7ff645482d5d10f2c15955ed812de183dead5557 |
|
01-Dec-2016 |
Tobias Thierer <tobiast@google.com> |
Switch Base64 decoder implementation used in harmony SSL tests These tests were using libcore.io.Base64.decode() for test data. Since that class is deprecated and the tests weren't depending on the subtle behavioral differences relative java.util.Base64.getDecoder().decode(), this CL converts them to the latter implementation. Bug: 31292683 Test: cts-tradefed run cts -m CtsLibcoreTestCases -a arm64-v8a Change-Id: I635625ceae9d6a33a4909101e65363d7052263d9
|
d4b7449574ca69c713a3f2230cb77c68ee642700 |
|
06-Sep-2016 |
Tobias Thierer <tobiast@google.com> |
Fix 58 libcore tests that caught "expected" Exceptions but didn't fail() There were no changes to code under test. 4 tests were fixed to not tolerate an Exception: - libcore.java.util.CollectionsTest#test_CheckedMap_computeIfAbsent In the case where a Map key is present, computeIfAbsent() should not invoke the Function, so there is no return value that could be expected to undergo a type check. The test was changed to verify that the Function is not invoked. - libcore.java.util.concurrent.CopyOnWriteArrayListTest#test_sort and libcore.java.util.concurrent.CopyOnWriteArrayListTest#test_subList_sort were sorting an already-sorted non-null list that contains no null values. That logic was redundant and wouldn't have thrown. - org.apache.harmony.security.tests.java.security.KeyStoreSpiTest#test_engineEntryInstanceOf (the KeyStoreSpi documentation does not mention whether NPE is thrown; the implementation comes from upstream and does not throw it). 54 tests were fixed by adding the missing fail(); No attempt was made to make this failure message specific to the exception that was expected. Tests that weren't changed: - SSLEngineTest.test_wrap_ByteBuffer_ByteBuffer_04() requires investigation to determine the correct resolution. A comment was added to the test linking to the corresponding bug. - ThreadTest#testParkUntilWithUnderflowValue() and OldNodeChangeEventTest.waitForEvent() were not touched since it was not clear whether that it was relevent to the test whether or not the exception was thrown. - two cases in Files2Test#test_setPosixFilePermissions_NPE() are left to a future CL Test: (croot; make vogar core-tests) \ && git diff aosp/master --name-only | xargs vogar `cparg core-tests` Change-Id: I1cd3220a06e331ff049b146b393f78879f07211a
|
c46a2ea848e7a62cd5ee24216e446ad7b9ba7629 |
|
19-Feb-2015 |
Piotr Jastrzebski <haaawk@google.com> |
Add missing package names for legacy harmony tests. Change-Id: I7bf808c7df0c432eb64b101506a7a419521c2e14
|
cff384fb32d139d2bda7c3dc872108fe9d86d6b1 |
|
09-Jul-2015 |
Narayan Kamath <narayan@google.com> |
Get rid of as much of dalvik.annotation.* as we can. Some of it is public API (why oh why..) so we're forced to keep it around. Change-Id: Ic1e059d7162c442a1a996eb1510d0f7f84add34e
|
a951e74897193848ee88ed1c5b0a1d536f403d70 |
|
04-Dec-2014 |
Narayan Kamath <narayan@google.com> |
Fix failing socket tests. Use bogus numeric addresses instead of invalid hostnames to make InetAddress.getAllByName fail before we ask a DNS server. This guards against flakiness caused by evil DNS servers. Note that from an API perspective, Socket* and SSLSocket* don't care about *why* getAllByName failed, just that it did. This change also fixes a test that was asserting that the exception we throw contains a useful error message. Again, this test was changed to use a broken numeric address. bug: 18575971 (cherry picked from commit 83b5554108e839b0c78178a029b65fddce8983f0) Change-Id: Ia1c521c05634ce798a61b060f4d3fbbc164bb9fc
|
83b5554108e839b0c78178a029b65fddce8983f0 |
|
04-Dec-2014 |
Narayan Kamath <narayan@google.com> |
Fix failing socket tests. Use bogus numeric addresses instead of invalid hostnames to make InetAddress.getAllByName fail before we ask a DNS server. This guards against flakiness caused by evil DNS servers. Note that from an API perspective, Socket* and SSLSocket* don't care about *why* getAllByName failed, just that it did. This change also fixes a test that was asserting that the exception we throw contains a useful error message. Again, this test was changed to use a broken numeric address. bug: 18575971 Change-Id: I0b49e5e66717dc29a3401f45af0ef2d5bb955219
|
e82be46881b2ab9d64b35c0ee53fc359b5cfd18a |
|
04-Dec-2014 |
Narayan Kamath <narayan@google.com> |
Rename packages to match directory structure. Doesn't make a difference to the CTS test runner (which runs these tests anyway) but vogar complains. bug: 18575971 Change-Id: Ifd323300f9c70a8625c215d8ff7d48ddd18083bd
|
ab762bb740405d0fefcccf4a0899a234f995be13 |
|
15-Nov-2013 |
Narayan Kamath <narayan@google.com> |
Move tests from tests/api to harmony-tests. (Take 2) This reverts commit d5e281743e55e1daa297fd8043b2d9b6e8e1bed2. The makefile issue that caused the original build break has been resolved.
|
d5e281743e55e1daa297fd8043b2d9b6e8e1bed2 |
|
15-Nov-2013 |
Narayan Kamath <narayan@google.com> |
Revert "Move tests from tests/api to harmony-tests." This reverts commit 81bf28ad31131815d0a36a43a0eca3c29aefdfcd. Breaks build Change-Id: I3f1562921ffe8fdbec36971dd65db398c27c92db
|
81bf28ad31131815d0a36a43a0eca3c29aefdfcd |
|
15-Nov-2013 |
Narayan Kamath <narayan@google.com> |
Move tests from tests/api to harmony-tests. The vast majority of cleaned up harmony tests are now in harmony-tests/src/test/org/apache/harmony/tests/ Change-Id: I56c2e484ff434b5618cf6751d602ae9f0db96b30
|