From 770021df6a414e3398331ed4985ba3fa282b1323 Mon Sep 17 00:00:00 2001
From: Mark Poliakov <mark@alnyan.me>
Date: Sat, 1 Mar 2025 01:37:31 +0200
Subject: [PATCH] sysutils: basic grep-like utility

---
 userspace/sysutils/Cargo.toml  |   6 ++
 userspace/sysutils/src/grep.rs | 102 +++++++++++++++++++++++++++++++++
 userspace/sysutils/src/tst.rs  |   6 +-
 xtask/src/build/userspace.rs   |   1 +
 4 files changed, 114 insertions(+), 1 deletion(-)
 create mode 100644 userspace/sysutils/src/grep.rs

diff --git a/userspace/sysutils/Cargo.toml b/userspace/sysutils/Cargo.toml
index e5680c71..e8804082 100644
--- a/userspace/sysutils/Cargo.toml
+++ b/userspace/sysutils/Cargo.toml
@@ -24,6 +24,8 @@ serde_json.workspace = true
 sha2.workspace = true
 chrono.workspace = true
 
+# Own regex implementation?
+regex = "1.11.1"
 pci-ids = { version = "0.2.5" }
 
 init = { path = "../init" }
@@ -97,6 +99,10 @@ path = "src/dd.rs"
 name = "view"
 path = "src/view.rs"
 
+[[bin]]
+name = "grep"
+path = "src/grep.rs"
+
 [[bin]]
 name = "chmod"
 path = "src/chmod.rs"
diff --git a/userspace/sysutils/src/grep.rs b/userspace/sysutils/src/grep.rs
new file mode 100644
index 00000000..0acfe9cf
--- /dev/null
+++ b/userspace/sysutils/src/grep.rs
@@ -0,0 +1,102 @@
+#![feature(seek_stream_len, slice_split_once, slice_internals)]
+#![allow(internal_features)]
+use std::{
+    fs::File,
+    io::{self, stdout, Seek, Write},
+    path::{Path, PathBuf},
+    process::ExitCode,
+};
+
+use clap::Parser;
+use cross::mem::FileMapping;
+use regex::bytes::Regex;
+
+#[derive(Debug, thiserror::Error)]
+enum Error {
+    #[error("{0}")]
+    Io(#[from] io::Error),
+    #[error("Invalid expression: {0}")]
+    Regex(#[from] regex::Error),
+}
+
+trait Input {
+    fn next_line(&mut self) -> io::Result<Option<&[u8]>>;
+}
+
+struct FileInput {
+    mmap: FileMapping,
+    position: usize,
+}
+
+impl FileInput {
+    pub fn open<P: AsRef<Path>>(path: P) -> io::Result<Self> {
+        let mut file = File::open(path)?;
+        let size = file.stream_len()?.try_into().unwrap();
+        // TODO handle non-mmap()able files
+        let mmap = FileMapping::map(file, size)?;
+        Ok(Self { mmap, position: 0 })
+    }
+}
+
+impl Input for FileInput {
+    fn next_line(&mut self) -> io::Result<Option<&[u8]>> {
+        if self.position >= self.mmap.len() {
+            return Ok(None);
+        }
+
+        if let Some(nl) = core::slice::memchr::memchr(b'\n', &self.mmap[self.position..]) {
+            let line = &self.mmap[self.position..self.position + nl];
+            self.position += line.len() + 1;
+            Ok(Some(line))
+        } else {
+            let slice = &self.mmap[self.position..];
+            self.position = self.mmap.len();
+            Ok(Some(slice))
+        }
+    }
+}
+
+fn run_on<I: Input>(input: &mut I, regex: &Regex) -> io::Result<bool> {
+    let mut any = false;
+    loop {
+        let Some(line) = input.next_line()? else {
+            break;
+        };
+
+        if regex.is_match(line) {
+            any = true;
+            stdout().write_all(line).ok();
+            stdout().write_all(b"\n").ok();
+        }
+    }
+    Ok(any)
+}
+
+#[derive(Debug, Parser)]
+struct Args {
+    expression: String,
+    filename: PathBuf,
+}
+
+fn run(args: &Args) -> Result<bool, Error> {
+    let regex = Regex::new(&args.expression)?;
+    let mut input = FileInput::open(&args.filename)?;
+    let any = run_on(&mut input, &regex)?;
+    Ok(any)
+}
+
+fn main() -> ExitCode {
+    let args = Args::parse();
+    match run(&args) {
+        Ok(true) => {
+            ExitCode::SUCCESS
+        }
+        Ok(false) => {
+            ExitCode::FAILURE
+        }
+        Err(error) => {
+            eprintln!("{error}");
+            ExitCode::FAILURE
+        }
+    }
+}
diff --git a/userspace/sysutils/src/tst.rs b/userspace/sysutils/src/tst.rs
index f328e4d9..f77cc528 100644
--- a/userspace/sysutils/src/tst.rs
+++ b/userspace/sysutils/src/tst.rs
@@ -1 +1,5 @@
-fn main() {}
+fn main() {
+    for i in 0..8 {
+        println!("\x1B[3{i}mHello\x1B[1molleH\x1B[0m");
+    }
+}
diff --git a/xtask/src/build/userspace.rs b/xtask/src/build/userspace.rs
index b3376c8b..41ddf620 100644
--- a/xtask/src/build/userspace.rs
+++ b/xtask/src/build/userspace.rs
@@ -40,6 +40,7 @@ const PROGRAMS: &[(&str, &str)] = &[
     ("dd", "bin/dd"),
     ("random", "bin/random"),
     ("view", "bin/view"),
+    ("grep", "bin/grep"),
     ("chmod", "bin/chmod"),
     ("sha256sum", "bin/sha256sum"),
     ("sysmon", "bin/sysmon"),