cmp: avoid using advanced rust formatting for -l

Octal conversion and simple integer to string both show up in profiling.
This change improves comparing ~36M completely different files wth both
-l and -b by ~11-13%.
This commit is contained in:
Gustavo Noronha Silva
2024-09-24 22:15:12 -03:00
parent 50057412bd
commit 2e681301b4
3 changed files with 40 additions and 8 deletions
Generated
+7
View File
@@ -128,6 +128,7 @@ dependencies = [
"assert_cmd",
"chrono",
"diff",
"itoa",
"predicates",
"pretty_assertions",
"regex",
@@ -190,6 +191,12 @@ dependencies = [
"cc",
]
[[package]]
name = "itoa"
version = "1.0.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "49f1f14873335454500d59611f1cf4a4b0f786f9ac11f4312a78e4cf2566695b"
[[package]]
name = "js-sys"
version = "0.3.69"
+1
View File
@@ -17,6 +17,7 @@ path = "src/main.rs"
[dependencies]
chrono = "0.4.38"
diff = "0.1.13"
itoa = "1.0.11"
regex = "1.10.4"
same-file = "1.0.6"
unicode-width = "0.2.0"
+32 -8
View File
@@ -483,6 +483,24 @@ fn is_ascii_printable(byte: u8) -> bool {
c.is_ascii() && !c.is_ascii_control()
}
#[inline]
fn format_octal(byte: u8, buf: &mut [u8; 3]) -> &str {
*buf = [b' ', b' ', b'0'];
let mut num = byte;
let mut idx = 2; // Start at the last position in the buffer
// Generate octal digits
while num > 0 {
buf[idx] = b'0' + num % 8;
num /= 8;
idx = idx.saturating_sub(1);
}
// SAFETY: the operations we do above always land within ascii range.
unsafe { std::str::from_utf8_unchecked(&buf[..]) }
}
#[inline]
fn format_byte(byte: u8) -> String {
let mut byte = byte;
@@ -520,15 +538,20 @@ fn report_verbose_diffs(diffs: Vec<(usize, u8, u8)>, params: &Params) -> Result<
// Obtain the width of the first column from the last byte offset.
let width = format!("{}", offset).len();
let mut at_byte_buf = itoa::Buffer::new();
let mut from_oct = [0u8; 3]; // for octal conversions
let mut to_oct = [0u8; 3];
if params.print_bytes {
for (at_byte, from_byte, to_byte) in diffs {
let at_byte_str = at_byte_buf.format(at_byte);
writeln!(
stdout,
"{:>width$} {:>3o} {:4} {:>3o} {}",
at_byte,
from_byte,
"{:>width$} {} {:4} {} {}",
at_byte_str,
format_octal(from_byte, &mut from_oct),
format_byte(from_byte),
to_byte,
format_octal(to_byte, &mut to_oct),
format_byte(to_byte),
)
.map_err(|e| {
@@ -540,12 +563,13 @@ fn report_verbose_diffs(diffs: Vec<(usize, u8, u8)>, params: &Params) -> Result<
}
} else {
for (at_byte, from_byte, to_byte) in diffs {
let at_byte_str = at_byte_buf.format(at_byte);
writeln!(
stdout,
"{:>width$} {:>3o} {:>3o}",
at_byte,
from_byte,
to_byte,
"{:>width$} {} {}",
at_byte_str,
format_octal(from_byte, &mut from_oct),
format_octal(to_byte, &mut to_oct),
width = width
)
.map_err(|e| {