testrunner: Fix test name parsing

A recent change removed compact-dex-level plumbing from the test
scripts, where the cdex option group was dropped from test config.
However, parsing of test strings (provided via the --test argument) was
only partially updated, so while testrunner correctly regex matches the
updated string format, it later fails due to expecting the removed
extra option group to still exist.

This CL fixes the issue, and amends the inline comments containing
out-of-date test name examples.

Test: testrunner.py -t test-art-target-run-test-debug-prebuild-jit-\
                       no-relocate-ntrace-cms-checkjni-picimage-\
                       ndebuggable-no-jvmti-001-HelloWorld64

Change-Id: I5d1c8c9b250dd279536e1bba76ef4fab7d573377

1 files changed, 3 insertions, 3 deletions

diff --git a/test/testrunner/testrunner.py b/test/testrunner/testrunner.py
index c6ebe36dab..c92f18369b 100755
--- a/test/testrunner/testrunner.py
+++ b/test/testrunner/testrunner.py
@@ -35,7 +35,7 @@ dependencies:
 There are various options to invoke the script which are:
 -t: Either the test name as in art/test or the test name including the variant
     information. Eg, "-t 001-HelloWorld",
-    "-t test-art-host-run-test-debug-prebuild-optimizing-relocate-ntrace-cms-checkjni-picimage-ndebuggable-001-HelloWorld32"
+    "-t test-art-host-run-test-debug-prebuild-optimizing-relocate-ntrace-cms-checkjni-picimage-ndebuggable-no-jvmti-001-HelloWorld32"
 -j: Number of thread workers to be used. Eg - "-j64"
 --dry-run: Instead of running the test name, just print its name.
 --verbose
@@ -996,7 +996,7 @@ def extract_test_name(test_name):
     test_name_matcher = re.compile(regex)
   match = test_name_matcher.match(test_name)
   if match:
-    return list(match.group(i) for i in range(1,15))
+    return list(match.groups())
   raise ValueError(test_name + " is not a valid test")
 
 def parse_test_name(test_name):
@@ -1004,7 +1004,7 @@ def parse_test_name(test_name):
   It supports two types of test_name:
   1) Like 001-HelloWorld. In this case, it will just verify if the test actually
   exists and if it does, it returns the testname.
-  2) Like test-art-host-run-test-debug-prebuild-interpreter-no-relocate-ntrace-cms-checkjni-pointer-ids-picimage-ndebuggable-001-HelloWorld32
+  2) Like test-art-host-run-test-debug-prebuild-interpreter-no-relocate-ntrace-cms-checkjni-picimage-ndebuggable-no-jvmti-001-HelloWorld32
   In this case, it will parse all the variants and check if they are placed
   correctly. If yes, it will set the various VARIANT_TYPES to use the
   variants required to run the test. Again, it returns the test_name