Revert documentation format changes

Scalafmt 3.8.4 introduced some formatting bugs that are reverted with this pull request.

Issue in Scalafmt:
- scalameta/scalafmt#4716

Bad formatting got merged to cats-effect in:
- #4235

core/js/src/main/scala/cats/effect/IOApp.scala

@@ -77,7 +77,7 @@ import scala.util.Try
  *   }
  * }}}
  *
- * It is valid to define `val run` rather than `def run` because `IO` 's evaluation is lazy: it
+ * It is valid to define `val run` rather than `def run` because `IO`'s evaluation is lazy: it
  * will only run when the `main` method is invoked by the runtime.
  *
  * In the event that the process receives an interrupt signal (`SIGINT`) due to Ctrl-C (or any
@@ -121,7 +121,7 @@ import scala.util.Try
  *
  * However, with that said, there really is no substitute to benchmarking your own application.
  * Every application and scenario is unique, and you will always get the absolute best results
- * by performing your own tuning rather than trusting someone else's defaults. `IOApp` 's
+ * by performing your own tuning rather than trusting someone else's defaults. `IOApp`'s
  * defaults are very ''good'', but they are not perfect in all cases. One common example of this
  * is applications which maintain network or file I/O worker threads which are under heavy load
  * in steady-state operations. In such a performance profile, it is usually better to reduce the

core/jvm/src/main/scala/cats/effect/IOApp.scala

@@ -80,7 +80,7 @@ import java.util.concurrent.atomic.AtomicInteger
  *   }
  * }}}
  *
- * It is valid to define `val run` rather than `def run` because `IO` 's evaluation is lazy: it
+ * It is valid to define `val run` rather than `def run` because `IO`'s evaluation is lazy: it
  * will only run when the `main` method is invoked by the runtime.
  *
  * In the event that the process receives an interrupt signal (`SIGINT`) due to Ctrl-C (or any
@@ -124,7 +124,7 @@ import java.util.concurrent.atomic.AtomicInteger
  *
  * However, with that said, there really is no substitute to benchmarking your own application.
  * Every application and scenario is unique, and you will always get the absolute best results
- * by performing your own tuning rather than trusting someone else's defaults. `IOApp` 's
+ * by performing your own tuning rather than trusting someone else's defaults. `IOApp`'s
  * defaults are very ''good'', but they are not perfect in all cases. One common example of this
  * is applications which maintain network or file I/O worker threads which are under heavy load
  * in steady-state operations. In such a performance profile, it is usually better to reduce the

core/jvm/src/main/scala/cats/effect/IOPlatform.scala

@@ -57,7 +57,7 @@ abstract private[effect] class IOPlatform[+A] extends Serializable { self: IO[A]
    * The first line will run `program` for at most five seconds, interrupt the calculation, and
    * run the finalizers for as long as they need to complete. The second line will run `program`
    * for at most five seconds and then immediately release the latch, without interrupting
-   * `program` 's ongoing execution.
+   * `program`'s ongoing execution.
    *
    * In other words, this function probably doesn't do what you think it does, and you probably
    * don't want to use it outside of tests.

core/jvm/src/main/scala/cats/effect/unsafe/FiberMonitor.scala

@@ -36,10 +36,9 @@ import scala.concurrent.ExecutionContext
  *      `IO#evalOn`).
  *   1. Because `java.util.Collections.synchronizedMap` is a simple wrapper around any map which
  *      just synchronizes the access to the map through the built in JVM `synchronized`
- *      mechanism, we need several instances of these synchronized `WeakHashMap` s just to
- *      reduce contention between threads. A particular instance is selected using a thread
- *      local source of randomness using an instance of
- *      `java.util.concurrent.ThreadLocalRandom`.
+ *      mechanism, we need several instances of these synchronized `WeakHashMap`s just to reduce
+ *      contention between threads. A particular instance is selected using a thread local
+ *      source of randomness using an instance of `java.util.concurrent.ThreadLocalRandom`.
  */
 private[effect] sealed class FiberMonitor(
     // A reference to the compute pool of the `IORuntime` in which this suspended fiber bag

core/jvm/src/main/scala/cats/effect/unsafe/TimerHeap.scala

@@ -40,7 +40,7 @@ import java.util.concurrent.atomic.AtomicInteger
  *
  * In general, this heap is not threadsafe and modifications (insertion/removal) may only be
  * performed on its owner WorkerThread. The exception is that the callback value of nodes may be
- * `null` ed by other threads and published via data race.
+ * `null`ed by other threads and published via data race.
  *
  * Other threads may traverse the heap with the `steal` method during which they may `null` some
  * callbacks. This is entirely subject to data races.

core/jvm/src/main/scala/cats/effect/unsafe/WorkerThread.scala

@@ -157,7 +157,7 @@ private[effect] final class WorkerThread[P <: AnyRef](
    * continues with the worker thread run loop.
    *
    * @param fiber
-   *   the fiber that `cede` s/`autoCede`s
+   *   the fiber that `cede`s/`autoCede`s
    */
   def reschedule(fiber: Runnable): Unit = {
     if ((cedeBypass eq null) && queue.isEmpty()) {

core/native/src/main/scala/cats/effect/IOApp.scala

@@ -75,7 +75,7 @@ import scala.scalanative.meta.LinktimeInfo._
  *   }
  * }}}
  *
- * It is valid to define `val run` rather than `def run` because `IO` 's evaluation is lazy: it
+ * It is valid to define `val run` rather than `def run` because `IO`'s evaluation is lazy: it
  * will only run when the `main` method is invoked by the runtime.
  *
  * In the event that the process receives an interrupt signal (`SIGINT`) due to Ctrl-C (or any
@@ -119,7 +119,7 @@ import scala.scalanative.meta.LinktimeInfo._
  *
  * However, with that said, there really is no substitute to benchmarking your own application.
  * Every application and scenario is unique, and you will always get the absolute best results
- * by performing your own tuning rather than trusting someone else's defaults. `IOApp` 's
+ * by performing your own tuning rather than trusting someone else's defaults. `IOApp`'s
  * defaults are very ''good'', but they are not perfect in all cases. One common example of this
  * is applications which maintain network or file I/O worker threads which are under heavy load
  * in steady-state operations. In such a performance profile, it is usually better to reduce the

core/shared/src/main/scala/cats/effect/IO.scala

@@ -186,7 +186,7 @@ sealed abstract class IO[+A] private () extends IOPlatform[A] {
   /**
    * Materializes any sequenced exceptions into value space, where they may be handled.
    *
-   * This is analogous to the `catch` clause in `try` /`catch`, being the inverse of
+   * This is analogous to the `catch` clause in `try`/`catch`, being the inverse of
    * `IO.raiseError`. Thus:
    *
    * {{{
@@ -1310,7 +1310,7 @@ object IO extends IOCompanionPlatform with IOLowPriorityImplicits with TuplePara
    * The effect returns `Either[Option[IO[Unit]], A]` where:
    *   - right side `A` is an immediate result of computation (callback invocation will be
    *     dropped);
-   *   - left side `Option[IO[Unit]] ` is an optional finalizer to be run in the event that the
+   *   - left side `Option[IO[Unit]]` is an optional finalizer to be run in the event that the
    *     fiber running `asyncCheckAttempt(k)` is canceled.
    *
    * For example, here is a simplified version of `IO.fromCompletableFuture`:

core/shared/src/main/scala/cats/effect/ResourceApp.scala

@@ -45,7 +45,7 @@ import cats.syntax.all._
  * [[https://tpolecat.github.io/skunk/ Skunk]] and [[https://http4s.org Http4s]], but otherwise
  * it represents a relatively typical example of what the main class for a realistic Cats Effect
  * application might look like. Notably, the whole thing is enclosed in `Resource`, which is
- * `use` d at the very end. This kind of pattern is so common that `ResourceApp` defines a
+ * `use`d at the very end. This kind of pattern is so common that `ResourceApp` defines a
  * special trait which represents it. We can rewrite the above example:
  *
  * {{{

core/shared/src/main/scala/cats/effect/SyncIO.scala

@@ -87,7 +87,7 @@ sealed abstract class SyncIO[+A] private () extends Serializable {
   /**
    * Materializes any sequenced exceptions into value space, where they may be handled.
    *
-   * This is analogous to the `catch` clause in `try` /`catch`, being the inverse of
+   * This is analogous to the `catch` clause in `try`/`catch`, being the inverse of
    * `SyncIO.raiseError`. Thus:
    *
    * {{{

core/shared/src/main/scala/cats/effect/unsafe/IORuntimeBuilder.scala

@@ -20,7 +20,7 @@ package unsafe
 import scala.concurrent.ExecutionContext
 
 /**
- * Builder object for creating custom `IORuntime` s. Useful for creating [[IORuntime]] based on
+ * Builder object for creating custom `IORuntime`s. Useful for creating [[IORuntime]] based on
  * the default one but with some wrappers around execution contexts or custom shutdown hooks.
  */
 final class IORuntimeBuilder protected (

kernel/shared/src/main/scala/cats/effect/kernel/GenConcurrent.scala

@@ -37,7 +37,7 @@ trait GenConcurrent[F[_], E] extends GenSpawn[F, E] {
    * and cache the result. If `get` is sequenced multiple times `fa` will only be evaluated
    * once.
    *
-   * If all `get` s are canceled prior to `fa` completing, it will be canceled and evaluated
+   * If all `get`s are canceled prior to `fa` completing, it will be canceled and evaluated
    * again the next time `get` is sequenced.
    */
   def memoize[A](fa: F[A]): F[F[A]] = {

kernel/shared/src/main/scala/cats/effect/kernel/MonadCancel.scala

@@ -227,7 +227,7 @@ trait MonadCancel[F[_], E] extends MonadError[F, E] {
    * Indicates the default "root scope" semantics of the `F` in question. For types which do
    * ''not'' implement auto-cancelation, this value may be set to `CancelScope.Uncancelable`,
    * which behaves as if all values `F[A]` are wrapped in an implicit "outer" `uncancelable`
-   * which cannot be polled. Most `IO` -like types will define this to be `Cancelable`.
+   * which cannot be polled. Most `IO`-like types will define this to be `Cancelable`.
    */
   def rootCancelScope: CancelScope
 

kernel/shared/src/main/scala/cats/effect/kernel/Ref.scala

@@ -321,7 +321,7 @@ object Ref {
     F.delay(unsafe(a))
 
   /**
-   * Creates an instance focused on a component of another `Ref` 's value. Delegates every get
+   * Creates an instance focused on a component of another `Ref`'s value. Delegates every get
    * and modification to underlying `Ref`, so both instances are always in sync.
    *
    * Example:

kernel/shared/src/main/scala/cats/effect/kernel/Resource.scala

@@ -443,7 +443,7 @@ sealed abstract class Resource[F[_], +A] extends Serializable {
     Resource.makeCase(F.unit)((_, ec) => f(ec)).flatMap(_ => this)
 
   /**
-   * Given a `Resource`, possibly built by composing multiple `Resource` s monadically, returns
+   * Given a `Resource`, possibly built by composing multiple `Resource`s monadically, returns
    * the acquired resource, as well as a cleanup function that takes an
    * [[Resource.ExitCase exit case]] and runs all the finalizers for releasing it.
    *
@@ -530,7 +530,7 @@ sealed abstract class Resource[F[_], +A] extends Serializable {
   }
 
   /**
-   * Given a `Resource`, possibly built by composing multiple `Resource` s monadically, returns
+   * Given a `Resource`, possibly built by composing multiple `Resource`s monadically, returns
    * the acquired resource, as well as an action that runs all the finalizers for releasing it.
    *
    * If the outer `F` fails or is interrupted, `allocated` guarantees that the finalizers will

testkit/shared/src/main/scala/cats/effect/testkit/TestControl.scala

@@ -37,9 +37,9 @@ import java.util.concurrent.atomic.AtomicReference
  * [[TestControl.executeEmbed]]).
  *
  * In other words, `TestControl` is sort of like a "handle" to the runtime internals within the
- * context of a specific `IO` 's execution. It makes it possible for users to manipulate and
+ * context of a specific `IO`'s execution. It makes it possible for users to manipulate and
  * observe the execution of the `IO` under test from an external vantage point. It is important
- * to understand that the ''outer'' `IO` s (e.g. those returned by the [[tick]] or [[results]]
+ * to understand that the ''outer'' `IO`s (e.g. those returned by the [[tick]] or [[results]]
  * methods) are ''not'' running under the test control environment, and instead they are meant
  * to be run by some outer runtime. Interactions between the outer runtime and the inner runtime
  * (potentially via mechanisms like [[cats.effect.std.Queue]] or
@@ -204,7 +204,7 @@ final class TestControl[A] private (
    * Analogous to, though critically not the same as, running an [[IO]] on a single-threaded
    * production runtime.
    *
-   * This function will terminate for `IO` s which deadlock ''asynchronously'', but any program
+   * This function will terminate for `IO`s which deadlock ''asynchronously'', but any program
    * which runs in a loop without fully suspending will cause this function to run indefinitely.
    * Also note that any `IO` which interacts with some external asynchronous scheduler (such as
    * NIO) will be considered deadlocked for the purposes of this runtime.
@@ -288,8 +288,8 @@ object TestControl {
   /**
    * Executes a given [[IO]] under fully mocked runtime control. Produces a `TestControl` which
    * can be used to manipulate the mocked runtime and retrieve the results. Note that the outer
-   * `IO` (and the `IO` s produced by the `TestControl`) do ''not'' evaluate under mocked
-   * runtime control and must be evaluated by some external harness, usually some test framework
+   * `IO` (and the `IO`s produced by the `TestControl`) do ''not'' evaluate under mocked runtime
+   * control and must be evaluated by some external harness, usually some test framework
    * integration.
    *
    * A simple example (returns an `IO` which must, itself, be run) using MUnit assertion syntax: