*
 * </ul>
 *
 * <p>When reading a property, the project searches the above scopes in order, and returns the value from the first
 * scope it finds the property in. If not found, an exception is thrown. See {@link #property(String)} for more details.</p>
 *
 * <p>When writing a property, the project searches the above scopes in order, and sets the property in the first scope

   * This is a big deal - we're talking factors of 2-8 in microbenchmarks, not just 10-20%. (This is
   * a hotspot optimizer bug, which should be fixed, but that's a separate, big project).
   *
   * The coding pattern above is heavily used in java.util, e.g. in ArrayList. There is a
   * RangeCheckMicroBenchmark in the JDK that was used to test this.
   *

If a build configuration input has changed, Gradle will not use the entry and will run the configuration phase again as above, saving the result for later reuse.

Build configuration inputs include:

- Init scripts
- Settings scripts
- Build scripts
- System properties used during the configuration phase

// rs or the modules selected in the graph of rs.
//
// The above invariants together imply the graph-pruning invariants for the
// go.mod file:
//
//  1. (The import invariant.) Every module that provides a package transitively
//     imported by any package or test in the main module is included as a root.
//     This follows by induction from (1) and (3) above. Transitively-imported

//
//    c. GC does concurrent sweeping in the background and in response
//    to allocation. See description below.
//
// 5. When sufficient allocation has taken place, replay the sequence
// starting with 1 above. See discussion of GC rate below.

// Concurrent sweep.
//
// The sweep phase proceeds concurrently with normal program execution.
// The heap is swept span-by-span both lazily (when a goroutine needs another span)

	r.SetSym(ctxt.PLT)
	r.SetSiz(8)
	r.SetOff(int32(glink.Size()))
	r.SetAdd(glink.Size())        // Adjust the offset to be relative to label 1 above.
	glink.AddUint64(ctxt.Arch, 0) // The offset to the PLT.

	// Resolve PC-rel offsets above now the final size of the stub is known.
	res0m1b := glink.Size() - 8 // res_0 - 1b
	glink.SetUint32(ctxt.Arch, 16, 0x38000000|uint32(uint16(-res0m1b)))

   * This is a big deal - we're talking factors of 2-8 in microbenchmarks, not just 10-20%. (This is
   * a hotspot optimizer bug, which should be fixed, but that's a separate, big project).
   *
   * The coding pattern above is heavily used in java.util, e.g. in ArrayList. There is a
   * RangeCheckMicroBenchmark in the JDK that was used to test this.
   *

				}
				if spec.Parallelism != nil && *spec.Parallelism > parallelismLimitForHighCompletions {

// %15 = stablehlo.broadcast_in_dim %14  // Operand = %4 if no `convert` above.
// %16 = stablehlo.subtract %13, %15  // q1 - z1
// %17 = stablehlo.convert %12  // i8->i32 cast for q2.
// %18 = stablehlo.convert %6  // [Optional] i8->i32 cast for z2.
// %19 = stablehlo.broadcast_in_dim %18  // Operand = %7 if no `convert` above.
// %20 = stablehlo.subtract %17, %19  // q2 - z2

which writes failures to testdata/fuzz).

The command-line package list, if present, must appear before any
flag not known to the go test command. Continuing the example above,
the package list would have to appear before -myflag, but could appear
on either side of -v.

When 'go test' runs in package list mode, 'go test' caches successful