Merge pull request #4918 from shinhs0506/core-size

shred: add support for octal and hex size

src/uu/head/src/parse.rs

@@ -114,7 +114,13 @@ pub fn parse_num(src: &str) -> Result<(u64, bool), ParseSizeError> {
         return Err(ParseSizeError::ParseFailure(src.to_string()));
     }
 
-    parse_size(size_string).map(|n| (n, all_but_last))
+    // remove leading zeros so that size is interpreted as decimal, not octal
+    let trimmed_string = size_string.trim_start_matches('0');
+    if trimmed_string.is_empty() {
+        Ok((0, all_but_last))
+    } else {
+        parse_size(trimmed_string).map(|n| (n, all_but_last))
+    }
 }
 
 #[cfg(test)]

src/uu/shred/src/shred.rs

@@ -19,6 +19,7 @@ use std::os::unix::prelude::PermissionsExt;
 use std::path::{Path, PathBuf};
 use uucore::display::Quotable;
 use uucore::error::{FromIo, UResult, USimpleError, UUsageError};
+use uucore::parse_size::parse_size;
 use uucore::{format_usage, help_about, help_section, help_usage, show, show_error, show_if_err};
 
 const ABOUT: &str = help_about!("shred.md");
@@ -238,7 +239,7 @@ pub fn uumain(args: impl uucore::Args) -> UResult<()> {
         .get_one::<String>(options::SIZE)
         .map(|s| s.to_string());
     let size = get_size(size_arg);
-    let exact = matches.get_flag(options::EXACT) && size.is_none(); // if -s is given, ignore -x
+    let exact = matches.get_flag(options::EXACT) || size.is_some();
     let zero = matches.get_flag(options::ZERO);
     let verbose = matches.get_flag(options::VERBOSE);
 
@@ -318,38 +319,17 @@ pub fn uu_app() -> Command {
         )
 }
 
-// TODO: Add support for all postfixes here up to and including EiB
-//       http://www.gnu.org/software/coreutils/manual/coreutils.html#Block-size
 fn get_size(size_str_opt: Option<String>) -> Option<u64> {
-    size_str_opt.as_ref()?;
-
-    let mut size_str = size_str_opt.as_ref().unwrap().clone();
-    // Immutably look at last character of size string
-    let unit = match size_str.chars().last().unwrap() {
-        'K' => {
-            size_str.pop();
-            1024u64
-        }
-        'M' => {
-            size_str.pop();
-            (1024 * 1024) as u64
-        }
-        'G' => {
-            size_str.pop();
-            (1024 * 1024 * 1024) as u64
-        }
-        _ => 1u64,
-    };
-
-    let coefficient = match size_str.parse::<u64>() {
-        Ok(u) => u,
-        Err(_) => {
-            show_error!("{}: Invalid file size", size_str_opt.unwrap().maybe_quote());
-            std::process::exit(1);
-        }
-    };
-
-    Some(coefficient * unit)
+    size_str_opt
+        .as_ref()
+        .and_then(|size| parse_size(size.as_str()).ok())
+        .or_else(|| {
+            if let Some(size) = size_str_opt {
+                show_error!("invalid file size: {}", size.quote());
+                std::process::exit(1);
+            }
+            None
+        })
 }
 
 fn pass_name(pass_type: &PassType) -> String {

src/uu/tail/src/args.rs

@@ -80,7 +80,7 @@ impl FilterMode {
                 Err(e) => {
                     return Err(USimpleError::new(
                         1,
-                        format!("invalid number of bytes: {e}"),
+                        format!("invalid number of bytes: '{e}'"),
                     ))
                 }
             }
@@ -415,16 +415,17 @@ fn parse_num(src: &str) -> Result<Signum, ParseSizeError> {
                 starting_with = true;
             }
         }
-    } else {
-        return Err(ParseSizeError::ParseFailure(src.to_string()));
     }
 
-    parse_size(size_string).map(|n| match (n, starting_with) {
-        (0, true) => Signum::PlusZero,
-        (0, false) => Signum::MinusZero,
-        (n, true) => Signum::Positive(n),
-        (n, false) => Signum::Negative(n),
-    })
+    match parse_size(size_string) {
+        Ok(n) => match (n, starting_with) {
+            (0, true) => Ok(Signum::PlusZero),
+            (0, false) => Ok(Signum::MinusZero),
+            (n, true) => Ok(Signum::Positive(n)),
+            (n, false) => Ok(Signum::Negative(n)),
+        },
+        Err(_) => Err(ParseSizeError::ParseFailure(size_string.to_string())),
+    }
 }
 
 pub fn parse_args(args: impl uucore::Args) -> UResult<Settings> {

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

@@ -3,7 +3,7 @@
 //  * 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
+// spell-checker:ignore (ToDO) hdsf ghead gtail ACDBK hexdigit
 
 use std::error::Error;
 use std::fmt;
@@ -25,6 +25,12 @@ pub struct Parser<'parser> {
     pub default_unit: Option<&'parser str>,
 }
 
+enum NumberSystem {
+    Decimal,
+    Octal,
+    Hexadecimal,
+}
+
 impl<'parser> Parser<'parser> {
     pub fn with_allow_list(&mut self, allow_list: &'parser [&str]) -> &mut Self {
         self.allow_list = Some(allow_list);
@@ -62,31 +68,26 @@ impl<'parser> Parser<'parser> {
     /// 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
+    /// assert_eq!(Ok(44251 * 1024), parse_size("0xACDBK"));
     /// ```
     pub fn parse(&self, size: &str) -> Result<u64, 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_ascii_digit()).collect();
-        let number: u64 = if numeric_string.is_empty() {
-            1
-        } else {
-            match numeric_string.parse() {
-                Ok(n) => n,
-                Err(_) => return Err(ParseSizeError::parse_failure(size)),
-            }
-        };
 
-        // 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.
-        //
-        // The lowercase "b" (used by `od`, `head`, `tail`, etc.) means
-        // "block" and the Posix block size is 512. The uppercase "B"
-        // means "byte".
+        let number_system: NumberSystem = self.determine_number_system(size);
+
+        // Split the size argument into numeric and unit parts
+        // For example, if the argument is "123K", the numeric part is "123", and
+        // the unit is "K"
+        let numeric_string: String = match number_system {
+            NumberSystem::Hexadecimal => size
+                .chars()
+                .take(2)
+                .chain(size.chars().skip(2).take_while(|c| c.is_ascii_hexdigit()))
+                .collect(),
+            _ => size.chars().take_while(|c| c.is_ascii_digit()).collect(),
+        };
         let mut unit: &str = &size[numeric_string.len()..];
 
         if let Some(default_unit) = self.default_unit {
@@ -115,6 +116,12 @@ impl<'parser> Parser<'parser> {
             }
         }
 
+        // Compute the factor the unit represents.
+        // empty string means the factor is 1.
+        //
+        // The lowercase "b" (used by `od`, `head`, `tail`, etc.) means
+        // "block" and the Posix block size is 512. The uppercase "B"
+        // means "byte".
         let (base, exponent): (u128, u32) = match unit {
             "" => (1, 0),
             "B" if self.capital_b_bytes => (1, 0),
@@ -142,10 +149,62 @@ impl<'parser> Parser<'parser> {
             Ok(n) => n,
             Err(_) => return Err(ParseSizeError::size_too_big(size)),
         };
+
+        // parse string into u64
+        let number: u64 = match number_system {
+            NumberSystem::Decimal => {
+                if numeric_string.is_empty() {
+                    1
+                } else {
+                    self.parse_number(&numeric_string, 10, size)?
+                }
+            }
+            NumberSystem::Octal => {
+                let trimmed_string = numeric_string.trim_start_matches('0');
+                self.parse_number(trimmed_string, 8, size)?
+            }
+            NumberSystem::Hexadecimal => {
+                let trimmed_string = numeric_string.trim_start_matches("0x");
+                self.parse_number(trimmed_string, 16, size)?
+            }
+        };
+
         number
             .checked_mul(factor)
             .ok_or_else(|| ParseSizeError::size_too_big(size))
     }
+
+    fn determine_number_system(&self, size: &str) -> NumberSystem {
+        if size.len() <= 1 {
+            return NumberSystem::Decimal;
+        }
+
+        if size.starts_with("0x") {
+            return NumberSystem::Hexadecimal;
+        }
+
+        let num_digits: usize = size
+            .chars()
+            .take_while(|c| c.is_ascii_digit())
+            .collect::<String>()
+            .len();
+        let all_zeros = size.chars().all(|c| c == '0');
+        if size.starts_with('0') && num_digits > 1 && !all_zeros {
+            return NumberSystem::Octal;
+        }
+
+        NumberSystem::Decimal
+    }
+
+    fn parse_number(
+        &self,
+        numeric_string: &str,
+        radix: u32,
+        original_size: &str,
+    ) -> Result<u64, ParseSizeError> {
+        u64::from_str_radix(numeric_string, radix)
+            .map_err(|_| ParseSizeError::ParseFailure(original_size.to_string()))
+    }
 }
 
 /// Parse a size string into a number of bytes.
@@ -336,7 +395,7 @@ mod tests {
 
     #[test]
     fn invalid_suffix() {
-        let test_strings = ["328hdsf3290", "5mib", "1e2", "1H", "1.2"];
+        let test_strings = ["5mib", "1eb", "1H"];
         for &test_string in &test_strings {
             assert_eq!(
                 parse_size(test_string).unwrap_err(),
@@ -450,4 +509,18 @@ mod tests {
         assert!(parser.parse("1B").is_err());
         assert!(parser.parse("B").is_err());
     }
+
+    #[test]
+    fn parse_octal_size() {
+        assert_eq!(Ok(63), parse_size("077"));
+        assert_eq!(Ok(528), parse_size("01020"));
+        assert_eq!(Ok(668 * 1024), parse_size("01234K"));
+    }
+
+    #[test]
+    fn parse_hex_size() {
+        assert_eq!(Ok(10), parse_size("0xA"));
+        assert_eq!(Ok(94722), parse_size("0x17202"));
+        assert_eq!(Ok(44251 * 1024), parse_size("0xACDBK"));
+    }
 }

tests/by-util/test_head.rs

@@ -189,6 +189,15 @@ fn test_no_such_file_or_directory() {
         .stderr_contains("cannot open 'no_such_file.toml' for reading: No such file or directory");
 }
 
+#[test]
+fn test_lines_leading_zeros() {
+    new_ucmd!()
+        .arg("--lines=010")
+        .pipe_in("\n\n\n\n\n\n\n\n\n\n\n\n")
+        .succeeds()
+        .stdout_is("\n\n\n\n\n\n\n\n\n\n");
+}
+
 /// Test that each non-existent files gets its own error message printed.
 #[test]
 fn test_multiple_nonexistent_files() {

tests/by-util/test_shred.rs

@@ -51,3 +51,14 @@ fn test_shred_force() {
     // file_a was deleted.
     assert!(!at.file_exists(file));
 }
+
+#[test]
+fn test_hex() {
+    let (at, mut ucmd) = at_and_ucmd!();
+
+    let file = "test_hex";
+
+    at.touch(file);
+
+    ucmd.arg("--size=0x10").arg(file).succeeds();
+}