Merge pull request #2297 from jhscheer/refactoring_parse_size

Reducing duplicate code by refactoring `parse_size()` of head, tail, truncate, stdbuf, etc.

src/uu/du/src/du.rs

@@ -1,9 +1,9 @@
-// This file is part of the uutils coreutils package.
+//  * This file is part of the uutils coreutils package.
-//
+//  *
-// (c) Derek Chiang <derekchiang93@gmail.com>
+//  * (c) Derek Chiang <derekchiang93@gmail.com>
-//
+//  *
-// For the full copyright and license information, please view the LICENSE
+//  * For the full copyright and license information, please view the LICENSE
-// file that was distributed with this source code.
+//  * file that was distributed with this source code.
 
 #[macro_use]
 extern crate uucore;
@@ -12,6 +12,7 @@ use chrono::prelude::DateTime;
 use chrono::Local;
 use clap::{crate_version, App, Arg};
 use std::collections::HashSet;
+use std::convert::TryFrom;
 use std::env;
 use std::fs;
 #[cfg(not(windows))]
@@ -28,6 +29,7 @@ use std::os::windows::io::AsRawHandle;
 use std::path::Path;
 use std::path::PathBuf;
 use std::time::{Duration, UNIX_EPOCH};
+use uucore::parse_size::{parse_size, ParseSizeError};
 use uucore::InvalidEncodingHandling;
 #[cfg(windows)]
 use winapi::shared::minwindef::{DWORD, LPVOID};
@@ -44,7 +46,7 @@ mod options {
     pub const NULL: &str = "0";
     pub const ALL: &str = "all";
     pub const APPARENT_SIZE: &str = "apparent-size";
-    pub const BLOCK_SIZE: &str = "B";
+    pub const BLOCK_SIZE: &str = "block-size";
     pub const BYTES: &str = "b";
     pub const TOTAL: &str = "c";
     pub const MAX_DEPTH: &str = "d";
@@ -227,63 +229,22 @@ fn get_file_info(path: &Path) -> Option<FileInfo> {
     result
 }
 
-fn unit_string_to_number(s: &str) -> Option<u64> {
+fn read_block_size(s: Option<&str>) -> usize {
-    let mut offset = 0;
+    if let Some(s) = s {
-    let mut s_chars = s.chars().rev();
+        parse_size(s)
-
+            .unwrap_or_else(|e| crash!(1, "{}", format_error_message(e, s, options::BLOCK_SIZE)))
-    let (mut ch, multiple) = match s_chars.next()? {
+    } else {
-        'B' | 'b' => ('B', 1000u64),
+        for env_var in &["DU_BLOCK_SIZE", "BLOCK_SIZE", "BLOCKSIZE"] {
-        ch => (ch, 1024u64),
+            if let Ok(env_size) = env::var(env_var) {
-    };
+                if let Ok(v) = parse_size(&env_size) {
-    if ch == 'B' {
+                    return v;
-        ch = s_chars.next()?;
-        offset += 1;
-    }
-    ch = ch.to_ascii_uppercase();
-
-    let unit = UNITS
-        .iter()
-        .rev()
-        .find(|&&(unit_ch, _)| unit_ch == ch)
-        .map(|&(_, val)| {
-            // we found a match, so increment offset
-            offset += 1;
-            val
-        })
-        .or_else(|| if multiple == 1024 { Some(0) } else { None })?;
-
-    let number = s[..s.len() - offset].parse::<u64>().ok()?;
-
-    Some(number * multiple.pow(unit))
-}
-
-fn translate_to_pure_number(s: &Option<&str>) -> Option<u64> {
-    match *s {
-        Some(s) => unit_string_to_number(s),
-        None => None,
-    }
-}
-
-fn read_block_size(s: Option<&str>) -> u64 {
-    match translate_to_pure_number(&s) {
-        Some(v) => v,
-        None => {
-            if let Some(value) = s {
-                show_error!("invalid --block-size argument '{}'", value);
-            };
-
-            for env_var in &["DU_BLOCK_SIZE", "BLOCK_SIZE", "BLOCKSIZE"] {
-                let env_size = env::var(env_var).ok();
-                if let Some(quantity) = translate_to_pure_number(&env_size.as_deref()) {
-                    return quantity;
                 }
             }
-
+        }
-            if env::var("POSIXLY_CORRECT").is_ok() {
+        if env::var("POSIXLY_CORRECT").is_ok() {
-                512
+            512
-            } else {
+        } else {
-                1024
+            1024
-            }
         }
     }
 }
@@ -449,7 +410,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         .arg(
             Arg::with_name(options::BLOCK_SIZE)
                 .short("B")
-                .long("block-size")
+                .long(options::BLOCK_SIZE)
                 .value_name("SIZE")
                 .help(
                     "scale sizes  by  SIZE before printing them. \
@@ -623,7 +584,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         }
     };
 
-    let block_size = read_block_size(matches.value_of(options::BLOCK_SIZE));
+    let block_size = u64::try_from(read_block_size(matches.value_of(options::BLOCK_SIZE))).unwrap();
 
     let multiplier: u64 = if matches.is_present(options::SI) {
         1000
@@ -759,31 +720,27 @@ Try '{} --help' for more information.",
     0
 }
 
+fn format_error_message(error: ParseSizeError, s: &str, option: &str) -> String {
+    // NOTE:
+    // GNU's du echos affected flag, -B or --block-size (-t or --threshold), depending user's selection
+    // GNU's du does distinguish between "invalid (suffix in) argument"
+    match error {
+        ParseSizeError::ParseFailure(_) => format!("invalid --{} argument '{}'", option, s),
+        ParseSizeError::SizeTooBig(_) => format!("--{} argument '{}' too large", option, s),
+    }
+}
+
 #[cfg(test)]
 mod test_du {
     #[allow(unused_imports)]
     use super::*;
 
-    #[test]
-    fn test_translate_to_pure_number() {
-        let test_data = [
-            (Some("10".to_string()), Some(10)),
-            (Some("10K".to_string()), Some(10 * 1024)),
-            (Some("5M".to_string()), Some(5 * 1024 * 1024)),
-            (Some("900KB".to_string()), Some(900 * 1000)),
-            (Some("BAD_STRING".to_string()), None),
-        ];
-        for it in test_data.iter() {
-            assert_eq!(translate_to_pure_number(&it.0.as_deref()), it.1);
-        }
-    }
-
     #[test]
     fn test_read_block_size() {
         let test_data = [
-            (Some("10".to_string()), 10),
+            (Some("1024".to_string()), 1024),
+            (Some("K".to_string()), 1024),
             (None, 1024),
-            (Some("BAD_STRING".to_string()), 1024),
         ];
         for it in test_data.iter() {
             assert_eq!(read_block_size(it.0.as_deref()), it.1);

src/uu/head/src/head.rs

@@ -1,3 +1,8 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 // spell-checker:ignore (vars) zlines
 
 use clap::{crate_version, App, Arg};
@@ -75,7 +80,7 @@ fn app<'a>() -> App<'a, 'a> {
         .arg(
             Arg::with_name(options::QUIET_NAME)
                 .short("q")
-                .long("--quiet")
+                .long("quiet")
                 .visible_alias("silent")
                 .help("never print headers giving file names")
                 .overrides_with_all(&[options::VERBOSE_NAME, options::QUIET_NAME]),
@@ -108,12 +113,7 @@ where
 {
     match parse::parse_num(src) {
         Ok((n, last)) => Ok((closure(n), last)),
-        Err(reason) => match reason {
+        Err(e) => Err(e.to_string()),
-            parse::ParseError::Syntax => Err(format!("'{}'", src)),
-            parse::ParseError::Overflow => {
-                Err(format!("'{}': Value too large for defined datatype", src))
-            }
-        },
     }
 }
 
@@ -466,7 +466,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     let args = match HeadOptions::get_from(args) {
         Ok(o) => o,
         Err(s) => {
-            crash!(EXIT_FAILURE, "head: {}", s);
+            crash!(EXIT_FAILURE, "{}", s);
         }
     };
     match uu_head(&args) {

src/uu/head/src/parse.rs

@@ -1,5 +1,10 @@
-use std::convert::TryFrom;
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 use std::ffi::OsString;
+use uucore::parse_size::{parse_size, ParseSizeError};
 
 #[derive(PartialEq, Debug)]
 pub enum ParseError {
@@ -92,88 +97,25 @@ pub fn parse_obsolete(src: &str) -> Option<Result<impl Iterator<Item = OsString>
 }
 /// Parses an -c or -n argument,
 /// the bool specifies whether to read from the end
-pub fn parse_num(src: &str) -> Result<(usize, bool), ParseError> {
+pub fn parse_num(src: &str) -> Result<(usize, bool), ParseSizeError> {
-    let mut num_start = 0;
+    let mut size_string = src.trim();
-    let (mut chars, all_but_last) = {
+    let mut all_but_last = false;
-        let mut chars = src.char_indices();
-        let (_, c) = chars.next().ok_or(ParseError::Syntax)?;
-        if c == '-' {
-            num_start += 1;
-            (chars, true)
-        } else {
-            (src.char_indices(), false)
-        }
-    };
-    let mut num_end = 0usize;
-    let mut last_char = 0 as char;
-    let mut num_count = 0usize;
-    for (n, c) in &mut chars {
-        if c.is_numeric() {
-            num_end = n;
-            num_count += 1;
-        } else {
-            last_char = c;
-            break;
-        }
-    }
 
-    let num = if num_count > 0 {
+    if let Some(c) = size_string.chars().next() {
-        Some(
+        if c == '+' || c == '-' {
-            src[num_start..=num_end]
+            // head: '+' is not documented (8.32 man pages)
-                .parse::<usize>()
+            size_string = &size_string[1..];
-                .map_err(|_| ParseError::Overflow)?,
+            if c == '-' {
-        )
+                all_but_last = true;
-    } else {
-        None
-    };
-
-    if last_char == 0 as char {
-        if let Some(n) = num {
-            Ok((n, all_but_last))
-        } else {
-            Err(ParseError::Syntax)
-        }
-    } else {
-        let base: u128 = match chars.next() {
-            Some((_, c)) => {
-                let b = match c {
-                    'B' if last_char != 'b' => 1000,
-                    'i' if last_char != 'b' => {
-                        if let Some((_, 'B')) = chars.next() {
-                            1024
-                        } else {
-                            return Err(ParseError::Syntax);
-                        }
-                    }
-                    _ => return Err(ParseError::Syntax),
-                };
-                if chars.next().is_some() {
-                    return Err(ParseError::Syntax);
-                } else {
-                    b
-                }
             }
-            None => 1024,
-        };
-        let mul = match last_char.to_lowercase().next().unwrap() {
-            'b' => 512,
-            'k' => base.pow(1),
-            'm' => base.pow(2),
-            'g' => base.pow(3),
-            't' => base.pow(4),
-            'p' => base.pow(5),
-            'e' => base.pow(6),
-            'z' => base.pow(7),
-            'y' => base.pow(8),
-            _ => return Err(ParseError::Syntax),
-        };
-        let mul = usize::try_from(mul).map_err(|_| ParseError::Overflow)?;
-        match num.unwrap_or(1).checked_mul(mul) {
-            Some(n) => Ok((n, all_but_last)),
-            None => Err(ParseError::Overflow),
         }
+    } else {
+        return Err(ParseSizeError::ParseFailure(src.to_string()));
     }
+
+    parse_size(size_string).map(|n| (n, all_but_last))
 }
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -191,44 +133,6 @@ mod tests {
         Some(Ok(src.iter().map(|s| s.to_string()).collect()))
     }
     #[test]
-    #[cfg(not(target_pointer_width = "128"))]
-    fn test_parse_overflow_x64() {
-        assert_eq!(parse_num("1Y"), Err(ParseError::Overflow));
-        assert_eq!(parse_num("1Z"), Err(ParseError::Overflow));
-        assert_eq!(parse_num("100E"), Err(ParseError::Overflow));
-        assert_eq!(parse_num("100000P"), Err(ParseError::Overflow));
-        assert_eq!(parse_num("1000000000T"), Err(ParseError::Overflow));
-        assert_eq!(
-            parse_num("10000000000000000000000"),
-            Err(ParseError::Overflow)
-        );
-    }
-    #[test]
-    #[cfg(target_pointer_width = "32")]
-    fn test_parse_overflow_x32() {
-        assert_eq!(parse_num("1T"), Err(ParseError::Overflow));
-        assert_eq!(parse_num("1000G"), Err(ParseError::Overflow));
-    }
-    #[test]
-    fn test_parse_bad_syntax() {
-        assert_eq!(parse_num("5MiB nonsense"), Err(ParseError::Syntax));
-        assert_eq!(parse_num("Nonsense string"), Err(ParseError::Syntax));
-        assert_eq!(parse_num("5mib"), Err(ParseError::Syntax));
-        assert_eq!(parse_num("biB"), Err(ParseError::Syntax));
-        assert_eq!(parse_num("-"), Err(ParseError::Syntax));
-        assert_eq!(parse_num(""), Err(ParseError::Syntax));
-    }
-    #[test]
-    fn test_parse_numbers() {
-        assert_eq!(parse_num("k"), Ok((1024, false)));
-        assert_eq!(parse_num("MiB"), Ok((1024 * 1024, false)));
-        assert_eq!(parse_num("-5"), Ok((5, true)));
-        assert_eq!(parse_num("b"), Ok((512, false)));
-        assert_eq!(parse_num("-2GiB"), Ok((2 * 1024 * 1024 * 1024, true)));
-        assert_eq!(parse_num("5M"), Ok((5 * 1024 * 1024, false)));
-        assert_eq!(parse_num("5MB"), Ok((5 * 1000 * 1000, false)));
-    }
-    #[test]
     fn test_parse_numbers_obsolete() {
         assert_eq!(obsolete("-5"), obsolete_result(&["-n", "5"]));
         assert_eq!(obsolete("-100"), obsolete_result(&["-n", "100"]));

src/uu/od/src/od.rs

@@ -43,6 +43,7 @@ use crate::partialreader::*;
 use crate::peekreader::*;
 use crate::prn_char::format_ascii_dump;
 use clap::{self, crate_version, AppSettings, Arg, ArgMatches};
+use uucore::parse_size::ParseSizeError;
 use uucore::InvalidEncodingHandling;
 
 const PEEK_BUFFER_SIZE: usize = 4; // utf-8 can be 4 bytes
@@ -128,13 +129,11 @@ impl OdOptions {
             }
         };
 
-        let mut skip_bytes = matches
+        let mut skip_bytes = matches.value_of(options::SKIP_BYTES).map_or(0, |s| {
-            .value_of(options::SKIP_BYTES)
+            parse_number_of_bytes(s).unwrap_or_else(|e| {
-            .map(|s| {
+                crash!(1, "{}", format_error_message(e, s, options::SKIP_BYTES))
-                parse_number_of_bytes(s).map_err(|_| format!("Invalid argument --skip-bytes={}", s))
             })
-            .transpose()?
+        });
-            .unwrap_or(0);
 
         let mut label: Option<usize> = None;
 
@@ -150,11 +149,14 @@ impl OdOptions {
 
         let formats = parse_format_flags(&args)?;
 
-        let mut line_bytes = match matches.value_of(options::WIDTH) {
+        let mut line_bytes = matches.value_of(options::WIDTH).map_or(16, |s| {
-            None => 16,
+            if matches.occurrences_of(options::WIDTH) == 0 {
-            Some(_) if matches.occurrences_of(options::WIDTH) == 0 => 16,
+                return 16;
-            Some(s) => s.parse::<usize>().unwrap_or(0),
+            };
-        };
+            parse_number_of_bytes(s)
+                .unwrap_or_else(|e| crash!(1, "{}", format_error_message(e, s, options::WIDTH)))
+        });
+
         let min_bytes = formats.iter().fold(1, |max, next| {
             cmp::max(max, next.formatter_item_info.byte_size)
         });
@@ -165,12 +167,11 @@ impl OdOptions {
 
         let output_duplicates = matches.is_present(options::OUTPUT_DUPLICATES);
 
-        let read_bytes = matches
+        let read_bytes = matches.value_of(options::READ_BYTES).map(|s| {
-            .value_of(options::READ_BYTES)
+            parse_number_of_bytes(s).unwrap_or_else(|e| {
-            .map(|s| {
+                crash!(1, "{}", format_error_message(e, s, options::READ_BYTES))
-                parse_number_of_bytes(s).map_err(|_| format!("Invalid argument --read-bytes={}", s))
             })
-            .transpose()?;
+        });
 
         let radix = match matches.value_of(options::ADDRESS_RADIX) {
             None => Radix::Octal,
@@ -251,7 +252,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 .short("S")
                 .long(options::STRINGS)
                 .help(
-                    "output strings of at least BYTES graphic chars. 3 is assumed when \
+                    "NotImplemented: output strings of at least BYTES graphic chars. 3 is assumed when \
                      BYTES is not specified.",
                 )
                 .default_value("3")
@@ -441,8 +442,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
 
     let od_options = match OdOptions::new(clap_matches, args) {
         Err(s) => {
-            show_usage_error!("{}", s);
+            crash!(1, "{}", s);
-            return 1;
         }
         Ok(o) => o,
     };
@@ -624,3 +624,13 @@ fn open_input_peek_reader(
     let pr = PartialReader::new(mf, skip_bytes, read_bytes);
     PeekReader::new(pr)
 }
+
+fn format_error_message(error: ParseSizeError, s: &str, option: &str) -> String {
+    // NOTE:
+    // GNU's od echos affected flag, -N or --read-bytes (-j or --skip-bytes, etc.), depending user's selection
+    // GNU's od does distinguish between "invalid (suffix in) argument"
+    match error {
+        ParseSizeError::ParseFailure(_) => format!("invalid --{} argument '{}'", option, s),
+        ParseSizeError::SizeTooBig(_) => format!("--{} argument '{}' too large", option, s),
+    }
+}

src/uu/od/src/parse_nrofbytes.rs

@@ -1,14 +1,17 @@
-pub fn parse_number_of_bytes(s: &str) -> Result<usize, &'static str> {
+use uucore::parse_size::{parse_size, ParseSizeError};
+
+pub fn parse_number_of_bytes(s: &str) -> Result<usize, ParseSizeError> {
     let mut start = 0;
     let mut len = s.len();
-    let mut radix = 10;
+    let mut radix = 16;
     let mut multiply = 1;
 
     if s.starts_with("0x") || s.starts_with("0X") {
         start = 2;
-        radix = 16;
     } else if s.starts_with('0') {
         radix = 8;
+    } else {
+        return parse_size(&s[start..]);
     }
 
     let mut ends_with = s.chars().rev();
@@ -56,78 +59,33 @@ pub fn parse_number_of_bytes(s: &str) -> Result<usize, &'static str> {
                 Some('P') => 1000 * 1000 * 1000 * 1000 * 1000,
                 #[cfg(target_pointer_width = "64")]
                 Some('E') => 1000 * 1000 * 1000 * 1000 * 1000 * 1000,
-                _ => return Err("parse failed"),
+                _ => return Err(ParseSizeError::ParseFailure(s.to_string())),
             }
         }
         _ => {}
     }
 
-    match usize::from_str_radix(&s[start..len], radix) {
+    let factor = match usize::from_str_radix(&s[start..len], radix) {
-        Ok(i) => Ok(i * multiply),
+        Ok(f) => f,
-        Err(_) => Err("parse failed"),
+        Err(e) => return Err(ParseSizeError::ParseFailure(e.to_string())),
-    }
+    };
-}
+    factor
-
+        .checked_mul(multiply)
-#[allow(dead_code)]
+        .ok_or_else(|| ParseSizeError::SizeTooBig(s.to_string()))
-fn parse_number_of_bytes_str(s: &str) -> Result<usize, &'static str> {
-    parse_number_of_bytes(&String::from(s))
 }
 
 #[test]
 fn test_parse_number_of_bytes() {
-    // normal decimal numbers
-    assert_eq!(0, parse_number_of_bytes_str("0").unwrap());
-    assert_eq!(5, parse_number_of_bytes_str("5").unwrap());
-    assert_eq!(999, parse_number_of_bytes_str("999").unwrap());
-    assert_eq!(2 * 512, parse_number_of_bytes_str("2b").unwrap());
-    assert_eq!(2 * 1024, parse_number_of_bytes_str("2k").unwrap());
-    assert_eq!(4 * 1024, parse_number_of_bytes_str("4K").unwrap());
-    assert_eq!(2 * 1048576, parse_number_of_bytes_str("2m").unwrap());
-    assert_eq!(4 * 1048576, parse_number_of_bytes_str("4M").unwrap());
-    assert_eq!(1073741824, parse_number_of_bytes_str("1G").unwrap());
-    assert_eq!(2000, parse_number_of_bytes_str("2kB").unwrap());
-    assert_eq!(4000, parse_number_of_bytes_str("4KB").unwrap());
-    assert_eq!(2000000, parse_number_of_bytes_str("2mB").unwrap());
-    assert_eq!(4000000, parse_number_of_bytes_str("4MB").unwrap());
-    assert_eq!(2000000000, parse_number_of_bytes_str("2GB").unwrap());
-
     // octal input
-    assert_eq!(8, parse_number_of_bytes_str("010").unwrap());
+    assert_eq!(8, parse_number_of_bytes("010").unwrap());
-    assert_eq!(8 * 512, parse_number_of_bytes_str("010b").unwrap());
+    assert_eq!(8 * 512, parse_number_of_bytes("010b").unwrap());
-    assert_eq!(8 * 1024, parse_number_of_bytes_str("010k").unwrap());
+    assert_eq!(8 * 1024, parse_number_of_bytes("010k").unwrap());
-    assert_eq!(8 * 1048576, parse_number_of_bytes_str("010m").unwrap());
+    assert_eq!(8 * 1_048_576, parse_number_of_bytes("010m").unwrap());
 
     // hex input
-    assert_eq!(15, parse_number_of_bytes_str("0xf").unwrap());
+    assert_eq!(15, parse_number_of_bytes("0xf").unwrap());
-    assert_eq!(15, parse_number_of_bytes_str("0XF").unwrap());
+    assert_eq!(15, parse_number_of_bytes("0XF").unwrap());
-    assert_eq!(27, parse_number_of_bytes_str("0x1b").unwrap());
+    assert_eq!(27, parse_number_of_bytes("0x1b").unwrap());
-    assert_eq!(16 * 1024, parse_number_of_bytes_str("0x10k").unwrap());
+    assert_eq!(16 * 1024, parse_number_of_bytes("0x10k").unwrap());
-    assert_eq!(16 * 1048576, parse_number_of_bytes_str("0x10m").unwrap());
+    assert_eq!(16 * 1_048_576, parse_number_of_bytes("0x10m").unwrap());
-
-    // invalid input
-    parse_number_of_bytes_str("").unwrap_err();
-    parse_number_of_bytes_str("-1").unwrap_err();
-    parse_number_of_bytes_str("1e2").unwrap_err();
-    parse_number_of_bytes_str("xyz").unwrap_err();
-    parse_number_of_bytes_str("b").unwrap_err();
-    parse_number_of_bytes_str("1Y").unwrap_err();
-    parse_number_of_bytes_str("∞").unwrap_err();
-}
-
-#[test]
-#[cfg(target_pointer_width = "64")]
-fn test_parse_number_of_bytes_64bits() {
-    assert_eq!(1099511627776, parse_number_of_bytes_str("1T").unwrap());
-    assert_eq!(1125899906842624, parse_number_of_bytes_str("1P").unwrap());
-    assert_eq!(
-        1152921504606846976,
-        parse_number_of_bytes_str("1E").unwrap()
-    );
-
-    assert_eq!(2000000000000, parse_number_of_bytes_str("2TB").unwrap());
-    assert_eq!(2000000000000000, parse_number_of_bytes_str("2PB").unwrap());
-    assert_eq!(
-        2000000000000000000,
-        parse_number_of_bytes_str("2EB").unwrap()
-    );
 }

src/uu/sleep/Cargo.toml

@@ -16,7 +16,7 @@ path = "src/sleep.rs"
 
 [dependencies]
 clap = "2.33"
-uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["parse_time"] }
+uucore = { version=">=0.0.8", package="uucore", path="../../uucore" }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
 
 [[bin]]

src/uu/sort/src/sort.rs

@@ -43,6 +43,7 @@ use std::ops::Range;
 use std::path::Path;
 use std::path::PathBuf;
 use unicode_width::UnicodeWidthStr;
+use uucore::parse_size::{parse_size, ParseSizeError};
 use uucore::InvalidEncodingHandling;
 
 static NAME: &str = "sort";
@@ -162,32 +163,29 @@ pub struct GlobalSettings {
 }
 
 impl GlobalSettings {
-    /// Interpret this `&str` as a number with an optional trailing si unit.
+    /// Parse a SIZE string into a number of bytes.
-    ///
+    /// A size string comprises an integer and an optional unit.
-    /// If there is no trailing si unit, the implicit unit is K.
+    /// The unit may be k, K, m, M, g, G, t, T, P, E, Z, Y (powers of 1024), or b which is 1.
-    /// The suffix B causes the number to be interpreted as a byte count.
+    /// Default is K.
-    fn parse_byte_count(input: &str) -> usize {
+    fn parse_byte_count(input: &str) -> Result<usize, ParseSizeError> {
-        const SI_UNITS: &[char] = &['B', 'K', 'M', 'G', 'T', 'P', 'E', 'Z', 'Y'];
+        // GNU sort (8.32)   valid: 1b,        k, K, m, M, g, G, t, T, P, E, Z, Y
+        // GNU sort (8.32) invalid:  b, B, 1B,                         p, e, z, y
+        const ALLOW_LIST: &[char] = &[
+            'b', 'k', 'K', 'm', 'M', 'g', 'G', 't', 'T', 'P', 'E', 'Z', 'Y',
+        ];
+        let mut size_string = input.trim().to_string();
 
-        let input = input.trim();
+        if size_string.ends_with(|c: char| ALLOW_LIST.contains(&c) || c.is_digit(10)) {
-
+            // b 1, K 1024 (default)
-        let (num_str, si_unit) =
+            if size_string.ends_with(|c: char| c.is_digit(10)) {
-            if input.ends_with(|c: char| SI_UNITS.contains(&c.to_ascii_uppercase())) {
+                size_string.push('K');
-                let mut chars = input.chars();
+            } else if size_string.ends_with('b') {
-                let si_suffix = chars.next_back().unwrap().to_ascii_uppercase();
+                size_string.pop();
-                let si_unit = SI_UNITS.iter().position(|&c| c == si_suffix).unwrap();
+            }
-                let num_str = chars.as_str();
+            parse_size(&size_string)
-                (num_str, si_unit)
+        } else {
-            } else {
+            Err(ParseSizeError::ParseFailure("invalid suffix".to_string()))
-                (input, 1)
+        }
-            };
-
-        let num_usize: usize = num_str
-            .trim()
-            .parse()
-            .unwrap_or_else(|e| crash!(1, "failed to parse buffer size `{}`: {}", num_str, e));
-
-        num_usize.saturating_mul(1000usize.saturating_pow(si_unit as u32))
     }
 
     fn out_writer(&self) -> BufWriter<Box<dyn Write>> {
@@ -1176,8 +1174,10 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
 
     settings.buffer_size = matches
         .value_of(OPT_BUF_SIZE)
-        .map(GlobalSettings::parse_byte_count)
+        .map_or(DEFAULT_BUF_SIZE, |s| {
-        .unwrap_or(DEFAULT_BUF_SIZE);
+            GlobalSettings::parse_byte_count(s)
+                .unwrap_or_else(|e| crash!(2, "{}", format_error_message(e, s, OPT_BUF_SIZE)))
+        });
 
     settings.tmp_dir = matches
         .value_of(OPT_TMP_DIR)
@@ -1587,6 +1587,16 @@ fn open(path: impl AsRef<OsStr>) -> Box<dyn Read + Send> {
     }
 }
 
+fn format_error_message(error: ParseSizeError, s: &str, option: &str) -> String {
+    // NOTE:
+    // GNU's sort echos affected flag, -S or --buffer-size, depending user's selection
+    // GNU's sort does distinguish between "invalid (suffix in) argument"
+    match error {
+        ParseSizeError::ParseFailure(_) => format!("invalid --{} argument '{}'", option, s),
+        ParseSizeError::SizeTooBig(_) => format!("--{} argument '{}' too large", option, s),
+    }
+}
+
 #[cfg(test)]
 mod tests {
 
@@ -1676,4 +1686,48 @@ mod tests {
         // How big is a selection? Constant cost all lines pay when we need selections.
         assert_eq!(std::mem::size_of::<Selection>(), 24);
     }
+
+    #[test]
+    fn test_parse_byte_count() {
+        let valid_input = [
+            ("0", 0),
+            ("50K", 50 * 1024),
+            ("50k", 50 * 1024),
+            ("1M", 1024 * 1024),
+            ("100M", 100 * 1024 * 1024),
+            #[cfg(not(target_pointer_width = "32"))]
+            ("1000G", 1000 * 1024 * 1024 * 1024),
+            #[cfg(not(target_pointer_width = "32"))]
+            ("10T", 10 * 1024 * 1024 * 1024 * 1024),
+            ("1b", 1),
+            ("1024b", 1024),
+            ("1024Mb", 1024 * 1024 * 1024), // NOTE: This might not be how GNU `sort` behaves for 'Mb'
+            ("1", 1024),                    // K is default
+            ("50", 50 * 1024),
+            ("K", 1024),
+            ("k", 1024),
+            ("m", 1024 * 1024),
+            #[cfg(not(target_pointer_width = "32"))]
+            ("E", 1024 * 1024 * 1024 * 1024 * 1024 * 1024),
+        ];
+        for (input, expected_output) in &valid_input {
+            assert_eq!(
+                GlobalSettings::parse_byte_count(input),
+                Ok(*expected_output)
+            );
+        }
+
+        // SizeTooBig
+        let invalid_input = ["500E", "1Y"];
+        for input in &invalid_input {
+            #[cfg(not(target_pointer_width = "128"))]
+            assert!(GlobalSettings::parse_byte_count(input).is_err());
+        }
+
+        // ParseFailure
+        let invalid_input = ["nonsense", "1B", "B", "b", "p", "e", "z", "y"];
+        for input in &invalid_input {
+            assert!(GlobalSettings::parse_byte_count(input).is_err());
+        }
+    }
 }

src/uu/split/src/split.rs

@@ -13,11 +13,13 @@ extern crate uucore;
 mod platform;
 
 use clap::{crate_version, App, Arg};
+use std::convert::TryFrom;
 use std::env;
 use std::fs::File;
 use std::io::{stdin, BufRead, BufReader, BufWriter, Read, Write};
 use std::path::Path;
 use std::{char, fs::remove_file};
+use uucore::parse_size::parse_size;
 
 static NAME: &str = "split";
 
@@ -231,10 +233,9 @@ struct LineSplitter {
 impl LineSplitter {
     fn new(settings: &Settings) -> LineSplitter {
         LineSplitter {
-            lines_per_split: settings
+            lines_per_split: settings.strategy_param.parse().unwrap_or_else(|_| {
-                .strategy_param
+                crash!(1, "invalid number of lines: ‘{}’", settings.strategy_param)
-                .parse()
+            }),
-                .unwrap_or_else(|e| crash!(1, "invalid number of lines: {}", e)),
         }
     }
 }
@@ -276,40 +277,14 @@ struct ByteSplitter {
 
 impl ByteSplitter {
     fn new(settings: &Settings) -> ByteSplitter {
-        // These multipliers are the same as supported by GNU coreutils.
+        let size_string = &settings.strategy_param;
-        let modifiers: Vec<(&str, u128)> = vec![
+        let size_num = match parse_size(size_string) {
-            ("K", 1024u128),
+            Ok(n) => n,
-            ("M", 1024 * 1024),
+            Err(e) => crash!(1, "invalid number of bytes: {}", e.to_string()),
-            ("G", 1024 * 1024 * 1024),
+        };
-            ("T", 1024 * 1024 * 1024 * 1024),
-            ("P", 1024 * 1024 * 1024 * 1024 * 1024),
-            ("E", 1024 * 1024 * 1024 * 1024 * 1024 * 1024),
-            ("Z", 1024 * 1024 * 1024 * 1024 * 1024 * 1024 * 1024),
-            ("Y", 1024 * 1024 * 1024 * 1024 * 1024 * 1024 * 1024 * 1024),
-            ("KB", 1000),
-            ("MB", 1000 * 1000),
-            ("GB", 1000 * 1000 * 1000),
-            ("TB", 1000 * 1000 * 1000 * 1000),
-            ("PB", 1000 * 1000 * 1000 * 1000 * 1000),
-            ("EB", 1000 * 1000 * 1000 * 1000 * 1000 * 1000),
-            ("ZB", 1000 * 1000 * 1000 * 1000 * 1000 * 1000 * 1000),
-            ("YB", 1000 * 1000 * 1000 * 1000 * 1000 * 1000 * 1000 * 1000),
-        ];
-
-        // This sequential find is acceptable since none of the modifiers are
-        // suffixes of any other modifiers, a la Huffman codes.
-        let (suffix, multiplier) = modifiers
-            .iter()
-            .find(|(suffix, _)| settings.strategy_param.ends_with(suffix))
-            .unwrap_or(&("", 1));
-
-        // Try to parse the actual numeral.
-        let n = &settings.strategy_param[0..(settings.strategy_param.len() - suffix.len())]
-            .parse::<u128>()
-            .unwrap_or_else(|e| crash!(1, "invalid number of bytes: {}", e));
 
         ByteSplitter {
-            bytes_per_split: n * multiplier,
+            bytes_per_split: u128::try_from(size_num).unwrap(),
         }
     }
 }

src/uu/stdbuf/src/stdbuf.rs

@@ -19,6 +19,7 @@ use std::path::PathBuf;
 use std::process::Command;
 use tempfile::tempdir;
 use tempfile::TempDir;
+use uucore::parse_size::parse_size;
 use uucore::InvalidEncodingHandling;
 
 static ABOUT: &str =
@@ -55,7 +56,7 @@ const STDBUF_INJECT: &[u8] = include_bytes!(concat!(env!("OUT_DIR"), "/libstdbuf
 enum BufferType {
     Default,
     Line,
-    Size(u64),
+    Size(usize),
 }
 
 struct ProgramOptions {
@@ -104,41 +105,6 @@ fn preload_strings() -> (&'static str, &'static str) {
     crash!(1, "Command not supported for this operating system!")
 }
 
-fn parse_size(size: &str) -> Option<u64> {
-    let ext = size.trim_start_matches(|c: char| c.is_digit(10));
-    let num = size.trim_end_matches(char::is_alphabetic);
-    let mut recovered = num.to_owned();
-    recovered.push_str(ext);
-    if recovered != size {
-        return None;
-    }
-    let buf_size: u64 = match num.parse().ok() {
-        Some(m) => m,
-        None => return None,
-    };
-    let (power, base): (u32, u64) = match ext {
-        "" => (0, 0),
-        "KB" => (1, 1024),
-        "K" => (1, 1000),
-        "MB" => (2, 1024),
-        "M" => (2, 1000),
-        "GB" => (3, 1024),
-        "G" => (3, 1000),
-        "TB" => (4, 1024),
-        "T" => (4, 1000),
-        "PB" => (5, 1024),
-        "P" => (5, 1000),
-        "EB" => (6, 1024),
-        "E" => (6, 1000),
-        "ZB" => (7, 1024),
-        "Z" => (7, 1000),
-        "YB" => (8, 1024),
-        "Y" => (8, 1000),
-        _ => return None,
-    };
-    Some(buf_size * base.pow(power))
-}
-
 fn check_option(matches: &ArgMatches, name: &str) -> Result<BufferType, ProgramOptionsError> {
     match matches.value_of(name) {
         Some(value) => match value {
@@ -151,11 +117,10 @@ fn check_option(matches: &ArgMatches, name: &str) -> Result<BufferType, ProgramO
                     Ok(BufferType::Line)
                 }
             }
-            x => {
+            x => parse_size(x).map_or_else(
-                let size = parse_size(x)
+                |e| crash!(125, "invalid mode {}", e),
-                    .ok_or_else(|| ProgramOptionsError(format!("invalid mode {}", x)))?;
+                |m| Ok(BufferType::Size(m)),
-                Ok(BufferType::Size(size))
+            ),
-            }
         },
         None => Ok(BufferType::Default),
     }

src/uu/tail/src/tail.rs

@@ -5,7 +5,6 @@
 //  *
 //  * For the full copyright and license information, please view the LICENSE
 //  * file that was distributed with this source code.
-//  *
 
 // spell-checker:ignore (ToDO) seekable seek'd tail'ing ringbuffer ringbuf
 
@@ -21,19 +20,18 @@ use chunks::ReverseChunks;
 
 use clap::{App, Arg};
 use std::collections::VecDeque;
-use std::error::Error;
 use std::fmt;
 use std::fs::File;
 use std::io::{stdin, stdout, BufRead, BufReader, Read, Seek, SeekFrom, Write};
 use std::path::Path;
 use std::thread::sleep;
 use std::time::Duration;
+use uucore::parse_size::{parse_size, ParseSizeError};
 use uucore::ringbuffer::RingBuffer;
 
 pub mod options {
     pub mod verbosity {
         pub static QUIET: &str = "quiet";
-        pub static SILENT: &str = "silent";
         pub static VERBOSE: &str = "verbose";
     }
     pub static BYTES: &str = "bytes";
@@ -42,13 +40,12 @@ pub mod options {
     pub static PID: &str = "pid";
     pub static SLEEP_INT: &str = "sleep-interval";
     pub static ZERO_TERM: &str = "zero-terminated";
+    pub static ARG_FILES: &str = "files";
 }
 
-static ARG_FILES: &str = "files";
-
 enum FilterMode {
-    Bytes(u64),
+    Bytes(usize),
-    Lines(u64, u8), // (number of lines, delimiter)
+    Lines(usize, u8), // (number of lines, delimiter)
 }
 
 struct Settings {
@@ -78,12 +75,14 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     let app = App::new(executable!())
         .version(crate_version!())
         .about("output the last part of files")
+        // TODO: add usage
         .arg(
             Arg::with_name(options::BYTES)
                 .short("c")
                 .long(options::BYTES)
                 .takes_value(true)
                 .allow_hyphen_values(true)
+                .overrides_with_all(&[options::BYTES, options::LINES])
                 .help("Number of bytes to print"),
         )
         .arg(
@@ -98,6 +97,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 .long(options::LINES)
                 .takes_value(true)
                 .allow_hyphen_values(true)
+                .overrides_with_all(&[options::BYTES, options::LINES])
                 .help("Number of lines to print"),
         )
         .arg(
@@ -110,13 +110,10 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
             Arg::with_name(options::verbosity::QUIET)
                 .short("q")
                 .long(options::verbosity::QUIET)
+                .visible_alias("silent")
+                .overrides_with_all(&[options::verbosity::QUIET, options::verbosity::VERBOSE])
                 .help("never output headers giving file names"),
         )
-        .arg(
-            Arg::with_name(options::verbosity::SILENT)
-                .long(options::verbosity::SILENT)
-                .help("synonym of --quiet"),
-        )
         .arg(
             Arg::with_name(options::SLEEP_INT)
                 .short("s")
@@ -128,6 +125,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
             Arg::with_name(options::verbosity::VERBOSE)
                 .short("v")
                 .long(options::verbosity::VERBOSE)
+                .overrides_with_all(&[options::verbosity::QUIET, options::verbosity::VERBOSE])
                 .help("always output headers giving file names"),
         )
         .arg(
@@ -137,7 +135,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 .help("Line delimiter is NUL, not newline"),
         )
         .arg(
-            Arg::with_name(ARG_FILES)
+            Arg::with_name(options::ARG_FILES)
                 .multiple(true)
                 .takes_value(true)
                 .min_values(1),
@@ -171,38 +169,21 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         }
     }
 
-    match matches.value_of(options::LINES) {
+    let mode_and_beginning = if let Some(arg) = matches.value_of(options::BYTES) {
-        Some(n) => {
+        match parse_num(arg) {
-            let mut slice: &str = n;
+            Ok((n, beginning)) => (FilterMode::Bytes(n), beginning),
-            if slice.as_bytes().first() == Some(&b'+') {
+            Err(e) => crash!(1, "invalid number of bytes: {}", e.to_string()),
-                settings.beginning = true;
-                slice = &slice[1..];
-            }
-            match parse_size(slice) {
-                Ok(m) => settings.mode = FilterMode::Lines(m, b'\n'),
-                Err(e) => {
-                    show_error!("{}", e.to_string());
-                    return 1;
-                }
-            }
         }
-        None => {
+    } else if let Some(arg) = matches.value_of(options::LINES) {
-            if let Some(n) = matches.value_of(options::BYTES) {
+        match parse_num(arg) {
-                let mut slice: &str = n;
+            Ok((n, beginning)) => (FilterMode::Lines(n, b'\n'), beginning),
-                if slice.as_bytes().first() == Some(&b'+') {
+            Err(e) => crash!(1, "invalid number of lines: {}", e.to_string()),
-                    settings.beginning = true;
-                    slice = &slice[1..];
-                }
-                match parse_size(slice) {
-                    Ok(m) => settings.mode = FilterMode::Bytes(m),
-                    Err(e) => {
-                        show_error!("{}", e.to_string());
-                        return 1;
-                    }
-                }
-            }
         }
+    } else {
+        (FilterMode::Lines(10, b'\n'), false)
     };
+    settings.mode = mode_and_beginning.0;
+    settings.beginning = mode_and_beginning.1;
 
     if matches.is_present(options::ZERO_TERM) {
         if let FilterMode::Lines(count, _) = settings.mode {
@@ -211,11 +192,10 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     }
 
     let verbose = matches.is_present(options::verbosity::VERBOSE);
-    let quiet = matches.is_present(options::verbosity::QUIET)
+    let quiet = matches.is_present(options::verbosity::QUIET);
-        || matches.is_present(options::verbosity::SILENT);
 
     let files: Vec<String> = matches
-        .values_of(ARG_FILES)
+        .values_of(options::ARG_FILES)
         .map(|v| v.map(ToString::to_string).collect())
         .unwrap_or_default();
 
@@ -264,92 +244,6 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     0
 }
 
-#[derive(Debug, PartialEq, Eq)]
-pub enum ParseSizeErr {
-    ParseFailure(String),
-    SizeTooBig(String),
-}
-
-impl Error for ParseSizeErr {
-    fn description(&self) -> &str {
-        match *self {
-            ParseSizeErr::ParseFailure(ref s) => &*s,
-            ParseSizeErr::SizeTooBig(ref s) => &*s,
-        }
-    }
-}
-
-impl fmt::Display for ParseSizeErr {
-    fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
-        let s = match self {
-            ParseSizeErr::ParseFailure(s) => s,
-            ParseSizeErr::SizeTooBig(s) => s,
-        };
-        write!(f, "{}", s)
-    }
-}
-
-impl ParseSizeErr {
-    fn parse_failure(s: &str) -> ParseSizeErr {
-        ParseSizeErr::ParseFailure(format!("invalid size: '{}'", s))
-    }
-
-    fn size_too_big(s: &str) -> ParseSizeErr {
-        ParseSizeErr::SizeTooBig(format!(
-            "invalid size: '{}': Value too large to be stored in data type",
-            s
-        ))
-    }
-}
-
-pub type ParseSizeResult = Result<u64, ParseSizeErr>;
-
-pub fn parse_size(mut size_slice: &str) -> Result<u64, ParseSizeErr> {
-    let mut base = if size_slice.as_bytes().last() == Some(&b'B') {
-        size_slice = &size_slice[..size_slice.len() - 1];
-        1000u64
-    } else {
-        1024u64
-    };
-
-    let exponent = match size_slice.as_bytes().last() {
-        Some(unit) => match unit {
-            b'K' | b'k' => 1u64,
-            b'M' => 2u64,
-            b'G' => 3u64,
-            b'T' => 4u64,
-            b'P' => 5u64,
-            b'E' => 6u64,
-            b'Z' | b'Y' => {
-                return Err(ParseSizeErr::size_too_big(size_slice));
-            }
-            b'b' => {
-                base = 512u64;
-                1u64
-            }
-            _ => 0u64,
-        },
-        None => 0u64,
-    };
-    if exponent != 0 {
-        size_slice = &size_slice[..size_slice.len() - 1];
-    }
-
-    let mut multiplier = 1u64;
-    for _ in 0u64..exponent {
-        multiplier *= base;
-    }
-    if base == 1000u64 && exponent == 0u64 {
-        // sole B is not a valid suffix
-        Err(ParseSizeErr::parse_failure(size_slice))
-    } else {
-        let value: Option<i64> = size_slice.parse().ok();
-        value
-            .map(|v| Ok((multiplier as i64 * v.abs()) as u64))
-            .unwrap_or_else(|| Err(ParseSizeErr::parse_failure(size_slice)))
-    }
-}
-
 fn follow<T: Read>(readers: &mut [BufReader<T>], filenames: &[String], settings: &Settings) {
     assert!(settings.follow);
     let mut last = readers.len() - 1;
@@ -463,7 +357,7 @@ fn bounded_tail(file: &mut File, settings: &Settings) {
 /// If any element of `iter` is an [`Err`], then this function panics.
 fn unbounded_tail_collect<T, E>(
     iter: impl Iterator<Item = Result<T, E>>,
-    count: u64,
+    count: usize,
     beginning: bool,
 ) -> VecDeque<T>
 where
@@ -508,3 +402,22 @@ fn print_byte<T: Write>(stdout: &mut T, ch: u8) {
         crash!(1, "{}", err);
     }
 }
+
+fn parse_num(src: &str) -> Result<(usize, bool), ParseSizeError> {
+    let mut size_string = src.trim();
+    let mut starting_with = false;
+
+    if let Some(c) = size_string.chars().next() {
+        if c == '+' || c == '-' {
+            // tail: '-' is not documented (8.32 man pages)
+            size_string = &size_string[1..];
+            if c == '+' {
+                starting_with = true;
+            }
+        }
+    } else {
+        return Err(ParseSizeError::ParseFailure(src.to_string()));
+    }
+
+    parse_size(size_string).map(|n| (n, starting_with))
+}

src/uu/timeout/Cargo.toml

@@ -17,7 +17,7 @@ path = "src/timeout.rs"
 [dependencies]
 clap = "2.33"
 libc = "0.2.42"
-uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["parse_time", "process", "signals"] }
+uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["process", "signals"] }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
 
 

src/uu/truncate/src/truncate.rs

@@ -11,19 +11,21 @@
 extern crate uucore;
 
 use clap::{crate_version, App, Arg};
+use std::convert::TryFrom;
 use std::fs::{metadata, OpenOptions};
 use std::io::ErrorKind;
 use std::path::Path;
+use uucore::parse_size::{parse_size, ParseSizeError};
 
 #[derive(Debug, Eq, PartialEq)]
 enum TruncateMode {
-    Absolute(u64),
+    Absolute(usize),
-    Extend(u64),
+    Extend(usize),
-    Reduce(u64),
+    Reduce(usize),
-    AtMost(u64),
+    AtMost(usize),
-    AtLeast(u64),
+    AtLeast(usize),
-    RoundDown(u64),
+    RoundDown(usize),
-    RoundUp(u64),
+    RoundUp(usize),
 }
 
 impl TruncateMode {
@@ -38,7 +40,7 @@ impl TruncateMode {
     /// let fsize = 10;
     /// assert_eq!(mode.to_size(fsize), 15);
     /// ```
-    fn to_size(&self, fsize: u64) -> u64 {
+    fn to_size(&self, fsize: usize) -> usize {
         match self {
             TruncateMode::Absolute(size) => *size,
             TruncateMode::Extend(size) => fsize + size,
@@ -58,10 +60,9 @@ pub mod options {
     pub static NO_CREATE: &str = "no-create";
     pub static REFERENCE: &str = "reference";
     pub static SIZE: &str = "size";
+    pub static ARG_FILES: &str = "files";
 }
 
-static ARG_FILES: &str = "files";
-
 fn get_usage() -> String {
     format!("{0} [OPTION]... [FILE]...", executable!())
 }
@@ -113,21 +114,28 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
             Arg::with_name(options::REFERENCE)
             .short("r")
             .long(options::REFERENCE)
+            .required_unless(options::SIZE)
             .help("base the size of each file on the size of RFILE")
             .value_name("RFILE")
         )
         .arg(
             Arg::with_name(options::SIZE)
             .short("s")
-            .long("size")
+            .long(options::SIZE)
+            .required_unless(options::REFERENCE)
             .help("set or adjust the size of each file according to SIZE, which is in bytes unless --io-blocks is specified")
             .value_name("SIZE")
         )
-        .arg(Arg::with_name(ARG_FILES).multiple(true).takes_value(true).min_values(1))
+        .arg(Arg::with_name(options::ARG_FILES)
+             .value_name("FILE")
+             .multiple(true)
+             .takes_value(true)
+             .required(true)
+             .min_values(1))
         .get_matches_from(args);
 
     let files: Vec<String> = matches
-        .values_of(ARG_FILES)
+        .values_of(options::ARG_FILES)
         .map(|v| v.map(ToString::to_string).collect())
         .unwrap_or_default();
 
@@ -149,8 +157,8 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                     crash!(
                         1,
                         "cannot stat '{}': No such file or directory",
-                        reference.unwrap()
+                        reference.unwrap_or_else(|| "".to_string())
-                    );
+                    ); // TODO: fix '--no-create' see test_reference and test_truncate_bytes_size
                 }
                 _ => crash!(1, "{}", e.to_string()),
             }
@@ -172,10 +180,10 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
 ///
 /// If the file could not be opened, or there was a problem setting the
 /// size of the file.
-fn file_truncate(filename: &str, create: bool, size: u64) -> std::io::Result<()> {
+fn file_truncate(filename: &str, create: bool, size: usize) -> std::io::Result<()> {
     let path = Path::new(filename);
     let f = OpenOptions::new().write(true).create(create).open(path)?;
-    f.set_len(size)
+    f.set_len(u64::try_from(size).unwrap())
 }
 
 /// Truncate files to a size relative to a given file.
@@ -206,9 +214,9 @@ fn truncate_reference_and_size(
             }
             _ => m,
         },
-        Err(_) => crash!(1, "Invalid number: ‘{}’", size_string),
+        Err(e) => crash!(1, "Invalid number: {}", e.to_string()),
     };
-    let fsize = metadata(rfilename)?.len();
+    let fsize = usize::try_from(metadata(rfilename)?.len()).unwrap();
     let tsize = mode.to_size(fsize);
     for filename in &filenames {
         file_truncate(filename, create, tsize)?;
@@ -232,7 +240,7 @@ fn truncate_reference_file_only(
     filenames: Vec<String>,
     create: bool,
 ) -> std::io::Result<()> {
-    let tsize = metadata(rfilename)?.len();
+    let tsize = usize::try_from(metadata(rfilename)?.len()).unwrap();
     for filename in &filenames {
         file_truncate(filename, create, tsize)?;
     }
@@ -261,10 +269,10 @@ fn truncate_size_only(
 ) -> std::io::Result<()> {
     let mode = match parse_mode_and_size(size_string) {
         Ok(m) => m,
-        Err(_) => crash!(1, "Invalid number: ‘{}’", size_string),
+        Err(e) => crash!(1, "Invalid number: {}", e.to_string()),
     };
     for filename in &filenames {
-        let fsize = metadata(filename).map(|m| m.len()).unwrap_or(0);
+        let fsize = usize::try_from(metadata(filename)?.len()).unwrap();
         let tsize = mode.to_size(fsize);
         file_truncate(filename, create, tsize)?;
     }
@@ -290,7 +298,7 @@ fn truncate(
         }
         (Some(rfilename), None) => truncate_reference_file_only(&rfilename, filenames, create),
         (None, Some(size_string)) => truncate_size_only(&size_string, filenames, create),
-        (None, None) => crash!(1, "you must specify either --reference or --size"),
+        (None, None) => crash!(1, "you must specify either --reference or --size"), // this case cannot happen anymore because it's handled by clap
     }
 }
 
@@ -317,114 +325,35 @@ fn is_modifier(c: char) -> bool {
 /// ```rust,ignore
 /// assert_eq!(parse_mode_and_size("+123"), (TruncateMode::Extend, 123));
 /// ```
-fn parse_mode_and_size(size_string: &str) -> Result<TruncateMode, ()> {
+fn parse_mode_and_size(size_string: &str) -> Result<TruncateMode, ParseSizeError> {
     // Trim any whitespace.
-    let size_string = size_string.trim();
+    let mut size_string = size_string.trim();
 
     // Get the modifier character from the size string, if any. For
     // example, if the argument is "+123", then the modifier is '+'.
-    let c = size_string.chars().next().unwrap();
+    if let Some(c) = size_string.chars().next() {
-    let size_string = if is_modifier(c) {
+        if is_modifier(c) {
-        &size_string[1..]
+            size_string = &size_string[1..];
+        }
+        parse_size(size_string).map(match c {
+            '+' => TruncateMode::Extend,
+            '-' => TruncateMode::Reduce,
+            '<' => TruncateMode::AtMost,
+            '>' => TruncateMode::AtLeast,
+            '/' => TruncateMode::RoundDown,
+            '%' => TruncateMode::RoundUp,
+            _ => TruncateMode::Absolute,
+        })
     } else {
-        size_string
+        Err(ParseSizeError::ParseFailure(size_string.to_string()))
-    };
+    }
-    parse_size(size_string).map(match c {
-        '+' => TruncateMode::Extend,
-        '-' => TruncateMode::Reduce,
-        '<' => TruncateMode::AtMost,
-        '>' => TruncateMode::AtLeast,
-        '/' => TruncateMode::RoundDown,
-        '%' => TruncateMode::RoundUp,
-        _ => TruncateMode::Absolute,
-    })
-}
-
-/// Parse a size string into a number of bytes.
-///
-/// A size string comprises an integer and an optional unit. The unit
-/// may be K, M, G, T, P, E, Z, or Y (powers of 1024) or KB, MB,
-/// etc. (powers of 1000).
-///
-/// # Errors
-///
-/// This function returns an error if the string does not begin with a
-/// numeral, or if the unit is not one of the supported units described
-/// in the preceding section.
-///
-/// # Examples
-///
-/// ```rust,ignore
-/// assert_eq!(parse_size("123").unwrap(), 123);
-/// assert_eq!(parse_size("123K").unwrap(), 123 * 1024);
-/// assert_eq!(parse_size("123KB").unwrap(), 123 * 1000);
-/// ```
-fn parse_size(size: &str) -> Result<u64, ()> {
-    // Get the numeric part of the size argument. For example, if the
-    // argument is "123K", then the numeric part is "123".
-    let numeric_string: String = size.chars().take_while(|c| c.is_digit(10)).collect();
-    let number: u64 = numeric_string.parse().map_err(|_| ())?;
-
-    // Get the alphabetic units part of the size argument and compute
-    // the factor it represents. For example, if the argument is "123K",
-    // then the unit part is "K" and the factor is 1024. This may be the
-    // empty string, in which case, the factor is 1.
-    let n = numeric_string.len();
-    let (base, exponent): (u64, u32) = match &size[n..] {
-        "" => (1, 0),
-        "K" | "k" => (1024, 1),
-        "M" | "m" => (1024, 2),
-        "G" | "g" => (1024, 3),
-        "T" | "t" => (1024, 4),
-        "P" | "p" => (1024, 5),
-        "E" | "e" => (1024, 6),
-        "Z" | "z" => (1024, 7),
-        "Y" | "y" => (1024, 8),
-        "KB" | "kB" => (1000, 1),
-        "MB" | "mB" => (1000, 2),
-        "GB" | "gB" => (1000, 3),
-        "TB" | "tB" => (1000, 4),
-        "PB" | "pB" => (1000, 5),
-        "EB" | "eB" => (1000, 6),
-        "ZB" | "zB" => (1000, 7),
-        "YB" | "yB" => (1000, 8),
-        _ => return Err(()),
-    };
-    let factor = base.pow(exponent);
-    Ok(number * factor)
 }
 
 #[cfg(test)]
 mod tests {
     use crate::parse_mode_and_size;
-    use crate::parse_size;
     use crate::TruncateMode;
 
-    #[test]
-    fn test_parse_size_zero() {
-        assert_eq!(parse_size("0").unwrap(), 0);
-        assert_eq!(parse_size("0K").unwrap(), 0);
-        assert_eq!(parse_size("0KB").unwrap(), 0);
-    }
-
-    #[test]
-    fn test_parse_size_without_factor() {
-        assert_eq!(parse_size("123").unwrap(), 123);
-    }
-
-    #[test]
-    fn test_parse_size_kilobytes() {
-        assert_eq!(parse_size("123K").unwrap(), 123 * 1024);
-        assert_eq!(parse_size("123KB").unwrap(), 123 * 1000);
-    }
-
-    #[test]
-    fn test_parse_size_megabytes() {
-        assert_eq!(parse_size("123").unwrap(), 123);
-        assert_eq!(parse_size("123M").unwrap(), 123 * 1024 * 1024);
-        assert_eq!(parse_size("123MB").unwrap(), 123 * 1000 * 1000);
-    }
-
     #[test]
     fn test_parse_mode_and_size() {
         assert_eq!(parse_mode_and_size("10"), Ok(TruncateMode::Absolute(10)));

src/uucore/Cargo.toml

@@ -44,7 +44,6 @@ entries = ["libc"]
 fs = ["libc"]
 fsext = ["libc", "time"]
 mode = ["libc"]
-parse_time = []
 perms = ["libc"]
 process = ["libc"]
 ringbuffer = []

src/uucore/src/lib/features.rs

@@ -6,8 +6,6 @@ pub mod encoding;
 pub mod fs;
 #[cfg(feature = "fsext")]
 pub mod fsext;
-#[cfg(feature = "parse_time")]
-pub mod parse_time;
 #[cfg(feature = "ringbuffer")]
 pub mod ringbuffer;
 #[cfg(feature = "zero-copy")]

src/uucore/src/lib/lib.rs

@@ -19,10 +19,10 @@ pub extern crate winapi;
 
 //## internal modules
 
-mod macros; // crate macros (macro_rules-type; exported to `crate::...`)
-
 mod features; // feature-gated code modules
+mod macros; // crate macros (macro_rules-type; exported to `crate::...`)
 mod mods; // core cross-platform modules
+mod parser; // string parsing modules
 
 // * cross-platform modules
 pub use crate::mods::backup_control;
@@ -31,6 +31,10 @@ pub use crate::mods::os;
 pub use crate::mods::panic;
 pub use crate::mods::ranges;
 
+// * string parsing modules
+pub use crate::parser::parse_size;
+pub use crate::parser::parse_time;
+
 // * feature-gated modules
 #[cfg(feature = "encoding")]
 pub use crate::features::encoding;
@@ -38,8 +42,6 @@ pub use crate::features::encoding;
 pub use crate::features::fs;
 #[cfg(feature = "fsext")]
 pub use crate::features::fsext;
-#[cfg(feature = "parse_time")]
-pub use crate::features::parse_time;
 #[cfg(feature = "ringbuffer")]
 pub use crate::features::ringbuffer;
 #[cfg(feature = "zero-copy")]

src/uucore/src/lib/parser.rs

@@ -0,0 +1,2 @@
+pub mod parse_size;
+pub mod parse_time;

src/uucore/src/lib/parser/parse_size.rs

@@ -0,0 +1,324 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
+// spell-checker:ignore (ToDO) hdsf ghead gtail
+
+use std::convert::TryFrom;
+use std::error::Error;
+use std::fmt;
+
+/// Parse a size string into a number of bytes.
+///
+/// A size string comprises an integer and an optional unit. The unit
+/// may be K, M, G, T, P, E, Z or Y (powers of 1024), or KB, MB,
+/// etc. (powers of 1000), or b which is 512.
+/// Binary prefixes can be used, too: KiB=K, MiB=M, and so on.
+///
+/// # Errors
+///
+/// Will return `ParseSizeError` if it’s not possible to parse this
+/// string into a number, e.g. if the string does not begin with a
+/// numeral, or if the unit is not one of the supported units described
+/// in the preceding section.
+///
+/// # Examples
+///
+/// ```rust
+/// use uucore::parse_size::parse_size;
+/// assert_eq!(Ok(123), parse_size("123"));
+/// assert_eq!(Ok(9 * 1000), parse_size("9kB")); // kB is 1000
+/// assert_eq!(Ok(2 * 1024), parse_size("2K")); // K is 1024
+/// ```
+pub fn parse_size(size: &str) -> Result<usize, ParseSizeError> {
+    if size.is_empty() {
+        return Err(ParseSizeError::parse_failure(size));
+    }
+    // Get the numeric part of the size argument. For example, if the
+    // argument is "123K", then the numeric part is "123".
+    let numeric_string: String = size.chars().take_while(|c| c.is_digit(10)).collect();
+    let number: usize = if !numeric_string.is_empty() {
+        match numeric_string.parse() {
+            Ok(n) => n,
+            Err(_) => return Err(ParseSizeError::parse_failure(size)),
+        }
+    } else {
+        1
+    };
+
+    // Get the alphabetic units part of the size argument and compute
+    // the factor it represents. For example, if the argument is "123K",
+    // then the unit part is "K" and the factor is 1024. This may be the
+    // empty string, in which case, the factor is 1.
+    let unit = &size[numeric_string.len()..];
+    let (base, exponent): (u128, u32) = match unit {
+        "" => (1, 0),
+        "b" => (512, 1), // (`od`, `head` and `tail` use "b")
+        "KiB" | "kiB" | "K" | "k" => (1024, 1),
+        "MiB" | "miB" | "M" | "m" => (1024, 2),
+        "GiB" | "giB" | "G" | "g" => (1024, 3),
+        "TiB" | "tiB" | "T" | "t" => (1024, 4),
+        "PiB" | "piB" | "P" | "p" => (1024, 5),
+        "EiB" | "eiB" | "E" | "e" => (1024, 6),
+        "ZiB" | "ziB" | "Z" | "z" => (1024, 7),
+        "YiB" | "yiB" | "Y" | "y" => (1024, 8),
+        "KB" | "kB" => (1000, 1),
+        "MB" | "mB" => (1000, 2),
+        "GB" | "gB" => (1000, 3),
+        "TB" | "tB" => (1000, 4),
+        "PB" | "pB" => (1000, 5),
+        "EB" | "eB" => (1000, 6),
+        "ZB" | "zB" => (1000, 7),
+        "YB" | "yB" => (1000, 8),
+        _ => return Err(ParseSizeError::parse_failure(size)),
+    };
+    let factor = match usize::try_from(base.pow(exponent)) {
+        Ok(n) => n,
+        Err(_) => return Err(ParseSizeError::size_too_big(size)),
+    };
+    number
+        .checked_mul(factor)
+        .ok_or_else(|| ParseSizeError::size_too_big(size))
+}
+
+#[derive(Debug, PartialEq, Eq)]
+pub enum ParseSizeError {
+    ParseFailure(String), // Syntax
+    SizeTooBig(String),   // Overflow
+}
+
+impl Error for ParseSizeError {
+    fn description(&self) -> &str {
+        match *self {
+            ParseSizeError::ParseFailure(ref s) => &*s,
+            ParseSizeError::SizeTooBig(ref s) => &*s,
+        }
+    }
+}
+
+impl fmt::Display for ParseSizeError {
+    fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
+        let s = match self {
+            ParseSizeError::ParseFailure(s) => s,
+            ParseSizeError::SizeTooBig(s) => s,
+        };
+        write!(f, "{}", s)
+    }
+}
+
+impl ParseSizeError {
+    fn parse_failure(s: &str) -> ParseSizeError {
+        // stderr on linux (GNU coreutils 8.32)
+        // has to be handled in the respective uutils because strings differ, e.g.:
+        //
+        // `NUM`
+        // head:     invalid number of bytes: ‘1fb’
+        // tail:     invalid number of bytes: ‘1fb’
+        //
+        // `SIZE`
+        // split:    invalid number of bytes: ‘1fb’
+        // truncate: Invalid number: ‘1fb’
+        //
+        // `MODE`
+        // stdbuf:   invalid mode ‘1fb’
+        //
+        // `SIZE`
+        // sort:     invalid suffix in --buffer-size argument '1fb'
+        // sort:     invalid --buffer-size argument 'fb'
+        //
+        // `SIZE`
+        // du:       invalid suffix in --buffer-size argument '1fb'
+        // du:       invalid suffix in --threshold argument '1fb'
+        // du:       invalid --buffer-size argument 'fb'
+        // du:       invalid --threshold argument 'fb'
+        //
+        // `BYTES`
+        // od:       invalid suffix in --read-bytes argument '1fb'
+        // od:       invalid --read-bytes argument  argument 'fb'
+        //                   --skip-bytes
+        //                   --width
+        //                   --strings
+        // etc.
+        ParseSizeError::ParseFailure(format!("‘{}’", s))
+    }
+
+    fn size_too_big(s: &str) -> ParseSizeError {
+        // stderr on linux (GNU coreutils 8.32)
+        // has to be handled in the respective uutils because strings differ, e.g.:
+        //
+        // head:     invalid number of bytes: ‘1Y’: Value too large for defined data type
+        // tail:     invalid number of bytes: ‘1Y’: Value too large for defined data type
+        // split:    invalid number of bytes: ‘1Y’: Value too large for defined data type
+        // truncate:          Invalid number: ‘1Y’: Value too large for defined data type
+        // stdbuf:               invalid mode ‘1Y’: Value too large for defined data type
+        // sort:     -S argument '1Y' too large
+        // du:       -B argument '1Y' too large
+        // od:       -N argument '1Y' too large
+        // etc.
+        //
+        // stderr on macos (brew - GNU coreutils 8.32) also differs for the same version, e.g.:
+        // ghead:   invalid number of bytes: ‘1Y’: Value too large to be stored in data type
+        // gtail:   invalid number of bytes: ‘1Y’: Value too large to be stored in data type
+        ParseSizeError::SizeTooBig(format!("‘{}’: Value too large for defined data type", s))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    fn variant_eq(a: &ParseSizeError, b: &ParseSizeError) -> bool {
+        std::mem::discriminant(a) == std::mem::discriminant(b)
+    }
+
+    #[test]
+    fn all_suffixes() {
+        // Units  are  K,M,G,T,P,E,Z,Y  (powers  of 1024) or KB,MB,... (powers of 1000).
+        // Binary prefixes can be used, too: KiB=K, MiB=M, and so on.
+        let suffixes = [
+            ('K', 1u32),
+            ('M', 2u32),
+            ('G', 3u32),
+            ('T', 4u32),
+            ('P', 5u32),
+            ('E', 6u32),
+            #[cfg(target_pointer_width = "128")]
+            ('Z', 7u32), // ParseSizeError::SizeTooBig on x64
+            #[cfg(target_pointer_width = "128")]
+            ('Y', 8u32), // ParseSizeError::SizeTooBig on x64
+        ];
+
+        for &(c, exp) in &suffixes {
+            let s = format!("2{}B", c); // KB
+            assert_eq!(Ok((2 * (1000_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("2{}", c); // K
+            assert_eq!(Ok((2 * (1024_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("2{}iB", c); // KiB
+            assert_eq!(Ok((2 * (1024_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("2{}iB", c.to_lowercase()); // kiB
+            assert_eq!(Ok((2 * (1024_u128).pow(exp)) as usize), parse_size(&s));
+
+            // suffix only
+            let s = format!("{}B", c); // KB
+            assert_eq!(Ok(((1000_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("{}", c); // K
+            assert_eq!(Ok(((1024_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("{}iB", c); // KiB
+            assert_eq!(Ok(((1024_u128).pow(exp)) as usize), parse_size(&s));
+            let s = format!("{}iB", c.to_lowercase()); // kiB
+            assert_eq!(Ok(((1024_u128).pow(exp)) as usize), parse_size(&s));
+        }
+    }
+
+    #[test]
+    #[cfg(not(target_pointer_width = "128"))]
+    fn overflow_x64() {
+        assert!(parse_size("10000000000000000000000").is_err());
+        assert!(parse_size("1000000000T").is_err());
+        assert!(parse_size("100000P").is_err());
+        assert!(parse_size("100E").is_err());
+        assert!(parse_size("1Z").is_err());
+        assert!(parse_size("1Y").is_err());
+
+        assert!(variant_eq(
+            &parse_size("1Z").unwrap_err(),
+            &ParseSizeError::SizeTooBig(String::new())
+        ));
+
+        assert_eq!(
+            ParseSizeError::SizeTooBig("‘1Y’: Value too large for defined data type".to_string()),
+            parse_size("1Y").unwrap_err()
+        );
+    }
+
+    #[test]
+    #[cfg(target_pointer_width = "32")]
+    fn overflow_x32() {
+        assert!(variant_eq(
+            &parse_size("1T").unwrap_err(),
+            &ParseSizeError::SizeTooBig(String::new())
+        ));
+        assert!(variant_eq(
+            &parse_size("1000G").unwrap_err(),
+            &ParseSizeError::SizeTooBig(String::new())
+        ));
+    }
+
+    #[test]
+    fn invalid_syntax() {
+        let test_strings = [
+            "328hdsf3290",
+            "5MiB nonsense",
+            "5mib",
+            "biB",
+            "-",
+            "+",
+            "",
+            "-1",
+            "1e2",
+            "∞",
+        ];
+        for &test_string in &test_strings {
+            assert_eq!(
+                parse_size(test_string).unwrap_err(),
+                ParseSizeError::ParseFailure(format!("‘{}’", test_string))
+            );
+        }
+    }
+
+    #[test]
+    fn b_suffix() {
+        assert_eq!(Ok(3 * 512), parse_size("3b")); // b is 512
+    }
+
+    #[test]
+    fn no_suffix() {
+        assert_eq!(Ok(1234), parse_size("1234"));
+        assert_eq!(Ok(0), parse_size("0"));
+        assert_eq!(Ok(5), parse_size("5"));
+        assert_eq!(Ok(999), parse_size("999"));
+    }
+
+    #[test]
+    fn kilobytes_suffix() {
+        assert_eq!(Ok(123 * 1000), parse_size("123KB")); // KB is 1000
+        assert_eq!(Ok(9 * 1000), parse_size("9kB")); // kB is 1000
+        assert_eq!(Ok(2 * 1024), parse_size("2K")); // K is 1024
+        assert_eq!(Ok(0), parse_size("0K"));
+        assert_eq!(Ok(0), parse_size("0KB"));
+        assert_eq!(Ok(1000), parse_size("KB"));
+        assert_eq!(Ok(1024), parse_size("K"));
+        assert_eq!(Ok(2000), parse_size("2kB"));
+        assert_eq!(Ok(4000), parse_size("4KB"));
+    }
+
+    #[test]
+    fn megabytes_suffix() {
+        assert_eq!(Ok(123 * 1024 * 1024), parse_size("123M"));
+        assert_eq!(Ok(123 * 1000 * 1000), parse_size("123MB"));
+        assert_eq!(Ok(1024 * 1024), parse_size("M"));
+        assert_eq!(Ok(1000 * 1000), parse_size("MB"));
+        assert_eq!(Ok(2 * 1_048_576), parse_size("2m"));
+        assert_eq!(Ok(4 * 1_048_576), parse_size("4M"));
+        assert_eq!(Ok(2_000_000), parse_size("2mB"));
+        assert_eq!(Ok(4_000_000), parse_size("4MB"));
+    }
+
+    #[test]
+    fn gigabytes_suffix() {
+        assert_eq!(Ok(1_073_741_824), parse_size("1G"));
+        assert_eq!(Ok(2_000_000_000), parse_size("2GB"));
+    }
+
+    #[test]
+    #[cfg(target_pointer_width = "64")]
+    fn x64() {
+        assert_eq!(Ok(1_099_511_627_776), parse_size("1T"));
+        assert_eq!(Ok(1_125_899_906_842_624), parse_size("1P"));
+        assert_eq!(Ok(1_152_921_504_606_846_976), parse_size("1E"));
+        assert_eq!(Ok(2_000_000_000_000), parse_size("2TB"));
+        assert_eq!(Ok(2_000_000_000_000_000), parse_size("2PB"));
+        assert_eq!(Ok(2_000_000_000_000_000_000), parse_size("2EB"));
+    }
+}

tests/by-util/test_du.rs

@@ -1,3 +1,8 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 // spell-checker:ignore (paths) sublink subwords
 
 use crate::common::util::*;
@@ -73,6 +78,23 @@ fn _du_basics_subdir(s: &str) {
     }
 }
 
+#[test]
+fn test_du_invalid_size() {
+    new_ucmd!()
+        .arg("--block-size=1fb4t")
+        .arg("/tmp")
+        .fails()
+        .code_is(1)
+        .stderr_only("du: invalid --block-size argument '1fb4t'");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .arg("--block-size=1Y")
+        .arg("/tmp")
+        .fails()
+        .code_is(1)
+        .stderr_only("du: --block-size argument '1Y' too large");
+}
+
 #[test]
 fn test_du_basics_bad_name() {
     new_ucmd!()

tests/by-util/test_head.rs

@@ -1,4 +1,9 @@
-// spell-checker:ignore (words) bogusfile emptyfile
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
+// spell-checker:ignore (words) bogusfile emptyfile abcdefghijklmnopqrstuvwxyz abcdefghijklmnopqrstu
 
 use crate::common::util::*;
 
@@ -244,3 +249,61 @@ hello
 ",
         );
 }
+
+#[test]
+fn test_head_invalid_num() {
+    new_ucmd!()
+        .args(&["-c", "1024R", "emptyfile.txt"])
+        .fails()
+        .stderr_is("head: invalid number of bytes: ‘1024R’");
+    new_ucmd!()
+        .args(&["-n", "1024R", "emptyfile.txt"])
+        .fails()
+        .stderr_is("head: invalid number of lines: ‘1024R’");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["-c", "1Y", "emptyfile.txt"])
+        .fails()
+        .stderr_is("head: invalid number of bytes: ‘1Y’: Value too large for defined data type");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["-n", "1Y", "emptyfile.txt"])
+        .fails()
+        .stderr_is("head: invalid number of lines: ‘1Y’: Value too large for defined data type");
+    #[cfg(target_pointer_width = "32")]
+    {
+        let sizes = ["1000G", "10T"];
+        for size in &sizes {
+            new_ucmd!()
+                .args(&["-c", size])
+                .fails()
+                .code_is(1)
+                .stderr_only(format!(
+                    "head: invalid number of bytes: ‘{}’: Value too large for defined data type",
+                    size
+                ));
+        }
+    }
+}
+
+#[test]
+fn test_head_num_with_undocumented_sign_bytes() {
+    // tail: '-' is not documented (8.32 man pages)
+    // head: '+' is not documented (8.32 man pages)
+    const ALPHABET: &str = "abcdefghijklmnopqrstuvwxyz";
+    new_ucmd!()
+        .args(&["-c", "5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("abcde");
+    new_ucmd!()
+        .args(&["-c", "-5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("abcdefghijklmnopqrstu");
+    new_ucmd!()
+        .args(&["-c", "+5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("abcde");
+}

tests/by-util/test_od.rs

@@ -1,3 +1,8 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 extern crate unindent;
 
 use self::unindent::*;
@@ -804,3 +809,40 @@ fn test_traditional_only_label() {
             ",
         ));
 }
+
+#[test]
+fn test_od_invalid_bytes() {
+    const INVALID_SIZE: &str = "1fb4t";
+    const BIG_SIZE: &str = "1Y";
+
+    // NOTE:
+    // GNU's od (8.32) with option '--width' does not accept 'Y' as valid suffix.
+    // According to the man page it should be valid in the same way it is valid for
+    // '--read-bytes' and '--skip-bytes'.
+
+    let options = [
+        "--read-bytes",
+        "--skip-bytes",
+        "--width",
+        // "--strings", // TODO: consider testing here once '--strings' is implemented
+    ];
+    for option in &options {
+        new_ucmd!()
+            .arg(format!("{}={}", option, INVALID_SIZE))
+            .arg("file")
+            .fails()
+            .code_is(1)
+            .stderr_only(format!(
+                "od: invalid {} argument '{}'",
+                option, INVALID_SIZE
+            ));
+
+        #[cfg(not(target_pointer_width = "128"))]
+        new_ucmd!()
+            .arg(format!("{}={}", option, BIG_SIZE))
+            .arg("file")
+            .fails()
+            .code_is(1)
+            .stderr_only(format!("od: {} argument '{}' too large", option, BIG_SIZE));
+    }
+}

tests/by-util/test_sort.rs

@@ -1,3 +1,8 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 // spell-checker:ignore (words) ints
 
 use crate::common::util::*;
@@ -21,9 +26,7 @@ fn test_helper(file_name: &str, possible_args: &[&str]) {
 
 #[test]
 fn test_buffer_sizes() {
-    let buffer_sizes = [
+    let buffer_sizes = ["0", "50K", "50k", "1M", "100M"];
-        "0", "50K", "50k", "1M", "100M", "1000G", "10T", "500E", "1Y",
-    ];
     for buffer_size in &buffer_sizes {
         new_ucmd!()
             .arg("-n")
@@ -32,6 +35,20 @@ fn test_buffer_sizes() {
             .arg("ext_sort.txt")
             .succeeds()
             .stdout_is_fixture("ext_sort.expected");
+
+        #[cfg(not(target_pointer_width = "32"))]
+        {
+            let buffer_sizes = ["1000G", "10T"];
+            for buffer_size in &buffer_sizes {
+                new_ucmd!()
+                    .arg("-n")
+                    .arg("-S")
+                    .arg(buffer_size)
+                    .arg("ext_sort.txt")
+                    .succeeds()
+                    .stdout_is_fixture("ext_sort.expected");
+            }
+        }
     }
 }
 
@@ -43,11 +60,39 @@ fn test_invalid_buffer_size() {
             .arg("-S")
             .arg(invalid_buffer_size)
             .fails()
+            .code_is(2)
             .stderr_only(format!(
-                "sort: failed to parse buffer size `{}`: invalid digit found in string",
+                "sort: invalid --buffer-size argument '{}'",
                 invalid_buffer_size
             ));
     }
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .arg("-n")
+        .arg("-S")
+        .arg("1Y")
+        .arg("ext_sort.txt")
+        .fails()
+        .code_is(2)
+        .stderr_only("sort: --buffer-size argument '1Y' too large");
+
+    #[cfg(target_pointer_width = "32")]
+    {
+        let buffer_sizes = ["1000G", "10T"];
+        for buffer_size in &buffer_sizes {
+            new_ucmd!()
+                .arg("-n")
+                .arg("-S")
+                .arg(buffer_size)
+                .arg("ext_sort.txt")
+                .fails()
+                .code_is(2)
+                .stderr_only(format!(
+                    "sort: --buffer-size argument '{}' too large",
+                    buffer_size
+                ));
+        }
+    }
 }
 
 #[test]
@@ -804,7 +849,7 @@ fn sort_multiple() {
 #[test]
 fn sort_empty_chunk() {
     new_ucmd!()
-        .args(&["-S", "40B"])
+        .args(&["-S", "40b"])
         .pipe_in("a\na\n")
         .succeeds()
         .stdout_is("a\na\n");
@@ -844,7 +889,7 @@ fn test_compress_fail() {
 #[test]
 fn test_merge_batches() {
     new_ucmd!()
-        .args(&["ext_sort.txt", "-n", "-S", "150B"])
+        .args(&["ext_sort.txt", "-n", "-S", "150b"])
         .succeeds()
         .stdout_only_fixture("ext_sort.expected");
 }

tests/by-util/test_split.rs

@@ -1,3 +1,8 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
 extern crate rand;
 extern crate regex;
 
@@ -285,3 +290,53 @@ fn test_filter_command_fails() {
     ucmd.args(&["--filter=/a/path/that/totally/does/not/exist", name])
         .fails();
 }
+
+#[test]
+fn test_split_lines_number() {
+    // Test if stdout/stderr for '--lines' option is correct
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+    at.touch("file");
+
+    scene
+        .ucmd()
+        .args(&["--lines", "2", "file"])
+        .succeeds()
+        .no_stderr()
+        .no_stdout();
+    scene
+        .ucmd()
+        .args(&["--lines", "2fb", "file"])
+        .fails()
+        .code_is(1)
+        .stderr_only("split: invalid number of lines: ‘2fb’");
+}
+
+#[test]
+fn test_split_invalid_bytes_size() {
+    new_ucmd!()
+        .args(&["-b", "1024R"])
+        .fails()
+        .code_is(1)
+        .stderr_only("split: invalid number of bytes: ‘1024R’");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["-b", "1Y"])
+        .fails()
+        .code_is(1)
+        .stderr_only("split: invalid number of bytes: ‘1Y’: Value too large for defined data type");
+    #[cfg(target_pointer_width = "32")]
+    {
+        let sizes = ["1000G", "10T"];
+        for size in &sizes {
+            new_ucmd!()
+                .args(&["-b", size])
+                .fails()
+                .code_is(1)
+                .stderr_only(format!(
+                    "split: invalid number of bytes: ‘{}’: Value too large for defined data type",
+                    size
+                ));
+        }
+    }
+}

tests/by-util/test_stdbuf.rs

@@ -57,8 +57,18 @@ fn test_stdbuf_line_buffering_stdin_fails() {
 #[cfg(not(target_os = "windows"))]
 #[test]
 fn test_stdbuf_invalid_mode_fails() {
-    new_ucmd!()
+    let options = ["--input", "--output", "--error"];
-        .args(&["-i", "1024R", "head"])
+    for option in &options {
-        .fails()
+        new_ucmd!()
-        .stderr_is("stdbuf: invalid mode 1024R\nTry 'stdbuf --help' for more information.");
+            .args(&[*option, "1024R", "head"])
+            .fails()
+            .code_is(125)
+            .stderr_only("stdbuf: invalid mode ‘1024R’");
+        #[cfg(not(target_pointer_width = "128"))]
+        new_ucmd!()
+            .args(&[*option, "1Y", "head"])
+            .fails()
+            .code_is(125)
+            .stderr_contains("stdbuf: invalid mode ‘1Y’: Value too large for defined data type");
+    }
 }

tests/by-util/test_tail.rs

@@ -1,6 +1,12 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
+// spell-checker:ignore (ToDO) abcdefghijklmnopqrstuvwxyz efghijklmnopqrstuvwxyz vwxyz emptyfile
+
 extern crate tail;
 
-use self::tail::parse_size;
 use crate::common::util::*;
 use std::char::from_digit;
 use std::io::Write;
@@ -236,41 +242,6 @@ fn test_bytes_big() {
     }
 }
 
-#[test]
-fn test_parse_size() {
-    // No suffix.
-    assert_eq!(Ok(1234), parse_size("1234"));
-
-    // kB is 1000
-    assert_eq!(Ok(9 * 1000), parse_size("9kB"));
-
-    // K is 1024
-    assert_eq!(Ok(2 * 1024), parse_size("2K"));
-
-    let suffixes = [
-        ('M', 2u32),
-        ('G', 3u32),
-        ('T', 4u32),
-        ('P', 5u32),
-        ('E', 6u32),
-    ];
-
-    for &(c, exp) in &suffixes {
-        let s = format!("2{}B", c);
-        assert_eq!(Ok(2 * (1000_u64).pow(exp)), parse_size(&s));
-
-        let s = format!("2{}", c);
-        assert_eq!(Ok(2 * (1024_u64).pow(exp)), parse_size(&s));
-    }
-
-    // Sizes that are too big.
-    assert!(parse_size("1Z").is_err());
-    assert!(parse_size("1Y").is_err());
-
-    // Bad number
-    assert!(parse_size("328hdsf3290").is_err()); // spell-checker:disable-line
-}
-
 #[test]
 fn test_lines_with_size_suffix() {
     const FILE: &str = "test_lines_with_size_suffix.txt";
@@ -320,12 +291,11 @@ fn test_multiple_input_files_with_suppressed_headers() {
 
 #[test]
 fn test_multiple_input_quiet_flag_overrides_verbose_flag_for_suppressing_headers() {
-    // TODO: actually the later one should win, i.e. -qv should lead to headers being printed, -vq to them being suppressed
     new_ucmd!()
         .arg(FOOBAR_TXT)
         .arg(FOOBAR_2_TXT)
-        .arg("-q")
         .arg("-v")
+        .arg("-q")
         .run()
         .stdout_is_fixture("foobar_multiple_quiet.expected");
 }
@@ -388,3 +358,61 @@ fn test_positive_zero_lines() {
         .succeeds()
         .stdout_is("a\nb\nc\nd\ne\n");
 }
+
+#[test]
+fn test_tail_invalid_num() {
+    new_ucmd!()
+        .args(&["-c", "1024R", "emptyfile.txt"])
+        .fails()
+        .stderr_is("tail: invalid number of bytes: ‘1024R’");
+    new_ucmd!()
+        .args(&["-n", "1024R", "emptyfile.txt"])
+        .fails()
+        .stderr_is("tail: invalid number of lines: ‘1024R’");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["-c", "1Y", "emptyfile.txt"])
+        .fails()
+        .stderr_is("tail: invalid number of bytes: ‘1Y’: Value too large for defined data type");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["-n", "1Y", "emptyfile.txt"])
+        .fails()
+        .stderr_is("tail: invalid number of lines: ‘1Y’: Value too large for defined data type");
+    #[cfg(target_pointer_width = "32")]
+    {
+        let sizes = ["1000G", "10T"];
+        for size in &sizes {
+            new_ucmd!()
+                .args(&["-c", size])
+                .fails()
+                .code_is(1)
+                .stderr_only(format!(
+                    "tail: invalid number of bytes: ‘{}’: Value too large for defined data type",
+                    size
+                ));
+        }
+    }
+}
+
+#[test]
+fn test_tail_num_with_undocumented_sign_bytes() {
+    // tail: '-' is not documented (8.32 man pages)
+    // head: '+' is not documented (8.32 man pages)
+    const ALPHABET: &str = "abcdefghijklmnopqrstuvwxyz";
+    new_ucmd!()
+        .args(&["-c", "5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("vwxyz");
+    new_ucmd!()
+        .args(&["-c", "-5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("vwxyz");
+    new_ucmd!()
+        .args(&["-c", "+5"])
+        .pipe_in(ALPHABET)
+        .succeeds()
+        .stdout_is("efghijklmnopqrstuvwxyz");
+}

tests/by-util/test_truncate.rs

@@ -1,3 +1,10 @@
+//  * This file is part of the uutils coreutils package.
+//  *
+//  * For the full copyright and license information, please view the LICENSE
+//  * file that was distributed with this source code.
+
+// spell-checker:ignore (words) RFILE
+
 use crate::common::util::*;
 use std::io::{Seek, SeekFrom, Write};
 
@@ -45,9 +52,18 @@ fn test_reference() {
     let at = &scene.fixtures;
     let mut file = at.make_file(FILE2);
 
-    scene.ucmd().arg("-s").arg("+5KB").arg(FILE1).run();
+    // manpage: "A FILE argument that does not exist is created."
+    // TODO: 'truncate' does not create the file in this case,
+    //        but should because '--no-create' wasn't specified.
+    at.touch(FILE1); // TODO: remove this when 'no-create' is fixed
+    scene.ucmd().arg("-s").arg("+5KB").arg(FILE1).succeeds();
 
-    scene.ucmd().arg("--reference").arg(FILE1).arg(FILE2).run();
+    scene
+        .ucmd()
+        .arg("--reference")
+        .arg(FILE1)
+        .arg(FILE2)
+        .succeeds();
 
     file.seek(SeekFrom::End(0)).unwrap();
     let actual = file.seek(SeekFrom::Current(0)).unwrap();
@@ -231,11 +247,18 @@ fn test_size_and_reference() {
     );
 }
 
+#[test]
+fn test_error_filename_only() {
+    // truncate: you must specify either ‘--size’ or ‘--reference’
+    new_ucmd!().args(&["file"]).fails().stderr_contains(
+        "error: The following required arguments were not provided:
+    --reference <RFILE>
+    --size <SIZE>",
+    );
+}
+
 #[test]
 fn test_invalid_numbers() {
-    // TODO For compatibility with GNU, `truncate -s 0X` should cause
-    // the same error as `truncate -s 0X file`, but currently it returns
-    // a different error.
     new_ucmd!()
         .args(&["-s", "0X", "file"])
         .fails()
@@ -265,3 +288,36 @@ fn test_reference_with_size_file_not_found() {
         .fails()
         .stderr_contains("cannot stat 'a': No such file or directory");
 }
+
+#[test]
+fn test_truncate_bytes_size() {
+    // TODO: this should succeed without error, uncomment when '--no-create' is fixed
+    // new_ucmd!()
+    //     .args(&["--no-create", "--size", "K", "file"])
+    //     .succeeds();
+    new_ucmd!()
+        .args(&["--size", "1024R", "file"])
+        .fails()
+        .code_is(1)
+        .stderr_only("truncate: Invalid number: ‘1024R’");
+    #[cfg(not(target_pointer_width = "128"))]
+    new_ucmd!()
+        .args(&["--size", "1Y", "file"])
+        .fails()
+        .code_is(1)
+        .stderr_only("truncate: Invalid number: ‘1Y’: Value too large for defined data type");
+    #[cfg(target_pointer_width = "32")]
+    {
+        let sizes = ["1000G", "10T"];
+        for size in &sizes {
+            new_ucmd!()
+                .args(&["--size", size, "file"])
+                .fails()
+                .code_is(1)
+                .stderr_only(format!(
+                    "truncate: Invalid number: ‘{}’: Value too large for defined data type",
+                    size
+                ));
+        }
+    }
+}