If your PR includes any changes to the Gradle Public API, it will cause the binary compatibility check to fail.
The binary compatibility check runs as a part of the broader sanity check.
The latter runs on every PR and is a prerequisite for merging.

If you run the sanity check locally with the `./gradlew sanityCheck`, you can see the binary compatibility error in the output.
It looks like the following:

```

    final N node;

    /**
     * The successors left to be visited, or {@code null} if we just added this {@code
     * NodeAndRemainingSuccessors} instance to the stack. In the latter case, we'll compute the
     * successors if we determine that we need them after we've performed the initial processing of
     * the node.
     */
    @Nullable Queue<N> remainingSuccessors;

     * the current count of the element in the multiset, depending on the implementation. Note that
     * in the former case, this method can never return zero, while in the latter, it will return
     * zero if all occurrences of the element were since removed from the multiset.
     *
     * @return the count of the element; never negative
     */
    int getCount();

    /**

			85, 120, 0, 0, // tag 30805 size 0
		},
	}
	testValidHeader(&h, t)
}

// Just benchmarking how fast the Zip64 test above is. Not related to
// our zip performance, since the test above disabled CRC32 and flate.
func BenchmarkZip64Test(b *testing.B) {
	for i := 0; i < b.N; i++ {
		testZip64(b, 1<<26)
	}
}

func BenchmarkZip64TestSizes(b *testing.B) {
	for _, size := range []int64{1 << 12, 1 << 20, 1 << 26} {

     * the current count of the element in the multiset, depending on the implementation. Note that
     * in the former case, this method can never return zero, while in the latter, it will return
     * zero if all occurrences of the element were since removed from the multiset.
     *
     * @return the count of the element; never negative
     */
    int getCount();

    /**

modification follow.  Pay close attention to the difference between a
"work based on the library" and a "work that uses the library".  The
former contains code derived from the library, whereas the latter must
be combined with the library in order to run.

                  GNU LESSER GENERAL PUBLIC LICENSE
   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION

	return partID > globalMaxPartID
}

// profilerWrapper is created because pkg/profiler doesn't
// provide any API to calculate the profiler file path in the
// disk since the name of this latter is randomly generated.
type profilerWrapper struct {
	// Profile recorded at start of benchmark.
	records map[string][]byte
	stopFn  func() ([]byte, error)
	ext     string
}

	ADD $-8589869055, R5            // 3fe0fffe63ff00017bff83e463ff00017cbf2a14 or 0602000138a50001

	//TODO: this compiles to add r5,r6,r0. It should be addi r5,r6,0.
	//      this is OK since r0 == $0, but the latter is preferred.
	ADD $0, R6, R5             	// 7ca60214

        //TODO: the assembler rewrites these into ADDIS $19, R5, Rx and ADD $-10617, Rx, Rx, but the test only sees the first ADDIS

		}

		// NOTE: if not using LDAP, then internal IDP or open ID is
		// being used - in the former, group info is enforced when
		// generated credentials are used to make requests, and in the
		// latter, a group notion is not supported.
	}

	newCred, updatedAt, err := globalIAMSys.NewServiceAccount(ctx, targetUser, targetGroups, opts)
	if err != nil {
		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)

//                 (i.e. `noutputs`) or be null. In the former case,
//                 the names should match the regular expression for ArgDef
//                 names - "[a-z][a-z0-9_]*". In the latter case,
//                 names for outputs will be generated automatically.
//  opts - various options for the function, e.g. XLA's inlining control.
//  description - optional human-readable description of this function.