| e31b37859051d3902e06e4ba384995df7188917f |
|
02-Dec-2017 |
Hans Boehm <hboehm@google.com> |
Add reachabilityFence, ReachabilitySensitive
Add the @ReachabilitySensitive annotation itself, and use it
in libcore itself, where needed.
Import OpenJDK9+181's reachabilityFence, so that it can be used in
preference to the annotation where it makes sense. Java 9's
stopgap reachabilityFence implementation doesn't work for us,
so replace it with a different stopgap.
The libcore changes correct many situations in which finalizers
could previously run prematurely according to the JLS. Or they
would correct them if we had a real implementation for
@ReachabilitySensitive.
The ZipFile constructor could close the file before notifying the
CloseGuard that it was open.
Aside from the ZipFile change, there should be no functional
change, yet.
Note that the ExemptionMechanism.java fix is the same as upstream.
Conspicuously missing:
- Any code, e.g. in ART or D8, that actually looks for the annotation.
- Uses in frameworks code, where it's probably most critical.
Bug: 63934467
Bug: 70906684
Bug: 28342794
Test: Build & boot AOSP
Change-Id: I652f0625b39b2ba9ac901d4f63e6aa6cf3b74af5
|
| 59b140db4c4bd7617792256cd5d52713409137bc |
|
12-Jan-2018 |
Hans Boehm <hboehm@google.com> |
Avoid duplicate free in CharsetXcoderICU on OOME
NativeAllocationRegistry registration attempts that throw should
immediately deallocate, and not register a Cleaner.
CharsetEncoder/CharsetDecoder should not deallocate the native
object on OOME when NativeAllocationRegistry already did.
Bug: 71750393
Bug: 70631114
Test: Build & boot AOSP, run art host tests,
vogar --mode host libcore/luni/src/test/java/libcore/libcore/util/NativeAllocationRegistryTest.java
Change-Id: If35db36387b53f14ce1eeba09029f7a10fd7bbb5
|
| de1819e35f05acf0acbbaace42c2e54797383ad7 |
|
24-Apr-2017 |
Narayan Kamath <narayan@google.com> |
CharsetEncoderICU: Don't call updateCallback from <init>.
Changed for correctness and for consistency with CharsetDecoderICU.
The existing code is incorrect because we'd end up deleting the same
allocation twice (once from NativeAllocationRegistry's reference queue,
and once from the finally block) if updateCallback throws.
Related to b/37315864.
Test: run cts -m CtsLibcoreTestCases
Change-Id: I5425093ed7fbdc4aa69f5b9108428399b3c8c555
|
| 1bdc6bc1c72b033ed860360e69fc9f1f1121579b |
|
22-Apr-2016 |
Narayan Kamath <narayan@google.com> |
CharsetEncoder/Decoder: Use NativeAllocationRegistry.
Works around issues with finalizer ordering.
bug: 26076560
(cherry picked from commit c672ce800bbc27cb79f78d9204a51ec06ef1ec3d)
Change-Id: I434842966fe2cda70aa955250162c562ce0f8991
|
| c672ce800bbc27cb79f78d9204a51ec06ef1ec3d |
|
22-Apr-2016 |
Narayan Kamath <narayan@google.com> |
CharsetEncoder/Decoder: Use NativeAllocationRegistry.
Works around issues with finalizer ordering.
bug: 26076560
Change-Id: I4977ff9e2f6234283bd23dfc39a574d4854f6655
|
| 4e4b93ffe988b0a597835d852489573b57e0f3fe |
|
30-Jul-2014 |
Narayan Kamath <narayan@google.com> |
Don't account buffer.arrayOffset() twice.
data[OUTPUT_OFFSET] from the perspective of native
code must always be the offset from the start of the
buffer passed in to it.
Change-Id: I4a5c9595aebd105c57ce2239dc95cf85d5bd6256
|
| 41c3d56dfdc92bd5710bdaf901f45d6c4da8fd0d |
|
30-Jul-2014 |
Narayan Kamath <narayan@google.com> |
Don't account buffer.arrayOffset() twice.
data[OUTPUT_OFFSET] from the perspective of native
code must always be the offset from the start of the
buffer passed in to it.
bug: 16449607
(cherry picked from commit 4e4b93ffe988b0a597835d852489573b57e0f3fe)
Change-Id: Iddd3a835b72ae122c10e802bad92baffd0a85283
|
| 91bb19d2ef9822cfb50fd1191cdb3ee06e2939b9 |
|
27-Nov-2013 |
Narayan Kamath <narayan@google.com> |
Fix a CharsetEncoder bug.
When malformed or unmappable characters span input
buffers, we'd end up setting a negative position on the
buffer.
Also fix up a few test cases which were wrong in
several ways.
- One test was simply checking for the wrong sort
of exception (unmappable vs malformed)
- Another test was expecting encode() to throw an
error (and ignoring flush) but the API allows flush
to throw an error instead of encode.
bug: 10729779
Change-Id: I6560b749ca2445651d61ca651f8a5e388cf1c1b0
|
| f15b280ac1c2e32f7becdb314440c2999dc8d204 |
|
26-Nov-2013 |
Narayan Kamath <narayan@google.com> |
Fix an ASCIICharsetEncoderTest testcase.
Deal with truncated char sequences correctly.
bug: 10729779
Change-Id: I232b0ec899084270a6d621e6c7292ef8b94f8f7d
|
| 5ec6bf8d033754e06a463adb091d2c0afd0755ac |
|
08-Aug-2013 |
Elliott Hughes <enh@google.com> |
Fix various Charset tests, clean up the implementation.
Bug: 10211558
Bug: 10211378
Change-Id: Ib3f97430f62163c0459c53e0c282ae0ca840e0af
|
| 5ec69b20ab9b3e2dcbe225d548168b09afbbbac2 |
|
24-Jul-2012 |
Elliott Hughes <enh@google.com> |
Clean up ICU error reporting.
Throw all ICU exceptions from the native side, and include the name of the
function that failed.
Bug: 5037042
Change-Id: I49b5493a7e7365226d612479294b5b8f47ec4e8c
|
| 4fd8ac297725190c6d81e9538b1faf7cfb0f5bb8 |
|
08-Feb-2011 |
Elliott Hughes <enh@google.com> |
Fix build.
The *ICU classes shouldn't be public.
Change-Id: I1ca75c696c52f75b70d6aef29888a5a48ef7b32a
|
| 3664d8839f0ba794f428119ee7f7304a66861da5 |
|
08-Feb-2011 |
Elliott Hughes <enh@google.com> |
Clean up the CharsetDecoder/CharsetEncoder implementation a bit more.
Moving CharsetEncoderICU into java.nio.charset lets us use a cleaner workaround
for the constructor-calling-abstract method API bug.
Change-Id: I2312580b2c27711e9d4c88fe4dc057eec13c12e0
|