const MAX_CTXT: u32 = CTXT_TAG - 1;
|
||||
|
||||
/// Dummy span, both position and length are zero, syntax context is zero as well.
|
||||
pub const DUMMY_SP: Span = Span { base_or_index: 0, len_or_tag: 0, ctxt_or_zero: 0 };
|
||||
pub const DUMMY_SP: Span = Span { base_or_index: 0, len_or_tag: 0, ctxt_or_tag: 0 };
|
||||
|
||||
impl Span {
|
||||
#[inline]
|
||||
@ -91,12 +97,13 @@ pub fn new(
|
||||
|
||||
if len <= MAX_LEN && ctxt2 <= MAX_CTXT && parent.is_none() {
|
||||
// Inline format.
|
||||
Span { base_or_index: base, len_or_tag: len as u16, ctxt_or_zero: ctxt2 as u16 }
|
||||
Span { base_or_index: base, len_or_tag: len as u16, ctxt_or_tag: ctxt2 as u16 }
|
||||
} else {
|
||||
// Interned format.
|
||||
let index =
|
||||
with_span_interner(|interner| interner.intern(&SpanData { lo, hi, ctxt, parent }));
|
||||
Span { base_or_index: index, len_or_tag: LEN_TAG, ctxt_or_zero: 0 }
|
||||
let ctxt_or_tag = if ctxt2 <= MAX_CTXT { ctxt2 } else { CTXT_TAG } as u16;
|
||||
Span { base_or_index: index, len_or_tag: LEN_TAG, ctxt_or_tag }
|
||||
}
|
||||
}
|
||||
|
||||
@ -119,16 +126,29 @@ pub fn data_untracked(self) -> SpanData {
|
||||
SpanData {
|
||||
lo: BytePos(self.base_or_index),
|
||||
hi: BytePos(self.base_or_index + self.len_or_tag as u32),
|
||||
ctxt: SyntaxContext::from_u32(self.ctxt_or_zero as u32),
|
||||
ctxt: SyntaxContext::from_u32(self.ctxt_or_tag as u32),
|
||||
parent: None,
|
||||
}
|
||||
} else {
|
||||
// Interned format.
|
||||
debug_assert!(self.ctxt_or_zero == 0);
|
||||
let index = self.base_or_index;
|
||||
with_span_interner(|interner| interner.spans[index as usize])
|
||||
}
|
||||
}
|
||||
|
||||
/// This function is used as a fast path when decoding the full `SpanData` is not necessary.
|
||||
#[inline]
|
||||
pub fn ctxt(self) -> SyntaxContext {
|
||||
let ctxt_or_tag = self.ctxt_or_tag as u32;
|
||||
if ctxt_or_tag <= MAX_CTXT {
|
||||
// Inline format or interned format with inline ctxt.
|
||||
SyntaxContext::from_u32(ctxt_or_tag)
|
||||
} else {
|
||||
// Interned format.
|
||||
let index = self.base_or_index;
|
||||
with_span_interner(|interner| interner.spans[index as usize].ctxt)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(Default)]
|
||||
|
Loading…
Reference in New Issue
Block a user