Merge pull request #2109 from Kobzol/add-regex-automata-0.4.8

Add `regex-automata-0.4.8` benchmark

Author: Kobzol

collector/compile-benchmarks/README.md

@@ -44,6 +44,8 @@ They mostly consist of real-world crates.
 - **nalgebra-0.33.0**: A linear algebra library. It exercises the new trait solver
   in different ways than the old solver.
 - **regex-1.5.5**: A regular expression parser. Used by many Rust programs.
+- **regex-automata-0.4.8**: A regular expression matching engine. Used by `regex`, which is used by
+  many Rust programs.
 - **ripgrep-13.0.0**: A line-oriented search tool. A widely-used utility, and a
   binary crate.
 - **ripgrep-14.1.1**: A line-oriented search tool. A widely-used utility, and a

collector/compile-benchmarks/REUSE.toml

@@ -195,6 +195,11 @@ path = "regex-1.5.5/**"
 SPDX-FileCopyrightText = "The Rust Project Developers (see https://thanks.rust-lang.org)"
 SPDX-License-Identifier = "MIT OR Apache-2.0"
 
+[[annotations]]
+path = "regex-automata-0.4.8/**"
+SPDX-FileCopyrightText = "The Rust Project Developers (see https://thanks.rust-lang.org)"
+SPDX-License-Identifier = "MIT OR Apache-2.0"
+
 [[annotations]]
 path = "regression-31157/**"
 SPDX-FileCopyrightText = "The Rust Project Developers (see https://thanks.rust-lang.org)"

collector/compile-benchmarks/regex-automata-0.4.8/.cargo_vcs_info.json

@@ -0,0 +1,6 @@
+{
+  "git": {
+    "sha1": "58e16f50f07729bf856570d1a8be0de0b4d5e9e0"
+  },
+  "path_in_vcs": "regex-automata"
+}

collector/compile-benchmarks/regex-automata-0.4.8/0-println.patch

@@ -0,0 +1,12 @@
+diff --git a/src/dfa/accel.rs b/src/dfa/accel.rs
+index c0ba18ea..009d534b 100644
+--- a/src/dfa/accel.rs
++++ b/src/dfa/accel.rs
+@@ -186,6 +186,7 @@ impl<'a> Accels<&'a [AccelTy]> {
+     pub fn from_bytes_unchecked(
+         mut slice: &'a [u8],
+     ) -> Result<(Accels<&'a [AccelTy]>, usize), DeserializeError> {
++        std::println!("testing");
+         let slice_start = slice.as_ptr().as_usize();
+ 
+         let (accel_len, _) =

collector/compile-benchmarks/regex-automata-0.4.8/1-reorder-enum-variant.patch

@@ -0,0 +1,17 @@
+diff --git a/src/nfa/thompson/builder.rs b/src/nfa/thompson/builder.rs
+index 6b69e878..24564db9 100644
+--- a/src/nfa/thompson/builder.rs
++++ b/src/nfa/thompson/builder.rs
+@@ -55,10 +55,10 @@ enum State {
+     /// that `Sparse` is used for via `Union`. But this creates a more bloated
+     /// NFA with more epsilon transitions than is necessary in the special case
+     /// of character classes.
+-    Sparse { transitions: Vec<Transition> },
++    Look { look: Look, next: StateID },
+     /// A conditional epsilon transition satisfied via some sort of
+     /// look-around.
+-    Look { look: Look, next: StateID },
++    Sparse { transitions: Vec<Transition> },
+     /// An empty state that records the start of a capture location. This is an
+     /// unconditional epsilon transition like `Empty`, except it can be used to
+     /// record position information for a capture group when using the NFA for

collector/compile-benchmarks/regex-automata-0.4.8/2-replace-array-elem.patch

@@ -0,0 +1,13 @@
+diff --git a/src/util/alphabet.rs b/src/util/alphabet.rs
+index e0e4d2fc..de443a1f 100644
+--- a/src/util/alphabet.rs
++++ b/src/util/alphabet.rs
+@@ -798,7 +798,7 @@ impl ByteSet {
+     /// Return true if and only if this set is empty.
+     #[cfg_attr(feature = "perf-inline", inline(always))]
+     pub(crate) fn is_empty(&self) -> bool {
+-        self.bits.0 == [0, 0]
++        self.bits.0 == [0, 1]
+     }
+ 
+     /// Deserializes a byte set from the given slice. If the slice is of

collector/compile-benchmarks/regex-automata-0.4.8/3-add-empty-block.patch

@@ -0,0 +1,13 @@
+diff --git a/src/meta/limited.rs b/src/meta/limited.rs
+index 5653adc9..b6b9d379 100644
+--- a/src/meta/limited.rs
++++ b/src/meta/limited.rs
+@@ -49,6 +49,8 @@ pub(crate) fn dfa_try_search_half_rev(
+ ) -> Result<Option<HalfMatch>, RetryError> {
+     use crate::dfa::Automaton;
+ 
++    {}
++
+     let mut mat = None;
+     let mut sid = dfa.start_state_reverse(input)?;
+     if input.start() == input.end() {