Switched PCRE backend to **much** better PCRE2 library. (Allows for concurrency directly, etc.)

TODO: Port this (XXX: or the other way round, maybe? Import `args.rs` module to here??) to `cli-refactor` branch

Fortune for rematch's current commit: Small curse − 小凶

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "rematch"
-version = "0.1.0"
+version = "0.1.0+1"
 authors = ["Avril <flanchan@cumallover.me>"]
 edition = "2018"
 
@@ -8,13 +8,15 @@ edition = "2018"
 
 [profile.release]
 opt-level = 3
-lto = "fat"
+lto = true
 codegen-units = 1
 panic = "unwind"
+strip = true
 
 [features]
-perl = ["pcre"]
+default= ["perl"]
+perl = ["dep:pcre2"]
 
 [dependencies]
 regex = "1"
-pcre = { version = "0.2.3", optional = true }
+pcre2 = { version = "0.2.9", optional = true }

src/main.rs

@@ -3,6 +3,18 @@
 mod re;
 mod text;
 
+fn print_group<S: ?Sized, G, T>(to: &mut S, g: G, group: usize) -> std::io::Result<()>
+where S: std::io::Write,
+      G: IntoIterator<Item = Option<T>>,
+      T: std::borrow::Borrow<str>
+{
+    match g.into_iter().nth(group) {
+	Some(None) => writeln!(to, ""),
+	Some(Some(g)) => writeln!(to, "{}", g.borrow()),
+	None => Ok(()),
+    }
+}
+
 fn main() -> Result<(), Box<dyn std::error::Error>>
 {
     let args: Vec<String> = std::env::args().collect();
@@ -16,10 +28,14 @@ fn main() -> Result<(), Box<dyn std::error::Error>>
 	let text = &args[1];
 	let group: usize = args[3].parse().expect("Invalid group number.");
 
+	use std::io::Write;
+	let mut stdout = std::io::stdout();
+	
 	if text == "-" {
-	    text::stdin_lines(|text| -> Result<bool, re::Error> {
+	    text::stdin_lines(|text| -> Result<bool, Box<dyn std::error::Error + Send + Sync + 'static>> {
+		let mut stdout = stdout.lock();
 		match re.exec(&text)? {
-		    Some(g) if g.len() > group => println!("{}", &g[group]),
+		    Some(g) if g.len() > group => print_group(&mut stdout, g, group)?, //println!("{}", &g[group]),
 		    _ => (),
 		}
 		Ok(true)
@@ -27,10 +43,11 @@ fn main() -> Result<(), Box<dyn std::error::Error>>
 	} else {
 
 	    match re.exec(&text)? {
-		Some(g) if g.len() > group => println!("{}", &g[group]),
+		Some(g) if g.len() > group => print_group(&mut stdout, g, group)?,//println!("{}", &g.nth(group).unwrap().map(|x| x.as_ref()).unwrap_or("")),
 		_ => (),
 	    }
 	}
+	stdout.flush().unwrap();
     }
     Ok(())
 }

src/re.rs

@@ -1,23 +1,23 @@
-#![allow(unused_imports)]
-
 use std::{
     error,
-    fmt::{
-	self,
-	Write,
-    },
-    sync::{
-	Arc,
-	Mutex,
-    }
+    fmt,
+    borrow::Cow,
 };
 
-pub type Groups = Vec<String>;
+pub type FrozenVec<T> = Box<[T]>;
+pub type FrozenString = Box<str>;
+
+// NOTE: Currently unused, as we use `to_utf8_lossy()` for PCRE2 `byte`-matching (XXX: Should we change?)
+// TODO: to return some kind of `Either<&'s str, impl bytes::Buf + 's>` type, which would use `str` on non-PCRE, but opaque `bytes::Buf` on PCRE?)
+pub type FrozenBytes = FrozenVec<u8>; 
+
+pub type Groups<String = FrozenString> = FrozenVec<Option<String>>;
 
+#[derive(Debug, Clone)]
 pub struct Regex
 {
     #[cfg(feature="perl")]
-    internal: Arc<Mutex<pcre::Pcre>>,
+    internal: pcre2::bytes::Regex,
     #[cfg(not(feature = "perl"))]
     internal: regex::Regex,
 }
@@ -50,24 +50,18 @@ impl Regex {
     pub fn compile(string: impl AsRef<str>) -> Result<Self, Error> 
     {
 	#[cfg(feature = "perl")]
-	return Ok(Self{internal: Arc::new(Mutex::new(pcre::Pcre::compile(string.as_ref())?))});
+	return Ok(Self{internal: pcre2::bytes::RegexBuilder::new().build(string.as_ref())?});
 	#[cfg(not(feature = "perl"))]
 	return Ok(Self{internal: regex::Regex::new(string.as_ref())?});
     }
 
-    pub fn exec(&self, string: impl AsRef<str>) -> Result<Option<Groups>, Error>
+    pub fn exec<'s>(&self, string: &'s str) -> Result<Option<Groups<Cow<'s, str>>>, Error>
     {
 	#[cfg(feature = "perl")]
 	return {
-	    let mut re = self.internal.lock().unwrap();
-	    Ok(match re.exec(string.as_ref()) {
+	    Ok(match self.internal.captures(string.as_ref())? {
 		Some(m) => {
-		    let len = m.string_count();
-		    let mut output = Vec::with_capacity(len);
-		    for i in 0..len {
-			output.push(m.group(i).to_owned());
-		    }
-		    Some(output)
+		    Some((0..m.len()).map(move |i| m.get(i).map(|x| String::from_utf8_lossy(x.as_bytes()) )).collect())
 		},
 		None => None,
 	    })
@@ -76,14 +70,7 @@ impl Regex {
 	return {
 	    Ok(match self.internal.captures(string.as_ref()) {
 		Some(m) => {
-		    let mut output = Vec::with_capacity(m.len());
-		    for i in 0..m.len() {
-			let ma = m.get(i).unwrap();
-			let mut op = String::with_capacity(ma.range().len());
-			write!(op, "{}", ma.as_str())?;
-			output.push(op);
-		    }
-		    Some(output)
+		    Some((0..m.len()).map(move |i| m.get(i).map(|x| Cow::Borrowed(x.as_str()) )).collect())
 		},
 		None => None,
 	    })
@@ -99,7 +86,7 @@ impl From<fmt::Error> for Error
     }
 }
 
-#[cfg(not(feature = "perl"))]
+//#[cfg(not(feature = "perl"))]
 impl From<regex::Error> for Error
 {
     fn from(er: regex::Error) -> Self
@@ -109,9 +96,9 @@ impl From<regex::Error> for Error
 }
 
 #[cfg(feature = "perl")]
-impl From<pcre::CompilationError> for Error
+impl From<pcre2::Error> for Error
 {
-    fn from(er: pcre::CompilationError) -> Self
+    fn from(er: pcre2::Error) -> Self
     {
 	Self::Compile(format!("{}", er))
     }