feat: support --files-from/--files0-from

Read NUL- or newline-terminated paths from FILE or stdin. Wire options through `config` and `main`. Add tests and update the man page, README, and completion scripts.
2025-12-05 20:40:11 -08:00 · 2025-08-30 10:38:17 +00:00
parent 14efddfd05
commit 7974e2eaf0
15 changed files with 137 additions and 39 deletions
--- a/README.md
+++ b/README.md
@@ -102,7 +102,8 @@ Usage: dust -S (Custom Stack size - Use if you see: 'fatal runtime error: stack
 Usage: dust --skip-total (No total row will be displayed)
 Usage: dust -z 40000/30MB/20kib (Exclude output files/directories below size 40000 bytes / 30MB / 20KiB)
 Usage: dust -j (Prints JSON representation of directories, try: dust -j  | jq)
-Usage: dust --files0-from=FILE (Reads null-terminated file paths from FILE); If FILE is - then read from stdin
+Usage: dust --files0-from=FILE (Read NUL-terminated file paths from FILE; if FILE is '-', read from stdin)
+Usage: dust --files-from=FILE (Read newline-terminated file paths from FILE; if FILE is '-', read from stdin)
 Usage: dust --collapse=node-modules will keep the node-modules folder collapsed in display instead of recursively opening it
 ```

--- a/completions/_dust
+++ b/completions/_dust
@@ -62,7 +62,8 @@ tb\:"terabyte (TB)"))' \
 '--atime=[just like -mtime, but based on file access time]:ATIME:_default' \
 '-y+[just like -mtime, but based on file change time]:CTIME:_default' \
 '--ctime=[just like -mtime, but based on file change time]:CTIME:_default' \
-'--files0-from=[run dust on NUL-terminated file names specified in file; if argument is -, then read names from standard input]:FILES0_FROM:_files' \
+'(--files-from)--files0-from=[Read NUL-terminated paths from FILE (use \`-\` for stdin)]:FILES0_FROM:_files' \
+'(--files0-from)--files-from=[Read newline-terminated paths from FILE (use \`-\` for stdin)]:FILES_FROM:_files' \
 '*--collapse=[Keep these directories collapsed]:COLLAPSE:_files' \
 '-m+[Directory '\''size'\'' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time]:FILETIME:((a\:"last accessed time"
 c\:"last changed time"
--- a/completions/_dust.ps1
+++ b/completions/_dust.ps1
@@ -50,7 +50,8 @@ Register-ArgumentCompleter -Native -CommandName 'dust' -ScriptBlock {
            [CompletionResult]::new('--atime', '--atime', [CompletionResultType]::ParameterName, 'just like -mtime, but based on file access time')
            [CompletionResult]::new('-y', '-y', [CompletionResultType]::ParameterName, 'just like -mtime, but based on file change time')
            [CompletionResult]::new('--ctime', '--ctime', [CompletionResultType]::ParameterName, 'just like -mtime, but based on file change time')
-            [CompletionResult]::new('--files0-from', '--files0-from', [CompletionResultType]::ParameterName, 'run dust on NUL-terminated file names specified in file; if argument is -, then read names from standard input')
+            [CompletionResult]::new('--files0-from', '--files0-from', [CompletionResultType]::ParameterName, 'Read NUL-terminated paths from FILE (use `-` for stdin)')
+            [CompletionResult]::new('--files-from', '--files-from', [CompletionResultType]::ParameterName, 'Read newline-terminated paths from FILE (use `-` for stdin)')
            [CompletionResult]::new('--collapse', '--collapse', [CompletionResultType]::ParameterName, 'Keep these directories collapsed')
            [CompletionResult]::new('-m', '-m', [CompletionResultType]::ParameterName, 'Directory ''size'' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time')
            [CompletionResult]::new('--filetime', '--filetime', [CompletionResultType]::ParameterName, 'Directory ''size'' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time')
--- a/completions/dust.bash
+++ b/completions/dust.bash
@@ -23,7 +23,7 @@ _dust() {

    case "${cmd}" in
        dust)
-            opts="-d -T -n -p -X -I -L -x -s -r -c -C -b -B -z -R -f -i -v -e -t -w -P -D -F -o -S -j -M -A -y -m -h -V --depth --threads --config --number-of-lines --full-paths --ignore-directory --ignore-all-in-file --dereference-links --limit-filesystem --apparent-size --reverse --no-colors --force-colors --no-percent-bars --bars-on-right --min-size --screen-reader --skip-total --filecount --ignore-hidden --invert-filter --filter --file-types --terminal-width --no-progress --print-errors --only-dir --only-file --output-format --stack-size --output-json --mtime --atime --ctime --files0-from --collapse --filetime --help --version [PATH]..."
+            opts="-d -T -n -p -X -I -L -x -s -r -c -C -b -B -z -R -f -i -v -e -t -w -P -D -F -o -S -j -M -A -y -m -h -V --depth --threads --config --number-of-lines --full-paths --ignore-directory --ignore-all-in-file --dereference-links --limit-filesystem --apparent-size --reverse --no-colors --force-colors --no-percent-bars --bars-on-right --min-size --screen-reader --skip-total --filecount --ignore-hidden --invert-filter --filter --file-types --terminal-width --no-progress --print-errors --only-dir --only-file --output-format --stack-size --output-json --mtime --atime --ctime --files0-from --files-from --collapse --filetime --help --version [PATH]..."
            if [[ ${cur} == -* || ${COMP_CWORD} -eq 1 ]] ; then
                COMPREPLY=( $(compgen -W "${opts}" -- "${cur}") )
                return 0
@@ -182,6 +182,10 @@ _dust() {
                    COMPREPLY=($(compgen -f "${cur}"))
                    return 0
                    ;;
+                --files-from)
+                    COMPREPLY=($(compgen -f "${cur}"))
+                    return 0
+                    ;;
                --collapse)
                    COMPREPLY=($(compgen -f "${cur}"))
                    return 0
--- a/completions/dust.elv
+++ b/completions/dust.elv
@@ -47,7 +47,8 @@ set edit:completion:arg-completer[dust] = {|@words|
            cand --atime 'just like -mtime, but based on file access time'
            cand -y 'just like -mtime, but based on file change time'
            cand --ctime 'just like -mtime, but based on file change time'
-            cand --files0-from 'run dust on NUL-terminated file names specified in file; if argument is -, then read names from standard input'
+            cand --files0-from 'Read NUL-terminated paths from FILE (use `-` for stdin)'
+            cand --files-from 'Read newline-terminated paths from FILE (use `-` for stdin)'
            cand --collapse 'Keep these directories collapsed'
            cand -m 'Directory ''size'' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time'
            cand --filetime 'Directory ''size'' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time'
--- a/completions/dust.fish
+++ b/completions/dust.fish
@@ -22,7 +22,8 @@ complete -c dust -s S -l stack-size -d 'Specify memory to use as stack size - us
 complete -c dust -s M -l mtime -d '+/-n matches files modified more/less than n days ago , and n matches files modified exactly n days ago, days are rounded down.That is +n => (−∞, curr−(n+1)), n => [curr−(n+1), curr−n), and -n => (𝑐𝑢𝑟𝑟−𝑛, +∞)' -r
 complete -c dust -s A -l atime -d 'just like -mtime, but based on file access time' -r
 complete -c dust -s y -l ctime -d 'just like -mtime, but based on file change time' -r
-complete -c dust -l files0-from -d 'run dust on NUL-terminated file names specified in file; if argument is -, then read names from standard input' -r -F
+complete -c dust -l files0-from -d 'Read NUL-terminated paths from FILE (use `-` for stdin)' -r -F
+complete -c dust -l files-from -d 'Read newline-terminated paths from FILE (use `-` for stdin)' -r -F
 complete -c dust -l collapse -d 'Keep these directories collapsed' -r -F
 complete -c dust -s m -l filetime -d 'Directory \'size\' is max filetime of child files instead of disk size. while a/c/m for last accessed/changed/modified time' -r -f -a "a\t'last accessed time'
 c\t'last changed time'
--- a/man-page/dust.1
+++ b/man-page/dust.1
@@ -4,7 +4,7 @@
 .SH NAME
 Dust \- Like du but more intuitive
 .SH SYNOPSIS
-\fBdust\fR [\fB\-d\fR|\fB\-\-depth\fR] [\fB\-T\fR|\fB\-\-threads\fR] [\fB\-\-config\fR] [\fB\-n\fR|\fB\-\-number\-of\-lines\fR] [\fB\-p\fR|\fB\-\-full\-paths\fR] [\fB\-X\fR|\fB\-\-ignore\-directory\fR] [\fB\-I\fR|\fB\-\-ignore\-all\-in\-file\fR] [\fB\-L\fR|\fB\-\-dereference\-links\fR] [\fB\-x\fR|\fB\-\-limit\-filesystem\fR] [\fB\-s\fR|\fB\-\-apparent\-size\fR] [\fB\-r\fR|\fB\-\-reverse\fR] [\fB\-c\fR|\fB\-\-no\-colors\fR] [\fB\-C\fR|\fB\-\-force\-colors\fR] [\fB\-b\fR|\fB\-\-no\-percent\-bars\fR] [\fB\-B\fR|\fB\-\-bars\-on\-right\fR] [\fB\-z\fR|\fB\-\-min\-size\fR] [\fB\-R\fR|\fB\-\-screen\-reader\fR] [\fB\-\-skip\-total\fR] [\fB\-f\fR|\fB\-\-filecount\fR] [\fB\-i\fR|\fB\-\-ignore\-hidden\fR] [\fB\-v\fR|\fB\-\-invert\-filter\fR] [\fB\-e\fR|\fB\-\-filter\fR] [\fB\-t\fR|\fB\-\-file\-types\fR] [\fB\-w\fR|\fB\-\-terminal\-width\fR] [\fB\-P\fR|\fB\-\-no\-progress\fR] [\fB\-\-print\-errors\fR] [\fB\-D\fR|\fB\-\-only\-dir\fR] [\fB\-F\fR|\fB\-\-only\-file\fR] [\fB\-o\fR|\fB\-\-output\-format\fR] [\fB\-S\fR|\fB\-\-stack\-size\fR] [\fB\-j\fR|\fB\-\-output\-json\fR] [\fB\-M\fR|\fB\-\-mtime\fR] [\fB\-A\fR|\fB\-\-atime\fR] [\fB\-y\fR|\fB\-\-ctime\fR] [\fB\-\-files0\-from\fR] [\fB\-\-collapse\fR] [\fB\-m\fR|\fB\-\-filetime\fR] [\fB\-h\fR|\fB\-\-help\fR] [\fB\-V\fR|\fB\-\-version\fR] [\fIPATH\fR] 
+\fBdust\fR [\fB\-d\fR|\fB\-\-depth\fR] [\fB\-T\fR|\fB\-\-threads\fR] [\fB\-\-config\fR] [\fB\-n\fR|\fB\-\-number\-of\-lines\fR] [\fB\-p\fR|\fB\-\-full\-paths\fR] [\fB\-X\fR|\fB\-\-ignore\-directory\fR] [\fB\-I\fR|\fB\-\-ignore\-all\-in\-file\fR] [\fB\-L\fR|\fB\-\-dereference\-links\fR] [\fB\-x\fR|\fB\-\-limit\-filesystem\fR] [\fB\-s\fR|\fB\-\-apparent\-size\fR] [\fB\-r\fR|\fB\-\-reverse\fR] [\fB\-c\fR|\fB\-\-no\-colors\fR] [\fB\-C\fR|\fB\-\-force\-colors\fR] [\fB\-b\fR|\fB\-\-no\-percent\-bars\fR] [\fB\-B\fR|\fB\-\-bars\-on\-right\fR] [\fB\-z\fR|\fB\-\-min\-size\fR] [\fB\-R\fR|\fB\-\-screen\-reader\fR] [\fB\-\-skip\-total\fR] [\fB\-f\fR|\fB\-\-filecount\fR] [\fB\-i\fR|\fB\-\-ignore\-hidden\fR] [\fB\-v\fR|\fB\-\-invert\-filter\fR] [\fB\-e\fR|\fB\-\-filter\fR] [\fB\-t\fR|\fB\-\-file\-types\fR] [\fB\-w\fR|\fB\-\-terminal\-width\fR] [\fB\-P\fR|\fB\-\-no\-progress\fR] [\fB\-\-print\-errors\fR] [\fB\-D\fR|\fB\-\-only\-dir\fR] [\fB\-F\fR|\fB\-\-only\-file\fR] [\fB\-o\fR|\fB\-\-output\-format\fR] [\fB\-S\fR|\fB\-\-stack\-size\fR] [\fB\-j\fR|\fB\-\-output\-json\fR] [\fB\-M\fR|\fB\-\-mtime\fR] [\fB\-A\fR|\fB\-\-atime\fR] [\fB\-y\fR|\fB\-\-ctime\fR] [\fB\-\-files0\-from\fR] [\fB\-\-files\-from\fR] [\fB\-\-collapse\fR] [\fB\-m\fR|\fB\-\-filetime\fR] [\fB\-h\fR|\fB\-\-help\fR] [\fB\-V\fR|\fB\-\-version\fR] [\fIPATH\fR] 
 .SH DESCRIPTION
 Like du but more intuitive
 .SH OPTIONS
@@ -138,7 +138,10 @@ just like \-mtime, but based on file access time
 just like \-mtime, but based on file change time
 .TP
 \fB\-\-files0\-from\fR=\fIFILES0_FROM\fR
-run dust on NUL\-terminated file names specified in file; if argument is \-, then read names from standard input
+Read NUL\-terminated paths from FILE (use `\-` for stdin)
+.TP
+\fB\-\-files\-from\fR=\fIFILES_FROM\fR
+Read newline\-terminated paths from FILE (use `\-` for stdin)
 .TP
 \fB\-\-collapse\fR=\fICOLLAPSE\fR
 Keep these directories collapsed
--- a/src/cli.rs
+++ b/src/cli.rs
@@ -172,11 +172,14 @@ pub struct Cli {
    #[arg(short('y'), long, allow_hyphen_values(true))]
    pub ctime: Option<String>,

-    /// run dust on NUL-terminated file names specified in file; if argument is
-    /// -, then read names from standard input
-    #[arg(long, value_hint(ValueHint::AnyPath))]
+    /// Read NUL-terminated paths from FILE (use `-` for stdin).
+    #[arg(long, value_hint(ValueHint::AnyPath), conflicts_with("files_from"))]
    pub files0_from: Option<String>,

+    /// Read newline-terminated paths from FILE (use `-` for stdin).
+    #[arg(long, value_hint(ValueHint::AnyPath), conflicts_with("files0_from"))]
+    pub files_from: Option<String>,
+
    /// Keep these directories collapsed
    #[arg(long, value_hint(ValueHint::AnyPath))]
    pub collapse: Option<Vec<String>>,
--- a/src/config.rs
+++ b/src/config.rs
@@ -36,16 +36,25 @@ pub struct Config {
    pub output_json: Option<bool>,
    pub print_errors: Option<bool>,
    pub files0_from: Option<String>,
+    pub files_from: Option<String>,
 }

 impl Config {
-    pub fn get_files_from(&self, options: &Cli) -> Option<String> {
+    pub fn get_files0_from(&self, options: &Cli) -> Option<String> {
        let from_file = &options.files0_from;
        match from_file {
            None => self.files0_from.as_ref().map(|x| x.to_string()),
            Some(x) => Some(x.to_string()),
        }
    }
+
+    pub fn get_files_from(&self, options: &Cli) -> Option<String> {
+        let from_file = &options.files_from;
+        match from_file {
+            None => self.files_from.as_ref().map(|x| x.to_string()),
+            Some(x) => Some(x.to_string()),
+        }
+    }
    pub fn get_no_colors(&self, options: &Cli) -> bool {
        Some(true) == self.no_colors || options.no_colors
    }
--- a/src/main.rs
+++ b/src/main.rs
@@ -22,8 +22,9 @@ use progress::PIndicator;
 use regex::Error;
 use std::collections::HashSet;
 use std::env;
-use std::fs::read_to_string;
+use std::fs::{read, read_to_string};
 use std::io;
+use std::io::Read;
 use std::panic;
 use std::process;
 use std::sync::Arc;
@@ -127,34 +128,15 @@ fn main() {
    })
    .expect("Error setting Ctrl-C handler");

-    let target_dirs = match config.get_files_from(&options) {
-        Some(path) => {
-            if path == "-" {
-                let mut targets_to_add = io::stdin()
-                    .lines()
-                    .map_while(Result::ok)
-                    .collect::<Vec<String>>();
-
-                if targets_to_add.is_empty() {
-                    eprintln!("No input provided, defaulting to current directory");
-                    targets_to_add.push(".".to_owned());
-                }
-                targets_to_add
-            } else {
-                // read file
-                match read_to_string(path) {
-                    Ok(file_content) => file_content.lines().map(|x| x.to_string()).collect(),
-                    Err(e) => {
-                        eprintln!("Error reading file: {e}");
-                        vec![".".to_owned()]
-                    }
-                }
-            }
-        }
-        None => match options.params {
+    let target_dirs = if let Some(path) = config.get_files0_from(&options) {
+        read_paths_from_source(&path, true)
+    } else if let Some(path) = config.get_files_from(&options) {
+        read_paths_from_source(&path, false)
+    } else {
+        match options.params {
            Some(ref values) => values.clone(),
            None => vec![".".to_owned()],
-        },
+        }
    };

    let summarize_file_types = options.file_types;
@@ -398,6 +380,53 @@ fn print_any_errors(print_errors: bool, final_errors: &RuntimeErrors) {
    }
 }

+fn read_paths_from_source(path: &str, null_terminated: bool) -> Vec<String> {
+    let from_stdin = path == "-";
+
+    let result: Result<Vec<String>, Option<String>> = (|| {
+        // 1) read bytes
+        let bytes = if from_stdin {
+            let mut b = Vec::new();
+            io::stdin().lock().read_to_end(&mut b).map_err(|_| None)?;
+            b
+        } else {
+            read(path).map_err(|e| Some(e.to_string()))?
+        };
+
+        let text = std::str::from_utf8(&bytes).map_err(|e| {
+            if from_stdin {
+                None
+            } else {
+                Some(e.to_string())
+            }
+        })?;
+        let items: Vec<String> = if null_terminated {
+            text.split('\0')
+                .filter(|s| !s.is_empty())
+                .map(str::to_owned)
+                .collect()
+        } else {
+            text.lines().map(str::to_owned).collect()
+        };
+        if from_stdin && items.is_empty() {
+            return Err(None);
+        }
+        Ok(items)
+    })();
+
+    match result {
+        Ok(v) => v,
+        Err(None) => {
+            eprintln!("No files provided, defaulting to current directory");
+            vec![".".to_owned()]
+        }
+        Err(Some(msg)) => {
+            eprintln!("Failed to read file: {msg}");
+            vec![".".to_owned()]
+        }
+    }
+}
+
 fn init_rayon(stack: &Option<usize>, threads: &Option<usize>) -> rayon::ThreadPool {
    let stack_size = match stack {
        Some(s) => Some(*s),
--- a/tests/test_dir_files_from/a_file
+++ b/tests/test_dir_files_from/a_file
--- a/tests/test_dir_files_from/files0_from.txt
+++ b/tests/test_dir_files_from/files0_from.txt
--- a/tests/test_dir_files_from/files_from.txt
+++ b/tests/test_dir_files_from/files_from.txt
@@ -0,0 +1,2 @@
+tests/test_dir_files_from/a_file
+tests/test_dir_files_from/hello_file
--- a/tests/test_dir_files_from/hello_file
+++ b/tests/test_dir_files_from/hello_file
@@ -0,0 +1 @@
+hello
--- a/tests/test_flags.rs
+++ b/tests/test_flags.rs
@@ -104,6 +104,48 @@ pub fn test_ignore_all_in_file() {
    assert!(!output.contains(".secret"));
 }

+#[test]
+pub fn test_files_from_flag_file() {
+    let output = build_command(vec!["--files-from", "tests/test_dir_files_from/files_from.txt"]);
+    assert!(output.contains("a_file"));
+    assert!(output.contains("hello_file"));
+}
+
+#[test]
+pub fn test_files0_from_flag_file() {
+    let output = build_command(vec!["--files0-from", "tests/test_dir_files_from/files0_from.txt"]);
+    assert!(output.contains("a_file"));
+    assert!(output.contains("hello_file"));
+}
+
+#[test]
+pub fn test_files_from_flag_stdin() {
+    let mut cmd = Command::cargo_bin("dust").unwrap();
+    cmd.arg("-P").arg("--files-from").arg("-");
+    let input = b"tests/test_dir_files_from/a_file\ntests/test_dir_files_from/hello_file\n";
+    cmd.write_stdin(input.as_ref());
+    let finished = &cmd.unwrap();
+    let stderr = std::str::from_utf8(&finished.stderr).unwrap();
+    assert_eq!(stderr, "");
+    let output = std::str::from_utf8(&finished.stdout).unwrap();
+    assert!(output.contains("a_file"));
+    assert!(output.contains("hello_file"));
+}
+
+#[test]
+pub fn test_files0_from_flag_stdin() {
+    let mut cmd = Command::cargo_bin("dust").unwrap();
+    cmd.arg("-P").arg("--files0-from").arg("-");
+    let input = b"tests/test_dir_files_from/a_file\0tests/test_dir_files_from/hello_file\0";
+    cmd.write_stdin(input.as_ref());
+    let finished = &cmd.unwrap();
+    let stderr = std::str::from_utf8(&finished.stderr).unwrap();
+    assert_eq!(stderr, "");
+    let output = std::str::from_utf8(&finished.stdout).unwrap();
+    assert!(output.contains("a_file"));
+    assert!(output.contains("hello_file"));
+}
+
 #[test]
 pub fn test_with_bad_param() {
    let mut cmd = Command::cargo_bin("dust").unwrap();