# ADR-0003 - Avoid introducing Groovy types to public API

## Status

- ACCEPTED on 2024-01-12

## Context

Gradle's public API requires equal access from all JVM-based languages.
Kotlin, Groovy, Java, and other JVM-based languages should be able to use the Gradle API without relying on another language's standard library.

Historically, Gradle has shipped with some Groovy types in very prominent APIs.

and it must not allow a single write to write multiple values.
</p>

<p>
All the following examples assume that `*p` and `*q` refer to
memory locations accessible to multiple goroutines.
</p>

<p>
Not introducing data races into race-free programs means not moving
writes out of conditional statements in which they appear.
For example, a compiler must not invert the conditional in this program:
</p>

<pre>
*p = 1

//   - Exclude test files (*_test.go)
//   - go-gorm/gen's Generated files (*.gen.go)
func CallerFrame() runtime.Frame {
	// Preserve the original CallerFrame stack depth after introducing callerFrame().
	return callerFrame(4)
}

func callerFrame(skip int) runtime.Frame {
	pcs := [13]uintptr{}
	// skip is caller-path sensitive and should be selected by each public helper.

      }
    }
  }

  /**
   * ThreadLocal holding a single-slot int array holding hash code. Unlike the JDK8 version of this
   * class, we use a suboptimal int[] representation to avoid introducing a new type that can impede
   * class-unloading when ThreadLocals are not removed.
   */
  static final ThreadLocal<int @Nullable []> threadHashCode = new ThreadLocal<>();

  /** Generator of new random hash codes */

Symmetrically, the same problem exists on the side of Gradle maintainers.
It should be clear when introducing a property what contract users can expect and what we can afford to provide.
As features progress through the lifecycle of stabilization and, possibly later, deprecation, the corresponding properties can also adjust accordingly.

introduce a Go variable that points at the C variable (the linker can
be told to initialize this pointer).  For example, if the gmp library
provided

	mpz_t zero;

then cgo would rewrite a reference to C.zero by introducing

	var _C_zero *C.mpz_t

and then replacing all instances of C.zero with (*_C_zero).

Cgo's most interesting translation is for functions.  If xxx is a C

      }
    }
  }

  /**
   * ThreadLocal holding a single-slot int array holding hash code. Unlike the JDK8 version of this
   * class, we use a suboptimal int[] representation to avoid introducing a new type that can impede
   * class-unloading when ThreadLocals are not removed.
   */
  static final ThreadLocal<int @Nullable []> threadHashCode = new ThreadLocal<>();

  /** Generator of new random hash codes */

      // Fine: If it doesn't exist, then there's no chance that we're going to be asked to test it.
    }

    /*
     * These methods "should" call checkNotNull. However, I'm wary of accidentally introducing
     * anything that might slow down execution on such a hot path. Given that the methods are only
     * package-private, I feel OK with just not testing them for NPE.
     *

<div style="text-align: right; margin-right: 10%;">Kevin Glisson, Marc Vilanova, Forest Monsen – <strong>Netflix</strong> <a href="https://netflixtechblog.com/introducing-dispatch-da4b8a2a8072"><small>(Ref.)</small></a></div>

---

„_Ich bin hellauf begeistert von **FastAPI**. Es macht so viel Spaß!_“

<div style="text-align: right; margin-right: 10%;">Kevin Glisson, Marc Vilanova, Forest Monsen - <strong>Netflix</strong> <a href="https://netflixtechblog.com/introducing-dispatch-da4b8a2a8072"><small>(ref)</small></a></div>

---

"_Я в полном восторге от **FastAPI**. Это так весело!_"