lints and optimisations

Cargo.toml

@@ -8,3 +8,20 @@ edition = "2021"
 [dependencies]
 clap = { version = "4.4.3", features = ["derive"] }
 rayon = "1.7.0"
+
+# lints from https://github.com/0atman/noboilerplate/blob/main/scripts/37-functional-rust.md
+# these are deliberately extremely restrictive to avoid most of the easily-refactored
+# footguns as much as possible. unlike the github lints, this does not forbid unsafe
+# code
+[lints.clippy]
+enum_glob_use = "deny"
+pedantic = "deny"
+nursery = "deny"
+unwrap_used = "deny"
+
+[profile.release]
+opt-level = 'z' # size optimisation
+lto = true # link time optimisation
+codegen-units = 1 # fewer -> more optimisation
+panic = 'abort' # abort on panic
+strip = 'symbols' # strip symbols

src/directory.rs

@@ -1,7 +1,10 @@
 use std::fs::read_dir;
+use std::hint::unreachable_unchecked;
 use std::path::{Path, PathBuf};
 use std::io::{Result, ErrorKind};
 
+use crate::unit::Unit;
+
 #[derive(Debug, Clone)]
 pub struct Directory {
     name: PathBuf,
@@ -9,12 +12,17 @@ pub struct Directory {
     children: Vec<Directory>,
 }
 impl Directory {
-    #[inline(always)]
+    #[inline]
     pub const fn size(&self) -> u64 {
         self.size
     }
 
-    #[inline(always)]
+    #[inline]
+    pub fn scale(&self, unit: Unit) -> String {
+        unit.convert(self.size)
+    }
+
+    #[inline]
     pub fn path(&self) -> &Path {
         self.name.as_ref()
     }
@@ -24,9 +32,15 @@ impl Directory {
         P: AsRef<Path>
     {
         let path = path.as_ref();
-        let name = path.canonicalize()?.file_name().unwrap().into();
+        let name = path.canonicalize()?
+            .file_name()
+            // file_name() returns None if and only if the path ends in "..".
+            // due to the call to canonicalize(), this can never be the case,
+            // so this can be safely unwrapped
+            .unwrap_or_else(|| unsafe { unreachable_unchecked() })
+            .into();
 
-        let dir = match read_dir(&path) {
+        let dir = match read_dir(path) {
             Ok(dir) => dir,
             Err(io_error) => match io_error.kind() {
                 ErrorKind::NotADirectory => return Ok(Self {
@@ -37,11 +51,11 @@ impl Directory {
                 _ => return Err(io_error),
             }
         };
-        
+
         let mut size = 0;
         let mut children = Vec::new();
         for entry in dir {
-            let child = Directory::new(entry?.path())?;
+            let child = Self::new(entry?.path())?;
             size += child.size;
             children.push(child);
         }
@@ -53,19 +67,28 @@ impl Directory {
         })
     }
 
-    pub fn display(self) -> String {
+    pub fn display(self, unit: Unit) -> String {
         // since self.size is definitionally the greatest value, the tab length
         // is just the length of self.len, plus two for a tab width
         let tab_size = self.size.to_string().len() + 2;
-        self.vectorise().iter().map(|e| e.to_string(tab_size) + "\n").collect()
+        let mut result = self.vectorise(unit).iter().map(|e| e.stringify_tabbed(tab_size) + "\n").collect::<String>();
+
+        if ! result.is_empty() {
+            let final_newline_char_range = result.len()-2 .. result.len();
+            result.drain(final_newline_char_range);
+        }
+
+        result
     }
 
     /// TODO: make not recursive, take &self if possible,
     /// and maybe write directly to stdout to not use so much mem
-    fn vectorise(self) -> Vec<TreeEntry> {
+    fn vectorise(self, unit: Unit) -> Vec<TreeEntry> {
         let mut result = Vec::new();
 
-        result.push(TreeEntry(Vec::new(), self.name.display().to_string(), self.size));
+        result.push(TreeEntry::new(
+            Vec::new(), self.name.display().to_string(), self.size, unit
+        ));
 
         let mut new_entry_part = TreePart::First;
         let mut continue_part = TreePart::Wait;
@@ -78,13 +101,13 @@ impl Directory {
                 continue_part = TreePart::Blank;
             }
 
-            let subtree = child.vectorise();
+            let subtree = child.vectorise(unit);
 
             for mut item in subtree {
-                if item.0.len() == 0 {
-                    item.0.push(new_entry_part);
+                if item.parts.is_empty() {
+                    item.parts.push(new_entry_part);
                 } else {
-                    item.0.push(continue_part);
+                    item.parts.push(continue_part);
                 }
                 result.push(item);
             }
@@ -95,18 +118,29 @@ impl Directory {
 }
 
 #[derive(Debug)]
-struct TreeEntry(Vec<TreePart>, String, u64);
+struct TreeEntry {
+    parts: Vec<TreePart>,
+    path: String,
+    size: u64,
+    unit: Unit
+}
 impl TreeEntry {
-    fn to_string(&self, tab_size: usize) -> String {
-        let mut result = format!("{:<tab_size$}", self.2);
+    fn new(parts: Vec<TreePart>, path: String, size: u64, unit: Unit) -> Self {
+        Self {
+            parts, path, size, unit
+        }
+    }
+    
+    fn stringify_tabbed(&self, tab_size: usize) -> String {
+        let mut result = format!("{:<tab_size$}", self.unit.convert(self.size));
 
-        for part in self.0.iter().rev() {
-            result += &part.display().to_owned();
+        for part in self.parts.iter().rev() {
+            result += part.display();
         }
         // dont add the space to empty entries
         result += " ";
-        result += &self.1;
-        
+        result += &self.path;
+
         result
     }
 }

src/main.rs

@@ -1,13 +1,15 @@
 #![feature(io_error_more, fs_try_exists)]
 
-use clap::Parser;
-use clap::ArgAction;
+use clap::{Parser, ArgAction};
 
 mod directory;
+mod unit;
 use directory::Directory;
+use unit::Unit;
 
 use std::process::ExitCode;
 
+#[allow(clippy::struct_excessive_bools)]
 #[derive(Parser, Debug, Clone)]
 pub struct Args {
     #[arg(
@@ -16,43 +18,61 @@ pub struct Args {
         long_help = "keep going if an error occurs (ex. unreadable subdirectories in a readable directory)"
     )]
     persistant: bool,
-    
+
     #[arg(
         short, long,
         help = "minimize output",
-        long_help = "like -t, but does not print \"total: \" before the summary or the newline after. It also surpresses all error messages",
-        conflicts_with = "total_only",
+        long_help = "print nothing but the total size for all directories, without a newline. Also supresses all error messages",
+        conflicts_with = "total",
         default_value_t = false,
     )]
     minimal: bool,
 
+    #[arg(
+        short='T', long,
+        help = "display in tree",
+        default_value_t = false,
+        conflicts_with = "minimal",
+    )]
+    tree: bool,
+
     #[arg(
         short, long,
-        help = "only display the total size",
+        help = "display the total size",
         conflicts_with = "minimal",
         default_value_t = false,
     )]
-    total_only: bool,
+    total: bool,
 
     #[arg(
         short='2', long,
-        help = "print sizes in powers of 1024",
+        help = "alias for --unit 1024",
         default_value_t = false,
-        conflicts_with = "si"
+        conflicts_with_all = ["si","unit"],
     )]
     base_two: bool,
 
     #[arg(
         short='0', long,
-        help = "print sizes in powers of 1000",
+        help = "alias for --unit 1000",
         default_value_t = false,
-        conflicts_with = "base_two"
+        conflicts_with_all = ["base_two","unit"],
     )]
     si: bool,
 
+    #[arg(
+        short, long,
+        help = "unit to print in",
+        long_help = "printing unit (case insensitive): b = bytes, kb = kilobytes, ki = kibibytes, gb = gigabytes, gi = gibibytes, tb = terabytes, ti = tibibytes",
+        value_parser = Unit::parse,
+        default_value_t = Unit::Byte,
+        conflicts_with_all = ["base_two","si"],
+    )]
+    unit: Unit,
+
     #[arg(
         value_parser = validate_path,
-        help = "directories to summate",
+        help = "items to summate",
         action = ArgAction::Append,
         num_args = 1..
     )]
@@ -68,8 +88,14 @@ fn validate_path(s: &str) -> Result<String, String> {
 }
 
 fn main() -> ExitCode {
-    let args = Args::parse();
+    let mut args = Args::parse();
 
+    if args.base_two {
+        args.unit = Unit::Kibi;
+    } else if args.si {
+        args.unit = Unit::Kilo;
+    }
+    
     let mut total = 0;
     for path in args.path {
         let dir_structure = match Directory::new(path) {
@@ -81,24 +107,33 @@ fn main() -> ExitCode {
                 return ExitCode::FAILURE;
             }
         };
-        
+
         total += dir_structure.size();
 
-        if !args.minimal {
-            if args.total_only {
-                println!("{}: {}", dir_structure.path().to_str().unwrap(), dir_structure.size());
-            } else {
-                print!("{}", dir_structure.display());
-            }
+        if args.minimal {
+            // skip printing (this is a matter of indentation)
+            continue;
+        }
+
+        if args.tree {
+            println!("{}", dir_structure.display(args.unit));
+        } else {
+            println!(
+                "{}: {}",
+                dir_structure.path().display(),
+                dir_structure.scale(args.unit),
+            );
         }
     }
 
-    if args.total_only {
+    let total = args.unit.convert(total);
+
+    if args.total {
         println!("total: {total}");
     }
     else if args.minimal {
         print!("{total}");
     }
-    
+
     ExitCode::SUCCESS
 }