- 15 Aug, 2011 3 commits
-
-
Jesse Wilson authored
-
Jesse Wilson authored
Tests were failing in CTS because they were relying on the static 'st' member to be side effected by other tests. This also makes some quick cleanups to style and names. Change-Id: I93b9ce370c247c07f3521a8db3904ede86394ffe
-
Jesse Wilson authored
The CTS test harness doesn't sanitize the preferences between tests so we need to make sure it's clean in the test. Change-Id: Idc80ac23f97431d31c902ad6a985ee8010a008f8
-
- 14 Aug, 2011 9 commits
-
-
Jesse Wilson authored
-
Jesse Wilson authored
-
Jesse Wilson authored
This weird, surprising behavior used the entire call stack as context to lookup resources. This is the wrong solution to the problem; context class loader is the right solution. And the code is already using context class loader. This change prevents some tests from failing with an NPE when iterating the null class context. Change-Id: Ic2419dc22f66a4fa8cceaee4746ef1a6d7d4680c
-
Jesse Wilson authored
This test invisibly uses two threads - the test thread and the preference API's preference change thread. This test was relying on flush() taking long enough that the other thread would have posted its event before it returned. The new test just waits up to 100ms. This is still racy but it eliminated the false failures in my runs. Change-Id: I87799b36cb4e5e26d8ca04e4529d1db0b936acb3
-
Jesse Wilson authored
-
Jesse Wilson authored
Change-Id: I24d60b6db32409ed84a81f4a067e3c60e2817493
-
Jesse Wilson authored
-
Jesse Wilson authored
Change-Id: Ie0b62f598bb97cec76394e99654eaeaf7956ae6a
-
Elliott Hughes authored
I've raised a bug to enhance CheckJNI. For now, this was caught by a setSoTimeout test. Bug: 5159133 Change-Id: If28e81b97c2fdd870ebbcd36d9d0bc01be9d8cf1
-
- 12 Aug, 2011 5 commits
-
-
Jesse Wilson authored
We had a test URLStreamHandlerFactoryTest.testInstallCustomProtocolHandler that was failing when run in an APK without this fix. The new behavior is definitely what Android applications will want! Change-Id: I2266a8aea6da833022b538c1f424a3099684d992
-
Jesse Wilson authored
-
Jesse Wilson authored
-
Jesse Wilson authored
Change-Id: Icc35a44bac6481b549a96d0533770d11aff507da
-
Jesse Wilson authored
This was breaking several CTS tests. Change-Id: I76a87deeb9a1f7b056adb72447001cf5e23ec0b3
-
- 11 Aug, 2011 3 commits
-
-
Jeff Brown authored
am a3038e56: (-s ours) am 8f591ff0: Merge "Support wrapping app processes to inject debug instrumentation. Bug: 4437846" * commit 'a3038e56': Support wrapping app processes to inject debug instrumentation. Bug: 4437846
-
Jeff Brown authored
* commit '8f591ff0': Support wrapping app processes to inject debug instrumentation. Bug: 4437846
-
Elliott Hughes authored
Exception throwing is pretty slow, and BlockGuard's isLingerSocket causes us to throw whenever we close a non-socket fd. Avoid that by adding a cheap test for socket fds and only calling isLingerSocket on socket fds. Bug: 4972558 Change-Id: Ib18d6c0c091cb366f953d7467e83fad570bb3457
-
- 10 Aug, 2011 2 commits
-
-
Jeff Brown authored
-
Elliott Hughes authored
The Format, Private Use, and Surrogate categories were all claimed to be Cf. The error was only in the summary table. The javadoc for the individual categories was correct. Bug: http://code.google.com/p/android/issues/detail?id=18942 Change-Id: Ifde47a85fbc84dbec05408f6581331ab4a8d275a
-
- 05 Aug, 2011 1 commit
-
-
Philip Milne authored
The public (though hidden) method contentEquals(char[], int, int) was added to java.lang.String to avoid creating large numbers of Strings during XML parsing. That objective is preserved with this change, which uses a private static method with a similar implementation that goes through the regular String API. Change-Id: I808e1ebc4ef6812dd651682628167b4c078af334
-
- 04 Aug, 2011 1 commit
-
-
Sebastien Hertz authored
Bug: http://code.google.com/p/android/issues/detail?id=17405 Change-Id: Id9e9cef8c86f3627ab943a24996f9fa09471d72a
-
- 29 Jul, 2011 5 commits
-
-
Jesse Wilson authored
This was lost in a recent refactoring. Change-Id: I6f625060192350636de7998a980a5c0f6cdc16c3
-
Jesse Wilson authored
-
Jesse Wilson authored
-
Jesse Wilson authored
-
Jesse Wilson authored
-
- 27 Jul, 2011 1 commit
-
-
Jeff Brown authored
Bug: 4437846 Change-Id: Ib6cee4a89113afa93710e85f515254d8cc07690d
-
- 24 Jul, 2011 4 commits
-
-
Jesse Wilson authored
We were failing two tests in URLTest because we were accepting IPv4 addresses and host names. Note that in the RI, URL is extremely lenient on validating the characters in the parts of a URL. It permits any character in any part except for the delimiter characters like '/' or ':'. The lax IPv6 check 'does it contain a colon?' is consistent with the RI. Change-Id: I957f2ee610b72cf68a7773b4c02a4b6d5fcbb9de
-
Jesse Wilson authored
Bug: http://b/4181738 Change-Id: I6311789044e8b6b573b00aa6ca6fca31f004dc2f
-
Jesse Wilson authored
My intention in this test was that derived headers were passed to the cookie handler. I recently refactored the HTTP client and some headers are now set very late; I need to adjust this test's expectations to match. Change-Id: Id605a213defaa25c0355ee581997233383658420
-
Jesse Wilson authored
When we fully removed the security manager security theatre we broke 'equals' on some AccessControlContexts that were used in map keys. Now we don't include the AccessControlContexts in the map keys. This fixes this test: tests.api.javax.net.ssl.SSLSessionBindingListenerTest#test_valueUnbound Change-Id: I685416c65056c9c540bf75c4aab5e884b66a4394
-
- 20 Jul, 2011 3 commits
-
-
Jesse Wilson authored
-
Jesse Wilson authored
Some other side effects of this change: - we no longer make a time lookup when we don't need to - we time out after 'n' milliseconds, not 'n+1' milliseconds - we duplicate code from Thread.join(). I opened http://b/5053948 to dedup this. Bug: http://b/5050245 Change-Id: I20375119cfa06be559b6aa5c209375b3d3a541f9
-
Jesse Wilson authored
Bug: http://code.google.com/p/android/issues/detail?id=18508 Change-Id: Ib992a0a4b22e340446abab14e4f32df5efcd0b49
-
- 19 Jul, 2011 1 commit
-
-
Elliott Hughes authored
We were relying on the fd being invalid to prevent us from performing activity on a closed stream. Unfortunately, this didn't work with our clever scheme to avoid having to treat the two cases separately. Go back to a simple scheme with one FileDescriptor and a boolean that tells us whether or not we should close it. If we don't close it, we switch it out for an invalid FileDescriptor so we won't continue to act on the unowned FileDescriptor. Bug: 4361076 Change-Id: I1851926fd2847ea63d91a8e0680a7430ecb16453
-
- 17 Jul, 2011 2 commits
-
-
Brian Carlstrom authored
-
Brian Carlstrom authored
Bug: http://code.google.com/p/android/issues/detail?id=18566 Bug: 5038554 Change-Id: I34004448acc4fc2ea6f5adaf6a6409554c226c47
-