Merge branch 'main' into patch-4

Author: oech3

.github/workflows/GnuTests.yml

@@ -59,7 +59,7 @@ jobs:
         path: |
           gnu/config.cache
           gnu/src/getlimits
-        key: ${{ runner.os }}-gnu-config-${{ hashFiles('gnu/NEWS') }}-${{ hashFiles('gnu/configure') }}
+        key: ${{ runner.os }}-gnu-config-${{ hashFiles('gnu/NEWS') }}-${{ hashFiles('uutils/util/build-gnu.sh') }} # use build-gnu.sh for extremely safe caching
 
     #### Build environment setup
     - name: Install dependencies
@@ -111,7 +111,7 @@ jobs:
         path: |
           gnu/config.cache
           gnu/src/getlimits
-        key: ${{ runner.os }}-gnu-config-${{ hashFiles('gnu/NEWS') }}-${{ hashFiles('gnu/configure') }}
+        key: ${{ runner.os }}-gnu-config-${{ hashFiles('gnu/NEWS') }}-${{ hashFiles('uutils/util/build-gnu.sh') }}
 
     ### Run tests as user
     - name: Run GNU tests

src/uu/cat/locales/en-US.ftl

@@ -19,3 +19,4 @@ cat-error-unknown-filetype = unknown filetype: { $ft_debug }
 cat-error-is-directory = Is a directory
 cat-error-input-file-is-output-file = input file is output file
 cat-error-too-many-symbolic-links = Too many levels of symbolic links
+cat-error-no-such-device-or-address = No such device or address

src/uu/cat/locales/fr-FR.ftl

@@ -19,3 +19,4 @@ cat-error-unknown-filetype = type de fichier inconnu : { $ft_debug }
 cat-error-is-directory = Est un répertoire
 cat-error-input-file-is-output-file = le fichier d'entrée est le fichier de sortie
 cat-error-too-many-symbolic-links = Trop de niveaux de liens symboliques
+cat-error-no-such-device-or-address = Aucun appareil ou adresse de ce type

src/uu/cat/src/cat.rs

@@ -13,15 +13,10 @@ use memchr::memchr2;
 use std::ffi::OsString;
 use std::fs::{File, metadata};
 use std::io::{self, BufWriter, ErrorKind, IsTerminal, Read, Write};
-/// Unix domain socket support
-#[cfg(unix)]
-use std::net::Shutdown;
 #[cfg(unix)]
 use std::os::fd::AsFd;
 #[cfg(unix)]
 use std::os::unix::fs::FileTypeExt;
-#[cfg(unix)]
-use std::os::unix::net::UnixStream;
 use thiserror::Error;
 use uucore::display::Quotable;
 use uucore::error::UResult;
@@ -103,6 +98,9 @@ enum CatError {
     },
     #[error("{}", translate!("cat-error-is-directory"))]
     IsDirectory,
+    #[cfg(unix)]
+    #[error("{}", translate!("cat-error-no-such-device-or-address"))]
+    NoSuchDeviceOrAddress,
     #[error("{}", translate!("cat-error-input-file-is-output-file"))]
     OutputIsInput,
     #[error("{}", translate!("cat-error-too-many-symbolic-links"))]
@@ -395,15 +393,7 @@ fn cat_path(path: &OsString, options: &OutputOptions, state: &mut OutputState) -
         }
         InputType::Directory => Err(CatError::IsDirectory),
         #[cfg(unix)]
-        InputType::Socket => {
-            let socket = UnixStream::connect(path)?;
-            socket.shutdown(Shutdown::Write)?;
-            let mut handle = InputHandle {
-                reader: socket,
-                is_interactive: false,
-            };
-            cat_handle(&mut handle, options, state)
-        }
+        InputType::Socket => Err(CatError::NoSuchDeviceOrAddress),
         _ => {
             let file = File::open(path)?;
             if is_unsafe_overwrite(&file, &io::stdout()) {

src/uu/cksum/src/cksum.rs

@@ -216,7 +216,6 @@ pub fn uumain(args: impl uucore::Args) -> UResult<()> {
         algo_kind: algo,
         output_format,
         line_ending,
-        binary: false,
         no_names: false,
     };
 

src/uu/hashsum/src/hashsum.rs

@@ -229,7 +229,6 @@ pub fn uumain(mut args: impl uucore::Args) -> UResult<()> {
             /* base64: */ false,
         ),
         line_ending,
-        binary,
         no_names,
     };
 

src/uu/head/src/parse.rs

@@ -4,7 +4,8 @@
 // file that was distributed with this source code.
 
 use std::ffi::OsString;
-use uucore::parser::parse_size::{ParseSizeError, parse_size_u64_max};
+use uucore::parser::parse_signed_num::{SignPrefix, parse_signed_num_max};
+use uucore::parser::parse_size::ParseSizeError;
 
 #[derive(PartialEq, Eq, Debug)]
 pub struct ParseError;
@@ -107,30 +108,12 @@ fn process_num_block(
 }
 
 /// Parses an -c or -n argument,
-/// the bool specifies whether to read from the end
+/// the bool specifies whether to read from the end (all but last N)
 pub fn parse_num(src: &str) -> Result<(u64, bool), ParseSizeError> {
-    let mut size_string = src.trim();
-    let mut all_but_last = false;
-
-    if let Some(c) = size_string.chars().next() {
-        if c == '+' || c == '-' {
-            // head: '+' is not documented (8.32 man pages)
-            size_string = &size_string[1..];
-            if c == '-' {
-                all_but_last = true;
-            }
-        }
-    } else {
-        return Err(ParseSizeError::ParseFailure(src.to_string()));
-    }
-
-    // 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_u64_max(trimmed_string).map(|n| (n, all_but_last))
-    }
+    let result = parse_signed_num_max(src)?;
+    // head: '-' means "all but last N"
+    let all_but_last = result.sign == Some(SignPrefix::Minus);
+    Ok((result.value, all_but_last))
 }
 
 #[cfg(test)]

src/uu/tail/src/args.rs

@@ -13,7 +13,8 @@ use std::ffi::OsString;
 use std::io::IsTerminal;
 use std::time::Duration;
 use uucore::error::{UResult, USimpleError, UUsageError};
-use uucore::parser::parse_size::{ParseSizeError, parse_size_u64};
+use uucore::parser::parse_signed_num::{SignPrefix, parse_signed_num};
+use uucore::parser::parse_size::ParseSizeError;
 use uucore::parser::parse_time;
 use uucore::parser::shortcut_value_parser::ShortcutValueParser;
 use uucore::translate;
@@ -386,27 +387,15 @@ pub fn parse_obsolete(arg: &OsString, input: Option<&OsString>) -> UResult<Optio
 }
 
 fn parse_num(src: &str) -> Result<Signum, 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;
-            }
-        }
-    }
-
-    match parse_size_u64(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())),
+    let result = parse_signed_num(src)?;
+    // tail: '+' means "starting from line/byte N", default/'-' means "last N"
+    let is_plus = result.sign == Some(SignPrefix::Plus);
+
+    match (result.value, is_plus) {
+        (0, true) => Ok(Signum::PlusZero),
+        (0, false) => Ok(Signum::MinusZero),
+        (n, true) => Ok(Signum::Positive(n)),
+        (n, false) => Ok(Signum::Negative(n)),
     }
 }
 

src/uucore/src/lib/features/checksum/compute.rs

@@ -20,6 +20,11 @@ use crate::{show, translate};
 /// from it: 32 KiB.
 const READ_BUFFER_SIZE: usize = 32 * 1024;
 
+/// Necessary options when computing a checksum. Historically, these options
+/// included a `binary` field to differentiate `--binary` and `--text` modes on
+/// windows. Since the support for this feature is approximate in GNU, and it's
+/// deprecated anyway, it was decided in #9168 to ignore the difference when
+/// computing the checksum.
 pub struct ChecksumComputeOptions {
     /// Which algorithm to use to compute the digest.
     pub algo_kind: SizedAlgoKind,
@@ -30,9 +35,6 @@ pub struct ChecksumComputeOptions {
     /// Whether to finish lines with '\n' or '\0'.
     pub line_ending: LineEnding,
 
-    /// On windows, open files as binary instead of text
-    pub binary: bool,
-
     /// (non-GNU option) Do not print file names
     pub no_names: bool,
 }
@@ -42,6 +44,12 @@ pub struct ChecksumComputeOptions {
 /// On most linux systems, this is irrelevant, as there is no distinction
 /// between text and binary files. Refer to GNU's cksum documentation for more
 /// information.
+///
+/// As discussed in #9168, we decide to ignore the reading mode to compute the
+/// digest, both on Windows and UNIX. The reason for that is that this is a
+/// legacy feature that is poorly documented and used. This enum is kept
+/// nonetheless to still take into account the flags passed to cksum when
+/// generating untagged lines.
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum ReadingMode {
     Binary,
@@ -280,7 +288,9 @@ where
 
         let mut digest = options.algo_kind.create_digest();
 
-        let (digest_output, sz) = digest_reader(&mut digest, &mut file, options.binary)
+        // Always compute the "binary" version of the digest, i.e. on Windows,
+        // never handle CRLFs specifically.
+        let (digest_output, sz) = digest_reader(&mut digest, &mut file, /* binary: */ true)
             .map_err_context(|| translate!("checksum-error-failed-to-read-input"))?;
 
         // Encodes the sum if df is Base64, leaves as-is otherwise.

src/uucore/src/lib/features/parser/mod.rs

@@ -9,6 +9,8 @@ pub mod num_parser;
 #[cfg(any(feature = "parser", feature = "parser-glob"))]
 pub mod parse_glob;
 #[cfg(any(feature = "parser", feature = "parser-size"))]
+pub mod parse_signed_num;
+#[cfg(any(feature = "parser", feature = "parser-size"))]
 pub mod parse_size;
 #[cfg(any(feature = "parser", feature = "parser-num"))]
 pub mod parse_time;