Use OsString for process names

Previously the stat file on Linux was parsed under the assumption that
it contains valid UTF-8. This is not necessarily the case for two
reasons:
1. Process names don't need to be valid UTF-8 at all on Linux. They can
   contain invalid bytes.
2. Even if the process name was valid UTF-8, the assumption breaks
   because the process name is limited to 15 characters, which can cut a
   valid code point in half.

This causes the stat file to not be parseable, causing the entire
process not to show up in the list. So even if you aren't using the name
of the process at all, you can't find it as part of the list.

One solution is to lossily convert the process name to a string. However
this means that the [Unicode replacement character
`�`](https://www.fileformat.info/info/unicode/char/fffd/index.htm) is
now part of those process names. The character when encoded as UTF-8 is
3 bytes. This means that process names can now be longer than 15 bytes,
or in other words, doing a name based comparison on the first 15 bytes,
such as suggested in the documentation of the crate, is no longer easily
possible.

The solution is to just provide the name as an `OsString` instead,
keeping the bytes around as is.

.cirrus.yml

@@ -1,10 +1,10 @@
 task:
-  name: rust 1.70 on freebsd 13
+  name: rust 1.74 on freebsd 13
   freebsd_instance:
     image: freebsd-13-1-release-amd64
   setup_script:
     - curl https://sh.rustup.rs -sSf --output rustup.sh
-    - sh rustup.sh -y --profile=minimal --default-toolchain=1.70
+    - sh rustup.sh -y --profile=minimal --default-toolchain=1.74
     - . $HOME/.cargo/env
     - rustup --version
     - rustup component add clippy
@@ -37,14 +37,14 @@ task:
     - FREEBSD_CI=1 cargo test --lib -j1 -- --ignored
 
 task:
-  name: rust 1.70 on mac m1
+  name: rust 1.74 on mac m1
   macos_instance:
     image: ghcr.io/cirruslabs/macos-monterey-base:latest
   setup_script:
     - brew update
     - brew install curl
     - curl https://sh.rustup.rs -sSf --output rustup.sh
-    - sh rustup.sh -y --profile=minimal --default-toolchain=1.70
+    - sh rustup.sh -y --profile=minimal --default-toolchain=1.74
     - source $HOME/.cargo/env
     - rustup --version
     - rustup component add clippy

.github/workflows/CI.yml

@@ -64,7 +64,7 @@ jobs:
           - { os: 'ubuntu-latest', target: 'x86_64-linux-android', cross: true }
           - { os: 'ubuntu-latest', target: 'i686-linux-android', cross: true }
         toolchain:
-          - "1.70.0"  # minimum supported rust version
+          - "1.74.0"  # minimum supported rust version
           - stable
           - nightly
     steps:
@@ -133,7 +133,7 @@ jobs:
           - macos-latest
           - windows-latest
         toolchain:
-          - "1.70.0"  # minimum supported rust version
+          - "1.74.0"  # minimum supported rust version
           - stable
           - nightly
     steps:
@@ -195,7 +195,7 @@ jobs:
     strategy:
       matrix:
         toolchain:
-          - "1.70.0" # minimum supported rust version
+          - "1.74.0" # minimum supported rust version
           - stable
     steps:
       - uses: actions/checkout@v4

Cargo.toml

@@ -6,7 +6,7 @@ description = "Library to get system information such as processes, CPUs, disks,
 repository = "https://github.com/GuillaumeGomez/sysinfo"
 license = "MIT"
 readme = "README.md"
-rust-version = "1.70"
+rust-version = "1.74"
 exclude = ["/test-unknown"]
 categories = ["filesystem", "os", "api-bindings"]
 edition = "2018"
@@ -42,6 +42,7 @@ rustdoc-args = ["--generate-link-to-definition"]
 
 [dependencies]
 cfg-if = "1.0"
+memchr = "2.7.1"
 rayon = { version = "^1.8", optional = true }
 serde = { version = "^1.0.190", optional = true }
 
@@ -94,3 +95,4 @@ tempfile = "3.9"
 
 [dev-dependencies]
 serde_json = "1.0" # Used in documentation tests.
+bstr = "1.9.0" # Used in example

README.md

@@ -18,7 +18,7 @@ You can still use `sysinfo` on non-supported OSes, it'll simply do nothing and a
 empty values. You can check in your program directly if an OS is supported by checking the
 [`IS_SUPPORTED_SYSTEM`] constant.
 
-The minimum-supported version of `rustc` is **1.70**.
+The minimum-supported version of `rustc` is **1.74**.
 
 ## Usage
 
@@ -68,7 +68,7 @@ println!("NB CPUs: {}", sys.cpus().len());
 
 // Display processes ID, name na disk usage:
 for (pid, process) in sys.processes() {
-    println!("[{pid}] {} {:?}", process.name(), process.disk_usage());
+    println!("[{pid}] {:?} {:?}", process.name(), process.disk_usage());
 }
 
 // We display all disks' information:
@@ -182,8 +182,8 @@ virtual systems.
 
 Apple has restrictions as to which APIs can be linked into binaries that are distributed through the app store.
 By default, `sysinfo` is not compatible with these restrictions. You can use the `apple-app-store`
-feature flag to disable the Apple prohibited features. This also enables the `apple-sandbox` feature. 
-In the case of applications using the sandbox outside of the app store, the `apple-sandbox` feature 
+feature flag to disable the Apple prohibited features. This also enables the `apple-sandbox` feature.
+In the case of applications using the sandbox outside of the app store, the `apple-sandbox` feature
 can be used alone to avoid causing policy violations at runtime.
 
 ### How it works

examples/simple.rs

@@ -4,6 +4,7 @@
 #![allow(unused_must_use, non_upper_case_globals)]
 #![allow(clippy::manual_range_contains)]
 
+use bstr::ByteSlice;
 use std::io::{self, BufRead, Write};
 use std::str::FromStr;
 use sysinfo::{Components, Disks, Networks, Pid, Signal, System, Users};
@@ -242,7 +243,7 @@ fn interpret_input(
                     &mut io::stdout(),
                     "{}:{} status={:?}",
                     pid,
-                    proc_.name(),
+                    proc_.name().as_encoded_bytes().as_bstr(),
                     proc_.status()
                 );
             }
@@ -289,8 +290,12 @@ fn interpret_input(
                 };
             } else {
                 let proc_name = tmp[1];
-                for proc_ in sys.processes_by_name(proc_name) {
-                    writeln!(&mut io::stdout(), "==== {} ====", proc_.name());
+                for proc_ in sys.processes_by_name(proc_name.as_ref()) {
+                    writeln!(
+                        &mut io::stdout(),
+                        "==== {} ====",
+                        proc_.name().as_encoded_bytes().as_bstr()
+                    );
                     writeln!(&mut io::stdout(), "{proc_:?}");
                 }
             }

src/common.rs

@@ -414,7 +414,7 @@ impl System {
     ///
     /// let s = System::new_all();
     /// for (pid, process) in s.processes() {
-    ///     println!("{} {}", pid, process.name());
+    ///     println!("{} {:?}", pid, process.name());
     /// }
     /// ```
     pub fn processes(&self) -> &HashMap<Pid, Process> {
@@ -428,7 +428,7 @@ impl System {
     ///
     /// let s = System::new_all();
     /// if let Some(process) = s.process(Pid::from(1337)) {
-    ///     println!("{}", process.name());
+    ///     println!("{:?}", process.name());
     /// }
     /// ```
     pub fn process(&self, pid: Pid) -> Option<&Process> {
@@ -450,17 +450,18 @@ impl System {
     /// use sysinfo::System;
     ///
     /// let s = System::new_all();
-    /// for process in s.processes_by_name("htop") {
-    ///     println!("{} {}", process.pid(), process.name());
+    /// for process in s.processes_by_name("htop".as_ref()) {
+    ///     println!("{} {:?}", process.pid(), process.name());
     /// }
     /// ```
     pub fn processes_by_name<'a: 'b, 'b>(
         &'a self,
-        name: &'b str,
+        name: &'b OsStr,
     ) -> impl Iterator<Item = &'a Process> + 'b {
+        let finder = memchr::memmem::Finder::new(name.as_encoded_bytes());
         self.processes()
             .values()
-            .filter(move |val: &&Process| val.name().contains(name))
+            .filter(move |val: &&Process| finder.find(val.name().as_encoded_bytes()).is_some())
     }
 
     /// Returns an iterator of processes with exactly the given `name`.
@@ -478,13 +479,13 @@ impl System {
     /// use sysinfo::System;
     ///
     /// let s = System::new_all();
-    /// for process in s.processes_by_exact_name("htop") {
-    ///     println!("{} {}", process.pid(), process.name());
+    /// for process in s.processes_by_exact_name("htop".as_ref()) {
+    ///     println!("{} {:?}", process.pid(), process.name());
     /// }
     /// ```
     pub fn processes_by_exact_name<'a: 'b, 'b>(
         &'a self,
-        name: &'b str,
+        name: &'b OsStr,
     ) -> impl Iterator<Item = &'a Process> + 'b {
         self.processes()
             .values()
@@ -832,7 +833,7 @@ impl System {
 ///
 /// let s = System::new_all();
 /// if let Some(process) = s.process(Pid::from(1337)) {
-///     println!("{}", process.name());
+///     println!("{:?}", process.name());
 /// }
 /// ```
 pub struct Process {
@@ -900,10 +901,10 @@ impl Process {
     ///
     /// let s = System::new_all();
     /// if let Some(process) = s.process(Pid::from(1337)) {
-    ///     println!("{}", process.name());
+    ///     println!("{:?}", process.name());
     /// }
     /// ```
-    pub fn name(&self) -> &str {
+    pub fn name(&self) -> &OsStr {
         self.inner.name()
     }
 

src/debug.rs

@@ -73,14 +73,7 @@ impl fmt::Debug for Process {
 
 impl fmt::Debug for Components {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(
-            f,
-            "Components {{ {} }}",
-            self.iter()
-                .map(|x| format!("{x:?}"))
-                .collect::<Vec<_>>()
-                .join(", ")
-        )
+        f.debug_list().entries(self.iter()).finish()
     }
 }
 
@@ -109,14 +102,7 @@ impl fmt::Debug for Component {
 
 impl fmt::Debug for Networks {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(
-            f,
-            "Networks {{ {} }}",
-            self.iter()
-                .map(|x| format!("{x:?}"))
-                .collect::<Vec<_>>()
-                .join(", ")
-        )
+        f.debug_list().entries(self.iter()).finish()
     }
 }
 
@@ -141,27 +127,13 @@ impl fmt::Debug for NetworkData {
 
 impl fmt::Debug for Disks {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(
-            f,
-            "Disks {{ {} }}",
-            self.iter()
-                .map(|x| format!("{x:?}"))
-                .collect::<Vec<_>>()
-                .join(", ")
-        )
+        f.debug_list().entries(self.iter()).finish()
     }
 }
 
 impl fmt::Debug for Users {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(
-            f,
-            "Users {{ {} }}",
-            self.iter()
-                .map(|x| format!("{x:?}"))
-                .collect::<Vec<_>>()
-                .join(", ")
-        )
+        f.debug_list().entries(self.iter()).finish()
     }
 }
 

src/unix/apple/app_store/process.rs

@@ -1,5 +1,6 @@
 // Take a look at the license at the top of the repository in the LICENSE file.
 
+use std::ffi::OsStr;
 use std::path::Path;
 
 use crate::{DiskUsage, Gid, Pid, ProcessStatus, Signal, Uid};
@@ -11,8 +12,8 @@ impl ProcessInner {
         None
     }
 
-    pub(crate) fn name(&self) -> &str {
-        ""
+    pub(crate) fn name(&self) -> &OsStr {
+        OsStr::new("")
     }
 
     pub(crate) fn cmd(&self) -> &[String] {

src/unix/apple/cpu.rs

@@ -186,7 +186,7 @@ pub(crate) unsafe fn get_cpu_frequency() -> u64 {
 
     #[cfg(any(target_os = "ios", feature = "apple-sandbox"))]
     {
-        return 0;
+        0
     }
     #[cfg(not(any(target_os = "ios", feature = "apple-sandbox")))]
     {

src/unix/apple/disk.rs

@@ -269,6 +269,7 @@ unsafe fn get_dict_value<T, F: FnOnce(*const c_void) -> Option<T>>(
 ) -> Option<T> {
     #[cfg(target_os = "macos")]
     let _defined;
+    #[allow(clippy::infallible_destructuring_match)]
     let key = match key {
         DictKey::Extern(val) => val,
         #[cfg(target_os = "macos")]