| Age | Commit message (Collapse) | Author |
|---|
|
Also refactor the existing tests for readabilioty and Junit4.
Bug: 150645191
Test: atest org.apache.harmony.tests.javax.net.ssl.SSLContextSpiTest
Change-Id: I70b8522222108f9497867ada83ec15693f8c15ac
Merged-In: I70b8522222108f9497867ada83ec15693f8c15ac
(cherry picked from commit 46a3e9f6f83bf888b4e5656532d0ced82b22ed22)
|
|
Various test expectations in StandardNames are determined by
CpuFeatures.isAESHardwareAccelerated(). Currently this also
acts as a test for AES not being available when expected because
assorted tests will fail in that scenario.
This change makes CpuFeatures.isAESHardwareAccelerated() return
the actual state of hardware acceleration as determined by
BoringSSL and adds a test for AES being enabled on platforms
where we know we should expect it.
Once this change lands I propose to make the equivalent change
in Conscrypt's copy of CpuFeatures.isAESHardwareAccelerated()
but only run the test for expected acceleration on non-Android
platforms where there is no need to handle emulated environments.
This is intended as a cleaner replacement for
https://r.android.com/1214558 to avoid adding native test code.
Test results:
Taimen:
ABI arm64-v8a: PASSED
ABI armeabi-v7a: PASSED
Cuttlefish:
ABI x86: ASSUMPTION_FAILED
x86 AVD:
ABI x86: ASSUMPTION_FAILED
x86_64 AVD:
ABI x86_64: PASSED
ABI x86: ASSUMPTION_FAILED
ABI in the above results is the ABI reported by atest.
ASSUMPTION_FAILED means that isKnownToSupportHardwareAes()
returned false for this ABI/hardware combination and so
isAesHardwareAccelerated() was not tested.
PASSED means isKnownToSupportHardwareAes() and
isAesHardwareAccelerated() both returned true.
Not tested: Physical x86_64 hardware or 32 bit only ARM,
but the results above demonstrate that our test expectations
are correct.
Bug: 144885685
Test: atest -a CtsLibcoreTestCases:libcore.javax.crypto.HardwareAesTest
(See above for results on different platforms)
Change-Id: Ifec9b8f95eb71d4aa26fd0dc6df295a213b43a0b
|
|
isAESHardwareAccelerated is calling platform api
is64BitInstructionSet, on x86 and x86_64 platforms.
This leads to CtsVerifier apk to crash during
the KeyChain Storage Test.
This CL simply does the same thing that
is64BitInstructionSet is doing, without calling
that api.
The impact is only limited to x86 and x86_64 architecture.
Test:
lunch sdk_phone_x86_64-userdebug
install CtsVerifier.apk
and run the KeyChain Storage Test, and it should not
crash.
BUG: 139022588
Change-Id: I311b961abe73e71b5c004ba66b920b9bb714ffd0
Merged-In: I311b961abe73e71b5c004ba66b920b9bb714ffd0
|
|
This proof-of-concept isn't needed any more as real mappings are about
to be committed.
This reverts commit 5d994b9169d9f3bda15a7b8154c48ba08c4b0e42.
Bug: 130283485
Test: Treehugger
|
|
|
|
Update StandardNames to add new cipher.
Update intra-core API definition due to refactoring in Conscrypt.
This is safe because the classes that libcore will access are supplied
by Conscrypt via the Provider system, so libcore's use of classes will
always be consistent with which classes are in fact present.
Test: cts -m CtsLibcoreTestCases
Test: cts -m CtsLibcoreOkHttpTestCases
Change-Id: I7558cb91ba30b98728e723fea209cea12370e88e
|
|
This change should be reverted once we have real instances of this
pattern and no longer need the example.
This change causes libcore.java.lang.PackageTest to be run in pre- and
post-submit for all libcore/luni changes. It demonstrates three
things:
1. Introducing a new annotation-based category of test, @Example in
this case.
2. Applying the category to a test.
3. Adding a TEST_MAPPING which runs tests from that category.
Note that PackageTest#testGetPackages does not run, because it in
expectations/knownfailures.txt. (This is not to be confused with
PackageTest#testGetPackage, which does run.)
Google-internal: In treehugger run for this change, navigating through
aosp-master -> aosp_cf_x86_phone-userdebug -> test status ->
suite/test-mapping-presubmit-retry_cloud-tf, there is the test-mapping
presubmit run for this change for this change:
http://sponge/3531116c-6f5d-46d3-bdae-089c3c37e104. Under x86
CtsLibcoreTestCases it shows the expected four tests from PackageTest,
excluding testGetPackages.
Test: atest --test-mapping libcore/luni
Test: treehugger (see above)
Bug: 130283485
Change-Id: I8a3e929af2caf7aa22dd720750f407685d97f7a7
|
|
This change is needed to allow the apache-harmony version of
Support_Configuration to be removed.
Bug: 131738612
Test: atest CtsLibcoreTestCases
Change-Id: I18a1df7c529bb2231499bc03c3d9e9d9e4309d85
|
|
Continuing the long process of shrinking the core platform API
by switching users, particularly tests, to alternatives.
Bug: 124232146
Test: atest CtsLibcoreTestCases:libcore.android.system.OsTest
Change-Id: If1a473f4d7a120723602faa70129ead1bf8ebc64
|
|
Preparation for splitting core-tests-support into two parts, one that
depends only on core_current and can be used outside libcore and one
that can depend on core-all directly.
Deletes a lot of classes that are no longer used. Some of the classes
were only used by external/conscrypt tests and it has its own copies of
them. Some of them were copied over from harmony and are not used in
libcore tests. Some of them have been extracted into
external/mockwebserver.
Moves some classes that are only used internally by a couple of luni
tests to am appropriate luni/src/test/java directory to reduce the
number of split packages.
Tightens up visibility in a couple of classes.
Removes some unused members from a couple of classes:
* support/src/test/java/libcore/java/security/StandardNames.java
* support/src/test/java/libcore/javax/net/ssl/SSLConfigurationAsserts.java
Bug: 117547194
Test: make checkbuild && \
atest CtsKeystoreTestCases CtsNetTestCases conscrypt-tests \
FrameworksCoreTests KeyChainTests time_zone_distro-tests \
CtsLibcoreTestCases
Change-Id: I8b258e66911ece69d4eaefef56810e17dddb2583
|
|
These classes have only ever been used by Conscrypt, so they've been
moved to the Conscrypt repo and can be deleted from here.
Test: make checkbuild
Change-Id: I9c18e18f1ede89a0d882300b5970030042328bf5
|
|
Rather than depending on internals, switch core-tests-support
to using the core platform API.
core-tests-support had a static dependency on the platform
bouncycastle target which included a copy of the repackaged
bouncycastle in CTS. This copy would then not be used. This looks
like a mistake. After this change a copy of bouncycastle is included
under its original package.
TestIoUtils was relying on a hidden method in java.lang.Math. This
has been replaced with public API calls.
Bug: 113148576
Test: build
Test: Ran CTS: cts -m CtsLibcoreTestCases
Change-Id: Ib8045537c2c191aca3f416cc459f265c8d957dfc
|
|
Add testing support code that CTS can use instead of
private platform APIs.
Bug: 117535564
Bug: 113148576
Test: Run cts
Change-Id: I63ad197744075c14886a639bfba83392ac39854f
|
|
Despite its name isAESHardwareAccelerated is used in assertions in
StandardNames.java to determine whether to expect hardware AES
acceleration is available and so to determine the default list of
cyphers to expect from Conscrypt. I haven't renamed it in this CL but
it might make sense to do so.
b/112830137 was due to this method expecting hardware AES acceleration
to be available when running on any x86 CPU where /proc/cpuinfo reports
the "aes" flag. However this is *not* the case when running in a 32bit VM,
which was causing CTS failures on 32bit x86 builds.
Note that 32bit x86 is not a supported platform for Conscrypt, so it
doesn't realy make sense to add an equivalent change to the
CpuFeatures.java in Conscript's libcore-stub.
Bug: 112830137
Test: cts -m CtsLibcoreTestCases on both ARM and x86 targets
Change-Id: I4a2643082aa3a638106a77f6916cc3329498e58f
|
|
This mainly updates the tests to account for the existence of TLS 1.3.
In particular, add the new TLS 1.3-related constants to StandardNames
and update some old harmony tests to handle the fact that we now
report TLS 1.3 cipher suites as being supported.
Bug: 110403171
Test: cts -m CtsLibcoreTestCases
Test: cts -m CtsLibcoreOkHttpTestCases
Test: cts -m CtsLibcoreWycheproofConscryptTestCases
Change-Id: I1ff6aa5961438527b0eb882488a5dbfaaeaacc6c
|
|
BoringSSL has removed support for some old cipher suites, remove them
from the expected set.
Test: cts -m CtsLibcoreTestCases
Change-Id: Ia1680f357775dc2ea2f5b04974c11b4e991fbdfa
|
|
We're planning to remove support for SHA-1 signatures in TLS
certificates, so change our tests to use SHA-256 signatures instead.
Bug: 77901166
Test: cts -m CtsLibcoreTestCases
Change-Id: I6c85194a88fb9d0582267ed8e6056b593b9e68de
|
|
This class was introduced for use in libcore.io.DiskLruCacheTest in
commit e342455abc6b650f847d931728223411e58831ce (May 2011), but that
class was dropped in commit 4f97bfe0d9e6f4bdb43fff91f341f4722af16301
(Apr 2013), after switching to OkHttp. It now appears to be unused.
Based on the commit message at http://r.android.com/592273 ,
core-tests-support (which includes MockOs) should only use public
Android API, but MockOs imports libcore-internal classes
libcore.io.{Os,Libcore}.
Therefore, this CL drops the unused class MockOs.
Bug: 69899800
Test: Treehugger
Change-Id: I043178ecd482e1ba8bb4d068c03497191e2722eb
|
|
The default HostnameVerifier now ignores any CommonName
in the certificate provided by the server, even when no
subjectAltName is present.
Bug: 70278814
Test: CtsLibcoreTestCases (ran against an internal build
with this CL cherrypicked)
Change-Id: Ib6fa0c40d8903352e88d8812bf0c09ec1d8ef6be
|
|
into oreo-cts-dev
am: 5956489768
Change-Id: I9b022287e4c7e8c99c7975acef1ad6d87c206282
|
|
Just like Android Key Store, creating Samsung Knox Keystore for
exception handling.
Change-Id: I04ebad92a1065afe8037674eed6d51c913ebbdf8
Signed-off-by: Raj Mamadgi <r.mamadgi@samsung.com>
|
|
For the most part, this is just adjusting tests to account for the fact
that "ChaCha20" now refers to the plain stream cipher instead of
ChaCha20+Poly1305.
Test: cts -m CtsLibcoreTestCases
Change-Id: I5976027811c58910952a186f3580a3f5e561407d
|
|
Adds basic compliance tests for ChaCha20/Poly1305/NoPadding.
Adds test that AES/GCM/NoPadding can be initialized with GCM
AlgorithmParameters.
Adds reuse test for ChaCha20/Poly1305/NoPadding.
Updates StandardNames for new names.
Updates SSLSocketTest for changed Conscrypt implementation details.
Test: cts -m CtsLibcoreTestCases
Change-Id: I608e4bbcced678fdfac8b28d500f7fa8b4599319
|
|
Conscrypt changed to limit visibility to EVP_has_aes_hardware, which means
that in some circumstances CPUs that have AES hardware will show to tests
as not having it, resulting in the wrong expectations in tests.
Bug: 64542254
Bug: 64542486
Test: cts -m CtsLibcoreTestCases (on various devices)
Change-Id: I90cd5a44f7aa21049528ce44c1ee4d96489c9e61
|
|
Since DHE does not allow negotiation of the group used, it is pretty
broken. ECDHE at least allows the negotiation of the group which allows
its security to be maintained with configuration changes in the client
or server.
This tracks a change in Conscrypt merged in
906cfad7e08fd339be06441ff42960743f95053c
Test: cts-tradefed run cts -m CtsLibcoreOkHttpTestCases -a arm64-v8a
Test: cts-tradefed run cts -m CtsLibcoreTestCases -a arm64-v8a
Test: make docs
Test: visual inspection of docs output in web browser
Change-Id: Ic90297bf6b1c82af192a887797238ad250e3d1ce
|
|
This applies the same regexp replacement across libcore that
http://r.android.com/345826 had only applied to ojluni.
Most of the previous noncanonical spellings were a lowercased
"android-" that should be "Android-".
This CL was created by running the following command on top of
the above CL. No manual changes were made.
find . -name \*\.java | xargs sed -i \
-e 's/Android[- ]changed/Android-changed/ig' \
-e 's/Android-changed :/Android-changed:/g' \
-e 's/Android-changed \(BEGIN\|END\)/\1 Android-changed/g' \
-e 's/Android-changed - /Android-changed: /g' \
-e 's/Android[- ]removed/Android-removed/ig' \
-e 's/Android-removed :/Android-removed:/g' \
-e 's/Android-removed \(BEGIN\|END\)/\1 Android-removed/g' \
-e 's/Android-removed - /Android-removed: /g' \
-e 's/Android[- ]added/Android-added/ig' \
-e 's/Android-added :/Android-added:/g' \
-e 's/Android-added \(BEGIN\|END\)/\1 Android-added/g' \
-e 's/Android-added - /Android-added: /g' \
-e 's/----- \(BEGIN\|END\) android\( -----\)\?/\1 Android-changed/g' \
-e 's/\/\* \(BEGIN\|END\) Android-changed \*\//\/\/ \1 Android-changed/g'
Bug: 35841464
Test: make droid cts
Change-Id: I060c7236b7607763e5d27d60aa395d2507703a95
|
|
Removes suppression of legacy harmony tests that fail on
Android.
Changes:
1) Switches libcore InetAddressTest to parameterized Junit4 test
framework to make additions of test data easier / clearer.
2) Removes failing harmony test cases. This seemed like the
correct approach as the harmony tests expected completely
opposite behavior to the libcore tests but while the libcore
tests pass the harmony tests have been broken for over 3 years.
3) Adds test cases to libcore InetAddressTest to assert the
behavior of addresses removed from the harmony test cases.
4) Removes suppression of harmony test cases.
5) Handles knock-ons in SerializationTester; it assumed that all
tests extend TestCase, no longer the case with JUnit 4.
Bug: 11689863
Test: with cts and vogar
Change-Id: Iea6ba11fda18f058068e4eea1490d98ee7a3eb2d
|
|
Bug: 29631070
Test: cts run -m CtsLibcoreTestCases -t libcore.java.security.ProviderTest
Test: cts run -m CtsLibcoreTestCases -t libcore.javax.crypto.CipherTest
Change-Id: I9bd6ed1a17048748cc5006326720dbd6c2012d67
|
|
|
|
|
|
|
|
The previous API has been deprecated within Bouncycastle for a while.
Switch to the newer one to avoid the deprecation warnings.
Test: cts-tradefed run cts -m CtsLibcoreTestCases
Change-Id: I24b1340185876f90730d362019f202431c94d4a2
|
|
When compiled against the SDK, junit.framework is noted as deprecated.
Changing this to use JUnit4-style asserts doesn't break any users.
Test: cts-tradefed run cts -m CtsLibcoreTestCases
Change-Id: Ic480b16db8e0ef13b55e7f3c005c7d9c26da0114
|
|
Bug: 29631070
Test: run cts -m CtsLibcoreTestCases
Change-Id: I7174ec7f6c598d46dce1935385c723b96907d9d1
|
|
Bug: 29631070
Test: run cts -m CtsLibcoreTestCases
Change-Id: Iff5246f9053a8a1691038f5719145a541ebec5e2
|
|
SSL was rexposed in 3f3b9c8292773841760263f410b74ed0ab7aae6b as an alias
to TLS.
Bug: 34722987
Test: cts-tradefed run cts -m CtsLibcoreTestCases -t
libcore.java.security.ProviderTest
Change-Id: Icf784ca30ccec2c5546dc9af6e7a06ae8ffefd56
|
|
SSL was removed with SSLv3 however a lot of callers use
SSL instead of Default leading to breakage.
Test: libcore/run-libcore-tests
libcore/luni/src/test/java/libcore/javax/net/ssl/*
Bug:32584776
Bug:30977793
Change-Id: I7ce48488a28e2c62bb0c2a76261a11fbd5bb9da7
|
|
https://boringssl-review.googlesource.com/#/c/10923/
Test: CtsLibcoreTestCases
Change-Id: Ie13bcce8b1557e55921366643efbf71b7b75314b
|
|
StuckServer uses kernel implementation details that have changed on
version > 4.4, and causing tests depending on it to fail. Since we can't
find an alternative method to simulate a stuck server without special
permission, and the test itself is not reliable in that they don't fail
on non-stuck servers, remove them from Libcore tests.
Test: CtsLibcoreTestCases
Bug: 32551747
Change-Id: I88d837329d65040a9f8550fff790f8703ea773f3
|
|
Attributes.java
- header change
- Use of generics
- Updated documentation links
- Deprecated:
EXTENSION_INSTALLATION, IMPLEMENTATION_VENDOR_ID, IMPLEMENTATION_URL
JarEntry.java
- header change
- Updated javadoc
JarFile.java:
- Entry enumerator moved to seperate static inner class
- Streams implementation
- Non-crucial META-INF/* (not *.SF|*.DSA|*.RSA|*.EC|MANIFEST.MF) files are not
passed to JarVerifier in .intializeVerifier() call. Otherwise we could signal
the end of META-INF parsing (due to changes in JarVerifier).
- Refactoring of hasClassPathAttribute, divided into
many methods
- Android-changed: Commented out isKnownNotToHaveSpecialAttributes()
method that doesn't make sense on android
JarOutputStream.java
- Header change
- Replaced a use of "& 0xff" with a cleaner conversion to unsigned int
JarVerifier.java
- Use of generics
- beginEntry explicitly omits META-INF manifest and index signature
check (So they can be passed to beginEntry without side-effects)
- beginEntry process properly non-crucial META-INF/ entries (not
*.SF|*.DSA|*.RSA|*.EC|MANIFEST.MF) and checks their signatures
(added a test for that)
- Change from "Enhance signed jar verification": .entryNames
treats .findMatchingSigners(c) null result as empty array
(probably a bug fix).
Manifest.java
- Use of generics
- Replaced a use of "& 0xff" with a cleaner conversion to unsigned int
Pack200.java
- Copyright update
- javadoc formating changes
- addPropertyChangeListener & removePropertyChangeListener gained empty default
implementation and deprecated
SignatureFileVerifier.java:
- uncommented isSigningRelated (requried by JarFile)
Test: CtsLibcoreTestCases
Change-Id: I56c59fac628c97a8b7c6967a43c0e23c8b068120
|
|
AndroidKeyStore can be used with TLS client auth, but we don't test
anything similar with SSLSocketTest. Add a PrivateKey wrapper that
allows us to closely simulate the conditions which trigger the special
code in Conscrypt to do upcalls to Java to generate signatures with the
client private key.
Test: cts-tradefed run cts -m CtsLibcoreTestCases -a arm64-v8a
Test: cts-tradefed run cts -m CtsLibcoreOkHttpTestCases -a arm64-v8a
Bug: 31714503
Change-Id: I559db546ddd31f8efbe73fc70a91689ed6d7d7e5
|
|
|
|
When running these tests against an Android system when we're not built
as part of the system image (e.g., we have our own BouncyCastleProvider
instance), then we need to make sure our instance of
BouncyCastleProvider is initialized properly. The initialization happens
in its constructor.
In TestKeyStore the included version of BouncyCastleProvider does not
need to be inserted as a security provider since it's only used to
create X.509 certificates. However, BouncyCastle calls into itself for
some things like OID -> key type conversion. This relies on all of its
internal data structures being initialized properly. In particular there
is a keyInfoConverters field that must be populated to work correctly.
If keyInfoConverters is not populated, the generated certificate will
have a null PublicKey since it can't figure out what type of key it is.
Test: ran outside of CTS
Test: cts-tradefed run cts -m CtsLibcoreTestCases -a arm64-v8a
Bug: 31119038
Change-Id: I1bcc73a363cb9dac9e9decb9046d386d3394f1d9
|
|
This change addresses some post-commit review comments for
commit 44c55c5b639d6c1eea6eeae8414ef9df95b67bb7:
* Moved SocketTimeoutException into a separate test.
* Added more infomative comments regarding setDelay.
* Reduced method visibility, being public method is unnecessary.
Bug: 32363029
Test: URLConnectionTest#test_setReadTimeoutI_SocketTimeoutException
Change-Id: Ia13ba801cfa6e06fc7597f4f665fa2d002b688e1
|
|
This was failing because the test assumes that read will not complete
within 1ms, while it is still possible, especially on host tests.
This patch introduces a minimum time delay before the TestWebServer
sends data.
Bug: 32363029
Test: org.apache.harmony.luni.tests.java.net.URLConnectionTest#test_setReadTimeoutI under host tests
Change-Id: Ibb4e2125f99b0544d0f09331b4ab9b0448b14035
|
|
Changes to KeyUsageExtension, NetscapeCerTypeExtension and
ReasonFlags as to check as to improve checks of array bounds.
sun.security.provider.certpath.DistributionPointFetcher has a
similar change concerning array bounds, where the opportunity
to improve probably was spotted by the fact that it uses
ReasonFlags.
Since in the classes in sun.security.x509 the logic for toString
changed slightly, tests are added that passed both before and
after the change, as to check that the outcome is the same.
Bug: 29631070
Test: run cts -m CtsLibcoreTestCases
Change-Id: I2a2c10c59509063e648f238f82ac71b4f513cd71
|
|
|
|
|
|
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=136450318
Test: mmma -j32 libcore
Change-Id: I62edce18ddd1ebf46d2488d88376636f566e3f6f
|
|
This is used by the RI so add it to the list of known values.
Test: mmma -j32 libcore
Test: cts-tradefed run cts -m CtsLibcoreTestCases -a arm64-v8a
Change-Id: Id4792b5f42e309570d5c27accd0e51c432df5289
|
