Auto merge of #55466 - sinkuu:cleanup, r=petrochenkov

syntax: Use iterator and pattern APIs instead of `char_at`

Iterating over chars with the `char_at(str, i)` `i += ch.len_utf8()` loop seems unidiomatic.

Author: bors

src/libsyntax/lib.rs

@@ -173,7 +173,6 @@ pub mod parse;
 pub mod ptr;
 pub mod show_span;
 pub mod std_inject;
-pub mod str;
 pub use syntax_pos::edition;
 pub use syntax_pos::symbol;
 pub mod test;

src/libsyntax/parse/lexer/comments.rs

@@ -16,7 +16,6 @@ use syntax_pos::{BytePos, CharPos, Pos, FileName};
 use parse::lexer::{is_block_doc_comment, is_pattern_whitespace};
 use parse::lexer::{self, ParseSess, StringReader, TokenAndSpan};
 use print::pprust;
-use str::char_at;
 
 use std::io::Read;
 use std::usize;
@@ -207,28 +206,22 @@ fn read_line_comments(rdr: &mut StringReader,
 /// Otherwise returns Some(k) where k is first char offset after that leading
 /// whitespace.  Note k may be outside bounds of s.
 fn all_whitespace(s: &str, col: CharPos) -> Option<usize> {
-    let len = s.len();
-    let mut col = col.to_usize();
-    let mut cursor: usize = 0;
-
-    while col > 0 && cursor < len {
-        let ch = char_at(s, cursor);
+    let mut idx = 0;
+    for (i, ch) in s.char_indices().take(col.to_usize()) {
         if !ch.is_whitespace() {
             return None;
         }
-        cursor += ch.len_utf8();
-        col -= 1;
+        idx = i + ch.len_utf8();
     }
-
-    Some(cursor)
+    Some(idx)
 }
 
 fn trim_whitespace_prefix_and_push_line(lines: &mut Vec<String>, s: String, col: CharPos) {
     let len = s.len();
     let s1 = match all_whitespace(&s[..], col) {
         Some(col) => {
             if col < len {
-                (&s[col..len]).to_string()
+                s[col..len].to_string()
             } else {
                 String::new()
             }
@@ -247,20 +240,13 @@ fn read_block_comment(rdr: &mut StringReader,
     let mut lines: Vec<String> = Vec::new();
 
     // Count the number of chars since the start of the line by rescanning.
-    let mut src_index = rdr.src_index(rdr.source_file.line_begin_pos(rdr.pos));
+    let src_index = rdr.src_index(rdr.source_file.line_begin_pos(rdr.pos));
     let end_src_index = rdr.src_index(rdr.pos);
     assert!(src_index <= end_src_index,
         "src_index={}, end_src_index={}, line_begin_pos={}",
         src_index, end_src_index, rdr.source_file.line_begin_pos(rdr.pos).to_u32());
-    let mut n = 0;
-
-    while src_index < end_src_index {
-        let c = char_at(&rdr.src, src_index);
-        src_index += c.len_utf8();
-        n += 1;
-    }
 
-    let col = CharPos(n);
+    let col = CharPos(rdr.src[src_index..end_src_index].chars().count());
 
     rdr.bump();
     rdr.bump();

src/libsyntax/parse/lexer/mod.rs

@@ -13,12 +13,12 @@ use syntax_pos::{self, BytePos, CharPos, Pos, Span, NO_EXPANSION};
 use source_map::{SourceMap, FilePathMapping};
 use errors::{Applicability, FatalError, Diagnostic, DiagnosticBuilder};
 use parse::{token, ParseSess};
-use str::char_at;
 use symbol::{Symbol, keywords};
 use core::unicode::property::Pattern_White_Space;
 
 use std::borrow::Cow;
 use std::char;
+use std::iter;
 use std::mem::replace;
 use rustc_data_structures::sync::Lrc;
 
@@ -459,45 +459,42 @@ impl<'a> StringReader<'a> {
 
     /// Converts CRLF to LF in the given string, raising an error on bare CR.
     fn translate_crlf<'b>(&self, start: BytePos, s: &'b str, errmsg: &'b str) -> Cow<'b, str> {
-        let mut i = 0;
-        while i < s.len() {
-            let ch = char_at(s, i);
-            let next = i + ch.len_utf8();
+        let mut chars = s.char_indices().peekable();
+        while let Some((i, ch)) = chars.next() {
             if ch == '\r' {
-                if next < s.len() && char_at(s, next) == '\n' {
-                    return translate_crlf_(self, start, s, errmsg, i).into();
+                if let Some((lf_idx, '\n')) = chars.peek() {
+                    return translate_crlf_(self, start, s, *lf_idx, chars, errmsg).into();
                 }
                 let pos = start + BytePos(i as u32);
-                let end_pos = start + BytePos(next as u32);
+                let end_pos = start + BytePos((i + ch.len_utf8()) as u32);
                 self.err_span_(pos, end_pos, errmsg);
             }
-            i = next;
         }
         return s.into();
 
         fn translate_crlf_(rdr: &StringReader,
                            start: BytePos,
                            s: &str,
-                           errmsg: &str,
-                           mut i: usize)
+                           mut j: usize,
+                           mut chars: iter::Peekable<impl Iterator<Item = (usize, char)>>,
+                           errmsg: &str)
                            -> String {
             let mut buf = String::with_capacity(s.len());
-            let mut j = 0;
-            while i < s.len() {
-                let ch = char_at(s, i);
-                let next = i + ch.len_utf8();
+            // Skip first CR
+            buf.push_str(&s[.. j - 1]);
+            while let Some((i, ch)) = chars.next() {
                 if ch == '\r' {
                     if j < i {
                         buf.push_str(&s[j..i]);
                     }
+                    let next = i + ch.len_utf8();
                     j = next;
-                    if next >= s.len() || char_at(s, next) != '\n' {
+                    if chars.peek().map(|(_, ch)| *ch) != Some('\n') {
                         let pos = start + BytePos(i as u32);
                         let end_pos = start + BytePos(next as u32);
                         rdr.err_span_(pos, end_pos, errmsg);
                     }
                 }
-                i = next;
             }
             if j < s.len() {
                 buf.push_str(&s[j..]);
@@ -1858,6 +1855,11 @@ fn ident_continue(c: Option<char>) -> bool {
     (c > '\x7f' && c.is_xid_continue())
 }
 
+#[inline]
+fn char_at(s: &str, byte: usize) -> char {
+    s[byte..].chars().next().unwrap()
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

src/libsyntax/parse/mod.rs

@@ -19,7 +19,6 @@ use errors::{Handler, ColorConfig, Diagnostic, DiagnosticBuilder};
 use feature_gate::UnstableFeatures;
 use parse::parser::Parser;
 use ptr::P;
-use str::char_at;
 use symbol::Symbol;
 use tokenstream::{TokenStream, TokenTree};
 use diagnostics::plugin::ErrorMap;
@@ -436,9 +435,7 @@ fn raw_str_lit(lit: &str) -> String {
 
 // check if `s` looks like i32 or u1234 etc.
 fn looks_like_width_suffix(first_chars: &[char], s: &str) -> bool {
-    s.len() > 1 &&
-        first_chars.contains(&char_at(s, 0)) &&
-        s[1..].chars().all(|c| '0' <= c && c <= '9')
+    s.starts_with(first_chars) && s[1..].chars().all(|c| c.is_ascii_digit())
 }
 
 macro_rules! err {
@@ -645,11 +642,11 @@ fn integer_lit(s: &str, suffix: Option<Symbol>, diag: Option<(Span, &Handler)>)
     let orig = s;
     let mut ty = ast::LitIntType::Unsuffixed;
 
-    if char_at(s, 0) == '0' && s.len() > 1 {
-        match char_at(s, 1) {
-            'x' => base = 16,
-            'o' => base = 8,
-            'b' => base = 2,
+    if s.starts_with('0') && s.len() > 1 {
+        match s.as_bytes()[1] {
+            b'x' => base = 16,
+            b'o' => base = 8,
+            b'b' => base = 2,
             _ => { }
         }
     }

src/libsyntax/str.rs

@@ -38,7 +38,8 @@ pub fn lev_distance(a: &str, b: &str) -> usize {
             current = next;
             t_last = j;
         }
-    } dcol[t_last + 1]
+    }
+    dcol[t_last + 1]
 }
 
 /// Find the best match for a given word in the given iterator

src/libsyntax/util/lev_distance.rs

@@ -25,7 +25,6 @@ use syntax::parse::new_parser_from_source_str;
 use syntax::parse::parser::Parser;
 use syntax::parse::token;
 use syntax::ptr::P;
-use syntax::str::char_at;
 use syntax::parse::attr::*;
 use syntax::print::pprust;
 use std::fmt;