[Android] Get package_info by using apk_under_test package.

build/android/pylib/instrumentation/instrumentation_test_instance.py

Index: build/android/pylib/instrumentation/instrumentation_test_instance.py
diff --git a/build/android/pylib/instrumentation/instrumentation_test_instance.py b/build/android/pylib/instrumentation/instrumentation_test_instance.py
index cd8617cfc15f7bb84fa79a60caa1a71261dd2de9..34ce1320a174cb79346708a3dbe75c1edccd1105 100644
--- a/build/android/pylib/instrumentation/instrumentation_test_instance.py
+++ b/build/android/pylib/instrumentation/instrumentation_test_instance.py
@@ -276,8 +276,9 @@ class InstrumentationTestInstance(test_instance.TestInstance):
     self._test_runner = self._test_apk.GetInstrumentationName()
 
     self._package_info = None
+    package_under_test = self._apk_under_test.GetPackageName()
     for package_info in constants.PACKAGE_INFO.itervalues():
-      if self._test_package == package_info.test_package:
+      if package_under_test == package_info.package:

[rnephew (Reviews Here), 2016/04/15 20:35:36]

In this context:
  package_under_test  ==  chrome
  self._test_package == chome_test_whatever_apk
correct?
If so, does switching this from the test apk to the chrome apk make sense?

[mikecase (-- gone --), 2016/04/15 20:55:36]

Yeah, so for example. In the case of like ChromeTest apk.
package_info.package is the Chrome package
package_info.test_package is the ChromeTest package

Currently, we are determining which package_info to use by keying on the
"package_info.test_package" entry. I have no idea why we are currently doing it
this way. But I am removing the "package_info.test_package" entry all-together
and just keying on the "package_info.package" entry instead.

This makes is far more convenient when multiple test_apks have the same
apk_under_test.

         self._package_info = package_info
     if not self._package_info:
       logging.warning('Unable to find package info for %s', self._test_package)