RGBCube/uutils-coreutils
1
Fork 0
mirror of https://github.com/RGBCube/uutils-coreutils synced 2025-07-27 19:17:43 +00:00
Code Issues Activity

wc: Avoid unnecessary work in general loop

Browse source 
This gives a big speedup if e.g. only characters are being counted.
This commit is contained in:
Jan Verbeek 2021-08-25 18:11:54 +02:00 • committed by Michael Debertol
parent f50b004860
commit 1358aeecdd
1 changed files with 34 additions and 26 deletions
Download patch file Download diff file Expand all files Collapse all files

60
src/uu/wc/src/wc.rs
View file

@ -235,34 +235,42 @@ fn word_count_from_reader<T: WordCountable>(
match chunk {
Ok(text) => {
for ch in text.chars() {
if ch.is_whitespace() {
in_word = false;
} else if ch.is_ascii_control() {
// These count as characters but do not affect the word state
} else if !in_word {
in_word = true;
total.words += 1;
}
match ch {
'\n' => {
total.max_line_length = max(current_len, total.max_line_length);
current_len = 0;
total.lines += 1;
}
// '\x0c' = '\f'
'\r' | '\x0c' => {
total.max_line_length = max(current_len, total.max_line_length);
current_len = 0;
}
'\t' => {
current_len -= current_len % 8;
current_len += 8;
}
_ => {
current_len += ch.width().unwrap_or(0);
if settings.show_words {
if ch.is_whitespace() {
in_word = false;
} else if ch.is_ascii_control() {
// These count as characters but do not affect the word state
} else if !in_word {
in_word = true;
total.words += 1;
}
}
total.chars += 1;
if settings.show_max_line_length {
match ch {
'\n' => {
total.max_line_length = max(current_len, total.max_line_length);
current_len = 0;
}
// '\x0c' = '\f'
'\r' | '\x0c' => {
total.max_line_length = max(current_len, total.max_line_length);
current_len = 0;
}
'\t' => {
current_len -= current_len % 8;
current_len += 8;
}
_ => {
current_len += ch.width().unwrap_or(0);
}
}
}
if settings.show_lines && ch == '\n' {
total.lines += 1;
}
if settings.show_chars {
total.chars += 1;
}
}
total.bytes += text.len();
}