Rollup merge of #80358 - pierwill:edit_rustc_span, r=lcnr

Edit rustc_span documentation

Various changes to the `rustc_span` docs, including the following:

- Additions to top-level docs
- Edits to the source_map module docs
- Edits to documentation for `Span` and `SpanData`
- Added intra-docs links
- Documentation for Levenshtein distances
- Fixed missing punctuation

compiler/rustc_span/src/edition.rs

@@ -4,24 +4,25 @@ use std::str::FromStr;
 
 use rustc_macros::HashStable_Generic;
 
-/// The edition of the compiler (RFC 2052)
+/// The edition of the compiler. (See [RFC 2052](https://github.com/rust-lang/rfcs/blob/master/text/2052-epochs.md).)
 #[derive(Clone, Copy, Hash, PartialEq, PartialOrd, Debug, Encodable, Decodable, Eq)]
 #[derive(HashStable_Generic)]
 pub enum Edition {
-    // editions must be kept in order, oldest to newest
+    // When adding new editions, be sure to do the following:
+    //
+    // - update the `ALL_EDITIONS` const
+    // - update the `EDITION_NAME_LIST` const
+    // - add a `rust_####()` function to the session
+    // - update the enum in Cargo's sources as well
+    //
+    // Editions *must* be kept in order, oldest to newest.
     /// The 2015 edition
     Edition2015,
     /// The 2018 edition
     Edition2018,
-    // when adding new editions, be sure to update:
-    //
-    // - Update the `ALL_EDITIONS` const
-    // - Update the EDITION_NAME_LIST const
-    // - add a `rust_####()` function to the session
-    // - update the enum in Cargo's sources as well
 }
 
-// must be in order from oldest to newest
+// Must be in order from oldest to newest.
 pub const ALL_EDITIONS: &[Edition] = &[Edition::Edition2015, Edition::Edition2018];
 
 pub const EDITION_NAME_LIST: &str = "2015|2018";

compiler/rustc_span/src/lev_distance.rs

@@ -1,10 +1,16 @@
+//! Levenshtein distances.
+//!
+//! The [Levenshtein distance] is a metric for measuring the difference between two strings.
+//!
+//! [Levenshtein distance]: https://en.wikipedia.org/wiki/Levenshtein_distance
+
 use crate::symbol::Symbol;
 use std::cmp;
 
 #[cfg(test)]
 mod tests;
 
-/// Finds the Levenshtein distance between two strings
+/// Finds the Levenshtein distance between two strings.
 pub fn lev_distance(a: &str, b: &str) -> usize {
     // cases which don't require further computation
     if a.is_empty() {
@@ -35,14 +41,14 @@ pub fn lev_distance(a: &str, b: &str) -> usize {
     dcol[t_last + 1]
 }
 
-/// Finds the best match for a given word in the given iterator
+/// Finds the best match for a given word in the given iterator.
 ///
 /// As a loose rule to avoid the obviously incorrect suggestions, it takes
 /// an optional limit for the maximum allowable edit distance, which defaults
 /// to one-third of the given word.
 ///
-/// Besides Levenshtein, we use case insensitive comparison to improve accuracy on an edge case with
-/// a lower(upper)case letters mismatch.
+/// Besides Levenshtein, we use case insensitive comparison to improve accuracy
+/// on an edge case with a lower(upper)case letters mismatch.
 #[cold]
 pub fn find_best_match_for_name(
     name_vec: &[Symbol],
@@ -98,7 +104,7 @@ fn find_match_by_sorted_words(iter_names: &[Symbol], lookup: &str) -> Option<Sym
 
 fn sort_by_words(name: &str) -> String {
     let mut split_words: Vec<&str> = name.split('_').collect();
-    // We are sorting primitive &strs and can use unstable sort here
+    // We are sorting primitive &strs and can use unstable sort here.
     split_words.sort_unstable();
     split_words.join("_")
 }

compiler/rustc_span/src/lib.rs

@@ -1,4 +1,13 @@
-//! The source positions and related helper functions.
+//! Source positions and related helper functions.
+//!
+//! Important concepts in this module include:
+//!
+//! - the *span*, represented by [`SpanData`] and related types;
+//! - source code as represented by a [`SourceMap`]; and
+//! - interned strings, represented by [`Symbol`]s, with some common symbols available statically in the [`sym`] module.
+//!
+//! Unlike most compilers, the span contains not only the position in the source code, but also various other metadata,
+//! such as the edition and macro hygiene. This metadata is stored in [`SyntaxContext`] and [`ExpnData`].
 //!
 //! ## Note
 //!
@@ -124,7 +133,7 @@ pub enum RealFileName {
 
 impl RealFileName {
     /// Returns the path suitable for reading from the file system on the local host.
-    /// Avoid embedding this in build artifacts; see `stable_name` for that.
+    /// Avoid embedding this in build artifacts; see `stable_name()` for that.
     pub fn local_path(&self) -> &Path {
         match self {
             RealFileName::Named(p)
@@ -133,7 +142,7 @@ impl RealFileName {
     }
 
     /// Returns the path suitable for reading from the file system on the local host.
-    /// Avoid embedding this in build artifacts; see `stable_name` for that.
+    /// Avoid embedding this in build artifacts; see `stable_name()` for that.
     pub fn into_local_path(self) -> PathBuf {
         match self {
             RealFileName::Named(p)
@@ -143,7 +152,7 @@ impl RealFileName {
 
     /// Returns the path suitable for embedding into build artifacts. Note that
     /// a virtualized path will not correspond to a valid file system path; see
-    /// `local_path` for something that is more likely to return paths into the
+    /// `local_path()` for something that is more likely to return paths into the
     /// local host file system.
     pub fn stable_name(&self) -> &Path {
         match self {
@@ -173,7 +182,7 @@ pub enum FileName {
     /// Custom sources for explicit parser calls from plugins and drivers.
     Custom(String),
     DocTest(PathBuf, isize),
-    /// Post-substitution inline assembly from LLVM
+    /// Post-substitution inline assembly from LLVM.
     InlineAsm(u64),
 }
 
@@ -266,14 +275,17 @@ impl FileName {
     }
 }
 
+/// Represents a span.
+///
 /// Spans represent a region of code, used for error reporting. Positions in spans
-/// are *absolute* positions from the beginning of the source_map, not positions
-/// relative to `SourceFile`s. Methods on the `SourceMap` can be used to relate spans back
+/// are *absolute* positions from the beginning of the [`SourceMap`], not positions
+/// relative to [`SourceFile`]s. Methods on the `SourceMap` can be used to relate spans back
 /// to the original source.
-/// You must be careful if the span crosses more than one file - you will not be
+///
+/// You must be careful if the span crosses more than one file, since you will not be
 /// able to use many of the functions on spans in source_map and you cannot assume
-/// that the length of the `span = hi - lo`; there may be space in the `BytePos`
-/// range between files.
+/// that the length of the span is equal to `span.hi - span.lo`; there may be space in the
+/// [`BytePos`] range between files.
 ///
 /// `SpanData` is public because `Span` uses a thread-local interner and can't be
 /// sent to other threads, but some pieces of performance infra run in a separate thread.
@@ -384,7 +396,7 @@ impl Span {
         Span::new(lo, hi, SyntaxContext::root())
     }
 
-    /// Returns a new span representing an empty span at the beginning of this span
+    /// Returns a new span representing an empty span at the beginning of this span.
     #[inline]
     pub fn shrink_to_lo(self) -> Span {
         let span = self.data();
@@ -398,7 +410,7 @@ impl Span {
     }
 
     #[inline]
-    /// Returns true if hi == lo
+    /// Returns `true` if `hi == lo`.
     pub fn is_empty(&self) -> bool {
         let span = self.data();
         span.hi == span.lo
@@ -512,7 +524,7 @@ impl Span {
     }
 
     /// Checks if a span is "internal" to a macro in which `unsafe`
-    /// can be used without triggering the `unsafe_code` lint
+    /// can be used without triggering the `unsafe_code` lint.
     //  (that is, a macro marked with `#[allow_internal_unsafe]`).
     pub fn allows_unsafe(&self) -> bool {
         self.ctxt().outer_expn_data().allow_internal_unsafe
@@ -700,6 +712,7 @@ impl Span {
     }
 }
 
+/// A span together with some additional data.
 #[derive(Clone, Debug)]
 pub struct SpanLabel {
     /// The span we are going to include in the final snippet.
@@ -743,7 +756,7 @@ impl<D: Decoder> Decodable<D> for Span {
 /// any spans that are debug-printed during the closure's execution.
 ///
 /// Normally, the global `TyCtxt` is used to retrieve the `SourceMap`
-/// (see `rustc_interface::callbacks::span_debug1). However, some parts
+/// (see `rustc_interface::callbacks::span_debug1`). However, some parts
 /// of the compiler (e.g. `rustc_parse`) may debug-print `Span`s before
 /// a `TyCtxt` is available. In this case, we fall back to
 /// the `SourceMap` provided to this function. If that is not available,
@@ -994,9 +1007,9 @@ pub enum ExternalSource {
     Unneeded,
     Foreign {
         kind: ExternalSourceKind,
-        /// This SourceFile's byte-offset within the source_map of its original crate
+        /// This SourceFile's byte-offset within the source_map of its original crate.
         original_start_pos: BytePos,
-        /// The end of this SourceFile within the source_map of its original crate
+        /// The end of this SourceFile within the source_map of its original crate.
         original_end_pos: BytePos,
     },
 }
@@ -1099,7 +1112,7 @@ impl SourceFileHash {
     }
 }
 
-/// A single source in the `SourceMap`.
+/// A single source in the [`SourceMap`].
 #[derive(Clone)]
 pub struct SourceFile {
     /// The name of the file that the source came from. Source that doesn't
@@ -1580,7 +1593,7 @@ fn remove_bom(src: &mut String, normalized_pos: &mut Vec<NormalizedPos>) {
 
 /// Replaces `\r\n` with `\n` in-place in `src`.
 ///
-/// Returns error if there's a lone `\r` in the string
+/// Returns error if there's a lone `\r` in the string.
 fn normalize_newlines(src: &mut String, normalized_pos: &mut Vec<NormalizedPos>) {
     if !src.as_bytes().contains(&b'\r') {
         return;
@@ -1705,13 +1718,16 @@ macro_rules! impl_pos {
 }
 
 impl_pos! {
-    /// A byte offset. Keep this small (currently 32-bits), as AST contains
-    /// a lot of them.
+    /// A byte offset.
+    ///
+    /// Keep this small (currently 32-bits), as AST contains a lot of them.
     #[derive(Clone, Copy, PartialEq, Eq, Hash, PartialOrd, Ord, Debug)]
     pub struct BytePos(pub u32);
 
-    /// A character offset. Because of multibyte UTF-8 characters, a byte offset
-    /// is not equivalent to a character offset. The `SourceMap` will convert `BytePos`
+    /// A character offset.
+    ///
+    /// Because of multibyte UTF-8 characters, a byte offset
+    /// is not equivalent to a character offset. The [`SourceMap`] will convert [`BytePos`]
     /// values to `CharPos` values as necessary.
     #[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug)]
     pub struct CharPos(pub usize);
@@ -1835,8 +1851,9 @@ fn lookup_line(lines: &[BytePos], pos: BytePos) -> isize {
 }
 
 /// Requirements for a `StableHashingContext` to be used in this crate.
-/// This is a hack to allow using the `HashStable_Generic` derive macro
-/// instead of implementing everything in librustc_middle.
+///
+/// This is a hack to allow using the [`HashStable_Generic`] derive macro
+/// instead of implementing everything in rustc_middle.
 pub trait HashStableContext {
     fn hash_def_id(&mut self, _: DefId, hasher: &mut StableHasher);
     fn hash_crate_num(&mut self, _: CrateNum, hasher: &mut StableHasher);
@@ -1856,6 +1873,7 @@ where
     /// offsets into the `SourceMap`). Instead, we hash the (file name, line, column)
     /// triple, which stays the same even if the containing `SourceFile` has moved
     /// within the `SourceMap`.
+    ///
     /// Also note that we are hashing byte offsets for the column, not unicode
     /// codepoint offsets. For the purpose of the hash that's sufficient.
     /// Also, hashing filenames is expensive so we avoid doing it twice when the

compiler/rustc_span/src/source_map.rs

@@ -1,9 +1,11 @@
-//! The `SourceMap` tracks all the source code used within a single crate, mapping
+//! Types for tracking pieces of source code within a crate.
+//!
+//! The [`SourceMap`] tracks all the source code used within a single crate, mapping
 //! from integer byte positions to the original source code location. Each bit
 //! of source parsed during crate parsing (typically files, in-memory strings,
 //! or various bits of macro expansion) cover a continuous range of bytes in the
-//! `SourceMap` and are represented by `SourceFile`s. Byte positions are stored in
-//! `Span` and used pervasively in the compiler. They are absolute positions
+//! `SourceMap` and are represented by [`SourceFile`]s. Byte positions are stored in
+//! [`Span`] and used pervasively in the compiler. They are absolute positions
 //! within the `SourceMap`, which upon request can be converted to line and column
 //! information, source code snippets, etc.
 

compiler/rustc_span/src/span_encoding.rs

@@ -12,7 +12,7 @@ use rustc_data_structures::fx::FxIndexSet;
 
 /// A compressed span.
 ///
-/// `SpanData` is 12 bytes, which is a bit too big to stick everywhere. `Span`
+/// Whereas [`SpanData`] is 12 bytes, which is a bit too big to stick everywhere, `Span`
 /// is a form that only takes up 8 bytes, with less space for the length and
 /// context. The vast majority (99.9%+) of `SpanData` instances will fit within
 /// those 8 bytes; any `SpanData` whose fields don't fit into a `Span` are
@@ -42,13 +42,11 @@ use rustc_data_structures::fx::FxIndexSet;
 /// - `base` is 32 bits in both `Span` and `SpanData`, which means that `base`
 ///   values never cause interning. The number of bits needed for `base`
 ///   depends on the crate size. 32 bits allows up to 4 GiB of code in a crate.
-///   `script-servo` is the largest crate in `rustc-perf`, requiring 26 bits
-///   for some spans.
 /// - `len` is 15 bits in `Span` (a u16, minus 1 bit for the tag) and 32 bits
 ///   in `SpanData`, which means that large `len` values will cause interning.
 ///   The number of bits needed for `len` does not depend on the crate size.
-///   The most common number of bits for `len` are 0--7, with a peak usually at
-///   3 or 4, and then it drops off quickly from 8 onwards. 15 bits is enough
+///   The most common numbers of bits for `len` are from 0 to 7, with a peak usually
+///   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