Merge pull request #7061 from DaringCuteSeal/cp-stream-2

cp: implement copying from streams

Author: sylvestre

src/uu/cp/Cargo.toml

@@ -28,6 +28,7 @@ quick-error = { workspace = true }
 selinux = { workspace = true, optional = true }
 uucore = { workspace = true, features = [
   "backup-control",
+  "buf-copy",
   "entries",
   "fs",
   "fsxattr",

src/uu/cp/src/cp.rs

@@ -10,7 +10,7 @@ use std::collections::{HashMap, HashSet};
 #[cfg(not(windows))]
 use std::ffi::CString;
 use std::ffi::OsString;
-use std::fs::{self, File, Metadata, OpenOptions, Permissions};
+use std::fs::{self, Metadata, OpenOptions, Permissions};
 use std::io;
 #[cfg(unix)]
 use std::os::unix::ffi::OsStrExt;
@@ -1963,6 +1963,7 @@ fn print_paths(parents: bool, source: &Path, dest: &Path) {
 ///
 /// * `Ok(())` - The file was copied successfully.
 /// * `Err(CopyError)` - An error occurred while copying the file.
+#[allow(clippy::too_many_arguments)]
 fn handle_copy_mode(
     source: &Path,
     dest: &Path,
@@ -1971,15 +1972,10 @@ fn handle_copy_mode(
     source_metadata: &Metadata,
     symlinked_files: &mut HashSet<FileInformation>,
     source_in_command_line: bool,
+    source_is_fifo: bool,
+    #[cfg(unix)] source_is_stream: bool,
 ) -> CopyResult<()> {
-    let source_file_type = source_metadata.file_type();
-
-    let source_is_symlink = source_file_type.is_symlink();
-
-    #[cfg(unix)]
-    let source_is_fifo = source_file_type.is_fifo();
-    #[cfg(not(unix))]
-    let source_is_fifo = false;
+    let source_is_symlink = source_metadata.is_symlink();
 
     match options.copy_mode {
         CopyMode::Link => {
@@ -2016,6 +2012,8 @@ fn handle_copy_mode(
                 source_is_symlink,
                 source_is_fifo,
                 symlinked_files,
+                #[cfg(unix)]
+                source_is_stream,
             )?;
         }
         CopyMode::SymLink => {
@@ -2036,6 +2034,8 @@ fn handle_copy_mode(
                             source_is_symlink,
                             source_is_fifo,
                             symlinked_files,
+                            #[cfg(unix)]
+                            source_is_stream,
                         )?;
                     }
                     update_control::UpdateMode::ReplaceNone => {
@@ -2066,6 +2066,8 @@ fn handle_copy_mode(
                                 source_is_symlink,
                                 source_is_fifo,
                                 symlinked_files,
+                                #[cfg(unix)]
+                                source_is_stream,
                             )?;
                         }
                     }
@@ -2079,6 +2081,8 @@ fn handle_copy_mode(
                     source_is_symlink,
                     source_is_fifo,
                     symlinked_files,
+                    #[cfg(unix)]
+                    source_is_stream,
                 )?;
             }
         }
@@ -2305,6 +2309,18 @@ fn copy_file(
 
     let dest_permissions = calculate_dest_permissions(dest, &source_metadata, options, context)?;
 
+    #[cfg(unix)]
+    let source_is_fifo = source_metadata.file_type().is_fifo();
+    #[cfg(not(unix))]
+    let source_is_fifo = false;
+
+    #[cfg(unix)]
+    let source_is_stream = source_is_fifo
+        || source_metadata.file_type().is_char_device()
+        || source_metadata.file_type().is_block_device();
+    #[cfg(not(unix))]
+    let source_is_stream = false;
+
     handle_copy_mode(
         source,
         dest,
@@ -2313,6 +2329,9 @@ fn copy_file(
         &source_metadata,
         symlinked_files,
         source_in_command_line,
+        source_is_fifo,
+        #[cfg(unix)]
+        source_is_stream,
     )?;
 
     // TODO: implement something similar to gnu's lchown
@@ -2328,8 +2347,16 @@ fn copy_file(
 
     if options.dereference(source_in_command_line) {
         if let Ok(src) = canonicalize(source, MissingHandling::Normal, ResolveMode::Physical) {
-            copy_attributes(&src, dest, &options.attributes)?;
+            if src.exists() {
+                copy_attributes(&src, dest, &options.attributes)?;
+            }
         }
+    } else if source_is_stream && source.exists() {
+        // Some stream files may not exist after we have copied it,
+        // like anonymous pipes. Thus, we can't really copy its
+        // attributes. However, this is already handled in the stream
+        // copy function (see `copy_stream` under platform/linux.rs).
+        copy_attributes(source, dest, &options.attributes)?;
     } else {
         copy_attributes(source, dest, &options.attributes)?;
     }
@@ -2393,6 +2420,7 @@ fn handle_no_preserve_mode(options: &Options, org_mode: u32) -> u32 {
 
 /// Copy the file from `source` to `dest` either using the normal `fs::copy` or a
 /// copy-on-write scheme if --reflink is specified and the filesystem supports it.
+#[allow(clippy::too_many_arguments)]
 fn copy_helper(
     source: &Path,
     dest: &Path,
@@ -2401,6 +2429,7 @@ fn copy_helper(
     source_is_symlink: bool,
     source_is_fifo: bool,
     symlinked_files: &mut HashSet<FileInformation>,
+    #[cfg(unix)] source_is_stream: bool,
 ) -> CopyResult<()> {
     if options.parents {
         let parent = dest.parent().unwrap_or(dest);
@@ -2411,12 +2440,7 @@ fn copy_helper(
         return Err(Error::NotADirectory(dest.to_path_buf()));
     }
 
-    if source.as_os_str() == "/dev/null" {
-        /* workaround a limitation of fs::copy
-         * https://github.com/rust-lang/rust/issues/79390
-         */
-        File::create(dest).context(dest.display().to_string())?;
-    } else if source_is_fifo && options.recursive && !options.copy_contents {
+    if source_is_fifo && options.recursive && !options.copy_contents {
         #[cfg(unix)]
         copy_fifo(dest, options.overwrite, options.debug)?;
     } else if source_is_symlink {
@@ -2428,8 +2452,10 @@ fn copy_helper(
             options.reflink_mode,
             options.sparse_mode,
             context,
-            #[cfg(any(target_os = "linux", target_os = "android", target_os = "macos"))]
+            #[cfg(unix)]
             source_is_fifo,
+            #[cfg(unix)]
+            source_is_stream,
         )?;
 
         if !options.attributes_only && options.debug {

src/uu/cp/src/platform/linux.rs

@@ -12,6 +12,7 @@ use std::os::unix::fs::MetadataExt;
 use std::os::unix::fs::{FileTypeExt, OpenOptionsExt};
 use std::os::unix::io::AsRawFd;
 use std::path::Path;
+use uucore::buf_copy;
 
 use quick_error::ResultExt;
 
@@ -220,8 +221,9 @@ fn check_dest_is_fifo(dest: &Path) -> bool {
     }
 }
 
-/// Copy the contents of the given source FIFO to the given file.
-fn copy_fifo_contents<P>(source: P, dest: P) -> std::io::Result<u64>
+/// Copy the contents of a stream from `source` to `dest`. The `if_fifo` argument is used to
+/// determine if we need to modify the file's attributes before and after copying.
+fn copy_stream<P>(source: P, dest: P, is_fifo: bool) -> std::io::Result<u64>
 where
     P: AsRef<Path>,
 {
@@ -250,8 +252,14 @@ where
         .write(true)
         .mode(mode)
         .open(&dest)?;
-    let num_bytes_copied = std::io::copy(&mut src_file, &mut dst_file)?;
-    dst_file.set_permissions(src_file.metadata()?.permissions())?;
+
+    let num_bytes_copied = buf_copy::copy_stream(&mut src_file, &mut dst_file)
+        .map_err(|_| std::io::Error::from(std::io::ErrorKind::Other))?;
+
+    if is_fifo {
+        dst_file.set_permissions(src_file.metadata()?.permissions())?;
+    }
+
     Ok(num_bytes_copied)
 }
 
@@ -268,6 +276,7 @@ pub(crate) fn copy_on_write(
     sparse_mode: SparseMode,
     context: &str,
     source_is_fifo: bool,
+    source_is_stream: bool,
 ) -> CopyResult<CopyDebug> {
     let mut copy_debug = CopyDebug {
         offload: OffloadReflinkDebug::Unknown,
@@ -279,10 +288,9 @@ pub(crate) fn copy_on_write(
             copy_debug.sparse_detection = SparseDebug::Zeros;
             // Default SparseDebug val for SparseMode::Always
             copy_debug.reflink = OffloadReflinkDebug::No;
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Avoided;
-
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let mut copy_method = CopyMethod::Default;
                 let result = handle_reflink_never_sparse_always(source, dest);
@@ -300,10 +308,9 @@ pub(crate) fn copy_on_write(
         (ReflinkMode::Never, SparseMode::Never) => {
             copy_debug.reflink = OffloadReflinkDebug::No;
 
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Avoided;
-
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let result = handle_reflink_never_sparse_never(source);
                 if let Ok(debug) = result {
@@ -315,9 +322,9 @@ pub(crate) fn copy_on_write(
         (ReflinkMode::Never, SparseMode::Auto) => {
             copy_debug.reflink = OffloadReflinkDebug::No;
 
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Avoided;
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let mut copy_method = CopyMethod::Default;
                 let result = handle_reflink_never_sparse_auto(source, dest);
@@ -335,10 +342,9 @@ pub(crate) fn copy_on_write(
         (ReflinkMode::Auto, SparseMode::Always) => {
             copy_debug.sparse_detection = SparseDebug::Zeros; // Default SparseDebug val for
                                                               // SparseMode::Always
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Avoided;
-
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let mut copy_method = CopyMethod::Default;
                 let result = handle_reflink_auto_sparse_always(source, dest);
@@ -356,9 +362,9 @@ pub(crate) fn copy_on_write(
 
         (ReflinkMode::Auto, SparseMode::Never) => {
             copy_debug.reflink = OffloadReflinkDebug::No;
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Avoided;
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let result = handle_reflink_auto_sparse_never(source);
                 if let Ok(debug) = result {
@@ -369,9 +375,9 @@ pub(crate) fn copy_on_write(
             }
         }
         (ReflinkMode::Auto, SparseMode::Auto) => {
-            if source_is_fifo {
+            if source_is_stream {
                 copy_debug.offload = OffloadReflinkDebug::Unsupported;
-                copy_fifo_contents(source, dest).map(|_| ())
+                copy_stream(source, dest, source_is_fifo).map(|_| ())
             } else {
                 let mut copy_method = CopyMethod::Default;
                 let result = handle_reflink_auto_sparse_auto(source, dest);

src/uu/cp/src/platform/macos.rs

@@ -4,12 +4,14 @@
 // file that was distributed with this source code.
 // spell-checker:ignore reflink
 use std::ffi::CString;
-use std::fs::{self, File};
-use std::io;
+use std::fs::{self, File, OpenOptions};
 use std::os::unix::ffi::OsStrExt;
+use std::os::unix::fs::OpenOptionsExt;
 use std::path::Path;
 
 use quick_error::ResultExt;
+use uucore::buf_copy;
+use uucore::mode::get_umask;
 
 use crate::{CopyDebug, CopyResult, OffloadReflinkDebug, ReflinkMode, SparseDebug, SparseMode};
 
@@ -24,6 +26,7 @@ pub(crate) fn copy_on_write(
     sparse_mode: SparseMode,
     context: &str,
     source_is_fifo: bool,
+    source_is_stream: bool,
 ) -> CopyResult<CopyDebug> {
     if sparse_mode != SparseMode::Auto {
         return Err("--sparse is only supported on linux".to_string().into());
@@ -85,10 +88,23 @@ pub(crate) fn copy_on_write(
             }
             _ => {
                 copy_debug.reflink = OffloadReflinkDebug::Yes;
-                if source_is_fifo {
+                if source_is_stream {
                     let mut src_file = File::open(source)?;
-                    let mut dst_file = File::create(dest)?;
-                    io::copy(&mut src_file, &mut dst_file).context(context)?
+                    let mode = 0o622 & !get_umask();
+                    let mut dst_file = OpenOptions::new()
+                        .create(true)
+                        .write(true)
+                        .mode(mode)
+                        .open(dest)?;
+
+                    let context = buf_copy::copy_stream(&mut src_file, &mut dst_file)
+                        .map_err(|_| std::io::Error::from(std::io::ErrorKind::Other))
+                        .context(context)?;
+
+                    if source_is_fifo {
+                        dst_file.set_permissions(src_file.metadata()?.permissions())?;
+                    }
+                    context
                 } else {
                     fs::copy(source, dest).context(context)?
                 }

src/uu/cp/src/platform/mod.rs

@@ -2,6 +2,18 @@
 //
 // For the full copyright and license information, please view the LICENSE
 // file that was distributed with this source code.
+
+#[cfg(all(
+    unix,
+    not(any(target_os = "macos", target_os = "linux", target_os = "android"))
+))]
+mod other_unix;
+#[cfg(all(
+    unix,
+    not(any(target_os = "macos", target_os = "linux", target_os = "android"))
+))]
+pub(crate) use self::other_unix::copy_on_write;
+
 #[cfg(target_os = "macos")]
 mod macos;
 #[cfg(target_os = "macos")]
@@ -12,7 +24,13 @@ mod linux;
 #[cfg(any(target_os = "linux", target_os = "android"))]
 pub(crate) use self::linux::copy_on_write;
 
-#[cfg(not(any(target_os = "linux", target_os = "android", target_os = "macos")))]
+#[cfg(not(any(
+    unix,
+    any(target_os = "macos", target_os = "linux", target_os = "android")
+)))]
 mod other;
-#[cfg(not(any(target_os = "linux", target_os = "android", target_os = "macos")))]
+#[cfg(not(any(
+    unix,
+    any(target_os = "macos", target_os = "linux", target_os = "android")
+)))]
 pub(crate) use self::other::copy_on_write;