Issue #1622 port du to windows (#1788)

* Issue #1622 port `du` to windows

* Attempt to support Rust 1.32

Old version was getting "attributes are not yet allowed on `if`
expressions" on Rust 1.32

* Less #[cfg]

* Less duplicate code.

I need the return and the semicolon after if otherwise the second #[cfg]
leads to unexpected token complilation error

* More accurate size on disk calculations for windows

* Expect the same output on windows as with WSL

* Better matches output from du on WSL

* In the absence of feedback I'm disabling these tests on Windows.
They require `ln`. Windows does not ship with this utility.

* Use the coreutils version of `ln` to test `du`

`fn ccmd` is courtesy of @Artoria2e5

* Look up inodes (file ids) on Windows

* One more #[cfg(windows)] to prevent unreachable statement warning on linux

Cargo.lock

@@ -1735,6 +1735,7 @@ dependencies = [
  "time",
  "uucore",
  "uucore_procs",
+ "winapi 0.3.9",
 ]
 
 [[package]]

Cargo.toml

@@ -43,6 +43,7 @@ feat_common_core = [
     "df",
     "dircolors",
     "dirname",
+    "du",
     "echo",
     "env",
     "expand",
@@ -149,7 +150,6 @@ feat_require_unix = [
     "chmod",
     "chown",
     "chroot",
-    "du",
     "groups",
     "hostid",
     "id",

src/uu/du/Cargo.toml

@@ -18,6 +18,7 @@ path = "src/du.rs"
 time = "0.1.40"
 uucore = { version=">=0.0.7", package="uucore", path="../../uucore" }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
+winapi = { version="0.3", features=[] }
 
 [[bin]]
 name = "du"

src/uu/du/src/du.rs

@@ -15,9 +15,24 @@ use std::env;
 use std::fs;
 use std::io::{stderr, Result, Write};
 use std::iter;
+#[cfg(not(windows))]
 use std::os::unix::fs::MetadataExt;
+#[cfg(windows)]
+use std::os::windows::fs::MetadataExt;
+#[cfg(windows)]
+use std::os::windows::io::AsRawHandle;
 use std::path::PathBuf;
 use time::Timespec;
+#[cfg(windows)]
+use winapi::shared::minwindef::{DWORD, LPVOID};
+#[cfg(windows)]
+use winapi::um::fileapi::{FILE_STANDARD_INFO, FILE_ID_INFO};
+#[cfg(windows)]
+use winapi::um::minwinbase::{FileStandardInfo, FileIdInfo};
+#[cfg(windows)]
+use winapi::um::winbase::GetFileInformationByHandleEx;
+#[cfg(windows)]
+use winapi::um::winnt::FILE_ID_128;
 
 const NAME: &str = "du";
 const SUMMARY: &str = "estimate file space usage";
@@ -48,7 +63,7 @@ struct Stat {
     is_dir: bool,
     size: u64,
     blocks: u64,
-    inode: u64,
+    inode: Option<u128>,
     created: u64,
     accessed: u64,
     modified: u64,
@@ -57,19 +72,106 @@ struct Stat {
 impl Stat {
     fn new(path: PathBuf) -> Result<Stat> {
         let metadata = fs::symlink_metadata(&path)?;
-        Ok(Stat {
+
+        #[cfg(not(windows))]
+        return Ok(Stat {
             path,
             is_dir: metadata.is_dir(),
             size: metadata.len(),
             blocks: metadata.blocks() as u64,
-            inode: metadata.ino() as u64,
+            inode: Some(metadata.ino() as u128),
             created: metadata.mtime() as u64,
             accessed: metadata.atime() as u64,
             modified: metadata.mtime() as u64,
+        });
+
+        #[cfg(windows)]
+        let size_on_disk = get_size_on_disk(&path);
+        #[cfg(windows)]
+        let inode = get_inode(&path);
+        #[cfg(windows)]
+        Ok(Stat {
+            path,
+            is_dir: metadata.is_dir(),
+            size: metadata.len(),
+            blocks: size_on_disk / 1024 * 2,
+            inode: inode,
+            created: windows_time_to_unix_time(metadata.creation_time()),
+            accessed: windows_time_to_unix_time(metadata.last_access_time()),
+            modified: windows_time_to_unix_time(metadata.last_write_time()),
         })
     }
 }
 
+#[cfg(windows)]
+// https://doc.rust-lang.org/std/os/windows/fs/trait.MetadataExt.html#tymethod.creation_time
+// "The returned 64-bit value [...] which represents the number of 100-nanosecond intervals since January 1, 1601 (UTC)."
+fn windows_time_to_unix_time(win_time: u64) -> u64 {
+    win_time / 10_000 - 11_644_473_600_000
+}
+
+#[cfg(windows)]
+fn get_size_on_disk(path: &PathBuf) -> u64 {
+    let mut size_on_disk = 0;
+
+    // bind file so it stays in scope until end of function
+    // if it goes out of scope the handle below becomes invalid
+    let file = match fs::File::open(path) {
+        Ok(file) => file,
+        Err(_) => return size_on_disk, // opening directories will fail
+    };
+
+    let handle = file.as_raw_handle();
+
+    unsafe {
+        let mut file_info: FILE_STANDARD_INFO = core::mem::zeroed();
+        let file_info_ptr: *mut FILE_STANDARD_INFO = &mut file_info;
+
+        let success = GetFileInformationByHandleEx(
+            handle,
+            FileStandardInfo,
+            file_info_ptr as LPVOID,
+            std::mem::size_of::<FILE_STANDARD_INFO>() as DWORD,
+        );
+
+        if success != 0 {
+            size_on_disk = *file_info.AllocationSize.QuadPart() as u64;
+        }
+    }
+
+    size_on_disk
+}
+
+#[cfg(windows)]
+fn get_inode(path: &PathBuf) -> Option<u128> {
+    let mut inode = None;
+
+    let file = match fs::File::open(path) {
+        Ok(file) => file,
+        Err(_) => return inode,
+    };
+
+    let handle = file.as_raw_handle();
+
+    unsafe {
+        let mut file_info: FILE_ID_INFO = core::mem::zeroed();
+        let file_info_ptr: *mut FILE_ID_INFO = &mut file_info;
+
+        let success = GetFileInformationByHandleEx(
+            handle,
+            FileIdInfo,
+            file_info_ptr as LPVOID,
+            std::mem::size_of::<FILE_ID_INFO>() as DWORD,
+        );
+
+        if success != 0 {
+            inode = Some(std::mem::transmute::<FILE_ID_128, u128>(file_info.FileId));
+        }
+    }
+
+    inode
+}
+
 fn unit_string_to_number(s: &str) -> Option<u64> {
     let mut offset = 0;
     let mut s_chars = s.chars().rev();
@@ -137,7 +239,7 @@ fn du(
     mut my_stat: Stat,
     options: &Options,
     depth: usize,
-    inodes: &mut HashSet<u64>,
+    inodes: &mut HashSet<u128>,
 ) -> Box<dyn DoubleEndedIterator<Item = Stat>> {
     let mut stats = vec![];
     let mut futures = vec![];
@@ -164,10 +266,13 @@ fn du(
                         if this_stat.is_dir {
                             futures.push(du(this_stat, options, depth + 1, inodes));
                         } else {
-                            if inodes.contains(&this_stat.inode) {
-                                continue;
+                            if this_stat.inode.is_some() {
+                                let inode = this_stat.inode.unwrap();
+                                if inodes.contains(&inode) {
+                                    continue;
+                                }
+                                inodes.insert(inode);
                             }
-                            inodes.insert(this_stat.inode);
                             my_stat.size += this_stat.size;
                             my_stat.blocks += this_stat.blocks;
                             if options.all {
@@ -418,7 +523,7 @@ Try '{} --help' for more information.",
         let path = PathBuf::from(&path_str);
         match Stat::new(path) {
             Ok(stat) => {
-                let mut inodes: HashSet<u64> = HashSet::new();
+                let mut inodes: HashSet<u128> = HashSet::new();
 
                 let iter = du(stat, &options, 0, &mut inodes);
                 let (_, len) = iter.size_hint();

tests/by-util/test_du.rs

@@ -45,7 +45,11 @@ fn test_du_basics_subdir() {
 fn _du_basics_subdir(s: String) {
     assert_eq!(s, "4\tsubdir/deeper\n");
 }
-#[cfg(not(target_vendor = "apple"))]
+#[cfg(target_os = "windows")]
+fn _du_basics_subdir(s: String) {
+    assert_eq!(s, "0\tsubdir/deeper\n");
+}
+#[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_basics_subdir(s: String) {
     // MS-WSL linux has altered expected output
     if !is_wsl() {
@@ -71,7 +75,7 @@ fn test_du_basics_bad_name() {
 fn test_du_soft_link() {
     let ts = TestScenario::new("du");
 
-    let link = ts.cmd("ln").arg("-s").arg(SUB_FILE).arg(SUB_LINK).run();
+    let link = ts.ccmd("ln").arg("-s").arg(SUB_FILE).arg(SUB_LINK).run();
     assert!(link.success);
 
     let result = ts.ucmd().arg(SUB_DIR_LINKS).run();
@@ -85,7 +89,11 @@ fn _du_soft_link(s: String) {
     // 'macos' host variants may have `du` output variation for soft links
     assert!((s == "12\tsubdir/links\n") || (s == "16\tsubdir/links\n"));
 }
-#[cfg(not(target_vendor = "apple"))]
+#[cfg(target_os = "windows")]
+fn _du_soft_link(s: String) {
+    assert_eq!(s, "8\tsubdir/links\n");
+}
+#[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_soft_link(s: String) {
     // MS-WSL linux has altered expected output
     if !is_wsl() {
@@ -99,7 +107,7 @@ fn _du_soft_link(s: String) {
 fn test_du_hard_link() {
     let ts = TestScenario::new("du");
 
-    let link = ts.cmd("ln").arg(SUB_FILE).arg(SUB_LINK).run();
+    let link = ts.ccmd("ln").arg(SUB_FILE).arg(SUB_LINK).run();
     assert!(link.success);
 
     let result = ts.ucmd().arg(SUB_DIR_LINKS).run();
@@ -113,7 +121,11 @@ fn test_du_hard_link() {
 fn _du_hard_link(s: String) {
     assert_eq!(s, "12\tsubdir/links\n")
 }
-#[cfg(not(target_vendor = "apple"))]
+#[cfg(target_os = "windows")]
+fn _du_hard_link(s: String) {
+    assert_eq!(s, "8\tsubdir/links\n")
+}
+#[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_hard_link(s: String) {
     // MS-WSL linux has altered expected output
     if !is_wsl() {
@@ -137,7 +149,11 @@ fn test_du_d_flag() {
 fn _du_d_flag(s: String) {
     assert_eq!(s, "16\t./subdir\n20\t./\n");
 }
-#[cfg(not(target_vendor = "apple"))]
+#[cfg(target_os = "windows")]
+fn _du_d_flag(s: String) {
+    assert_eq!(s, "8\t./subdir\n8\t./\n");
+}
+#[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_d_flag(s: String) {
     // MS-WSL linux has altered expected output
     if !is_wsl() {

tests/common/util.rs

@@ -589,6 +589,14 @@ impl TestScenario {
         UCommand::new_from_tmp(bin, self.tmpd.clone(), true)
     }
 
+    /// Returns builder for invoking any uutils command. Paths given are treated
+    /// relative to the environment's unique temporary test directory.
+    pub fn ccmd<S: AsRef<OsStr>>(&self, bin: S) -> UCommand {
+        let mut cmd = self.cmd(&self.bin_path);
+        cmd.arg(bin);
+        cmd
+    }
+
     // different names are used rather than an argument
     // because the need to keep the environment is exceedingly rare.
     pub fn ucmd_keepenv(&self) -> UCommand {