perf: Improve dfa matching

.travis.yml

@@ -2,7 +2,7 @@ dist: trusty
 sudo: required
 language: rust
 rust:
-  - 1.20.0
+  - 1.21.0
   - stable
   - beta
   - nightly

README.md

@@ -210,7 +210,7 @@ recommended for general use.
 
 ### Minimum Rust version policy
 
-This crate's minimum supported `rustc` version is `1.20.0`.
+This crate's minimum supported `rustc` version is `1.21.0`.
 
 The current **tentative** policy is that the minimum Rust version required
 to use this crate can be increased in minor version updates. For example, if

bench/src/bench.rs

@@ -44,9 +44,9 @@ pub use ffi::re2::Regex;
 #[cfg(feature = "re-dphobos")]
 pub use ffi::d_phobos::Regex;
 #[cfg(feature = "re-rust")]
-pub use regex::Regex;
+pub use regex::{Regex, RegexSet};
 #[cfg(feature = "re-rust-bytes")]
-pub use regex::bytes::Regex;
+pub use regex::bytes::{Regex, RegexSet};
 #[cfg(feature = "re-tcl")]
 pub use ffi::tcl::Regex;
 
@@ -272,6 +272,56 @@ macro_rules! bench_captures {
     }
 }
 
+macro_rules! bench_is_match_set {
+    ($name:ident, $is_match:expr, $re:expr, $haystack:expr) => {
+        #[bench]
+        fn $name(b: &mut Bencher) {
+            use std::sync::Mutex;
+            lazy_static! {
+                static ref RE: Mutex<RegexSet> = Mutex::new($re);
+                static ref TEXT: Mutex<Text> = Mutex::new(text!($haystack));
+            };
+            let re = RE.lock().unwrap();
+            let text = TEXT.lock().unwrap();
+            b.bytes = text.len() as u64;
+            b.iter(|| {
+                if re.is_match(&text) != $is_match {
+                    if $is_match {
+                        panic!("expected match, got not match");
+                    } else {
+                        panic!("expected no match, got match");
+                    }
+                }
+            });
+        }
+    }
+}
+
+macro_rules! bench_matches_set {
+    ($name:ident, $is_match:expr, $re:expr, $haystack:expr) => {
+        #[bench]
+        fn $name(b: &mut Bencher) {
+            use std::sync::Mutex;
+            lazy_static! {
+                static ref RE: Mutex<RegexSet> = Mutex::new($re);
+                static ref TEXT: Mutex<Text> = Mutex::new(text!($haystack));
+            };
+            let re = RE.lock().unwrap();
+            let text = TEXT.lock().unwrap();
+            b.bytes = text.len() as u64;
+            b.iter(|| {
+                if re.matches(&text).matched_any() != $is_match {
+                    if $is_match {
+                        panic!("expected match, got not match");
+                    } else {
+                        panic!("expected no match, got match");
+                    }
+                }
+            });
+        }
+    }
+}
+
 mod ffi;
 mod misc;
 mod regexdna;

bench/src/misc.rs

@@ -14,6 +14,8 @@ use std::iter::repeat;
 
 use test::Bencher;
 
+#[cfg(any(feature = "re-rust", feature = "re-rust-bytes"))]
+use RegexSet;
 use {Regex, Text};
 
 #[cfg(not(feature = "re-onig"))]
@@ -278,3 +280,21 @@ bench_captures!(short_haystack_1000000x,
             repeat("aaaa").take(1000000).collect::<String>(),
             repeat("dddd").take(1000000).collect::<String>(),
             ));
+
+#[cfg(any(feature = "re-rust", feature = "re-rust-bytes"))]
+bench_is_match_set!(is_match_set,
+    true,
+    RegexSet::new(vec!["aaaaaaaaaaaaaaaaaaa", "abc579", "def.+", "e24fg", "a.*2c", "23."]).unwrap(),
+    format!("{}a482c{}",
+            repeat('a').take(10).collect::<String>(),
+            repeat('b').take(10).collect::<String>())
+    );
+
+#[cfg(any(feature = "re-rust", feature = "re-rust-bytes"))]
+bench_matches_set!(matches_set,
+    true,
+    RegexSet::new(vec!["aaaaaaaaaaaaaaaaaaa", "abc579", "def.+", "e24fg", "a.*2c", "23."]).unwrap(),
+    format!("{}a482c{}",
+            repeat('a').take(10).collect::<String>(),
+            repeat('b').take(10).collect::<String>())
+    );

ci/script.sh

@@ -3,7 +3,7 @@
 # This is the main CI script for testing the regex crate and its sub-crates.
 
 set -ex
-MSRV="1.20.0"
+MSRV="1.21.0"
 
 # If we're building on 1.20, then lazy_static 1.2 will fail to build since it
 # updated its MSRV to 1.24.1. In this case, we force the use of lazy_static 1.1