Fast charIn for ranges like 'A' to 'Z' (#360)

typelevel · Feb 10, 2022 · 371bd61 · 371bd61
1 parent 82ef549
commit 371bd61
Show file tree

Hide file tree

Showing 4 changed files with 53 additions and 10 deletions.
diff --git a/core/js/src/main/scala/cats/parse/BitSet.scala b/core/js/src/main/scala/cats/parse/BitSet.scala
@@ -56,4 +56,14 @@ object BitSetUtil {
 
     bs.flatMap { case (m, bs) => toIter(m, bs) }.toSet
   }
+
+  def bitSetForRange(count: Int): BitSet = {
+    val bs = new BitSet(count)
+    var cur = 0
+    while (cur < count) {
+      bs += cur
+      cur += 1
+    }
+    bs
+  }
 }
diff --git a/core/jvm/src/main/scala/cats/parse/BitSet.scala b/core/jvm/src/main/scala/cats/parse/BitSet.scala
@@ -61,4 +61,10 @@ object BitSetUtil {
 
     bs.flatMap { case (m, bs) => toIter(m, bs) }.toSet
   }
+
+  def bitSetForRange(count: Int): BitSet = {
+    val bs = new BitSet(count)
+    bs.flip(0, count)
+    bs
+  }
 }
diff --git a/core/shared/src/main/scala/cats/parse/Parser.scala b/core/shared/src/main/scala/cats/parse/Parser.scala
@@ -28,6 +28,7 @@ import cats.implicits._
 import scala.collection.immutable.SortedSet
 import scala.collection.mutable.ListBuffer
 import java.util.Arrays
+import scala.collection.immutable.NumericRange
 
 /** Parser0[A] attempts to extract an `A` value from the given input, potentially moving its offset
   * forward in the process.
@@ -1533,16 +1534,22 @@ object Parser {
   /** An empty iterable is the same as fail
     */
   def charIn(cs: Iterable[Char]): Parser[Char] =
-    if (cs.isEmpty) fail
-    else {
-      val ary = cs.toArray
-      Arrays.sort(ary)
-      rangesFor(ary) match {
-        case NonEmptyList((low, high), Nil) if low == Char.MinValue && high == Char.MaxValue =>
-          anyChar
-        case notAnyChar =>
-          Impl.CharIn(ary(0).toInt, BitSetUtil.bitSetFor(ary), notAnyChar)
-      }
+    cs match {
+      case _ if cs.isEmpty => fail
+      case Impl.CharsRange(Char.MinValue, Char.MaxValue) =>
+        anyChar
+      case Impl.CharsRange(start, end) =>
+        val bitSet = BitSetUtil.bitSetForRange(end.toInt - start.toInt + 1)
+        Impl.CharIn(start.toInt, bitSet, NonEmptyList.one(start -> end))
+      case _ =>
+        val ary = cs.toArray
+        Arrays.sort(ary)
+        rangesFor(ary) match {
+          case NonEmptyList((low, high), Nil) if low == Char.MinValue && high == Char.MaxValue =>
+            anyChar
+          case notAnyChar =>
+            Impl.CharIn(ary(0).toInt, BitSetUtil.bitSetFor(ary), notAnyChar)
+        }
     }
 
   /** Parse any single character in a set of characters as lower or upper case
@@ -2988,6 +2995,10 @@ object Parser {
         a
       }
     }
+    object CharsRange {
+      def unapply(range: NumericRange.Inclusive[Char]): Option[(Char, Char)] =
+        if (range.step == 1) Some(range.start -> range.end) else None
+    }
   }
 }
 

diff --git a/core/shared/src/test/scala/cats/parse/ParserTest.scala b/core/shared/src/test/scala/cats/parse/ParserTest.scala
@@ -1181,6 +1181,22 @@ class ParserTest extends munit.ScalaCheckSuite {
     }
   }
 
+  property("charIn range matches charIn list") {
+    forAll { (c1: Char, c2: Char) =>
+      val start = c1.min(c2)
+      val end = c1.max(c2)
+
+      val p1 = Parser.charIn(start to end)
+      val p2 = Parser.charIn((start to end).toList)
+
+      assertEquals(p1, p2)
+    }
+  }
+
+  property("charIn full range == anyChar") {
+    assertEquals(Parser.charIn(Char.MinValue to Char.MaxValue), Parser.anyChar)
+  }
+
   property("Parser.end gives the right error") {
     forAll { (str: String) =>
       Parser.end.parse(str) match {