Sort test result by status

Currently Bazel doesn't group the test result by their status.

For example, the failed tests are scattered among the passed tests. It's hard to figure out what tests are failing.
```
//tools/android:instrumentation_test_check_test PASSED in 0.8s
//tools/android:junction_test PASSED in 5.4s
//tools/android:resource_extractor_test FAILED in 0.7s
//tools/android:stubify_manifest_test PASSED in 3.1s
//tools/aquery_differ:aquery_differ_test PASSED in 1.9s
//tools/bash/runfiles:runfiles_test PASSED in 2.0s
//tools/build_defs/pkg:path_test FAILED in 0.9s
//tools/java/runfiles/testing:RunfilesTest PASSED in 1.4s
//tools/java/runfiles/testing:UtilTest FAILED in 2.1s
//tools/jdk:proguard_whitelister_test PASSED in 0.7s
```

Instead we should sort the test results by status:
```
//tools/android:instrumentation_test_check_test PASSED in 0.8s
//tools/android:junction_test PASSED in 5.4s
//tools/android:stubify_manifest_test PASSED in 3.1s
//tools/aquery_differ:aquery_differ_test PASSED in 1.9s
//tools/bash/runfiles:runfiles_test PASSED in 2.0s
//tools/java/runfiles/testing:RunfilesTest PASSED in 1.4s
//tools/jdk:proguard_whitelister_test PASSED in 0.7s
//tools/android:resource_extractor_test FAILED in 0.7s
//tools/java/runfiles/testing:UtilTest FAILED in 2.1s
//tools/build_defs/pkg:path_test FAILED in 0.9s
```

Start by locating the Bazel code that produces the test-result summary shown in the issue, then trace how individual statuses are collected and ordered. Use the example output to verify that results are grouped by status while retaining each result's existing details and that the displayed order matches the requested behavior.

Written by the indexing model from the issue text.