Rollup merge of rust-lang#141369 - yotamofek:pr/rustdoc/format_integer_with_underscore_sep, r=notriddle

Simplify `format_integer_with_underscore_sep`

Noticed that this helper fn only ever gets called with decimal-base-formatted ints, so can be simplified a lot by not trying to handle hex and octal radixes.
Second commit is completely unrelated, just simplified some code I wrote a while back 😁

Author: matthiaskrgr

src/librustdoc/clean/utils.rs

@@ -1,7 +1,7 @@
 use std::assert_matches::debug_assert_matches;
 use std::fmt::{self, Display, Write as _};
-use std::mem;
 use std::sync::LazyLock as Lazy;
+use std::{ascii, mem};
 
 use rustc_ast::tokenstream::TokenTree;
 use rustc_hir::def::{DefKind, Res};
@@ -24,7 +24,7 @@ use crate::clean::{
     clean_middle_ty, inline,
 };
 use crate::core::DocContext;
-use crate::display::Joined as _;
+use crate::display::{Joined as _, MaybeDisplay as _};
 
 #[cfg(test)]
 mod tests;
@@ -254,14 +254,7 @@ pub(crate) fn qpath_to_string(p: &hir::QPath<'_>) -> String {
     fmt::from_fn(|f| {
         segments
             .iter()
-            .map(|seg| {
-                fmt::from_fn(|f| {
-                    if seg.ident.name != kw::PathRoot {
-                        write!(f, "{}", seg.ident)?;
-                    }
-                    Ok(())
-                })
-            })
+            .map(|seg| (seg.ident.name != kw::PathRoot).then_some(seg.ident).maybe_display())
             .joined("::", f)
     })
     .to_string()
@@ -391,30 +384,12 @@ pub(crate) fn print_evaluated_const(
     })
 }
 
-fn format_integer_with_underscore_sep(num: &str) -> String {
-    let num_chars: Vec<_> = num.chars().collect();
-    let mut num_start_index = if num_chars.first() == Some(&'-') { 1 } else { 0 };
-    let chunk_size = match &num.as_bytes()[num_start_index..] {
-        [b'0', b'b' | b'x', ..] => {
-            num_start_index += 2;
-            4
-        }
-        [b'0', b'o', ..] => {
-            num_start_index += 2;
-            let remaining_chars = num_chars.len() - num_start_index;
-            if remaining_chars <= 6 {
-                // don't add underscores to Unix permissions like 0755 or 100755
-                return num.to_string();
-            }
-            3
-        }
-        _ => 3,
-    };
-
-    num_chars[..num_start_index]
-        .iter()
-        .chain(num_chars[num_start_index..].rchunks(chunk_size).rev().intersperse(&['_']).flatten())
-        .collect()
+fn format_integer_with_underscore_sep(num: u128, is_negative: bool) -> String {
+    let num = num.to_string();
+    let chars = num.as_ascii().unwrap();
+    let mut result = if is_negative { "-".to_string() } else { String::new() };
+    result.extend(chars.rchunks(3).rev().intersperse(&[ascii::Char::LowLine]).flatten());
+    result
 }
 
 fn print_const_with_custom_print_scalar<'tcx>(
@@ -428,7 +403,10 @@ fn print_const_with_custom_print_scalar<'tcx>(
     match (ct, ct.ty().kind()) {
         (mir::Const::Val(mir::ConstValue::Scalar(int), _), ty::Uint(ui)) => {
             let mut output = if with_underscores {
-                format_integer_with_underscore_sep(&int.to_string())
+                format_integer_with_underscore_sep(
+                    int.assert_scalar_int().to_bits_unchecked(),
+                    false,
+                )
             } else {
                 int.to_string()
             };
@@ -445,7 +423,10 @@ fn print_const_with_custom_print_scalar<'tcx>(
                 .size;
             let sign_extended_data = int.assert_scalar_int().to_int(size);
             let mut output = if with_underscores {
-                format_integer_with_underscore_sep(&sign_extended_data.to_string())
+                format_integer_with_underscore_sep(
+                    sign_extended_data.unsigned_abs(),
+                    sign_extended_data.is_negative(),
+                )
             } else {
                 sign_extended_data.to_string()
             };

src/librustdoc/clean/utils/tests.rs

@@ -2,40 +2,10 @@ use super::*;
 
 #[test]
 fn int_format_decimal() {
-    assert_eq!(format_integer_with_underscore_sep("12345678"), "12_345_678");
-    assert_eq!(format_integer_with_underscore_sep("123"), "123");
-    assert_eq!(format_integer_with_underscore_sep("123459"), "123_459");
-    assert_eq!(format_integer_with_underscore_sep("-12345678"), "-12_345_678");
-    assert_eq!(format_integer_with_underscore_sep("-123"), "-123");
-    assert_eq!(format_integer_with_underscore_sep("-123459"), "-123_459");
-}
-
-#[test]
-fn int_format_hex() {
-    assert_eq!(format_integer_with_underscore_sep("0xab3"), "0xab3");
-    assert_eq!(format_integer_with_underscore_sep("0xa2345b"), "0xa2_345b");
-    assert_eq!(format_integer_with_underscore_sep("0xa2e6345b"), "0xa2e6_345b");
-    assert_eq!(format_integer_with_underscore_sep("-0xab3"), "-0xab3");
-    assert_eq!(format_integer_with_underscore_sep("-0xa2345b"), "-0xa2_345b");
-    assert_eq!(format_integer_with_underscore_sep("-0xa2e6345b"), "-0xa2e6_345b");
-}
-
-#[test]
-fn int_format_binary() {
-    assert_eq!(format_integer_with_underscore_sep("0o12345671"), "0o12_345_671");
-    assert_eq!(format_integer_with_underscore_sep("0o123"), "0o123");
-    assert_eq!(format_integer_with_underscore_sep("0o123451"), "0o123451");
-    assert_eq!(format_integer_with_underscore_sep("-0o12345671"), "-0o12_345_671");
-    assert_eq!(format_integer_with_underscore_sep("-0o123"), "-0o123");
-    assert_eq!(format_integer_with_underscore_sep("-0o123451"), "-0o123451");
-}
-
-#[test]
-fn int_format_octal() {
-    assert_eq!(format_integer_with_underscore_sep("0b101"), "0b101");
-    assert_eq!(format_integer_with_underscore_sep("0b101101011"), "0b1_0110_1011");
-    assert_eq!(format_integer_with_underscore_sep("0b01101011"), "0b0110_1011");
-    assert_eq!(format_integer_with_underscore_sep("-0b101"), "-0b101");
-    assert_eq!(format_integer_with_underscore_sep("-0b101101011"), "-0b1_0110_1011");
-    assert_eq!(format_integer_with_underscore_sep("-0b01101011"), "-0b0110_1011");
+    assert_eq!(format_integer_with_underscore_sep(12345678, false), "12_345_678");
+    assert_eq!(format_integer_with_underscore_sep(123, false), "123");
+    assert_eq!(format_integer_with_underscore_sep(123459, false), "123_459");
+    assert_eq!(format_integer_with_underscore_sep(12345678, true), "-12_345_678");
+    assert_eq!(format_integer_with_underscore_sep(123, true), "-123");
+    assert_eq!(format_integer_with_underscore_sep(123459, true), "-123_459");
 }

src/librustdoc/lib.rs

@@ -3,6 +3,8 @@
     html_playground_url = "https://play.rust-lang.org/"
 )]
 #![feature(rustc_private)]
+#![feature(ascii_char)]
+#![feature(ascii_char_variants)]
 #![feature(assert_matches)]
 #![feature(box_patterns)]
 #![feature(debug_closure_helpers)]