| f71c7ea18fdeae6d70009209653ac8626b419292 |
|
13-Apr-2016 |
Sergio Giro <sgiro@google.com> |
JarFile: make getManifestFromReference and getManEntry synchronized
So that all modifications to the cached manifest entry and to the
underlying JarVerifier are synchronized.
computedHasClassPathAttribute and hasClassPathAttribute are just cached
booleans that don't change once set, unsynchronized access to them would
cause only for them to be computed twice in the worst case.
Also, removed dead code and corresponding test.
Bug: 27826114
Change-Id: Id3a9b74997cb462c625861c0284196f74dcb8e9a
|
| 458bfa01aa4dfbe5d4895afc887616a39056f7da |
|
03-Feb-2016 |
Yi Kong <yikong@google.com> |
Fix JarFileTest#test_getInputStreamLjava_util_jar_JarEntry_subtest0
This failure is due to OpenJDK has different behaviour when dealing
with JarEntry in getInputStream. If argument 'entry' is of type
JarEntry, it reloads it as JarFileEntry using its file name. Therefore
setting size on 'entry' would be ignored.
We initialize entry as JarFileEntry instead, so that getInputStream sees
our change to JarEntry's size.
Bug: 25996922
Change-Id: I4976db19c4b5ddb5802a3e52af7dc9a9df85aaa1
|
| 119d05942e760b078fd8ac1ea133918ec15b8dbe |
|
03-Feb-2016 |
Yi Kong <yikong@google.com> |
Fix JarFileTest#test_getInputStreamLjava_util_jar_JarEntry_subtest0
This failure is due to OpenJDK has different behaviour when dealing
with JarEntry in getInputStream. If argument 'entry' is of type
JarEntry, it reloads it as JarFileEntry using its file name. Therefore
setting size on 'entry' would be ignored.
We initialize entry as JarFileEntry instead, so that getInputStream sees
our change to JarEntry's size.
Bug: 25996922
Change-Id: I4976db19c4b5ddb5802a3e52af7dc9a9df85aaa1
|
| ef7f5a16547089a7cdba9e48d780720f606ff54a |
|
13-Nov-2014 |
Kenny Root <kroot@google.com> |
JarUtils: stop trying to build chain past candidates length
If the certs in the PKCS#7 bag are in a loop, it will go on forever
trying to build a chain. Instead just stop trying to build the chain
when our chain exceeds the length of the candidates.
Bug: 17972577
Change-Id: If4f92e3eeabe893612a618bab0068a0f8cf75ea9
|
| 00bf89dd858de6c7eaca555210ba429a89193722 |
|
07-Nov-2014 |
Kenny Root <kroot@google.com> |
Switch order of digest and digest encryption algorithm
Sometimes "digest encryption algorithm" would be "RSA" which would match
a Signature provider, but its default setup would be whatever the
provider chose. This works fine with newer algorithms that have a
specific OID for their signature format (e.g., ECDSA and SHA256), but
not with algorithms that just have a generic OID for all possible uses
(e.g., RSA). Stock Android never hits this problem, because nothing
registers a "Signature.RSA" provider, but Spongycastle does so using
JarURLClassLoader after inserting Spongycastle causes a problem.
Flip the order of tries to make this work more uniformly with more JAR
and provider combinations.
(cherry picked from commit b1da6d3df5f9cce6e6d77c63599eba62edb465d6)
Bug: 17790692
Bug: https://code.google.com/p/android/issues/detail?id=68562
Change-Id: I3bb07ea25d7bf1d55fa2466b204594179ac38932
|
| b1da6d3df5f9cce6e6d77c63599eba62edb465d6 |
|
07-Nov-2014 |
Kenny Root <kroot@google.com> |
Switch order of digest and digest encryption algorithm
Sometimes "digest encryption algorithm" would be "RSA" which would match
a Signature provider, but its default setup would be whatever the
provider chose. This works fine with newer algorithms that have a
specific OID for their signature format (e.g., ECDSA and SHA256), but
not with algorithms that just have a generic OID for all possible uses
(e.g., RSA). Stock Android never hits this problem, because nothing
registers a "Signature.RSA" provider, but Spongycastle does so using
JarURLClassLoader after inserting Spongycastle causes a problem.
Flip the order of tries to make this work more uniformly with more JAR
and provider combinations.
Bug: 17790692
Bug: https://code.google.com/p/android/issues/detail?id=68562
Change-Id: I3bb07ea25d7bf1d55fa2466b204594179ac38932
|
| cdd4aae90277f2182420526c14b9d77a9e7f1203 |
|
29-Apr-2014 |
Kenny Root <kroot@google.com> |
JarFile: remove tests for chainCheck hidden API, part 2
The hidden API was reverted since the fix is now in
PackageManagerService.
This was PS#2 of the original change, but didn't get uploaded.
Bug: 13678484
Change-Id: Ic488ec5cd7c0d1bf6fb2d7aba16deaee8b6f98a5
|
| fe821d0a3d96518107489e22b37f664632472c41 |
|
29-Apr-2014 |
Kenny Root <kroot@google.com> |
JarFile: remove tests for chainCheck hidden API
The hidden API was reverted since the fix is now in
PackageManagerService.
Bug: 13678484
Change-Id: I9b5478af9145df82eea01e768d02945f14524bf7
|
| a3d40b4064c0f4de77122eff9253da242877c938 |
|
23-Apr-2014 |
Kenny Root <kroot@google.com> |
JarFile: Add tests for ambiguous chains
Some chains using the old code will have incorrect splits between
CodeSigner paths.
Bug: 13678484
Change-Id: Ia77b7cbcde9394e0d48a6f082bbd1cdfd880d74b
|
| d32a5c2f44587946ad9d4de26117e45b610b0a6d |
|
23-Apr-2014 |
Kenny Root <kroot@google.com> |
Add tests for getCodeSigners
Change-Id: I2bf7449eb72319a547abfd439cabfcc41982d668
|
| 19745ce89add193c0e13c18979a3c9fb25bee475 |
|
18-Feb-2014 |
Narayan Kamath <narayan@google.com> |
DO NOT MERGE Fix another set of file related tests.
- Get rid of Support_PlatformFile, use File.createTempFile
instead.
- Replace numerous uses of "user.dir" / "user.home" etc. with
File.createTempFile
- Replace calls to Process / chmod with Libcore.os.chmod. The
calls wouldn't have worked anyway, since chmod on android
doesn't understand the incremental (+w, -x) syntax.
(cherry picked from commit 2b7c83942a5e28c53698232182193d5118028e6c)
Bug: 13763685
Bug: 13763900
Change-Id: Ic6944568bbee6a548da6dcb4364b882636f2959f
|
| 2b7c83942a5e28c53698232182193d5118028e6c |
|
18-Feb-2014 |
Narayan Kamath <narayan@google.com> |
Fix another set of file related tests.
- Get rid of Support_PlatformFile, use File.createTempFile
instead.
- Replace numerous uses of "user.dir" / "user.home" etc. with
File.createTempFile
- Replace calls to Process / chmod with Libcore.os.chmod. The
calls wouldn't have worked anyway, since chmod on android
doesn't understand the incremental (+w, -x) syntax.
Change-Id: Ic6944568bbee6a548da6dcb4364b882636f2959f
|
| d567f9025c4b94fc5e9b47f5702c1b48c2a45c96 |
|
17-Dec-2013 |
Narayan Kamath <narayan@google.com> |
Move tests from harmony/archive to libcore.
Also moves a couple of tests from luni/src/test to
harmony-tests/ where they belong.
Change-Id: I0b441b1fcbd355bd3c66551d16732a671fb64f54
|